Compensation method for positioning errors of semiconductor metrology apparatuses and related products

Abstract

This invention provides a method for compensating positioning errors in semiconductor measurement equipment and related products. The compensation method includes: continuously acquiring multiple motion parameters of the workpiece stage at intervals; inputting these motion parameters into a positioning error model; analyzing the previously input motion parameters using the positioning error model; and outputting the positioning error of the workpiece stage's current position. Based on the positioning error output by the positioning error model, motion compensation is performed on the workpiece stage until it moves to the target position. This method achieves real-time, dynamic compensation for the coupled positioning errors of multiple-axis motion systems, improving the positioning accuracy of semiconductor measurement equipment.

Description

Technical Field

[0001] This invention relates to the field of integrated circuit technology, and in particular to a method for compensating for positioning errors in semiconductor measurement equipment, a computer-readable storage medium, a computer program product, and a computer device. Background Technology

[0002] As a core inspection device in the integrated circuit manufacturing process, the critical dimension measurement scanning electron microscope relies on electron beam imaging technology to achieve nanometer-level resolution, and is specifically used for fine inspection of defects found in the process.

[0003] In CD-SEM equipment, the high-precision motion control platform plays the following core functions: it achieves high-precision positioning of wafer defect coordinates through the XY dual-axis motion system, and achieves precise macro-motion focusing through the Z-axis motion system. The XY dual-axis motion system and the Z-axis motion system work together to ensure that the electron beam can be accurately focused on the target detection area, providing a basic guarantee for subsequent high-resolution imaging.

[0004] However, during macro-motion focusing, the Z-axis motion system transmits vibrations through the mechanical structure, interfering with the linear motion accuracy of the XY dual-axis motion system. This causes positioning errors in the XY plane of the workpiece stage. Furthermore, the modes of the XY dual-axis motion system and the Z-axis motion system at different work positions can also cause positioning errors in the XY plane of the workpiece stage. Summary of the Invention

[0005] One object of the present invention is to provide a method for compensating positioning errors of semiconductor measurement equipment, a computer-readable storage medium, a computer program product, and a computer device, so as to compensate for multi-axis coupling errors generated by the semiconductor measurement equipment during the movement of the workpiece stage and improve the positioning accuracy of the workpiece stage.

[0006] Specifically, according to one aspect of the present invention, the present invention provides a method for compensating for positioning errors in a semiconductor measurement device, wherein the following steps are performed during the process of moving the workpiece stage of the semiconductor measurement device to a target position: Multiple motion parameters of the workpiece stage are continuously acquired at intervals; The multiple motion parameters are input into the positioning error model. The positioning error model analyzes the multiple motion parameters input in each iteration and outputs the positioning error of the current position of the workpiece stage. The positioning error model is trained to learn the mapping relationship between the time-series data of the multiple motion parameters and the positioning error, and outputs the positioning error corresponding to the most recent input data based on the input data in each iteration. Based on the positioning error output by the positioning error model, motion compensation is performed on the workpiece stage until it moves to the target position.

[0007] Optionally, the semiconductor measurement device includes a scanning electron microscope, which includes an XY-axis motion system and a Z-axis motion system. The XY-axis motion system is used to control the movement of the workpiece stage in the XY plane, and the Z-axis motion system is used to control the movement of the workpiece stage in the Z-axis direction. The Z-axis is parallel to the optical axis of the electron optical system of the scanning electron microscope.

[0008] The positioning error model is used to output the positioning error of the workpiece stage in the XY plane when it is driven by the Z-axis motion system to move along the Z-axis direction. The motion compensation of the workpiece stage based on the positioning error refers to compensation in the XY plane.

[0009] Optionally, the motion parameters include X-axis and Y-axis positions, Z-axis position, Z-axis velocity, and Z-axis acceleration.

[0010] Optionally, the positioning error model is constructed based on a recurrent neural network.

[0011] Optionally, the positioning error model is constructed based on a long short-term memory neural network.

[0012] Optionally, the positioning error model is trained using training data, and the steps for obtaining the training data include: Using the XY axis motion system, the workpiece stage is moved to multiple set positions in the XY plane; At each of the set positions, the Z-axis motion system is used to drive the workpiece stage to move within a set stroke range of the Z-axis, and the plurality of motion parameters are continuously acquired at intervals, as well as the measured position coordinates of the workpiece stage on the XY plane measured by the laser measuring instrument are continuously acquired at intervals.

[0013] Optionally, the laser measuring instrument measures the actual position coordinates under vacuum conditions.

[0014] According to another aspect of the present invention, a computer-readable storage medium is also provided, on which a computer program is stored, wherein the computer program, when executed by a processor, implements the steps of the method for compensating for positioning errors of any of the above-described semiconductor measurement devices.

[0015] According to another aspect of the present invention, a computer program product is also provided, comprising a computer program that, when executed by a processor, implements the steps of the method for compensating for positioning errors of any of the above-described semiconductor measurement devices.

[0016] According to another aspect of the present invention, a computer device is also provided, including a memory, a processor, and a computer program stored in the memory, wherein the processor executes the computer program to implement the steps of the semiconductor measurement device positioning error compensation method described above.

[0017] The method for compensating positioning errors in semiconductor metrology equipment of the present invention constructs a positioning error model to obtain the dynamic transmission relationship between multiple axis motion systems of the semiconductor metrology equipment, and obtains the mapping relationship between the time-series data of multiple motion parameters and the positioning error corresponding to the most recent input data. By continuously acquiring multiple motion parameters during the movement of the workpiece stage at intervals, inputting them into the positioning error model in real time, and using the positioning error output by the positioning error model to compensate for the motion of the workpiece stage, the method achieves real-time and dynamic compensation for coupling positioning errors caused by mechanical vibration, different workstation modes, etc. between multiple axis motion systems. This enables the semiconductor metrology equipment to accurately position and focus on the target position of the wafer, improves the positioning accuracy of the semiconductor metrology equipment, and provides a better metrology foundation for subsequent high-resolution imaging.

[0018] The above and other objects, advantages and features of the present invention will become more apparent to those skilled in the art from the following detailed description of specific embodiments of the invention in conjunction with the accompanying drawings. Attached Figure Description

[0019] The following sections will describe some specific embodiments of the invention in a detailed manner by way of example and not limitation, with reference to the accompanying drawings. The same reference numerals in the drawings denote the same or similar parts or portions. Those skilled in the art should understand that these drawings are not necessarily drawn to scale. In the drawings: Figure 1 This is a flowchart illustrating a method for compensating for positioning errors in a semiconductor measurement device according to an embodiment of the present invention. Figure 2 This is a schematic diagram of the process of obtaining training data using a compensation method according to an embodiment of the present invention; Figure 3 This is a schematic diagram of a computer program product according to an embodiment of the present invention; Figure 4 This is a schematic diagram of a computer-readable storage medium according to an embodiment of the present invention; Figure 5 This is a schematic diagram of a computer device according to an embodiment of the present invention. Detailed Implementation

[0020] The purpose of the semiconductor measurement equipment positioning error compensation method in this embodiment is to compensate for the multi-axis coupling error generated by the semiconductor measurement equipment during the movement of the workpiece stage, thereby improving the positioning accuracy of the workpiece stage.

[0021] Figure 1 This is a flowchart illustrating a method for compensating for positioning errors in a semiconductor measurement device according to an embodiment of the present invention. The method generally includes: During the process of moving the workpiece stage of the semiconductor measurement equipment to the target position, the following steps are performed: S100 continuously acquires multiple motion parameters of the workpiece stage at intervals; S200: Multiple motion parameters are input into the positioning error model. The positioning error model analyzes the multiple motion parameters input in each iteration and outputs the positioning error of the current position of the workpiece stage. The positioning error model is trained to learn the mapping relationship between the time series data of multiple motion parameters and the positioning error, and outputs the positioning error corresponding to the most recent input data based on the input data in each iteration. S300 performs motion compensation on the workpiece stage based on the positioning error output by the positioning error model until it moves to the target position.

[0022] In the semiconductor integrated circuit manufacturing process, after core processes such as exposure and etching, wafers need to be inspected using semiconductor metrology equipment (such as a Critical Dimension Scanning Electron Microscope, or CD-SEM) to detect critical dimensions (such as linewidth, hole depth, and spacing) on ​​the wafer surface. This allows for precise detection of defects generated during manufacturing, verification of process compliance, and assessment of chip performance and yield. As semiconductor manufacturing processes advance towards 7nm, 5nm, and even 3nm nodes, critical dimensions have shrunk to the nanometer scale, requiring semiconductor metrology equipment to achieve nanometer-level resolution imaging.

[0023] As the core component of semiconductor metrology equipment, the motion control platform needs to control the movement and positioning of the workpiece stage (i.e., the wafer stage) with extremely high precision to achieve high-precision positioning of wafer defect coordinates and ensure that the electron beam can be accurately focused on the target detection area, providing a basic guarantee for subsequent high-resolution imaging.

[0024] Motion control platforms typically include an XY-axis motion system (X / Y stage) and a Z-axis motion system (Z stage). The XY-axis motion system controls the movement of the stage in the XY plane, while the Z-axis motion system controls the movement of the stage along the Z-axis. Specifically, the XY-axis motion system is used to precisely position the electron beam at the target location on the wafer. The Z-axis, parallel to the optical axis of the scanning electron microscope's electron optics system, is used to control the movement of the stage along the Z-axis, enabling macro-focusing of the electron beam and ensuring precise focusing onto the target location on the wafer. The XY-axis and Z-axis motion systems work together to ensure the electron beam is precisely positioned and focused onto the target location on the wafer.

[0025] In other embodiments of this application, the motion control platform further includes one or more rotation systems to adjust the orientation angle of the wafer surface. The XY-axis motion system, Z-axis motion system, and rotation system work together to ensure that the electron beam can be precisely positioned and focused on the target location on the wafer.

[0026] When the motion control platform's various axis motion systems drive the workpiece stage, they transmit vibrations through mechanical structures, interfering with the motion and positioning accuracy of other axis motion systems. The modalities of each axis motion system at different workstations also contribute to workpiece stage positioning errors. For example, during macro-focusing, the Z-axis motion system transmits vibrations through its mechanical structure, interfering with the linear motion accuracy of the XY-axis motion system, causing positioning errors in the XY-plane. Furthermore, the modalities of the XY-axis and Z-axis motion systems at different workstations also result in positioning errors in the XY-plane. Similarly, when the XY-axis motion system controls the workpiece stage's movement in the XY-plane, it transmits vibrations through its mechanical structure, interfering with the linear motion accuracy of the Z-axis motion system, leading to Z-axis positioning errors.

[0027] To address the positioning errors caused by motion system coupling, the inventors proposed a dynamic error coupling modeling technique. Specifically, a positioning error model is first trained based on a deep learning algorithm. This model takes the time-series data of multiple motion parameters of the workpiece stage as input. By performing multiphysics coupling analysis on the input data, the dynamic transmission relationship between two or more axis motion systems is extracted and established, thus obtaining the error coupling effect between the two or more axis motion systems, and outputting the positioning error corresponding to the most recent time-series data.

[0028] Motion parameters may include multiple parameters such as the X-axis position coordinates, velocity, acceleration, and jerk of the workpiece stage; the Y-axis position coordinates, velocity, acceleration, and jerk; and the Z-axis position coordinates, velocity, acceleration, and jerk. These motion parameters can be acquired directly or indirectly through multiple sensors (such as linear encoders and encoders) of the motion control platform, and can be continuously acquired in real-time or at intervals to obtain time-series data of multiple motion parameters.

[0029] Localization error models can be built based on Recurrent Neural Networks (RNNs) or improved models of RNNs (such as Long Short-Term Memory Networks, LSTMs). An RNN is a neural network capable of processing sequential data. It achieves forward and backward information transfer through internal loop connections, possessing memory capabilities. This network uses shared parameters at each time step, enabling it to effectively process temporal data. RNNs typically contain input, hidden, and output layers, with recurrent connections in the hidden layers being a hallmark feature. Trained using the Backpropagation-Time-of-Temporal (BPTT) algorithm, RNNs can learn patterns in temporal data with multiple motion parameters, leading to variants such as bidirectional RNNs, LSTMs, and GRUs to address problems such as long sequence processing.

[0030] Specifically, in this embodiment, the positioning error model is trained to learn the mapping relationship between time-series data of multiple motion parameters and positioning errors, and outputs the positioning error corresponding to the most recent input data based on each input data. For example, the positioning error model can output the current positioning error of the workpiece stage in the XY plane and / or Z axis based on multiple sets of time-series data of motion parameters such as the X-axis position coordinates, velocity, and acceleration of the workpiece stage; the Y-axis position coordinates, velocity, and acceleration of the workpiece stage; and the Z-axis position coordinates, velocity, acceleration, and jerk.

[0031] After constructing the positioning error model, multiple motion parameters of the workpiece stage can be continuously acquired at intervals during the process of moving the workpiece stage from the starting position to the target position, thus obtaining time-series data of multiple motion parameters. The interval time can be, for example, 1s, 0.1s, 0.01s, etc. The shorter the interval time, the more timely the acquired time-series data. Those skilled in the art can reasonably set the length of the interval time based on the speed of the workpiece stage movement, positioning accuracy, etc.

[0032] Simultaneously, the acquired time-series data is continuously input into the positioning error model at intervals. Specifically, the positioning error model can be input immediately after acquiring a set of motion parameter data. Alternatively, it can be packaged and input into the positioning error model after acquiring a set number of sets of motion parameter data (e.g., two sets). The positioning error model analyzes the multiple motion parameters input in each iteration and outputs the positioning error of the workpiece stage corresponding to the most recent set of multiple motion parameters. When the interval is sufficiently short, the positioning error corresponding to the most recent set of multiple motion parameters is the positioning error of the current position of the workpiece stage. It should be understood that when the positioning error model is built based on an RNN, those skilled in the art can reasonably set the maximum sequence length of the time-series data based on the duration of the coupling effects between the motion systems of each axis (e.g., the duration of a single mechanical vibration).

[0033] Next, based on the positioning error output by the positioning error model each time, the workpiece stage is dynamically and in real time to compensate for the current positioning error of the workpiece stage until it moves to the target position.

[0034] The method for compensating positioning errors in semiconductor metrology equipment in this embodiment constructs a positioning error model to obtain the dynamic transmission relationship between multiple axis motion systems of the semiconductor metrology equipment, and obtains the mapping relationship between the time-series data of multiple motion parameters and the positioning error corresponding to the most recent input data. By continuously acquiring multiple motion parameters during the movement of the workpiece stage at intervals, inputting them into the positioning error model in real time, and using the positioning error output by the positioning error model each time to compensate for the motion of the workpiece stage, the coupling positioning errors between multiple axis motion systems caused by mechanical vibration, different workstation modes, etc., can be compensated in real time and dynamically. This enables the semiconductor metrology equipment to accurately position and focus on the target position of the wafer, improves the overall performance of the semiconductor metrology equipment, and provides a better metrology foundation for subsequent high-resolution imaging.

[0035] In some embodiments of the compensation method of the present invention, the semiconductor measurement device includes a scanning electron microscope, which includes an XY-axis motion system and a Z-axis motion system. The XY-axis motion system is used to control the movement of the workpiece stage in the XY plane, and the Z-axis motion system is used to control the movement of the workpiece stage in the Z-axis direction. The Z-axis is parallel to the optical axis of the electron optical system of the scanning electron microscope.

[0036] The positioning error model is used to output the positioning error of the workpiece stage in the XY plane when it is driven by the Z-axis motion system to move along the Z-axis direction. Motion compensation of the workpiece stage based on the positioning error refers to compensation in the XY plane.

[0037] In this embodiment, the Z-axis is parallel to the optical axis of the scanning electron microscope's electron optical system, while the XY plane is perpendicular to the optical axis of the electron optical system. That is, the XY-axis motion system is used to precisely position the electron beam at the target location on the wafer, and the Z-axis motion system is used to achieve macro-focusing of the electron beam, ensuring precise focusing of the electron beam at the target location on the wafer. When the Z-axis motion system controls the stage to move vertically along the Z-axis, mechanical coupling disturbances can affect the linear positioning accuracy of the XY-axis motion system, causing the electron beam to fail to accurately align with the target location on the wafer.

[0038] In this embodiment, the positioning error model focuses on the positioning error in the XY plane when the Z-axis motion system drives the workpiece stage to move along the Z-axis direction, thereby improving the positioning accuracy of the workpiece stage in the XY plane and enabling the electron beam to be accurately aligned with the target position of the wafer.

[0039] In some embodiments of the compensation method of the present invention, the motion parameters include X-axis and Y-axis position coordinates, Z-axis position coordinates, Z-axis velocity, and Z-axis acceleration.

[0040] The above motion parameters can fully reflect the motion state of the Z-axis motion system and the workpiece stage's position mode in the Z-axis and XY planes, which is sufficient to obtain the current positioning error of the workpiece stage in the XY plane with high accuracy based on the mapping relationship of the positioning error model.

[0041] In this embodiment, the positioning error model is also used to obtain the influence of the current mechanical vibration of the Z-axis motion system on the coupling error of the workpiece stage in the future, so as to more accurately compensate for the positioning error of the workpiece stage and improve the positioning accuracy of the workpiece stage in the XY plane.

[0042] In some embodiments of the compensation method of the present invention, the positioning error model is constructed based on a long short-term memory neural network.

[0043] LSTM is a special variant of RNN, featuring a gated structure. It uses the logic of gate units to determine whether data is updated or discarded, overcoming the drawbacks of RNNs such as excessive weight influence and the tendency for gradient vanishing or exploding. This allows the network to converge better and faster, effectively improving prediction accuracy. LSTM has three gates: the input gate, the forget gate, and the output gate, which determine the information to be remembered and forgotten at each time step. The input gate determines the intake of new information; the forget gate controls which information from the previous time step is retained in the current time step; and the output gate determines the output at the current time step.

[0044] Specifically, in this embodiment, LSTM can acquire and analyze the time-series data of all motion parameters of the workpiece stage from the starting position to the current position (the position corresponding to the most recently input multiple motion parameters), thereby simultaneously capturing the characteristics of mechanical vibrations that cause long-term and short-term effects, and thus more accurately predicting the positioning error of the workpiece stage and improving the positioning accuracy of the workpiece stage.

[0045] In some embodiments of the compensation method of the present invention, such as Figure 2 As shown, the positioning error model is trained using training data. The steps for obtaining the training data include: S211, using the XY axis motion system, moves the workpiece stage to multiple set positions in the XY plane; S213, at each set position, the Z-axis motion system drives the workpiece stage to move within the set stroke range of the Z-axis, and continuously acquires multiple motion parameters at intervals, as well as continuously acquires the measured position coordinates of the workpiece stage on the XY plane measured by the laser measuring instrument at intervals.

[0046] For example, the travel range of the workpiece stage in the XY plane can be divided into a grid, with a grid spacing of, for example, 10 mm. When acquiring training data, the XY-axis motion system is used to move the workpiece stage to each grid node in the XY plane in 10 mm increments. At each grid node, the Z-axis motion system is used to drive the workpiece stage from a low position to a high position on the Z-axis, and multiple motion parameters are continuously acquired at intervals during the ascent.

[0047] Laser measuring instruments can measure the coordinates of a position with high accuracy.

[0048] By comparing the XY plane coordinates obtained from sensors with the measured position coordinates obtained from a laser measuring instrument, the actual positioning error of the workpiece stage in the XY plane can be accurately obtained. During the training process, a loss function can be set based on the difference between the predicted positioning error and the actual positioning error of the positioning error model, prompting the positioning error model to output a predicted positioning error consistent with the actual positioning error.

[0049] In some embodiments of the compensation method of the present invention, the laser measuring instrument measures the actual position coordinates under vacuum conditions.

[0050] Current positioning error tests are typically conducted under normal pressure, while in practical applications, the motion systems of each axis and the workpiece stage need to operate in a vacuum environment. Compared to normal pressure, a vacuum environment induces micro-deformation effects in the structural materials (such as deformation of the tracks of each axis), resulting in significant differences between the actual positioning error characteristics and the results of normal pressure tests.

[0051] In this embodiment, by acquiring training data under vacuum conditions, the measured position coordinate data accurately reflects the micro-deformation effects of each axis motion system and the workpiece stage under actual working conditions, thereby ensuring that the positioning error output by the trained positioning error model is consistent with that under vacuum conditions, thus improving the accuracy of the positioning error model.

[0052] The flowchart provided in this embodiment is not intended to indicate that the operations of the method will be performed in any particular order, or that all operations of the method are included in every case. Furthermore, the method may include additional operations. Within the scope of the technical concept provided by the method in this embodiment, additional variations can be made to the above method.

[0053] It should be understood that in some embodiments, the components may be implemented using hardware, software, firmware, or a combination thereof. In the above embodiments, multiple steps or methods may be implemented using software or firmware stored in memory and executed by a suitable instruction execution system.

[0054] This invention also provides a computer program product 10, a computer-readable storage medium 20, and a computer device 30. Figure 3This is a schematic diagram of a computer program product 10 according to an embodiment of the present invention. Figure 4 This is a schematic diagram of a computer-readable storage medium 20 according to an embodiment of the present invention. Figure 5 This is a schematic diagram of a computer device 30 according to an embodiment of the present invention. The computer program product 10 includes a computer program 11, which, when executed by the processor 32, implements the steps of the method for compensating positioning errors of the semiconductor measurement device described above. A computer-readable storage medium 20 stores the computer program 11 thereon, which, when executed by the processor 32, implements the steps of the method for compensating positioning errors of the semiconductor measurement device described above. The computer device 30 may include a memory 31, a processor 32, and the computer program 11 stored in the memory 31 and running on the processor 32.

[0055] The computer program 11 used to perform the operations of this invention may be assembly instructions, Instruction Set Architecture (ISA) instructions, machine instructions, machine-dependent instructions, microcode, firmware instructions, state setting data, integrated circuit configuration data, or source code or object code written in any combination of one or more programming languages ​​and procedural programming languages. The computer program 11 may execute entirely on the user's computer, partially on the user's computer, as a standalone software package, partially on the user's computer and partially on a remote computer, or entirely on a remote computer or server. In the latter case, the remote computer may be connected to the user's computer via any type of network, including a Local Area Network (LAN) or Wide Area Network (WAN), or may be connected to an external computer (e.g., via the Internet using an Internet service provider). In some embodiments, to perform aspects of this invention, electronic circuits, including, for example, programmable logic circuits, Field-Programmable Gate Arrays (FPGAs), or Programmable Logic Arrays (PLAs), may execute computer-readable program instructions to personalize the electronic circuits by utilizing state information from computer-readable program instructions.

[0056] For the purposes of this embodiment, computer program product 10 is a related product that includes computer program 11.

[0057] For the purposes of this embodiment, the computer-readable storage medium 20 is a tangible device capable of holding and storing a computer program 11. It can be any device capable of containing, storing, communicating, propagating, or transmitting the computer program 11 for use by or in conjunction with an instruction execution system, apparatus, or device. More specific examples (a non-exhaustive list) of the computer-readable storage medium 20 include: portable computer disks, hard disks, random access memory (RAM), read-only memory (ROM), erasable programmable read-only memory (EPROM or flash memory), static random access memory (SRAM), portable optical disc read-only memory (CD-ROM), digital versatile disc (DVD), memory stick, floppy disk, mechanical encoding device, and any suitable combination thereof.

[0058] Computer device 30 can be, for example, a server, desktop computer, laptop computer, tablet computer, or smartphone. In some examples, computer device 30 can be a cloud computing node. Computer device 30 can be described in the general context of computer system executable instructions (such as program modules) executed by a computer system. Typically, program modules can include routines, programs, object programs, components, logic, data structures, etc., that perform specific tasks or implement specific abstract data types. Computer device 30 can be implemented in a distributed cloud computing environment where tasks are performed by remote processing devices linked through a communication network. In a distributed cloud computing environment, program modules can reside on local or remote computing system storage media, including storage devices.

[0059] Computer device 30 may include a processor 32 adapted to execute stored instructions and a memory 31 that provides temporary storage space for the operation of said instructions during operation. The processor 32 may be a single-core processor, a multi-core processor, a computing cluster, or any other configuration. The memory 31 may include random access memory (RAM), read-only memory, flash memory, or any other suitable storage system.

[0060] Computer device 30 may also include a network adapter / interface and an input / output (I / O) interface. The I / O interface allows external devices that can be connected to the computer device to input and output data. The network adapter / interface provides communication between the computer device and a network, typically represented as a communication network.

[0061] Therefore, those skilled in the art should recognize that although numerous exemplary embodiments of the present invention have been shown and described in detail herein, many other variations or modifications conforming to the principles of the present invention can be directly determined or derived from the disclosure of the present invention without departing from the spirit and scope of the invention. Thus, the scope of the present invention should be understood and construed as covering all such other variations or modifications.

Claims

1. A method for compensating for positioning errors in a semiconductor measurement device, characterized in that, During the process of moving the workpiece stage of the semiconductor measurement equipment to the target position, the following steps are performed: Multiple motion parameters of the workpiece stage are continuously acquired at intervals; The multiple motion parameters are input into the positioning error model. The positioning error model analyzes the multiple motion parameters input in each iteration and outputs the positioning error of the current position of the workpiece stage. The positioning error model is trained to learn the mapping relationship between the time-series data of the multiple motion parameters and the positioning error, and outputs the positioning error corresponding to the most recent input data based on the input data in each iteration. Based on the positioning error output by the positioning error model, motion compensation is performed on the workpiece stage until it moves to the target position.

2. The compensation method according to claim 1, characterized in that, The semiconductor measurement device includes a scanning electron microscope, which includes an XY-axis motion system and a Z-axis motion system. The XY-axis motion system is used to control the movement of the workpiece stage in the XY plane, and the Z-axis motion system is used to control the movement of the workpiece stage in the Z-axis direction. The Z-axis is parallel to the optical axis of the electron optical system of the scanning electron microscope. The positioning error model is used to output the positioning error of the workpiece stage in the XY plane when it is driven by the Z-axis motion system to move along the Z-axis direction. The motion compensation of the workpiece stage based on the positioning error refers to compensation in the XY plane.

3. The compensation method according to claim 2, characterized in that, The motion parameters include X-axis and Y-axis positions, Z-axis position, Z-axis velocity, and Z-axis acceleration.

4. The compensation method according to claim 2, characterized in that, The positioning error model is constructed based on a recurrent neural network.

5. The compensation method according to claim 2, characterized in that, The positioning error model is constructed based on a long short-term memory neural network.

6. The compensation method according to claim 4 or 5, characterized in that, The positioning error model is trained using training data, and the steps for obtaining the training data include: Using the XY axis motion system, the workpiece stage is moved to multiple set positions in the XY plane; At each of the set positions, the Z-axis motion system is used to drive the workpiece stage to move within a set stroke range of the Z-axis, and the plurality of motion parameters are continuously acquired at intervals, as well as the measured position coordinates of the workpiece stage on the XY plane measured by the laser measuring instrument are continuously acquired at intervals.

7. The compensation method according to claim 6, characterized in that, The laser measuring instrument obtains the actual position coordinates under vacuum conditions.

8. A computer-readable storage medium, characterized in that, It stores a computer program, which, when executed by a processor, implements the steps of the method for compensating for positioning errors of a semiconductor measurement device as described in any one of claims 1 to 7.

9. A computer program product, comprising a computer program, characterized in that, When the computer program is executed by the processor, it implements the steps of the method for compensating for positioning errors of a semiconductor measurement device as described in any one of claims 1 to 7.

10. A computer device, characterized in that, The device includes a memory, a processor, and a computer program stored in the memory, wherein the processor executes the computer program to implement the steps of the method for compensating for positioning errors of a semiconductor measurement device according to any one of claims 1 to 7.

Images