Abrasive Disk Inclination Angle Self-Adjustment Method, Device and Wafer Thinning Equipment

Abstract

The present application provides a method and apparatus for automatically adjusting the inclination angle of a polishing pad and a wafer thinning device, which relate to the field of wafer processing and manufacturing. The method includes: obtaining a first variable and a second variable that affect the thinning effect of a wafer to be thinned; inputting the first variable and the second variable into a pre-trained surface shape prediction model to obtain the predicted surface shape of the wafer to be thinned, where the predicted surface shape at least includes the thickness change of the wafer to be thinned after being thinned and processed by the wafer thinning device; if the predicted surface shape of the wafer to be thinned is inconsistent with the set target surface shape, inputting the predicted surface shape and the target surface shape of the wafer to be thinned into an inclination angle adjustment model to obtain the target inclination angle of the polishing pad; adjusting the inclination angle of the polishing pad from the initial inclination angle to the target inclination angle, and performing the thinning process of the wafer to be thinned by the wafer thinning device adjusted to the target inclination angle. This method can achieve precise automatic adjustment of the inclination angle of the polishing pad and improve the accuracy and quality of wafer thinning processing.

Description

Technical Field

[0001] The present application relates to the field of wafer processing and manufacturing, and particularly to a method and device for automatically adjusting the inclination angle of a polishing pad and a wafer thinning device. Background Art

[0002] In the field of semiconductor manufacturing, the thinning process of wafers is one of the key steps to improve device performance and reduce costs. The thinning techniques in related technologies mainly rely on the experience of equipment operators to adjust the inclination angle of the polishing pad to control the thickness uniformity of the wafers. However, this method has many deficiencies: on the one hand, manual adjustment lacks systematicness and precision, and it is difficult to achieve optimal control of the wafer surface shape; on the other hand, manual operation is time-consuming and laborious, seriously affecting production efficiency and the overall operation efficiency of the equipment. In addition, with the increasing requirements for wafer thickness uniformity and processing accuracy, the traditional empirical adjustment method has been difficult to meet the needs of modern production. Summary of the Invention

[0003] Based on this, in view of the above technical problems, it is necessary to provide a method and device for automatically adjusting the inclination angle of a polishing pad and a wafer thinning device, which can accurately and automatically adjust the inclination angle of the polishing pad and improve the accuracy and quality of wafer thinning processing.

[0004] In a first aspect, the present application provides a method for automatically adjusting the inclination angle of a polishing pad, the method comprising:

[0005] Obtaining a first variable and a second variable that affect the thinning effect of the wafer to be thinned; wherein, the first variable at least includes the operating parameters of the wafer thinning device, the wafer thinning device being configured to perform thinning processing on the wafer to be thinned, and the second variable includes a target parameter to be adjusted, the target parameter being the initial inclination angle of the polishing pad;

[0006] Inputting the first variable and the second variable into a pre-trained surface shape prediction model to obtain the predicted surface shape of the wafer to be thinned, the predicted surface shape at least including the thickness change of the wafer to be thinned after being thinned by the wafer thinning device;

[0007] If the predicted surface shape of the wafer to be thinned is inconsistent with the set target surface shape, inputting the predicted surface shape and the target surface shape of the wafer to be thinned into an inclination angle adjustment model to obtain the target inclination angle of the polishing pad;

[0008] Adjusting the inclination angle of the polishing pad in the wafer thinning device from the initial inclination angle to the target inclination angle, and performing thinning processing on the wafer to be thinned by the wafer thinning device adjusted to the target inclination angle.

[0009] In one embodiment, the inclination angle adjustment model adopts an optimization algorithm; inputting the predicted surface shape and the target surface shape of the wafer to be thinned into the inclination angle adjustment model to obtain the target inclination angle of the polishing pad, including:

[0010] Determine a set of inclination angle combinations. The set of inclination angle combinations includes multiple inclination angle combinations, and any one inclination angle combination includes a first angle and a second angle;

[0011] For each inclination angle combination, use the surface shape prediction model to determine the predicted surface shape corresponding to the inclination angle combination, and calculate the difference between the predicted surface shape of the inclination angle combination and the target surface shape, and define the difference as the fitness value;

[0012] If the fitness values do not meet the preset conditions, update the inclination angle combination within the search space and recalculate the fitness value;

[0013] Repeat the iteration until there is a fitness value that meets the preset conditions or reaches the maximum number of iterations, and output the optimal inclination angle combination. The optimal inclination angle combination is the inclination angle combination with the smallest fitness value.

[0014] In one embodiment, each inclination angle combination is assigned a velocity. The assigned velocity includes the unit angle of adjustment in the first angle direction and the unit angle of adjustment in the second angle direction; if the fitness values do not meet the preset conditions, update the inclination angle combination and recalculate the fitness value, including:

[0015] Update the inclination angle combination based on the set velocity update method and the set angle update method;

[0016] For each updated inclination angle combination, calculate the fitness value of the updated inclination angle combination.

[0017] In one embodiment, the set velocity update method is:

[0018] ;

[0019] Wherein, is the number of iterations, , , , , are hyperparameters, is the inclination angle combination at the historical optimal inclination angle in the th iteration, is the historical optimal inclination angle of the set of inclination angle combinations in the

[0020] The set angle update method is: ;

[0021] Wherein, is the position of the previous iteration, is the velocity of the next iteration.

[0022] In one embodiment, the operating parameters include the spindle speed, the carrier speed, and the spindle feed speed in the wafer thinning equipment;

[0023] The first variable further includes the wafer parameters of the wafer to be thinned and the auxiliary material parameters used in the wafer thinning process of the wafer to be thinned. The wafer parameters at least include the type of the wafer to be thinned, and the auxiliary material parameters at least include the type of the auxiliary material and the usage time of the auxiliary material.

[0024] In one embodiment, the surface profile prediction model adopts the extreme gradient boosting algorithm. Training the surface profile prediction model includes:

[0025] According to the wafer parameters, the auxiliary material parameters, the operating parameters, and the inclination angle of the polishing pad in each thinning process, as the input variables of the surface profile prediction model to be trained, and according to the number matching rule, find the surface profile measurement results of each wafer in the surface profile measuring device, as the output variables of the surface profile prediction model to be trained;

[0026] Adopt the extreme gradient boosting algorithm to establish the mapping relationship between the input and the output, and train the surface profile prediction model to be trained to obtain the surface profile prediction model;

[0027] Among them, the wafer type of the wafer to be thinned is input into the surface profile prediction model after one-hot encoding. Through one-hot encoding, the wafer type is converted into a unique binary feature vector, and only one value in the binary feature vector is 1.

[0028] In one embodiment, adopting the extreme gradient boosting algorithm to establish the mapping relationship between the input and the output, and training the surface profile prediction model to be trained to obtain the surface profile prediction model, includes:

[0029] Adopt the extreme gradient boosting algorithm to establish the mapping relationship between the input and the output to obtain the surface profile prediction model to be trained;

[0030] Predict the training set through the surface profile prediction model to be trained to obtain the prediction set of the training set;

[0031] Based on the deviation between the prediction set and the training set, continue to train the surface profile prediction model to be trained until the number of training times is reached or the deviation is less than the threshold, then stop training to obtain the surface profile prediction model.

[0032] In one embodiment, the present application further provides a wafer thinning equipment, and the wafer thinning equipment includes:

[0033] A polishing pad for thinning the wafer to be thinned;

[0034] A memory storing a computer program;

[0035] A processor, when executing a computer program, implements a method for automatically adjusting the inclination angle of a polishing pad.

[0036] In a second aspect, the present application further provides a device for automatically adjusting the inclination angle of a polishing pad, the device comprising:

[0037] An acquisition module, the acquisition module is used to acquire a first variable and a second variable that affect the thinning effect of the wafer to be thinned; wherein, the first variable at least includes the operating parameters of the wafer thinning equipment, and the wafer thinning equipment is configured to perform thinning processing on the wafer to be thinned, and the second variable includes the target parameter to be adjusted, and the target parameter is the initial inclination angle of the polishing pad;

[0038] A prediction module, the prediction module is used to input the first variable and the second variable into a pre-trained surface shape prediction model to obtain the predicted surface shape of the wafer to be thinned, and the predicted surface shape at least includes the thickness change of the wafer to be thinned after being thinned by the wafer thinning equipment;

[0039] An adjustment module, if the predicted surface shape of the wafer to be thinned is inconsistent with the set target surface shape, the adjustment module is used to input the predicted surface shape and the target surface shape of the wafer to be thinned into an inclination angle adjustment model to obtain the target inclination angle of the polishing pad; the adjustment module is further used to adjust the inclination angle of the polishing pad in the wafer thinning equipment from the initial inclination angle to the target inclination angle, and the wafer thinning equipment adjusted to the target inclination angle performs the thinning processing on the wafer to be thinned.

[0040] In a third aspect, the present application further provides a computer-readable storage medium, on which a computer program is stored, and when the computer program is executed by a processor, it implements the method for automatically adjusting the inclination angle of the polishing pad in the first aspect.

[0041] For the above-mentioned method for automatically adjusting the inclination angle of a polishing pad, the method acquires a first variable and a second variable that affect the thinning effect of the wafer to be thinned, wherein the first variable at least includes the operating parameters of the wafer thinning equipment, and the second variable is the initial inclination angle of the polishing pad. These two variables are input into a pre-trained surface shape prediction model to obtain the predicted surface shape of the wafer to be thinned, and the predicted surface shape at least includes the thickness change after wafer thinning. If the predicted surface shape is inconsistent with the set target surface shape, the predicted surface shape and the target surface shape are input into an inclination angle adjustment model to obtain the target inclination angle of the polishing pad. The inclination angle of the polishing pad is adjusted from the initial inclination angle to the target inclination angle, and the wafer thinning equipment after adjustment performs the thinning processing. This method can achieve precise automatic adjustment of the inclination angle of the polishing pad, improving the accuracy and quality of wafer thinning processing. Description of the Drawings

[0042] Figure 1 It is a flowchart of the method for automatically adjusting the inclination angle of a polishing pad in an embodiment;

[0043] Figure 2 It is a flowchart of obtaining the target inclination angle of a polishing pad by using an optimization algorithm in an embodiment;

[0044] Figure 3 Flow chart for updating the inclination angle combination and recalculating the fitness value in an embodiment;

[0045] Figure 4 Flow chart for obtaining the surface shape prediction model in an embodiment;

[0046] Figure 5 Structural diagram of a wafer thinning device in an embodiment;

[0047] Figure 6 Diagram of the inclination angle self - adjustment device of the grinding plate in an embodiment. Detailed implementation manners

[0048] In order to make the objectives, technical solutions and advantages of the present application clearer and more understandable, the present application will be further described in detail below with reference to 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.

[0049] In an embodiment, as Figure 1 shown, a method for self - adjusting the inclination angle of a grinding plate is provided, and the method includes the following steps:

[0050] Step 101: Obtain a first variable and a second variable that affect the thinning effect of the wafer to be thinned; wherein, the first variable includes at least the operating parameters of the wafer thinning device, the wafer thinning device is configured to perform thinning processing on the wafer to be thinned, and the second variable includes the target parameter to be adjusted, and the target parameter is the initial inclination angle of the grinding plate;

[0051] The first variable may include the operating parameters of the wafer thinning device, such as the spindle speed, the carrier plate speed, the spindle feed speed, etc. The operating parameters of the wafer thinning device can determine the basic conditions and processing rate of the thinning operation. Further, the first variable may also include the wafer type and auxiliary material information. Among them, the wafer type may be a code predefined according to the process to which the processing belongs and the wafer size; the auxiliary material information includes but is not limited to the type and life of the grinding wheel during processing.

[0052] The second variable may be the initial inclination angle of the grinding plate in the wafer thinning device, and the initial inclination angle of the grinding plate is a target parameter to be adjusted. By adjusting the initial inclination angle of the grinding plate, the thickness uniformity of the wafer can be optimized, and further the overall quality of the thinning effect can be improved. Exemplarily, appropriately increasing the initial inclination angle of the grinding plate can enhance the cutting effect of the abrasive on the edge area of the wafer, which helps to improve the thickness uniformity of the edge; while reducing the inclination angle may make the thinning of the central area more sufficient.

[0053] Step 102: Input the first variable and the second variable into a pre-trained surface shape prediction model to obtain the predicted surface shape of the wafer to be thinned. The predicted surface shape at least includes the thickness change of the wafer to be thinned after being thinned by a wafer thinning device.

[0054] It should be noted that the surface shape prediction model can be trained through a large amount of historical data. During the training process, various data in the thinning processing history are collected, including wafer type, auxiliary material information, process parameter settings, chuck tilt angle, etc. as input variables, and the corresponding surface shape measurement results as output variables. Further, a machine learning algorithm, such as the Extreme Gradient Boosting algorithm, can be used to establish the mapping relationship between the input and the output. The Extreme Gradient Boosting algorithm optimizes the objective function, iteratively constructs a tree model, and optimizes the structure of the tree through a greedy algorithm to obtain a powerful integrated model. During the training process, the model continuously learns the association between the input variables and the surface shape results to improve the prediction accuracy. After sufficient training and optimization, when the fitting error of the model reaches the minimum, the surface shape prediction model is obtained.

[0055] When the first variable and the second variable are input into this pre-trained surface shape prediction model, the surface shape prediction model can predict the surface shape of the wafer after thinning according to its internal algorithm and the knowledge learned during the previous training process. The predicted surface shape can include the thickness change of the wafer after being thinned, that is, the Total Thickness Variation (TTV). The surface shape prediction model can comprehensively consider the influence of operating parameters and the initial tilt angle on the thinning process, and through complex calculations and analyses, output a predicted surface shape result.

[0056] Exemplarily, if the predicted surface shape is consistent with the target surface shape, it indicates that the current operating parameters and the initial tilt angle setting are reasonable, and the thinning process can continue. On the contrary, if there is a deviation between the predicted surface shape and the target surface shape, the tilt angle of the polishing pad needs to be adjusted to ensure that the final wafer surface shape meets the requirements.

[0057] Step 103: If the predicted surface shape of the wafer to be thinned is inconsistent with the set target surface shape, input the predicted surface shape and the target surface shape of the wafer to be thinned into an inclination angle adjustment model to obtain the target inclination angle of the polishing pad.

[0058] It should be noted that the tilt angle adjustment model can be constructed based on the particle swarm optimization algorithm. To obtain the tilt angle adjustment model, a large amount of historical data needs to be collected, including the wafer surface shape results under different tilt angle combinations and the corresponding target surface shapes. The large amount of historical data is used to train the tilt angle adjustment model, enabling the tilt angle adjustment model to learn how to adjust the tilt angle according to the surface shape difference. During the training process, the tilt angle adjustment model can continuously try different tilt angle combinations, calculate the matching degree between the surface shape under each combination and the target surface shape, and optimize the tilt angle combination according to the matching degree. Through repeated iterations, the tilt angle adjustment model can learn the optimal tilt angle adjustment strategy, and then be able to accurately output the optimal target tilt angle of the polishing pad according to the input predicted surface shape and target surface shape.

[0059] When the predicted surface shape and the target surface shape are input into the pre-trained tilt angle adjustment model, the tilt angle adjustment model can quickly calculate a target tilt angle of the polishing pad that can make the predicted surface shape closer to the target surface shape based on the optimization strategy learned previously. This target tilt angle can accurately guide the wafer thinning equipment to automatically adjust the tilt angle of the polishing pad, so that in the subsequent thinning process, the actual surface shape of the wafer to be thinned is closer to or even reaches the target surface shape, effectively improving the processing quality and production efficiency of the wafer.

[0060] Step 104: Adjust the tilt angle of the polishing pad from the initial tilt angle to the target tilt angle, and the wafer thinning equipment adjusted to the target tilt angle performs the thinning process on the wafer to be thinned.

[0061] After determining the target tilt angle of the polishing pad, the tilt angle of the polishing pad can be adjusted. The process of adjusting the tilt angle of the polishing pad can be completed by the automatic adjustment system of the wafer thinning equipment. Further, based on the target tilt angle output by the tilt angle adjustment model, the tilt angle of the polishing pad is adjusted from the initial tilt angle to the target tilt angle. After the adjustment is completed, the wafer thinning equipment is in a new processing condition, and the wafer thinning equipment can perform the thinning process on the wafer to be thinned according to the established operating parameters and the new tilt angle setting of the polishing pad. During the processing, the polishing pad applies pressure to the wafer at the target tilt angle and performs polishing, so that the wafer can remove materials according to the expected surface shape change during the thinning process to achieve a processing effect closer to the target surface shape.

[0062] In this embodiment, the method obtains a first variable and a second variable that affect the thinning effect of the wafer to be thinned. The first variable at least includes the operating parameters of the wafer thinning equipment, and the second variable is the initial inclination angle of the grinding plate in the wafer thinning equipment. These two variables are input into a pre-trained surface shape prediction model to obtain the predicted surface shape of the wafer to be thinned, and the predicted surface shape at least includes the thickness change after wafer thinning. If the predicted surface shape is inconsistent with the set target surface shape, the predicted surface shape and the target surface shape are input into an inclination angle adjustment model to obtain the target inclination angle of the grinding plate. The inclination angle of the grinding plate is adjusted from the initial inclination angle to the target inclination angle, and the wafer thinning process is performed by the adjusted wafer thinning equipment. This method can achieve precise automatic adjustment of the inclination angle of the grinding plate and improve the accuracy and quality of wafer thinning processing.

[0063] In one embodiment, as Figure 2 shown, the inclination angle adjustment model adopts an optimization algorithm; inputting the predicted surface shape and the target surface shape of the wafer to be thinned into the inclination angle adjustment model to obtain the target inclination angle of the grinding plate includes the following steps:

[0064] Step 201: Determine an inclination angle combination set, which includes multiple inclination angle combinations. Any one inclination angle combination includes a first angle and a second angle;

[0065] There are usually multiple feet at the bottom of the grinding plate. By adjusting the height of the feet, the grinding plate can be tilted at different angles. The first angle and the second angle can respectively correspond to the tilt angles of the grinding plate in two orthogonal directions, such as tilting along the length direction and the width direction of the grinding plate. It should be noted that the inclination angle combination set can be obtained through the collection of historical data, and the historical data can come from previous

[0066] Step 202: For each inclination angle combination, use the surface shape prediction model to determine the predicted surface shape corresponding to the inclination angle combination, and calculate the difference between the predicted surface shape of the inclination angle combination and the target surface shape, and define the difference as the fitness value;

[0067] Take the first angle and the second angle in each inclination angle combination as input variables, and input them into the surface shape prediction model in combination with the first variable. The surface shape prediction model can predict the surface shape of the wafer after thinning processing according to its internal algorithm and the knowledge learned in the previous training process, and output the corresponding predicted surface shape.

[0068] After obtaining the predicted surface shape corresponding to each inclination angle combination, compare the predicted surface shape with the set target surface shape, calculate the difference between the two, and define the difference as the fitness value.

[0069] It should be noted that the magnitude of the fitness value can reflect the matching degree between the predicted wafer surface shape and the desired target surface shape under this inclination combination. The smaller the fitness value, the more it can indicate that this inclination combination can make the actual processed wafer surface shape closer to the target surface shape.

[0070] Step 203: If the fitness values do not all meet the preset conditions, update the inclination combination within the search space and recalculate the fitness values.

[0071] Compare these fitness values with the preset conditions. If it is found that the fitness values of all inclination combinations do not meet the preset conditions, it means that there is no suitable solution in the current set of inclination combinations that can make the predicted surface shape close to the target surface shape. Among them, the preset conditions can be various specific judgment criteria. For example, an absolute error threshold can be set, and this threshold requires that the difference in thickness change between the predicted surface shape and the target surface shape is less than a preset value, such as 0.01 mm.

[0072] If the fitness values do not all meet the preset conditions, it is necessary to update the inclination combination within the search space. Among them, the search space can refer to the set of all possible inclination combinations, including the angular ranges of the first angle and the second angle. It should be noted that the method of updating the inclination combination can be based on various optimization algorithms, such as the particle swarm optimization algorithm. Exemplarily, in the particle swarm optimization algorithm, each inclination combination can be regarded as a particle, and these particles move and adjust their positions in the search space to find a better solution. After the updated set of inclination combinations is generated, the surface shape prediction model can be used again to evaluate each new inclination combination, determine its corresponding predicted surface shape, and recalculate the fitness values.

[0073] Step 204: Repeat the iteration until there is a fitness value that meets the preset conditions or the maximum number of iterations is reached, and output the optimal inclination combination. The optimal inclination combination is the inclination combination with the smallest fitness value.

[0074] In each iteration, use the surface shape prediction model to evaluate each inclination combination in the current set of inclination combinations, obtain its corresponding predicted surface shape, and calculate the difference between the predicted surface shape and the target surface shape as the fitness value. If in the current iteration, there is a fitness value of a certain inclination combination that meets the preset conditions, that is, the difference between its predicted surface shape and the target surface shape is within the allowable range, the iteration process stops, and this inclination combination is the optimal inclination combination.

[0075] If no inclination angle combination meets the conditions in the current iteration, the set of inclination angle combinations is updated according to an optimization algorithm (such as the particle swarm optimization algorithm) to generate new inclination angle combinations, and their fitness values are recalculated. This process is repeated continuously until an inclination angle combination that meets the conditions is found or the maximum number of iterations is reached. When the maximum number of iterations is reached, even if no inclination angle combination meets the preset conditions, the inclination angle combination with the minimum fitness value among all iterations is output as the optimal inclination angle combination.

[0076] In this embodiment, precise adjustment of the inclination angle of the polishing pad can be achieved by this method, improving the quality and efficiency of wafer thinning processing.

[0077] In one embodiment, as Figure 3 shown, each inclination angle combination is assigned a velocity, and the assigned velocity includes the unit angle of adjustment in the first angle direction and the unit angle of adjustment in the second angle direction; if the fitness values do not meet the preset conditions, the inclination angle combination is updated, and the fitness value is recalculated, including the following steps:

[0078] Step 301: Update the inclination angle combination based on the set velocity update method and the set angle update method;

[0079] The velocity update method can adjust the velocity of each inclination angle combination (regarded as a particle). The velocity of each inclination angle can determine the moving direction and distance of the particle in the search space. The update of the velocity takes into account the best position of the particle itself (i.e., the optimal inclination angle combination in the history of this particle) and the best position of the entire population (i.e., the inclination angle combination with the minimum fitness value among all particles). By integrating the position information of its own best and the population best, the particle can determine a suitable velocity to more effectively explore the possible solution space in subsequent searches.

[0080] The angle update method can be to adjust the position (i.e., the angle value) of the inclination angle combination according to the updated velocity. The update of the position means that the new inclination angle combination (position) is the current position plus the updated velocity. In this way, each inclination angle combination moves in the search space to explore better solutions.

[0081] Step 302: For each updated inclination angle combination, calculate the fitness value of the updated inclination angle combination.

[0082] After each update of the inclination angle combination, each new inclination angle combination is input into the surface shape prediction model, and the wafer surface shape corresponding to this inclination angle combination is predicted by using this model. The predicted surface shape is compared with the target surface shape, and the difference between the two is calculated. This difference is the fitness value. By calculating the fitness value of each updated inclination angle combination, the advantages and disadvantages of different inclination angle combinations can be quantitatively evaluated, and then the optimal inclination angle combination can be selected in the subsequent iteration process to achieve precise adjustment of the inclination angle of the polishing pad.

[0083] In this embodiment, by updating the speed and angle, the fitness value is calculated, which can quantitatively evaluate the advantages and disadvantages of each inclination angle combination, so as to screen out the optimal combination during the iteration process and achieve precise adjustment of the inclination angle of the grinding disc.

[0084] In one embodiment, the set speed update method is:

[0085] ;

[0086] Wherein, is the number of iterations, , , , , are hyperparameters, is the inclination angle combination at the historical optimal inclination angle in the th iteration, is the historical optimal inclination angle of the inclination angle combination set in the th iteration;

[0087] The set angle update method is: ;

[0088] Wherein, is the position of the previous iteration, is the speed of the next iteration.

[0089] The update of the inclination angle combination is based on the set speed update method and angle update method. The speed update method comprehensively considers the historical optimal position of the particle and the historical optimal position of the population, and determines the moving direction and distance of the particle in the search space by adjusting parameters such as the inertia weight, learning factor, and random number. The speed update formula is: .

[0090] Wherein, is the number of iterations, , , , , are hyperparameters, is the inclination angle combination at the historical optimal inclination angle in the th iteration, is the historical optimal inclination angle of the inclination angle combination set in the th iteration. It should be noted that is used to balance the influence of the current speed of the particle on the next speed; , are learning factors, representing the learning abilities of particles with respect to their own historical optimal positions and the global historical optimal position respectively

[0091] The angle update method adjusts the position of the inclination combination according to the updated velocity, and the formula is:

[0092] 。

[0093] Wherein, is the position of the previous iteration, i.e., the current inclination combination; is the velocity of the next iteration, i.e., the new moving direction and distance calculated according to the velocity update formula.

[0094] The process of angle update can be understood as the particle moving from the current position to a new position according to its new velocity. This step is actually to generate a new inclination combination in the search space to explore possible better solutions.

[0095] In this embodiment, in this way, it is possible to systematically explore better inclination combinations within the search space, gradually approaching the optimal solution that makes the predicted surface shape closest to the target surface shape, thereby achieving precise adjustment of the inclination of the polishing pad and improving the precision and quality of wafer thinning processing.

[0096] In one embodiment, the operating parameters include the spindle speed, the carrier speed, and the spindle feed speed in the wafer thinning equipment;

[0097] The first variable further includes the wafer parameters of the wafer to be thinned and the auxiliary material parameters used during the wafer thinning process of the wafer to be thinned. The wafer parameters at least include the type of the wafer to be thinned, and the auxiliary material parameters at least include the type of the auxiliary material and the usage time of the auxiliary material.

[0098] The operating parameters include the spindle speed, the carrier speed, and the spindle feed speed in the wafer thinning equipment. The operating parameters can determine the basic conditions and processing rate of the thinning operation. The first variable further includes the wafer parameters of the wafer to be thinned and the auxiliary material parameters used during the wafer thinning process of the wafer to be thinned. The wafer parameters at least include the type of the wafer to be thinned. Different types of wafers may have different physical and chemical properties during the thinning process, such as hardness, thickness, material composition, etc. These properties can affect the thinning efficiency and effect. The auxiliary material parameters at least include the type of the auxiliary material and the usage time of the auxiliary material. The type of the auxiliary material can determine its action mode and effect on the wafer during the thinning process. For example, the particle size and hardness of the abrasive will affect the material removal rate and surface quality; the usage time of the auxiliary material will affect the stability of its performance. As the usage time increases, the auxiliary material may wear or fail, thus affecting the thinning effect.

[0099] In this embodiment, multiple parameters together constitute the first variable, which is closely related to the operating state of the wafer thinning equipment and has an important impact on the thinning process and the final wafer surface shape.

[0100] In one embodiment, the surface shape prediction model adopts the extreme gradient boosting algorithm. Based on the wafer parameters, auxiliary material parameters, operating parameters, and the inclination angle of the polishing pad during each thinning process, they are used as the input variables of the surface shape prediction model to be trained. According to the number matching rule, the surface shape measurement result of each wafer is obtained as the output variable of the surface shape prediction model to be trained. The extreme gradient boosting algorithm is used to establish the mapping relationship between the input and output, and the surface shape prediction model to be trained is trained to obtain the surface shape prediction model.

[0101] Among them, the wafer type of the wafer to be thinned is processed by one-hot encoding and then input into the surface shape prediction model. Through one-hot encoding, the wafer type is converted into a unique binary feature vector, and only one value in the binary feature vector is 1.

[0102] The surface shape prediction model can be trained using the extreme gradient boosting algorithm to accurately predict the surface shape of the wafer based on various parameters during each thinning process. The wafer parameters, auxiliary material parameters, operating parameters, and the inclination angle of the polishing pad are used as the input variables of the surface shape prediction model to be trained. These parameters comprehensively cover the key influencing factors in the thinning process. Among them, the wafer parameters include the type of the wafer to be thinned, etc., the auxiliary material parameters involve the type and usage time of the auxiliary materials, the operating parameters include the spindle speed, carrier speed, and spindle feed speed of the wafer thinning equipment, etc., and the inclination angle of the polishing pad is also an important influencing factor.

[0103] To obtain the output variable of the surface shape prediction model, that is, the surface shape measurement result of the wafer, the surface shape data corresponding to each wafer is obtained according to the number matching rule. The surface shape measurement result can reflect the thickness change of the wafer after thinning and is a key indicator for evaluating the processing quality.

[0104] The extreme gradient boosting algorithm is used to train the model. The extreme gradient boosting algorithm is an ensemble learning algorithm of decision trees based on gradient boosting. By optimizing the objective function, multiple decision trees are iteratively constructed, and the prediction results of these trees are combined to improve the prediction accuracy and generalization ability of the model. During the training process, the extreme gradient boosting algorithm learns the influence pattern of the input parameters on the surface shape according to the relationship between the input variables and the output variables in the training samples, so as to establish the mapping relationship between the input and output.

[0105] Specifically, the extreme gradient boosting algorithm uses the training set and the sample true values to train the first tree, and uses this tree to predict the training set to obtain the predicted values of each sample. Since there is a deviation between the predicted values and the true values, subtracting the two can obtain the "residuals". Then, the second tree is trained. At this time, the true values are no longer used, but the residuals are used as the standard answers. After the two trees are trained, the residuals of each sample can be obtained again, and the third tree is further trained, and so on. The total number of trees can be specified manually or the training can be stopped by monitoring certain metrics (such as the error on the validation set). When predicting new samples, each tree will have an output value, and adding up the output values of each tree gives the final predicted value of the sample.

[0106] Furthermore, for the categorical variable of the wafer type of the wafer to be thinned, one-hot encoding can be used for processing and then input into the model. One-hot encoding converts each wafer type into a unique binary feature vector, in which only one value is 1 and the rest are 0. For example, if there are three wafer types A, B, and C, then A can be encoded as [1, 0, 0], B as [0, 1, 0], and C as [0, 0, 1]. This processing method avoids introducing false order relationships between categories, enabling the model to more accurately identify and utilize the differences between different wafer types, and improving the accuracy of prediction.

[0107] In this embodiment, this method can obtain a well-trained surface shape prediction model, which can quickly and accurately predict the surface shape of the wafer to be thinned according to the input processing parameters and wafer type in actual production, so as to achieve precise control of the wafer thinning process.

[0108] In one embodiment, as Figure 4 shown, the extreme gradient boosting algorithm is used to establish the mapping relationship between the input and the output, and the surface shape prediction model to be trained is trained to obtain the surface shape prediction model, including the following steps:

[0109] Step 401: Use the extreme gradient boosting algorithm to establish the mapping relationship between the input and the output to obtain the surface shape prediction model to be trained;

[0110] The input variables include the operating parameters of the wafer thinning equipment, the wafer parameters of the wafer to be thinned, the auxiliary material parameters used in the thinning process, and the inclination angle of the polishing pad. The output variable is the surface shape measurement result of each wafer obtained according to the number matching rule, including at least the thickness change after wafer thinning.

[0111] During the training process, the extreme gradient boosting algorithm can learn the influence mode of the input parameters on the surface shape according to the relationship between the input variables and the output variables, establish the mapping relationship between the input and the output, and then obtain the surface shape prediction model to be trained.

[0112] It should be noted that the Extreme Gradient Boosting algorithm can initialize a decision tree and determine the structure and parameters of the tree by minimizing the objective function. The objective function usually includes a loss function (such as mean squared error) and a regularization term to balance the fitting ability and complexity of the surface shape prediction model. Further, based on the residuals between the prediction results of the current tree and the true outputs, the next tree is trained, and so on, until the preset number of trees or the convergence condition of the algorithm is reached.

[0113] Step 402: Predict the training set through the surface shape prediction model to be trained, and obtain a prediction set corresponding to the training set.

[0114] The training set can include the operating parameters of the wafer thinning equipment, the wafer parameters of the wafer to be thinned, the auxiliary material parameters used in the thinning process, and the inclination angle of the polishing pad as input variables. The surface shape measurement results of each wafer obtained according to the number matching rule are used as output variables.

[0115] Through the Extreme Gradient Boosting algorithm, the surface shape prediction model learns the mapping relationship between the input variables and the output variables. After training is completed, the surface shape prediction model is used to predict the training set, that is, the input variables of the training set are input into the model, and the surface shape prediction model can output the corresponding predicted surface shape results, which together form the prediction set. Further, the prediction set is compared with the output variables of the training set to evaluate the performance and accuracy of the surface shape prediction model, ensuring that the surface shape prediction model can accurately predict the surface shape after wafer thinning.

[0116] Step 403: Based on the deviation between the prediction set and the training set, continue to train the surface shape prediction model to be trained until the number of training times is reached or the deviation is less than the threshold, then stop training to obtain the surface shape prediction model.

[0117] The training set is used to conduct preliminary training on the surface shape prediction model. The surface shape prediction model is used to predict the training set to obtain a prediction set. The prediction set is compared with the output variables of the training set, and the deviation between the prediction set and the output variables of the training set is calculated. If the deviation is greater than the set threshold and the number of training times has not reached the preset maximum value, then continue to train the surface shape prediction model and adjust the parameters of the surface shape prediction model to reduce the deviation, where the threshold can be the maximum allowable value of the deviation between the prediction set and the training set. This process will be repeated continuously until the deviation is less than the threshold or the maximum number of training times is reached. At this time, the training of the surface shape prediction model is completed, and the final surface shape prediction model is obtained, which can more accurately predict the surface shape after wafer thinning.

[0118] In this embodiment, this method can effectively improve the prediction accuracy and generalization ability of the surface shape prediction model, ensuring that the surface shape prediction model can accurately predict the surface shape after wafer thinning in practical applications.

[0119] In one embodiment, as Figure 5 shown, a wafer thinning device 50 is provided, and the wafer thinning device 50 includes:

[0120] A grinding platen 51 for thinning the wafer to be thinned;

[0121] A memory 52 storing a computer program;

[0122] A processor 53 that implements the self-adjustment method of the inclination angle of the grinding platen when executing the computer program.

[0123] Specifically, a wafer thinning device 50 aims to realize the automatic adjustment of the inclination angle of the grinding platen to optimize the wafer thinning effect. The device mainly consists of a grinding platen 51, a memory 52, and a processor 53. The grinding platen 51 is a key component that directly contacts the wafer to be thinned and performs the thinning process. The precise adjustment of its inclination angle plays a decisive role in the flatness uniformity and processing quality of the wafer surface. The memory 52 is used to store a computer program containing the self-adjustment method of the inclination angle of the grinding platen, and this program is the core software support for realizing the automatic adjustment function. The processor 43 is the key hardware for executing the program in the memory. When the processor 53 runs the program, it will implement the self-adjustment method of the inclination angle of the grinding platen according to the preset algorithm and logic.

[0124] Based on the same concept, as Figure 6 shown, the present application also provides a device for self-adjusting the inclination angle of a grinding platen, and the device includes:

[0125] An acquisition module 601 for acquiring a first variable and a second variable that affect the thinning effect of the wafer to be thinned; wherein, the first variable at least includes the operating parameters of the wafer thinning device configured to perform thinning processing on the wafer to be thinned, and the second variable includes a target parameter to be adjusted, and the target parameter is the initial inclination angle of the grinding platen;

[0126] A prediction module 602 for inputting the first variable and the second variable into a pre-trained surface shape prediction model to obtain the predicted surface shape of the wafer to be thinned, and the predicted surface shape at least includes the thickness change of the wafer to be thinned after being thinned by the wafer thinning device;

[0127] An adjustment module 603. If the predicted surface shape of the wafer to be thinned is inconsistent with the set target surface shape, the adjustment module is used to input the predicted surface shape and the target surface shape of the wafer to be thinned into an inclination angle adjustment model to obtain the target inclination angle of the grinding platen; the adjustment module is also used to adjust the inclination angle of the grinding platen from the initial inclination angle to the target inclination angle, and the wafer thinning device adjusted to the target inclination angle performs the thinning processing on the wafer to be thinned.

[0128] In one embodiment, the adjustment module 603 inputs the predicted surface shape and the target surface shape of the wafer to be thinned into the tilt angle adjustment model to obtain the target tilt angle of the polishing pad, specifically for: determining a set of tilt angle combinations, where the set of tilt angle combinations includes multiple tilt angle combinations, and any tilt angle combination includes a first angle and a second angle; for each tilt angle combination, using the surface shape prediction model to determine the predicted surface shape corresponding to the tilt angle combination, and calculating the difference between the predicted surface shape of the tilt angle combination and the target surface shape, and defining the difference as the fitness value; if the fitness values do not satisfy the preset conditions, update the tilt angle combination within the search space and recalculate the fitness value; repeat the iteration until there is a fitness value that satisfies the preset conditions or the maximum number of iterations is reached, and output the optimal tilt angle combination, where the optimal tilt angle combination is the tilt angle combination with the smallest fitness value.

[0129] In one embodiment, the adjustment module 603 assigns a speed to each tilt angle combination, and the assigned speed includes the unit angle of adjustment in the first angle direction and the unit angle of adjustment in the second angle direction; if the fitness values do not satisfy the preset conditions, update the tilt angle combination and recalculate the fitness value, specifically for: updating the tilt angle combination based on the set speed update method and the set angle update method; for each updated tilt angle combination, calculate the fitness value of the updated tilt angle combination.

[0130] In one embodiment, the speed update method set by the adjustment module 603 is:

[0131] ;

[0132] Wherein, is the number of iterations, , , , , are hyperparameters, is the tilt angle combination In the historical optimal tilt angle in the iteration, is the historical optimal tilt angle of the set of tilt angle combinations in the iteration; the set angle update method is: ; wherein, is the position of the previous iteration, is the speed of the next iteration.

[0133] In one embodiment, the adjustment module 603 is specifically used for: the operating parameters include the spindle speed, the carrier speed, and the spindle feed speed in the wafer thinning equipment; the first variable further includes the wafer parameters of the wafer to be thinned and the auxiliary material parameters used in the wafer thinning process of the wafer to be thinned, and the wafer parameters at least include the type of the wafer to be thinned, and the auxiliary material parameters at least include the type of the auxiliary material and the usage time of the auxiliary material.

[0134] In one embodiment, the adjustment module 603 is specifically configured to: the surface shape prediction model adopts the extreme gradient boosting algorithm. The wafer parameters, auxiliary material parameters, operating parameters, and the inclination angle of the polishing pad during each thinning process are used as the input variables of the surface shape prediction model to be trained. According to the number matching rule, the surface shape measurement results of each wafer in the surface shape measurement device are found and used as the output variables of the surface shape prediction model to be trained. The extreme gradient boosting algorithm is used to establish the mapping relationship between the input and the output, and the surface shape prediction model to be trained is trained to obtain the surface shape prediction model. Among them, the wafer type of the wafer to be thinned is processed by one-hot encoding and then input into the surface shape prediction model. Through the one-hot encoding process, the wafer type is converted into a unique binary feature vector, and only one value in the binary feature vector is 1.

[0135] In one embodiment, the adjustment module 603 uses the extreme gradient boosting algorithm to establish the mapping relationship between the input and the output, and trains the surface shape prediction model to be trained to obtain the surface shape prediction model. Specifically, it is configured to: use the extreme gradient boosting algorithm to establish the mapping relationship between the input and the output to obtain the surface shape prediction model to be trained. The training set is predicted through the surface shape prediction model to be trained to obtain the prediction set of the training set. Based on the deviation between the prediction set and the training set, the surface shape prediction model to be trained is continuously trained until the training times are reached or the deviation is less than the threshold, and then the training is stopped to obtain the surface shape prediction model.

[0136] Based on the same concept, the present application also provides a computer-readable storage medium, on which a computer program is stored. When the computer program is executed by a processor, the method for self-adjusting the inclination angle of the polishing pad is implemented.

[0137] The technical features of the above embodiments can be combined arbitrarily. For the sake of concise description, 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, it should be considered as the scope described in this specification.

[0138] The above embodiments only represent several implementation manners of the present application. The description is relatively specific and detailed, but it cannot be understood as a limitation on the patent scope of the present application. It should be noted that for those of ordinary skill in the art, without departing from the concept of the present application, several deformations and improvements can be made, and these all belong to the protection scope of the present application. Therefore, the protection scope of the present application should be subject to the appended claims.

Claims

1. A method for self-adjusting the inclination angle of a grinding disc, characterized in that: The method comprises: Obtaining a first variable and a second variable that affect the thinning effect of the wafer to be thinned; wherein the first variable at least includes an operating parameter of a wafer thinning device, the wafer thinning device is configured to perform a thinning process on the wafer to be thinned, and the second variable includes an adjusted target parameter, the target parameter being an initial inclination angle of the grinding disc; Inputting the first variable and the second variable into a pre-trained surface shape prediction model to obtain a predicted surface shape of the wafer to be thinned, wherein the predicted surface shape at least includes a thickness change of the wafer to be thinned after thinning by the wafer thinning equipment; If the predicted surface shape of the wafer to be thinned is inconsistent with the set target surface shape, the predicted surface shape of the wafer to be thinned and the target surface shape are input into the inclination adjustment model to obtain the target inclination angle of the grinding disc; The inclination angle of the grinding disc is adjusted from the initial inclination angle to the target inclination angle, and the wafer thinning equipment adjusted to the target inclination angle performs the thinning process on the wafer to be thinned.

2. The grinding disc inclination self-adjusting method according to claim 1, characterized in that: The tilt angle adjustment model adopts an optimization algorithm; Inputting the predicted surface shape of the wafer to be thinned and the target surface shape into the inclination adjustment model to obtain the target inclination angle of the grinding disc comprises: Determine a tilt angle combination set, wherein the tilt angle combination set includes a plurality of tilt angle combinations, and any of the tilt angle combinations includes a first angle and a second angle; For each inclination angle combination, using the surface shape prediction model to determine a predicted surface shape corresponding to the inclination angle combination, and calculating a difference between the predicted surface shape of the inclination angle combination and the target surface shape, and defining the difference as a fitness value; If none of the fitness values ​​meet the preset conditions, the inclination angle combination is updated in the search space, and the fitness value is recalculated; Repeat the iteration until there is a fitness value that meets the preset condition or reaches the maximum number of iterations, and output the optimal inclination angle combination, where the optimal inclination angle combination is the inclination angle combination with the smallest fitness value.

3. The grinding disc inclination self-adjusting method according to claim 2, characterized in that: Each tilt angle combination is assigned a speed, and the assigned speed includes a unit angle adjusted in the first angle direction and a unit angle adjusted in the second angle direction; If the fitness values ​​do not meet the preset conditions, the inclination angle combination is updated and the fitness value is recalculated, including: Update the inclination angle combination based on a set speed update method and a set angle update method; For each updated inclination angle combination, the fitness value of the updated inclination angle combination is calculated.

4. The grinding disc inclination self-adjusting method according to claim 3, characterized in that: The speed update method of the setting is: ； in, is the number of iterations, , , , , is a hyperparameter, Combination of inclination angles exist The historical optimal inclination angle in the iteration, The inclination angle combination is The historical optimal inclination angle in the iteration; The setting angle update method is: ; in, is the position of the previous iteration, The speed of the next iteration.

5. The grinding disc inclination self-adjusting method according to claim 1, characterized in that: The operating parameters include spindle speed, carrier speed and spindle feed speed in the wafer thinning equipment; The first variable also includes wafer parameters of the wafer to be thinned and auxiliary material parameters used in the thinning process of the wafer to be thinned. The wafer parameters at least include the type of the wafer to be thinned, and the auxiliary material parameters at least include the type of auxiliary material and the usage time of the auxiliary material.

6. The grinding disc inclination self-adjusting method according to claim 5, characterized in that: The face shape prediction model adopts an extreme gradient boosting algorithm, and training the face shape prediction model includes: According to the wafer parameters, auxiliary material parameters, operating parameters and the inclination angle of the grinding disc in each thinning process, as the input variables of the surface shape prediction model to be trained, according to the number matching rule, the surface shape measurement result of each wafer in the surface shape measurement device is searched as the output variable of the surface shape prediction model to be trained; The extreme gradient boosting algorithm is used to establish a mapping relationship between input and output, and the face shape prediction model to be trained is trained to obtain the face shape prediction model; Among them, the wafer type of the wafer to be thinned is input into the surface prediction model after one-hot encoding processing, and the wafer type is converted into a unique binary feature vector through one-hot encoding processing, and only one value in the binary feature vector is 1.

7. The grinding disc inclination self-adjusting method according to claim 6, characterized in that: The extreme gradient boosting algorithm is used to establish a mapping relationship between input and output, and the face shape prediction model to be trained is trained to obtain the face shape prediction model, including: The extreme gradient boosting algorithm is used to establish a mapping relationship between input and output to obtain a face shape prediction model to be trained; Predicting a training set by the face shape prediction model to be trained to obtain a prediction set corresponding to the training set; Based on the deviation between the prediction set and the training set, the face shape prediction model to be trained is continuously trained until the number of training times is reached or the deviation is less than a threshold, then the training is stopped to obtain the face shape prediction model.

8. A wafer thinning device, characterized in that: The wafer thinning equipment comprises: A grinding disc, wherein the grinding disc is used to thin the wafer to be thinned; a memory storing a computer program; A processor, wherein when executing the computer program, the processor implements the grinding disc inclination self-adjustment method according to any one of claims 1 to 7.

9. A grinding disc inclination self-adjusting device, characterized in that: The device comprises: An acquisition module, the acquisition module is used to acquire a first variable and a second variable that affect the thinning effect of the wafer to be thinned; wherein the first variable at least includes an operating parameter of a wafer thinning device, the wafer thinning device is configured to perform a thinning process on the wafer to be thinned, and the second variable includes an adjusted target parameter, the target parameter being an initial inclination angle of the grinding disc; A prediction module, the prediction module is used to input the first variable and the second variable into a pre-trained surface shape prediction model to obtain a predicted surface shape of the wafer to be thinned, the predicted surface shape at least including a thickness change of the wafer to be thinned after thinning by the wafer thinning equipment; An adjustment module is provided. If the predicted surface shape of the wafer to be thinned is inconsistent with the set target surface shape, the adjustment module is used to input the predicted surface shape of the wafer to be thinned and the target surface shape into the inclination adjustment model to obtain the target inclination angle of the grinding disk; the adjustment module is also used to adjust the inclination angle of the grinding disk in the wafer thinning equipment from the initial inclination angle to the target inclination angle, and the wafer thinning equipment adjusted to the target inclination angle performs the thinning process of the wafer to be thinned.

10. A computer-readable storage medium having a computer program stored thereon, characterized in that: When the computer program is executed by a processor, the grinding disc inclination self-adjustment method according to any one of claims 1 to 7 is implemented.

Images