End effector for semiconductor wafer conveying manipulator

Abstract

The invention relates to the technical field of semiconductors, in particular to an end effector for a semiconductor wafer conveying manipulator. The tail end executor mainly comprises a tail end executor main body, a plurality of gas path adsorption modules with a negative pressure value independent control function and a control module. According to the invention, by arranging the plurality of gas path adsorption modules with the function of independently controlling the negative pressure value, multiple adsorption modes can be realized to adapt to the warping forms of different areas of the wafer; through the synergistic effect of the gas path adsorption module, the sensor array and the control module, according to the gap value between the target area and the end effector main body collected by the optical fiber sensor, the warping area of the wafer is automatically predicted and positioned through a machine learning classification model, and a negative pressure dynamic adjustment strategy is combined to adjust the warping area of the wafer. Therefore, differential adsorption force adjustment of different warping areas of the wafer is achieved, the response speed and adaptability of the end effector of the wafer conveying mechanical arm are remarkably improved, and the stability and reliability of wafer adsorption are effectively improved.

Description

Technical Field

[0001] The present invention relates to the field of semiconductor technology, and in particular, to an end effector for a semiconductor wafer transfer manipulator. Background Art

[0002] Currently, traditional end effectors of wafer transfer manipulators usually adopt a single air circuit structure for negative pressure adsorption control, and cannot perform differential negative pressure adjustment according to the warping conditions of different regions of the wafer. This single air circuit structure is prone to cause insecure adsorption or stress concentration when encountering deformations such as warping at the wafer edge or depression at the center, thereby leading to problems such as wafer slippage or rupture. In the prior art, related technical solutions attempt to detect the contact state between the wafer and the chuck through a pressure sensor, but in a complex electromagnetic interference environment, the signal of the pressure sensor is prone to distortion, and the warping area cannot be accurately located. In addition, the regulation logic of the adsorption force is usually a fixed threshold judgment, lacking the ability to dynamically predict the warping trend of the wafer, which reduces the response speed and adaptability of the end effector of the wafer transfer manipulator.

[0003] Therefore, this application provides an end effector for a semiconductor wafer transfer manipulator. Summary of the Invention

[0004] Based on this, it is necessary to provide an end effector for a semiconductor wafer transfer manipulator in view of the above technical problems.

[0005] The present invention provides an end effector for a semiconductor wafer transfer manipulator, including: an end effector main body for carrying a wafer; a plurality of air circuit adsorption modules with independent negative pressure value control functions, each of the air circuit adsorption modules is correspondingly provided with a sensor array, and each of the sensor arrays corresponds to a target area on the surface of a wafer, for real-time collecting the gap value between the target area and the end effector main body, and transmitting it to a control module, and dynamically regulating the negative pressure value of the corresponding target area according to the received dynamic negative pressure value output signal; a control module for constructing and training a machine learning classification model, preprocessing the received gap value between the target area and the end effector main body to generate a deformation feature array, inputting the deformation feature array into the trained machine learning classification model, outputting a wafer warping type judgment result, and generating a dynamic negative pressure value output signal for each air circuit adsorption module based on a negative pressure dynamic adjustment strategy, and transmitting it to the corresponding air circuit adsorption module.

[0006] Optionally, each of the air circuit adsorption modules includes at least one adsorption unit, and each adsorption unit includes a boss and an air cushion, the air cushion is hermetically fitted with the inner sunken hole of the boss, and a negative pressure hole is opened on the air cushion.

[0007] Optionally, the end effector body includes a main body portion, a connecting portion, and a cover plate. An air passage is provided inside the main body portion and is connected to each negative pressure hole in one-to-one correspondence. The rear end of the main body portion is fixedly connected to the connecting portion, and the main body portion and the connecting portion are hermetically fitted through the cover plate. An air pipeline is provided inside the connecting portion and is connected to the air passage in one-to-one correspondence. A regulating valve is provided on the air pipeline.

[0008] Optionally, each of the sensor arrays includes at least one fiber optic sensor, and one fiber optic sensor is disposed at the bottom of the side surface of each of the bosses for starting to collect the gap value between the target area and the end effector body in real time when the distance between the end effector body and the surface of the wafer is less than a preset distance threshold.

[0009] Optionally, the air adsorption module further includes a vacuum pump controlled by PWM. The vacuum pump is provided with a D / A conversion module. The vacuum pump is connected to the control module for receiving the dynamic negative pressure value output signal transmitted by the control module, and mapping the dynamic negative pressure value output signal to the rotation speed of the vacuum pump through the D / A conversion module to achieve dynamic regulation of the negative pressure value of the corresponding target area.

[0010] Optionally, constructing and training the machine learning classification model includes: constructing a historical gap value set, the historical gap value set containing a plurality of training samples, using a multi-class encoding technique to label the typical warping type to which each training sample belongs to obtain a labeled historical gap value set, where the typical warping types include edge warping, center depression, local deformation, and normal state; constructing an initial machine learning classification model using a lightweight CNN convolutional neural network; training the initial machine learning classification model based on the labeled historical gap value set, and combining with the cross-entropy loss function to obtain the trained machine learning classification model.

[0011] Optionally, preprocessing the received gap value between the target area and the end effector body to generate a deformation feature array includes: performing sliding window filtering processing on the received gap value between the target area and the end effector body to obtain gap change filtering processed data; based on the gap change filtering processed data, calculating the gap change rate and local curvature of each target area respectively, where the gap change rate is used to reflect the speed of local deformation of the wafer, and the local curvature is used to reflect the degree of bending of the wafer surface; combining the gap change rate and the local curvature to construct an initial deformation feature array; performing Z-score normalization processing on the initial deformation feature array to obtain the deformation feature array.

[0012] Optionally, the calculation formula for the gap change rate of each target area is: , where is the gap change rate of each target area, d( t+1 is the gap value collected by the i-th fiber optic sensor in the t+1 time for each target area, d ( t is the gap value collected by the i-th fiber optic sensor in the t time for each target area, and Δt is the time interval for collecting the gap value.

[0013] Optionally, the local curvature calculation formula for each target area is: , where is the local curvature of each target area, dx is the distance between two adjacent fiber optic sensors in each target area, is the second derivative of the gap value collected by the i-th fiber optic sensor in the t time for each target area.

[0014] Optionally, the generating of the dynamic negative pressure value output signals for each gas path adsorption module based on the negative pressure dynamic regulation strategy includes: constructing the negative pressure dynamic regulation strategy based on the mapping relationship between the typical warping types and the negative pressure regulation logic rules; determining the corresponding negative pressure regulation logic rules based on the output wafer warping type judgment result, and calculating and generating the dynamic negative pressure value output signals for each gas path adsorption module.

[0015] The advantages and beneficial effects of the present invention are as follows: A semiconductor wafer transfer robot end effector provided by the present invention can achieve multiple adsorption modes by setting multiple gas path adsorption modules with independent negative pressure value control functions to adapt to the warping forms of different regions of the wafer; through the synergistic effect among the gas path adsorption module, the sensor array and the control module, according to the gap value between the target area collected by the fiber optic sensor and the end effector main body, the warping area of the wafer is automatically predicted and located by the machine learning classification model, and combined with the negative pressure dynamic regulation strategy, so as to realize the differential adsorption force regulation of different warping areas of the wafer, which not only significantly improves the response speed and adaptability of the semiconductor wafer transfer robot end effector, but also effectively improves the stability and reliability of wafer adsorption. At the same time, using fiber optic sensors to replace traditional pressure sensors also has the advantages of anti-electromagnetic interference and higher accuracy. BRIEF DESCRIPTION OF THE DRAWINGS

[0016] Figure 1 is a schematic structural diagram of the end effector main body of the present invention mounted on a wafer transfer robot.

[0017] Figure 2 is a schematic structural diagram of the semiconductor wafer transfer robot end effector of the present invention.

[0018] Figure 3 is a schematic internal structural diagram of the end effector main body of the present invention.

[0019] Figure 4 This is a schematic structural diagram of the gas path adsorption module of the present invention disposed on the main body part.

[0020] Figure 5 This is another schematic structural diagram of the gas path adsorption module of the present invention disposed on the main body part.

[0021] Figure 6 This is another schematic structural diagram of the gas path adsorption module of the present invention disposed on the main body part.

[0022] Among them, the end effector body 1, the multi-joint mechanical motion structure 2, the wafer transfer robot 3, the main body part 4, the connecting part 5, the cover plate 6, the boss 7, the air cushion 8, the fiber optic sensor 9, the gas path channel 10, the tracheal joint 11, the gas path pipeline 12, the regulating valve 13, the gas path adsorption module a14, the gas path adsorption module b15, the gas path adsorption module c16, the gas path adsorption module d17, the gas path adsorption module e18, the gas path adsorption module f19, the gas path adsorption module g20, the gas path adsorption module h21. Detailed implementation manners

[0023] Hereinafter, embodiments of the present application will be described in detail. Examples of the embodiments are shown in the accompanying drawings, where the same or similar reference numerals denote the same or similar elements or elements having the same or similar functions throughout. The embodiments described below with reference to the accompanying drawings are exemplary only for explaining the present application and should not be construed as limiting the present application. And without conflict, the features in the following embodiments and the embodiments can be combined with each other. All other embodiments obtained by those of ordinary skill in the art based on the embodiments in the present application without creative efforts belong to the scope of protection of the present application.

[0024] In the description of the present application, it should be noted that, unless otherwise clearly defined and limited, the terms "installed", "connected", "connected" should be understood in a broad sense. For example, it can be a fixed connection, a detachable connection, or an integral connection; it can be a mechanical connection, an electrical connection; it can be directly connected, or indirectly connected through an intermediate medium, and it can be the communication inside two elements. For those of ordinary skill in the art, the specific meanings of the above terms in the present application can be understood according to specific situations.

[0025] The present application provides an end effector for a semiconductor wafer transfer manipulator, mainly including an end effector body, a plurality of gas path adsorption modules with independent negative pressure value control functions, and a control module.

[0026] In some optional implementation manners of the present application, the end effector body is used to carry a wafer.

[0027] In some optional implementation manners of the present application, referring to the attached Figure 2-3, the end effector body includes a body portion 4, a connecting portion 5, and a cover plate 6. An air passage 10 connected to each negative pressure hole is provided inside the body portion 4. The rear end of the body portion 4 is fixedly connected to the connecting portion 5, and the body portion 4 and the connecting portion 5 are hermetically fitted through the cover plate 6. An air pipeline 12 connected to each air passage 10 is provided inside the connecting portion 5, and a regulating valve 13 is provided on the air pipeline 12.

[0028] In some alternative implementation manners of the present application, referring to the attached Figure 1 , the end effector body 1 involved in the present application is arranged on the multi-joint mechanical motion structure 2 through a joint axis. Among them, the multi-joint mechanical motion structure 2 includes N robotic arms controlled by N + 1 joint axes rotating on a horizontal plane, where N is an integer greater than 0. At the same time, the multi-joint mechanical motion structure 2 is arranged on the wafer transfer robot 3 through a joint axis. The wafer transfer robot 3 is a body with its own vertical direction transmission. Through the coordinated action of the end effector body 1, the multi-joint mechanical motion structure 2, and the wafer transfer robot 3, it is used to transfer wafers.

[0029] In some alternative implementation manners of the present application, an air pipe joint 11 is also provided at the connection where the air passage 10 penetrates through the body portion 4 and is connected to the air pipeline 12, for connecting the air passage 10 and the air pipeline 12 in one-to-one correspondence.

[0030] In some alternative implementation manners of the present application, there are multiple air path adsorption modules with independent negative pressure value control functions. Each air path adsorption module is correspondingly provided with a sensor array, and each sensor array corresponds to a target area on the surface of a wafer, for real-time collecting the gap value between the target area and the end effector body, and transmitting it to the control module, and outputting a signal according to the received dynamic negative pressure value to dynamically adjust the negative pressure value of the corresponding target area.

[0031] In some alternative implementation manners of the present application, referring to the attached Figure 2 , each air path adsorption module includes at least one adsorption unit. The adsorption unit includes a boss 7 and an air cushion 8. The air cushion 8 is hermetically fitted with the inner sunken hole of the boss 7, and a negative pressure hole is provided on the air cushion 8.

[0032] It should be noted that, considering the comprehensive factors of reducing manufacturing costs, simplifying the device integration, the structural shape of the main body part 4, and avoiding extreme warping conditions (such as more than 5 mm), the number and layout positions of the gas path adsorption modules can be flexibly configured. Among them, multiple gas path adsorption modules can be arranged concentrically on the main body part 4 in a ring, or arranged in a circular array on the main body part 4, or arranged in other forms on the main body part 4, and can be flexibly selected according to the actual industrial production situation, which will not be limited here. At the same time, the adsorption units can be arranged in a circular array on the main body part 4, or arranged in other forms on the main body part 4, and can be flexibly selected according to the actual industrial production situation, which will not be limited here. As long as the strategy of combining the multi-gas path negative pressure dynamic regulation and real-time feedback mechanism provided by this application can achieve the purpose of differentially regulating the negative pressure of different areas of the wafer for warping conditions.

[0033] In some alternative implementation manners of the present application, referring to the attached Figure 4 , for example: the front adsorption surface of the main body part 4 is in a semi-circular structure, and the number of gas path adsorption modules can be set to three, specifically the gas path adsorption module a14, the gas path adsorption module b15, and the gas path adsorption module c16. The three gas path adsorption modules are arranged concentrically on the main body part 4 to divide the wafer surface into multiple target areas, such as the edge area, the center area, and the local area. The gas path adsorption module a14, the gas path adsorption module b15, and the gas path adsorption module c16 all include at least one adsorption unit. Each adsorption unit includes 1 boss 7 and 1 air cushion 8 embedded on the boss 7. Each air cushion 8 is provided with 1 negative pressure hole, and each negative pressure hole is respectively communicated with a gas path channel 10. Each gas path channel 10 is communicated with a gas path pipeline 12. Through the independent regulation of the three gas path adsorption modules, multiple adsorption modes can be realized, such as only the gas path adsorption module a14 working, only the gas path adsorption module b15 working, only the gas path adsorption module c16 working, the gas path adsorption module a14 and the gas path adsorption module b15 working together, the gas path adsorption module a14 and the gas path adsorption module c16 working together, the gas path adsorption module b15 and the gas path adsorption module c16 working together, and the gas path adsorption module a14, the gas path adsorption module b15, and the gas path adsorption module c16 working together, so as to adapt to the warping forms of different areas of the wafer.

[0034] In some alternative implementation manners of the present application, referring to the attached Figure 5, for example: the front adsorption surface of the body part 4 is circular in structure, and the number of the gas path adsorption modules can be set to three, specifically the gas path adsorption module d17, the gas path adsorption module e18 and the gas path adsorption module f19. The three gas path adsorption modules are concentrically arranged on the body part 4 and are used to divide the surface of the wafer into multiple target areas, such as the edge area, the central area and the local area. The gas path adsorption module d17, the gas path adsorption module e18 and the gas path adsorption module f19 all include at least one adsorption unit. Each adsorption unit includes 1 boss 7 and 1 air cushion 8 embedded on the boss 7. Each air cushion 8 is provided with 1 negative pressure hole. Each negative pressure hole is respectively communicated with a gas path channel 10. Each gas path channel 10 is communicated with a gas path pipeline 12. Through the independent regulation of the three gas path adsorption modules, multiple adsorption modes can be realized, such as only the gas path adsorption module d17 works, only the gas path adsorption module e18 works, only the gas path adsorption module f19 works, the gas path adsorption module d17 and the gas path adsorption module e18 work together, the gas path adsorption module d17 and the gas path adsorption module f19 work together, the gas path adsorption module e18 and the gas path adsorption module f19 work together, and the gas path adsorption module d17, the gas path adsorption module e18 and the gas path adsorption module f19 work together, so as to adapt to the warping forms of different areas of the wafer.

[0035] In some alternative implementation manners of the present application, with reference to the attached Figure 6 , for example: the body part 4 is in a Y-shaped structure, and the number of the gas path adsorption modules can be set to two, specifically the gas path adsorption module g20 and the gas path adsorption module h21. The two gas path adsorption modules are concentrically arranged on the body part 4 and are used to divide the surface of the wafer into multiple target areas, such as the edge area and the local area. The gas path adsorption module g20 and the gas path adsorption module h21 both include at least one adsorption unit (such as 3 adsorption units). Each adsorption unit includes 1 boss 7 and 1 air cushion 8 embedded on the boss 7. Each air cushion 8 is provided with 1 negative pressure hole. Each negative pressure hole is respectively communicated with a gas path channel 10. Each gas path channel 10 is communicated with a gas path pipeline 12. Through the independent regulation of the two gas path adsorption modules, multiple adsorption modes can be realized, such as only the gas path adsorption module g20 works, only the gas path adsorption module h21 works, and the gas path adsorption module g20 and the gas path adsorption module h21 work together, so as to adapt to the warping forms of different areas of the wafer.

[0036] In some alternative implementation manners of the present application, each of the sensor arrays includes a plurality of fiber optic sensors 9. One fiber optic sensor 9 is arranged at the bottom of the side surface of each boss 7 and is used to start real-time acquisition of the gap value between the target area and the end effector body when the distance between the end effector body and the surface of the wafer is less than a preset distance threshold.

[0037] In some alternative implementation manners of the present application, in a complex electromagnetic interference environment, in order to avoid the problems that the signal of the pressure sensor is prone to distortion and the warping area cannot be accurately located, the present application adopts a high-precision fiber optic sensor 9 (such as FBG). By arranging the fiber optic sensor 9 at the bottom of the side surface of the boss 7, the gap value between the target area and the end effector body can be collected in real time. When the distance between the end effector body and the wafer surface is less than a preset distance threshold (such as 1 mm), the fiber optic sensor 9 starts to collect time series data (i.e., the gap value between the target area and the end effector body), and the sampling frequency can be set to 1 kHz to ensure high-resolution capture of the deformation of the wafer surface. The fiber optic sensor 9 detects the change in the phase difference of the reflected light on the wafer surface to obtain the gap value between the wafer and the end effector body in real time.

[0038] In some alternative implementation manners of the present application, the air path adsorption module further includes a vacuum pump controlled by PWM. The vacuum pump is provided with a D / A conversion module. The vacuum pump is connected to the control module and is used to receive the dynamic negative pressure value output signal transmitted by the control module. The dynamic negative pressure value output signal is mapped to the rotation speed of the vacuum pump through the D / A conversion module to realize dynamic regulation of the negative pressure value of the corresponding target area.

[0039] In some alternative implementation manners of the present application, the present application also sets a pressure sensor on each air path pipeline 12 to monitor the negative pressure value generated at each target area in real time, and forms a closed-loop regulation and verification with the generated dynamic negative pressure values of each air path adsorption module to ensure the accuracy and reliability of dynamically regulating the negative pressure values at each target area.

[0040] In some alternative implementation manners of the present application, the control module is used to construct and train a machine learning classification model, preprocess the received gap value between the target area and the end effector body, generate a deformation feature array, input the deformation feature array into the trained machine learning classification model, output a wafer warping type judgment result, and generate a dynamic negative pressure value output signal for each air path adsorption module based on the negative pressure dynamic regulation strategy and transmit it to the corresponding air path adsorption module.

[0041] In some alternative implementation manners of the present application, constructing and training the machine learning classification model includes: constructing a historical gap value set, where the historical gap value set contains multiple training samples, and using a multi-class encoding technique to label the typical warping type to which each training sample belongs, obtaining a historical gap value set with labels, where the typical warping types include edge warping, center depression, local deformation, and normal state; constructing an initial machine learning classification model using a lightweight CNN convolutional neural network; training the initial machine learning classification model based on the historical gap value set with labels, and combining with a cross-entropy loss function to obtain a trained machine learning classification model.

[0042] In some alternative implementation manners of the present application, the present application uses a lightweight CNN convolutional neural network to construct an initial machine learning classification model. The lightweight CNN convolutional neural network is composed of an input layer, a one-dimensional convolutional layer, a max pooling layer, and a fully connected layer connected in sequence. Among them, the input layer is used to input a deformation feature array with a length of n, where the value of n corresponds to the number of fiber optic sensors; the one-dimensional convolutional layer contains 64 filters and 3 convolutional kernels, and the activation function uses ReLU; the pooling window of the max pooling layer is 2; the fully connected layer contains 128 neurons and uses the Softmax activation function to output the wafer warping type judgment result. During training, the cross-entropy loss function and the Adam optimizer are used to continuously optimize the machine learning classification model, and the learning rate can be set to 0.001. In addition, during the offline training phase of the machine learning classification model, GPU acceleration can be used, and finally it is deployed in the embedded FPGA of the control module to ensure that the inference delay is less than 10 ms.

[0043] In some alternative implementation manners of the present application, the constructed historical gap value set involved in the present application contains multiple training samples (i.e., the historical gap values between the target area and the end effector body). According to the preset typical warping types, a multi-class encoding technique is used to label the typical warping type to which each training sample belongs. The typical warping types include edge warping, center depression, local deformation, and normal state. Among them, edge warping means that the gap change rate in the wafer edge area is significantly higher than that in the center area, and can be defined by the formula as , σ represents the standard deviation of the deformation feature, represents the gap change rate in the wafer edge area; center depression means that the gap change rate in the center area is negative and the local curvature increases significantly, and can be defined by the formula as and , represents the gap change rate in the wafer center area, represents the local curvature in the wafer center area; local deformation means that the gap change rate in the acquisition area of a single fiber optic sensor changes suddenly, and can be defined by the formula as , Represents the gap change rate of the acquisition area of a single fiber optic sensor; the normal state is represented by the gap change rates of all areas being within ±σ; the label can use multi-class coding. For example: edge warping is coded as 0, central depression is coded as 1, local deformation is coded as 2, and the normal state is coded as 3.

[0044] In some alternative implementation manners of the present application, the preprocessing of the gap value between the received target area and the end effector body to generate a deformation feature array includes: performing sliding window filtering on the gap value between the received target area and the end effector body to obtain gap change filtered data; based on the gap change filtered data, calculating the gap change rate and local curvature of each target area respectively, where the gap change rate is used to reflect the speed of local deformation of the wafer, and the local curvature is used to reflect the degree of bending of the wafer surface; combining the gap change rate and the local curvature to construct an initial deformation feature array; performing Z-score normalization processing on the initial deformation feature array to obtain the deformation feature array.

[0045] In some alternative implementation manners of the present application, the present application adopts a sliding window filtering technique to perform sliding window filtering on the gap value between the target area collected by the fiber optic sensor and the end effector body, with a window length of N and a sampling frequency of , to eliminate high-frequency noise; in addition, detrending processing can also be performed on the data after sliding window filtering, specifically: removing the linear trend in the data through polynomial fitting to ensure the accuracy of subsequent feature extraction.

[0046] In some alternative implementation manners of the present application, the construction steps of the initial deformation feature array involved in the present application include: (1) calculating the gap change rate of each target area based on the gap change filtered data, which is used to reflect the speed of local deformation of the wafer, and the formula is: , where is the gap change rate of each target area, d ( t+1 ) is the gap value collected by the i-th fiber optic sensor in each target area at time t + 1, d ( t ) is the gap value collected by the i-th fiber optic sensor in each target area at time t, and Δt is the time interval for collecting the gap value; (2) calculating the local curvature of each target area based on the gap change filtered data, which is used to reflect the degree of bending of the wafer surface, and the formula is: , where is the local curvature of each target area, dx is the distance between two adjacent fiber optic sensors in each target area, is the second derivative of the gap value collected by the i-th fiber optic sensor in each target area at time t; (3) Combine the gap change rate and the local curvature to construct an initial deformation feature array, which can be expressed as: , where X represents the set of initial deformation feature arrays, respectively represent the gap change rates of target area 1, target area 2, and target area n, respectively represent the local curvatures of target area 1, target area 2, and target area n.

[0047] In some alternative implementation manners of the present application, perform Z-score normalization processing on the initial deformation feature array, and the formula is: , where, represents the deformation feature array after Z-score normalization processing, μ represents the mean value of the deformation features, which is used to eliminate the dimensional differences between different fiber optic sensors and obtain the deformation feature array.

[0048] In some alternative implementation manners of the present application, the dynamic negative pressure value output signals of each gas path adsorption module are generated based on the negative pressure dynamic adjustment strategy, including: constructing a negative pressure dynamic adjustment strategy based on the mapping relationship between the typical warping types and the negative pressure adjustment logic rules; determining the corresponding negative pressure adjustment logic rules based on the output wafer warping type judgment result, and calculating and generating the dynamic negative pressure value output signals of each gas path adsorption module.

[0049] In some alternative implementation manners of the present application, the negative pressure dynamic adjustment strategy involved in the present application is constructed based on the mapping relationship between the typical warping types and the negative pressure adjustment logic rules. Specifically: (1) For the edge warping situation, the negative pressure adjustment logic rules include: increasing the negative pressure value of the gas path adsorption module in the edge area and keeping the negative pressure value of the gas path adsorption module in the central area unchanged at the reference negative pressure value. "Increasing the negative pressure value of the gas path adsorption module in the edge area" can be defined by the formula as , represents the dynamic negative pressure value of the gas path adsorption module in the edge area, represents the reference negative pressure value, ω represents the proportionality coefficient, and ω for the edge warping situation is experimentally calibrated to 0.5, represents the gap change rate of the edge area; (2) For the central depression situation, the negative pressure adjustment logic rules include: increasing the negative pressure value of the gas path adsorption module in the central area and keeping the negative pressure value of the gas path adsorption module in the edge area unchanged at the reference negative pressure value. "Increasing the negative pressure value of the gas path adsorption module in the central area" can be defined by the formula as , represents the dynamic negative pressure value of the gas path adsorption module in the central area, It represents the gap change rate of the central region, and the ω experiment for the central depression case is calibrated to 0.7; (3) For the local deformation case, the negative pressure regulation logic rule includes: applying a pulsed negative pressure to the gas path adsorption module at the deformation region for a duration of 50 ms, which can be defined by the formula as , represents the dynamic negative pressure value of the gas path adsorption module at the deformation region, t represents the duration of applying the pulsed negative pressure to the gas path adsorption module at the deformation region, and the experiment is calibrated to 50 ms, represents the gap change rate of the acquisition region of a single fiber optic sensor, τ represents the pulse decay time constant, and τ is taken as 20 ms; (4) For the normal state case, the negative pressure regulation logic rule includes: applying a uniform reference negative pressure to all regions .

[0050] In some alternative implementation manners of the present application, according to the wafer warpage type judgment result output by the machine learning classification model, based on the negative pressure dynamic regulation strategy, the corresponding negative pressure regulation logic rule is determined. Combining with the formula corresponding to the negative pressure regulation logic rule, the dynamic negative pressure value of each gas path adsorption module is calculated, and the output signal of the dynamic negative pressure value is transmitted to the corresponding gas path adsorption module, and the negative pressure value of the corresponding target region is dynamically regulated through the corresponding gas path adsorption module.

[0051] In some alternative implementation manners of the present application, taking the transfer of 12-inch wafers as an example for illustration, the performance comparison between the actuator with a single gas path structure and the present application is shown in Table 1 for reference.

[0052] Table 1 Performance comparison results between the actuator with a single gas path structure and the present application As can be seen from Table 1, the end effector for semiconductor wafer transfer manipulator provided by the present application can achieve the purpose of differential negative pressure regulation for the warpage conditions of different regions of the wafer through the combination strategy of multi-gas path negative pressure dynamic regulation and real-time feedback mechanism, which not only significantly improves the response speed and adaptability of the end effector of the wafer transfer manipulator, but also effectively improves the stability and reliability of wafer adsorption.

[0053] Finally, it should be noted that: the above embodiments are only used to illustrate the technical solutions of the present invention and not to limit them. Although the present invention has been described in detail with reference to the above embodiments, those of ordinary skill in the art should understand that: the specific implementation manners of the present invention can still be modified or equivalently replaced, and any modification or equivalent replacement without departing from the spirit and scope of the present invention should be covered within the protection scope of the claims of the present invention.

Claims

1. An end effector for a semiconductor wafer transfer manipulator, characterized in that, Comprising: An end effector body for carrying a wafer; A plurality of gas path adsorption modules with independent negative pressure value control functions, the gas path adsorption modules are arranged on the end effector body, each gas path adsorption module is correspondingly provided with a sensor array, and each sensor array corresponds to a target area on the surface of a wafer, for real-time collecting the gap value between the target area and the end effector body, and transmitting it to the control module, and outputting a signal according to the received dynamic negative pressure value to dynamically adjust the negative pressure value of the corresponding target area; A control module for constructing and training a machine learning classification model, preprocessing the received gap value between the target area and the end effector body to generate a deformation feature array, inputting the deformation feature array into the trained machine learning classification model, outputting a wafer warping type judgment result, and generating a dynamic negative pressure value output signal for each gas path adsorption module based on the negative pressure dynamic adjustment strategy, and transmitting it to the corresponding gas path adsorption module.

2. The end effector of a semiconductor wafer transfer manipulator according to claim 1, wherein Each gas path adsorption module includes at least one adsorption unit, and each adsorption unit includes a boss and an air cushion, the air cushion is hermetically fitted with the inner sunken hole of the boss, and a negative pressure hole is opened on the air cushion.

3. The end effector of a semiconductor wafer transfer manipulator according to claim 2, characterized in that, The end effector body includes a main body part, a connecting part and a cover plate. An air path channel corresponding to the negative pressure hole one by one is arranged inside the main body part. The rear end of the main body part is fixedly connected with the connecting part, and the main body part and the connecting part are hermetically fitted through the cover plate. An air path pipeline corresponding to the air path channel one by one is arranged inside the connecting part, and a regulating valve is arranged on the air path pipeline.

4. The end effector of a semiconductor wafer transfer manipulator according to claim 2, wherein Each sensor array includes at least one fiber optic sensor, and a fiber optic sensor is arranged at the bottom of the side surface of each boss, for starting to collect the gap value between the target area and the end effector body in real time when the distance between the end effector body and the wafer surface is less than a preset distance threshold.

5. A semiconductor wafer transfer robot end effector according to claim 1, characterized in that, The gas path adsorption module further includes a vacuum pump controlled by PWM, the vacuum pump is provided with a D / A conversion module, and the vacuum pump is connected with the control module, for receiving the dynamic negative pressure value output signal transmitted by the control module, mapping the dynamic negative pressure value output signal to the rotation speed of the vacuum pump through the D / A conversion module, so as to realize the dynamic adjustment of the negative pressure value of the corresponding target area.

6. The end effector of a semiconductor wafer transfer manipulator according to claim 1, characterized in that The constructing and training the machine learning classification model includes: Constructing a historical gap value set, the historical gap value set contains a plurality of training samples, using a multi-classification coding technology to label the typical warping type to which each training sample belongs, and obtaining a historical gap value set with labels, wherein the typical warping types include edge warping, center depression, local deformation and normal state; Using a lightweight CNN convolutional neural network to construct an initial machine learning classification model; Training the initial machine learning classification model based on the historical gap value set with labels, and combining with the cross-entropy loss function to obtain the trained machine learning classification model.

7. A semiconductor wafer transfer robot end effector according to claim 1, characterized in that, The preprocessing the received gap value between the target area and the end effector body to generate a deformation feature array includes: Perform sliding window filtering on the received gap value between the target area and the end effector body to obtain gap change filtering processed data; Based on the gap change filtering processed data, calculate the gap change rate and local curvature of each target area respectively, where the gap change rate is used to reflect the speed of local deformation of the wafer, and the local curvature is used to reflect the degree of bending of the wafer surface; Combine the gap change rate and local curvature to construct an initial deformation feature array; Perform Z-score normalization on the initial deformation feature array to obtain a deformation feature array.

8. The end effector of a semiconductor wafer transfer manipulator according to claim 7, characterized in that, The calculation formula for the gap change rate of each target area is: Among them, is the gap change rate of each target area, d ( t+1 ) is the gap value collected by the i-th fiber optic sensor in each target area at time t + 1, d ( t ) is the gap value collected by the i-th fiber optic sensor in each target area at time t, and Δt is the time interval for collecting the gap value.

9. The end effector of a semiconductor wafer transfer manipulator according to claim 7, wherein, The calculation formula for the local curvature of each target area is: wherein, is the local curvature of each target area, dx is the spacing between two adjacent optical fiber sensors within each target area, is the second derivative of the gap value collected by the i-th optical fiber sensor at time t within each target area.

10. A semiconductor wafer transfer robot end effector according to claim 1, characterized in that, Based on the negative pressure dynamic regulation strategy, generate the dynamic negative pressure value output signals of each gas path adsorption module, including: Based on the mapping relationship between typical warping types and negative pressure regulation logic rules, construct a negative pressure dynamic regulation strategy; Based on the determined wafer warping type judgment result, determine the corresponding negative pressure regulation logic rule, and calculate and generate the dynamic negative pressure value output signals of each gas path adsorption module.

Images