Laser ruler ranging, collecting and analyzing system adaptive to lithography machine double workpiece table

Abstract

The application belongs to the technical field of lithography machine operation control, and specifically relates to a laser ruler ranging collection and analysis system suitable for a lithography machine double workpiece table, which comprises a double-frequency laser interferometer array module, a data collection and preprocessing module, a multi-source error decoupling analysis module, a synchronous control feedback module and an operation supervision terminal; the double-frequency laser interferometer array module is used to comprehensively measure the double workpiece table in six degrees of freedom, the data collection and preprocessing module is used for data collection and preprocessing, the multi-source error decoupling analysis module is used to output the decoupled net error, the synchronous control feedback module is used to generate a compensation signal according to the net error and to close-loop control a piezoelectric ceramic driver to adjust the position of the workpiece table, and the environment adaptive compensation module is used to correct the ranging error caused by laser wavelength drift, so that the real-time collection, decoupling analysis and closed-loop feedback of the double workpiece table displacement data are realized, the dynamic alignment accuracy of the lithography machine is significantly improved, the production control difficulty is reduced, and the automation and intelligent level are high.

Description

Technical Field

[0001] This invention relates to the field of lithography machine operation and control technology, specifically a laser ruler ranging acquisition and analysis system adapted to a dual-stage lithography machine. Background Technology

[0002] Photolithography machines are core equipment in semiconductor manufacturing. They use projection optical systems to precisely transfer circuit patterns onto silicon wafers. Their accuracy directly affects chip manufacturing process nodes. Modern high-end photolithography machines adopt a dual-platform architecture, with one stage responsible for silicon wafer exposure and the other simultaneously performing alignment and measurement, achieving parallel "exposure-measurement" operations and significantly improving production efficiency.

[0003] However, the high-speed and high-acceleration motion of the dual workpiece stage of the lithography machine places higher demands on the displacement measurement system. Traditional laser interferometers and grating rulers have gradually revealed their limitations in terms of dynamic response and error compensation. In particular, when the workpiece stage structure deforms or the gas film gap changes, it is difficult to ensure measurement accuracy in real time. It is impossible to realize real-time acquisition, decoupling analysis and closed-loop feedback of displacement data of the dual workpiece stage of the lithography machine. This is not conducive to improving the dynamic alignment accuracy of the lithography machine and reducing the difficulty of production control. The level of intelligence and automation is low.

[0004] To address the aforementioned technical shortcomings, a solution is proposed. Summary of the Invention

[0005] The purpose of this invention is to provide a laser ruler ranging acquisition and analysis system adapted to the dual-stage of a lithography machine. This system solves the problems of existing technologies, such as difficulty in ensuring measurement accuracy in real time, inability to achieve real-time acquisition, decoupling analysis and closed-loop feedback of displacement data of the dual-stage of the lithography machine, which is not conducive to improving the dynamic alignment accuracy of the lithography machine and reducing the difficulty of production control, and low level of intelligence and automation.

[0006] To achieve the above objectives, the present invention provides the following technical solution:

[0007] A laser ruler ranging acquisition and analysis system adapted to a dual-stage lithography machine includes a dual-frequency laser interferometer array module, a data acquisition and preprocessing module, a multi-source error decoupling analysis module, a synchronous control feedback module, and an operation monitoring terminal.

[0008] The dual-frequency laser interferometer array module deploys several sets of dual-frequency laser interferometers to form a comprehensive measurement of the dual workpiece stage in six degrees of freedom, and transmits the measured raw displacement data to the data acquisition and preprocessing module through a low-noise differential signal line;

[0009] The data acquisition and preprocessing module is used to acquire and preprocess data, and inputs the preprocessed data into the multi-source error decoupling analysis module;

[0010] The multi-source error decoupling analysis module uses a Kalman filter to fuse laser ranging data and accelerometer data, estimate the optimal state vector of the system in real time, and correct nonlinear errors by combining a pre-stored finite element analysis compensation table, and outputs the decoupled net error to the synchronous control feedback module.

[0011] The synchronous control feedback module generates a compensation signal based on the net error, uses closed-loop control to adjust the position of the piezoelectric ceramic actuator, and sends the adjustment information to the operation monitoring terminal.

[0012] Furthermore, each dual-frequency laser interferometer in the dual-frequency laser interferometer array module adopts an orthogonal polarization optical path design for the reference light and the measurement light to eliminate optical path refraction error, and the measurement repeatability accuracy is less than or equal to 0.1nm;

[0013] Furthermore, by using MEMS micromirrors to dynamically adjust the position of the laser spot and compensate for the optical path offset caused by the tilt of the workpiece stage in real time, the laser spot is always precisely focused on the center of the reflector.

[0014] Furthermore, the multi-source error decoupling analysis module is connected to the environment adaptive compensation module, which monitors environmental parameters in real time and corrects the ranging error caused by laser wavelength drift.

[0015] Furthermore, the environmental adaptive compensation module deploys a distributed temperature and humidity sensor network, with eight sensor nodes deployed around the dual workpiece stage for environmental monitoring, and data is aggregated via the I2C bus.

[0016] Furthermore, a long short-term memory network is used to fit the nonlinear relationship between environmental parameters and ranging error, dynamically correcting the laser wavelength. The corrected environmental parameters are then input into the multi-source error decoupling analysis module as boundary condition optimization error model.

[0017] Furthermore, the data acquisition and preprocessing module is based on the Xilinx Zynq-7000 series FPGA, realizing eight-channel synchronous sampling, a sampling rate of 10MHz, and an ADC resolution of 24 bits;

[0018] Furthermore, the wavelet threshold denoising algorithm is applied to perform db4 wavelet transform on the sampled data, and a threshold is set to filter out high-frequency vibration noise, thereby improving the signal-to-noise ratio by 40dB.

[0019] Furthermore, the synchronous control feedback module is designed with a feedforward-feedback hybrid control algorithm to shorten the response delay, and adopts predictive control theory to pre-compensate for acceleration abrupt changes during the exchange process by pre-storing the workpiece stage motion trajectory database, thereby reducing trajectory tracking errors.

[0020] Furthermore, the operation monitoring terminal communicates with the application effect evaluation module. The application effect evaluation module is used to set a production period of duration P1, evaluate the system's performance during the production period, and generate an excellent application effect signal or an application effect alarm signal accordingly. The excellent application effect signal or the application effect alarm signal is sent to the operation monitoring terminal. When the operation monitoring terminal receives the application effect alarm signal, it issues a corresponding warning.

[0021] Furthermore, the specific operation process of the application effect evaluation module is as follows:

[0022] The average hourly silicon wafer output of the lithography machine after using the system during the production period is statistically analyzed and compared with the average hourly silicon wafer output before using the system. Based on this, the yield improvement ratio is collected. The yield improvement ratio is compared with the preset yield improvement ratio threshold. If the yield improvement ratio does not exceed the preset yield improvement ratio threshold, an application effect alarm signal is generated.

[0023] If the yield increase ratio exceeds the preset yield increase ratio threshold, all quality inspection information of the products produced by the lithography machine during the production period is obtained, and products that fail the quality inspection are marked as defective objects. The number of defective objects during the production period is obtained and its ratio is calculated with the total number of products produced during the production period to obtain the defect ratio value. The defect ratio value is compared with the preset defect ratio threshold. If the defect ratio value exceeds the preset defect ratio threshold, an application effect alarm signal is generated. If the defect ratio value does not exceed the preset defect ratio threshold, an application effect excellent signal is generated.

[0024] Furthermore, the application effect evaluation module communicates with the test and verification decision module. The application effect evaluation module sends the application effect alarm signal to the test and verification decision module. When the test and verification decision module receives the application effect alarm signal, it tests the system performance, analyzes the test data to verify the system's operating performance, and generates a performance verification qualified signal or a performance verification abnormal signal. The performance verification qualified signal or performance verification abnormal signal is then sent to the operation monitoring terminal. When the operation monitoring terminal receives the performance verification abnormal signal, it issues a corresponding warning.

[0025] The test items involved in the test verification decision module include measurement accuracy test, dynamic response speed test and environmental adaptability test. Among them, in the measurement accuracy test, a high-precision standard is used to compare with the system output data to collect its synchronization error information, and the repeatability of the system to measure known displacement under static conditions is tested to collect its standard deviation information.

[0026] In the dynamic response speed test, the sudden acceleration of the workpiece stage was simulated, and the delay time from the detection of the error to the output of the compensation signal was tested. In the high-speed scanning mode, the tracking error of the sine trajectory of the test system was tested. In the environmental adaptability test, the distance measurement error drift of the test system was tested under the conditions of temperature and humidity fluctuation and air pressure change.

[0027] Furthermore, the specific operation process of the test and verification decision module is as follows:

[0028] All test data from the testing process are acquired and categorized. The corresponding test data is compared with the corresponding preset data requirements. If the corresponding test data does not meet the corresponding preset data requirements, a verification symbol WY-1 is assigned.

[0029] The test anomaly value is obtained by calculating the ratio of the number of times the corresponding type of data is assigned the verification symbol WY-1 and the total number of the corresponding type of data. The test anomaly value is compared with the corresponding preset test anomaly threshold. If the test anomaly value exceeds the corresponding preset test anomaly threshold, the corresponding type of data is marked as a non-standard type.

[0030] If the test anomaly value does not exceed the corresponding preset test anomaly threshold, when assigning the verification symbol WY-1, the deviation value of the corresponding test data from the corresponding preset data requirement is marked as the numerical deviation value, the average value of all numerical deviation values ​​of the corresponding type of data is calculated to obtain the numerical risk performance value, and the numerical deviation value with the largest value corresponding to the corresponding type of data is marked as the numerical risk amplitude value.

[0031] The test evaluation value is obtained by weighted summation of test anomaly values, risk performance values, and risk amplitude values. The test evaluation value is then compared with the corresponding preset test evaluation threshold. If the test evaluation value exceeds the corresponding preset test evaluation threshold, the corresponding data type is marked as a non-standard type. If a non-standard type exists, a performance verification anomaly signal is generated.

[0032] If there are no non-standard types, then set a set of preset performance impact weight values ​​for each type of data, calculate the ratio of the numerical evaluation value of the corresponding type of data to the corresponding preset numerical evaluation threshold to obtain the test analysis value, multiply the test analysis value by the corresponding preset performance impact weight value, and mark the product result as the test feature value.

[0033] The verification decision value is obtained by summing the test feature values ​​of all types of data. The verification decision value is then compared with the preset verification decision threshold. If the verification decision value exceeds the preset verification decision threshold, a performance verification anomaly signal is generated; if the verification decision value does not exceed the preset verification decision threshold, a performance verification pass signal is generated.

[0034] Compared with the prior art, the beneficial effects of the present invention are:

[0035] 1. In this invention, a dual-frequency laser interferometer array module is used to perform comprehensive measurement of the two workpiece stages in six degrees of freedom. A data acquisition and preprocessing module performs data acquisition and preprocessing. A multi-source error decoupling analysis module outputs the net error after decoupling. A compensation signal is generated based on the net error and a closed-loop control is used to adjust the position of the workpiece stage by the piezoelectric ceramic actuator. An environmental adaptive compensation module corrects the ranging error caused by laser wavelength drift. This enables real-time acquisition, decoupling analysis, and closed-loop feedback of the displacement data of the two workpiece stages, significantly improving the dynamic alignment accuracy of the lithography machine and reducing the difficulty of production management. The level of automation and intelligence is high.

[0036] 2. In this invention, the application effect evaluation module evaluates the system's performance during the production period. When an application effect alarm signal is generated, the cause is investigated, and the system performance is tested through the test verification decision module. The test data is analyzed to verify the system's operating performance, so as to remind managers to take timely system optimization and other improvement measures to ensure subsequent productivity and production quality and reduce enterprise production losses. Attached Figure Description

[0037] To facilitate understanding by those skilled in the art, the present invention will be further described below with reference to the accompanying drawings;

[0038] Figure 1 This is a system block diagram of Embodiment 1 of the present invention;

[0039] Figure 2 This is a system block diagram of Embodiments 2 and 3 of the present invention. Detailed Implementation

[0040] The technical solutions of the embodiments of the present invention will be clearly and completely described below with reference to the accompanying drawings. Obviously, the described embodiments are only some embodiments of the present invention, and not all embodiments. Based on the embodiments of the present invention, all other embodiments obtained by those skilled in the art without creative effort are within the scope of protection of the present invention.

[0041] Example 1: As Figure 1 As shown, the laser ruler ranging acquisition and analysis system adapted to the dual workpiece stage of a lithography machine proposed in this invention includes a dual-frequency laser interferometer array module, a data acquisition and preprocessing module, a multi-source error decoupling analysis module, an environmental adaptive compensation module, a synchronous control feedback module, and an operation monitoring terminal.

[0042] The dual-frequency laser interferometer array module, acting as the system's "sensing organ," measures the displacement data of the dual workpiece stage in six degrees of freedom (X / Y / Z / θx / θy / θz) in real time using the principle of laser interferometry. By deploying several sets (preferably eight sets) of dual-frequency laser interferometers, a comprehensive measurement of the dual workpiece stage in six degrees of freedom (i.e., X / Y / Z / θx / θy / θz directions, where θ represents rotation) is achieved. The raw displacement data obtained from the measurement is then transmitted to the data acquisition and preprocessing module via a low-noise differential signal line.

[0043] It should be noted that each dual-frequency laser interferometer in the dual-frequency laser interferometer array module adopts an orthogonal polarization optical path design for the reference light and the measurement light (i.e., two orthogonally polarized laser beams, the reference light and the measurement light, are used, and the reflected light is separated by a polarization beam splitter), which eliminates optical path refraction error, improves measurement stability, and achieves a measurement repeatability accuracy of less than or equal to 0.1 nm.

[0044] That is, a dual-frequency laser (e.g., f1=193.4nm, f2=193.5nm) is emitted by a laser, which is split into a reference beam and a measurement beam by a polarizing beam splitter. The measurement beam returns through a reflector on the workpiece stage and interferes with the reference beam to generate a beat frequency signal. The displacement is calculated by a counter.

[0045] Furthermore, by using MEMS (Micro-Electro-Mechanical Systems) micromirrors to dynamically adjust the position of the laser spot, the optical path offset caused by the tilt of the workpiece stage (such as in the θx / θy direction) is compensated in real time, thereby ensuring that the laser spot is always accurately focused on the center of the reflector.

[0046] The data acquisition and preprocessing module is used for data acquisition and preprocessing, and inputs the preprocessed data into the multi-source error decoupling analysis module. It should be noted that the data acquisition and preprocessing module is based on Xilinx Zynq-7000 series FPGA, realizing eight-channel synchronous sampling, sampling rate of 10MHz, ADC resolution of 24-bit, ensuring that the data is not distorted when the dual workpiece stage moves at high speed (e.g., 1m / s).

[0047] Furthermore, the wavelet threshold denoising algorithm is applied to perform db4 wavelet transform on the sampled data, and a threshold is set to filter out high-frequency vibration noise (such as signals above 100Hz from ground vibration), while retaining effective displacement information, thereby improving the signal-to-noise ratio by 40dB.

[0048] The multi-source error decoupling analysis module uses a Kalman filter to fuse laser ranging data and accelerometer data, estimate the optimal state vector of the system in real time, and correct nonlinear errors by combining a pre-stored finite element analysis compensation table. The decoupled net error is then output to the synchronous control feedback module.

[0049] The synchronous control feedback module generates a compensation signal based on the net error, uses closed-loop control to adjust the position of the piezoelectric ceramic actuator, and sends the adjustment information to the operation monitoring terminal.

[0050] It should be noted that the synchronous control feedback module uses a feedforward-feedback hybrid control algorithm to shorten the response delay. By combining feedforward control (based on trajectory prediction) and feedback control (based on real-time error), the response delay is reduced from 100μs in pure feedback to 50μs. Furthermore, predictive control theory is adopted to pre-compensate for acceleration abrupt changes during the exchange process (such as acceleration step when switching from measurement mode to exposure mode) by pre-storing the workpiece stage motion trajectory database, thereby reducing trajectory tracking errors.

[0051] The environmental adaptive compensation module monitors environmental parameters (temperature, humidity, air pressure) in real time and corrects the ranging error caused by laser wavelength drift. It deploys a distributed temperature and humidity sensor network, with eight sensor nodes (including eight SHT31 temperature and humidity sensors with a spatial resolution of 100mm) around the dual workpiece stage for environmental monitoring and data aggregation via I2C bus.

[0052] Furthermore, a long short-term memory network is used to fit the nonlinear relationship between environmental parameters and ranging error, dynamically correcting the laser wavelength (e.g., the wavelength drifts by about 1 pm for every 1°C increase in temperature), and inputting the corrected environmental parameters into the multi-source error decoupling analysis module as boundary condition optimization error model.

[0053] This invention combines a dual-frequency laser interferometer with an error decoupling algorithm, enabling a dual-stage synchronization error of ≤2nm, supporting mass production of processes below 7nm. Feedforward control reduces trajectory tracking delay by 60%, adapting to high-speed scanning requirements and improving yield. Adaptive compensation reduces ranging errors caused by temperature drift. Modular design enables real-time acquisition, decoupling analysis, and closed-loop feedback of dual-stage displacement data, achieving a complete link from raw data acquisition to closed-loop control. This significantly improves the dynamic alignment accuracy of the lithography machine, resulting in high system integration, reduced maintenance time and lithography machine downtime costs, and a high level of automation and intelligence.

[0054] Example 2: Figure 2 As shown, the difference between this embodiment and embodiment one is that the operation monitoring terminal is connected to the application effect evaluation module. The application effect evaluation module is used to set a production period of duration P1, evaluate the system application effect during the production period, generate an excellent application effect signal or an application effect alarm signal, and send the excellent application effect signal or the application effect alarm signal to the operation monitoring terminal.

[0055] When the monitoring terminal receives an application effect alarm signal, it issues a corresponding warning to remind managers to investigate the cause and take appropriate improvement measures to ensure subsequent productivity and product quality, thereby reducing production losses for the enterprise. The specific operation process of the application effect evaluation module is as follows:

[0056] The average hourly silicon wafer output of the lithography machine after using the system during the production period is statistically analyzed and compared with the average hourly silicon wafer output before using the system. Based on this, the yield improvement ratio is collected. The yield improvement ratio is compared with the preset yield improvement ratio threshold. If the yield improvement ratio does not exceed the preset yield improvement ratio threshold, it indicates that the yield performance during the production period is not good and the overall system application effect needs to be improved. Then, an application effect alarm signal is generated.

[0057] If the yield increase ratio exceeds the preset yield increase ratio threshold, it indicates that the yield performance during the production period is good. Then, all quality inspection information of the products produced by the lithography machine during the production period is obtained, and the products that fail the quality inspection are marked as defective objects. The number of defective objects during the production period is obtained and its ratio with the total number of products produced during the production period is calculated to obtain the defective ratio value.

[0058] The output difference ratio is compared with the preset output difference ratio threshold. If the output difference ratio exceeds the preset output difference ratio threshold, it indicates that the quality of the products produced during the production period is poor and the overall system application effect is not good. In this case, an application effect alarm signal is generated. If the output difference ratio does not exceed the preset output difference ratio threshold, it indicates that the quality of the products produced during the production period is good and the overall system application effect is good. In this case, an application effect excellent signal is generated.

[0059] Example 3: Figure 2 As shown, the difference between this embodiment and Embodiment 1 and Embodiment 2 is that the application effect evaluation module is connected to the test verification decision module. The application effect evaluation module sends the application effect alarm signal to the test verification decision module. When the test verification decision module receives the application effect alarm signal, it tests the system performance, analyzes the test data to verify the system's operating performance, and generates a performance verification qualified signal or a performance verification abnormal signal accordingly.

[0060] Furthermore, the system sends either a pass signal or an error signal to the operation monitoring terminal. When the operation monitoring terminal receives an error signal, it issues a corresponding warning to remind administrators to optimize the system in a timely manner, improve system performance, and thus ensure the subsequent system operation effect.

[0061] The test verification decision module includes test items such as measurement accuracy test, dynamic response speed test, and environmental adaptability test, and performs several tests for each test item. In the measurement accuracy test, a high-precision standard is used to compare the system output data to collect its synchronization error information (preferably, to verify whether the synchronization error of the dual workpiece stage is ≤2nm), and the system's measurement repeatability for known displacements (such as 10μm step displacement) is tested under static conditions to collect its standard deviation information (preferably, the standard deviation is required to be ≤0.5nm).

[0062] In the dynamic response speed test, the sudden acceleration of the workpiece stage is simulated (e.g., from 0 m / s² to 10 m / s²). The delay time from the detection of the error to the output of the compensation signal of the test system is measured (preferably, verified to be ≤50 μs). In the high-speed scanning mode (1 m / s), the tracking error of the test system for the sinusoidal trajectory (frequency 50 Hz) is measured (preferably, the amplitude error is required to be ≤5 nm). In the environmental adaptability test, the distance measurement error drift of the test system is measured under the conditions of temperature and humidity fluctuation (±2℃ / ±10%RH) and air pressure change (±50 Pa) (preferably, verified to be ≤1 nm / h).

[0063] Specifically, the test verification decision module operates as follows: it acquires and classifies all test data during the test process (including types such as synchronization error, repeatability standard deviation, and delay time), compares the corresponding test data with the corresponding preset data requirements, and assigns the verification symbol WY-1 if the corresponding test data does not meet the corresponding preset data requirements.

[0064] The test anomaly value is obtained by calculating the ratio of the number of times the corresponding type of data is assigned the verification symbol WY-1 and the total number of the corresponding type of data. The test anomaly value is compared with the corresponding preset test anomaly threshold. If the test anomaly value exceeds the corresponding preset test anomaly threshold, it indicates that the data condition of the corresponding type of data is abnormal and the corresponding related system performance is poor. Then the corresponding type of data is marked as a non-standard type.

[0065] If the test anomaly value does not exceed the corresponding preset test anomaly threshold, when assigning the verification symbol WY-1, the deviation value of the corresponding test data from the corresponding preset data requirement is marked as the numerical deviation value, the average value of all numerical deviation values ​​of the corresponding type of data is calculated to obtain the numerical risk performance value, and the numerical deviation value with the largest value corresponding to the corresponding type of data is marked as the numerical risk amplitude value.

[0066] The data verification evaluation value is obtained by weighted summation of test anomaly values, data risk performance values, and data risk amplitude values. Specifically, each of the test anomaly values, data risk performance values, and data risk amplitude values ​​is assigned a corresponding preset weight coefficient, and then each of these values ​​is multiplied by its respective preset weight coefficient. The sum of these three products is then marked as the data verification evaluation value. Furthermore, the larger the data verification evaluation value, the more abnormal the data condition of the corresponding data type, and the worse the performance of the corresponding related system.

[0067] The numerical evaluation value is compared with the corresponding preset numerical evaluation threshold. If the numerical evaluation value exceeds the corresponding preset numerical evaluation threshold, it indicates that the data condition of the corresponding data type is abnormal and the corresponding system performance is poor. Then, the corresponding data type is marked as a non-standard type. If there is a non-standard type, it indicates that the system performance is initially poor. Then, a performance verification abnormality signal is generated.

[0068] If there are no non-standard types, then each type of data is assigned a set of preset performance influence weight values ​​with values ​​greater than zero. The higher the importance of the corresponding type of data, the larger the value of the preset performance influence weight value. The test analysis value is obtained by calculating the ratio between the numerical evaluation value of the corresponding type of data and the corresponding preset numerical evaluation threshold. The test analysis value is multiplied by the corresponding preset performance influence weight value, and the product result is marked as the test feature value.

[0069] The verification decision value is obtained by summing the test feature values ​​of all types of data. The verification decision value is then compared with the preset verification decision threshold. If the verification decision value exceeds the preset verification decision threshold, it indicates that the overall system performance is poor, and a performance verification anomaly signal is generated. If the verification decision value does not exceed the preset verification decision threshold, it indicates that the overall system performance is good, and a performance verification pass signal is generated.

[0070] The working principle of this invention is as follows: In use, a dual-frequency laser interferometer array module forms a comprehensive measurement of the dual workpiece stages in six degrees of freedom. A data acquisition and preprocessing module performs data acquisition and preprocessing. A multi-source error decoupling analysis module estimates the optimal state vector of the system in real time and corrects nonlinear errors. The decoupled net error is output to the synchronous control feedback module. The synchronous control feedback module generates a compensation signal based on the net error. A closed-loop control piezoelectric ceramic actuator adjusts the position of the workpiece stage. An environmental adaptive compensation module monitors environmental parameters in real time and corrects the ranging error caused by laser wavelength drift. Based on a modular design, real-time acquisition, decoupling analysis, and closed-loop feedback of the dual workpiece stage displacement data are realized, significantly improving the dynamic alignment accuracy of the lithography machine. The system has a high degree of integration, which helps to shorten maintenance time and reduce the downtime cost of the lithography machine. It also has a high level of automation and intelligence.

[0071] The preferred embodiments of the present invention disclosed above are merely illustrative of the invention. The threshold values ​​in the technical solutions are for result comparison and analysis to determine whether a result is good or bad. The value of the threshold is determined by a combination of large-scale model analysis of the sample data and human experience, and can also be appropriately adjusted based on seasonal or common-sense influencing factors. The preferred embodiments do not exhaustively describe all details, nor do they limit the invention to specific implementation methods. Clearly, many modifications and variations can be made based on the content of this specification. This specification selects and specifically describes these embodiments to better explain the principles and practical applications of the invention, thereby enabling those skilled in the art to better understand and utilize the invention. The invention is limited only by the claims and their full scope and equivalents.

Claims

1. A laser ruler ranging and acquisition analysis system adapted to a dual-stage lithography machine, characterized in that, It includes a dual-frequency laser interferometer array module, a data acquisition and preprocessing module, a multi-source error decoupling analysis module, a synchronous control feedback module, and an operation monitoring terminal. The dual-frequency laser interferometer array module deploys several sets of dual-frequency laser interferometers to form a comprehensive measurement of the two workpiece stages in six degrees of freedom, and transmits the measured raw displacement data to the data acquisition and preprocessing module through a low-noise differential signal line. The data acquisition and preprocessing module is used to acquire and preprocess data, and inputs the preprocessed data into the multi-source error decoupling analysis module; The multi-source error decoupling analysis module estimates the optimal state vector of the system in real time through a Kalman filter, and at the same time corrects the nonlinear error by combining the pre-stored finite element analysis compensation table, and outputs the decoupled net error to the synchronous control feedback module. The synchronous control feedback module generates a compensation signal based on the net error, uses closed-loop control to adjust the position of the piezoelectric ceramic actuator, and sends the adjustment information to the operation monitoring terminal. The operation monitoring terminal communicates with the application effect evaluation module. The application effect evaluation module is used to set a production period of duration P1, evaluate the system's application effect during the production period, and generate an excellent application effect signal or an application effect alarm signal accordingly. The excellent application effect signal or the application effect alarm signal is sent to the operation monitoring terminal. When the operation monitoring terminal receives the application effect alarm signal, it issues a corresponding warning. The specific operation process of the application effect evaluation module is as follows: The average hourly silicon wafer output of the lithography machine after using the system during the production period is statistically analyzed and compared with the average hourly silicon wafer output before using the system. Based on this, the yield improvement ratio is collected. The yield improvement ratio is compared with the preset yield improvement ratio threshold. If the yield improvement ratio does not exceed the preset yield improvement ratio threshold, an application effect alarm signal is generated. If the yield increase ratio exceeds the preset yield increase ratio threshold, all quality inspection information of the products produced by the lithography machine during the production period is obtained, and products that fail the quality inspection are marked as defective objects. The number of defective objects during the production period is obtained and its ratio with the total number of products produced during the production period is calculated to obtain the defective ratio value. The defective ratio value is compared with the preset defective ratio threshold. If the defective ratio value exceeds the preset defective ratio threshold, an application effect alarm signal is generated. If the defective ratio value does not exceed the preset defective ratio threshold, an excellent application effect signal is generated. The application effect evaluation module communicates with the test and verification decision module. The application effect evaluation module sends the application effect alarm signal to the test and verification decision module. When the test and verification decision module receives the application effect alarm signal, it tests the system performance, analyzes the test data to verify the system's operating performance, and generates a performance verification qualified signal or a performance verification abnormal signal. The performance verification qualified signal or performance verification abnormal signal is then sent to the operation monitoring terminal. When the operation monitoring terminal receives the performance verification abnormal signal, it issues a corresponding warning. The test items involved in the test verification decision module include measurement accuracy test, dynamic response speed test and environmental adaptability test. Among them, in the measurement accuracy test, a high-precision standard is used to compare with the system output data to collect its synchronization error information, and the repeatability of the system to measure known displacement under static conditions is tested to collect its standard deviation information. In the dynamic response speed test, the sudden acceleration of the workpiece stage was simulated, and the delay time from the detection of the error to the output of the compensation signal was tested. In the high-speed scanning mode, the tracking error of the sine trajectory of the test system was tested. In the environmental adaptability test, the distance measurement error drift of the test system was tested under the conditions of temperature and humidity fluctuation and air pressure change. The specific operation process of the test verification decision module is as follows: All test data from the testing process are acquired and categorized. The corresponding test data is compared with the corresponding preset data requirements. If the corresponding test data does not meet the corresponding preset data requirements, a verification symbol WY-1 is assigned. The test anomaly value is obtained by calculating the ratio of the number of times the corresponding type of data is assigned the verification symbol WY-1 and the total number of the corresponding type of data. The test anomaly value is compared with the corresponding preset test anomaly threshold. If the test anomaly value exceeds the corresponding preset test anomaly threshold, the corresponding type of data is marked as a non-standard type. If the test anomaly value does not exceed the corresponding preset test anomaly threshold, when assigning the verification symbol WY-1, the deviation value of the corresponding test data from the corresponding preset data requirement is marked as the numerical deviation value, the average value of all numerical deviation values ​​of the corresponding type of data is calculated to obtain the numerical risk performance value, and the numerical deviation value with the largest value corresponding to the corresponding type of data is marked as the numerical risk amplitude value. The test evaluation value is obtained by weighted summation of test anomaly values, risk performance values, and risk amplitude values. The test evaluation value is then compared with the corresponding preset test evaluation threshold. If the test evaluation value exceeds the corresponding preset test evaluation threshold, the corresponding data type is marked as a non-standard type. If a non-standard type exists, a performance verification anomaly signal is generated. If there are no non-standard types, then set a set of preset performance impact weight values ​​for each type of data, calculate the ratio of the numerical evaluation value of the corresponding type of data to the corresponding preset numerical evaluation threshold to obtain the test analysis value, multiply the test analysis value by the corresponding preset performance impact weight value, and mark the product result as the test feature value. The verification decision value is obtained by summing the test feature values ​​of all types of data. The verification decision value is then compared with the preset verification decision threshold. If the verification decision value exceeds the preset verification decision threshold, a performance verification anomaly signal is generated; if the verification decision value does not exceed the preset verification decision threshold, a performance verification pass signal is generated.

2. The laser ruler ranging acquisition and analysis system adapted to a dual-stage lithography machine according to claim 1, characterized in that, Each dual-frequency laser interferometer in the dual-frequency laser interferometer array module adopts an orthogonal polarization optical path design for the reference light and the measurement light to eliminate optical path refraction error, and the measurement repeatability accuracy is less than or equal to 0.1nm; Furthermore, by using MEMS micromirrors to dynamically adjust the position of the laser spot and compensate for the optical path offset caused by the tilt of the workpiece stage in real time, the laser spot is always precisely focused on the center of the reflector.

3. The laser ruler ranging acquisition and analysis system adapted to a dual-workpiece stage of a lithography machine according to claim 2, characterized in that, The multi-source error decoupling analysis module communicates with the environment adaptive compensation module, which monitors environmental parameters in real time and corrects ranging errors caused by laser wavelength drift.

4. The laser ruler ranging acquisition and analysis system adapted to a dual-workpiece stage of a lithography machine according to claim 3, characterized in that, The environmental adaptive compensation module deploys a distributed temperature and humidity sensor network, with eight sensor nodes deployed around the dual workpiece stage for environmental monitoring, and data is aggregated via I2C bus; Furthermore, a long short-term memory network is used to fit the nonlinear relationship between environmental parameters and ranging error, dynamically correcting the laser wavelength. The corrected environmental parameters are then input into the multi-source error decoupling analysis module as boundary condition optimization error model.

5. The laser ruler ranging acquisition and analysis system adapted to a dual-stage lithography machine according to claim 1, characterized in that, The data acquisition and preprocessing module is based on the Xilinx Zynq-7000 series FPGA, which realizes eight-channel synchronous sampling, a sampling rate of 10MHz, and an ADC resolution of 24 bits. It also applies a wavelet threshold denoising algorithm to perform db4 wavelet transform on the sampled data and sets a threshold to filter out high-frequency vibration noise.

6. The laser ruler ranging acquisition and analysis system adapted to a dual-stage lithography machine according to claim 1, characterized in that, The synchronous control feedback module is designed with a feedforward-feedback hybrid control algorithm to shorten the response delay. It also adopts predictive control theory and pre-compensates for acceleration abrupt changes during the exchange process by pre-storing the workpiece stage motion trajectory database, thereby reducing trajectory tracking errors.

Images