A method for analyzing position detection signals

Abstract

The present disclosure provides a method for analyzing a position detection signal, comprising: obtaining the signal characteristics of the position detection signal; performing time-frequency analysis on the position detection signal using wavelet transform according to the signal characteristics to obtain time-frequency information; and analyzing the distortion information and graphic mark information of the position detection signal according to the time-frequency information. In the present disclosure, wavelet transform is used to perform wavelet decomposition on the obtained position detection signal, and the time domain and frequency domain characteristics of the signal are obtained at the same time. For the frequency domain analysis, the distribution of the position detection signal in a specific frequency range can be obtained, and the frequency characteristics of the position detection signal can be obtained. For the time domain analysis, the overall changes of the position detection signal in the entire test cycle can be obtained, and the moments at which certain detection signal distortions caused by external interference and defects and deformations of the position detection mark are located, as well as the spatial positions corresponding to the marks, can be obtained.

Description

Technical Field

[0001] The present disclosure relates to the field of semiconductor technology, and in particular to a method for analyzing position detection signals. Background Art

[0002] In the field of semiconductor manufacturing and testing technology, it is necessary to detect the position of the sample being processed or tested.

[0003] Currently, the industry often uses detection technologies based on optical methods or graphic methods. For the first detection method, it is generally necessary to design and process specific graphic marks on the sample. By detecting the diffraction light signal formed by the graphic mark, it is converted into a photoelectric signal and obtained through a subsequent series of signal processing and calculations. The relative position of the graphic mark is then solved to calculate the position information of the sample being measured.

[0004] The most commonly used analysis method is the Fourier transform. While this method can accurately analyze the distribution of various components in the detection signal in the frequency domain, it requires the signal to be stationary. Actual detection signals rarely meet this requirement. Furthermore, the Fourier transform cannot express the distribution of specific components of the signal being analyzed in the time-frequency domain.

[0005] While improvements to the Fourier transform, such as the short-time Fourier transform, address this shortcoming, they still can't effectively resolve the conflict between the time and frequency domains. Position detection signals based on graphic markers are non-stationary due to external interference, defects in the markers, and deformations. Analysis using only the Fourier transform method has significant limitations. Summary of the Invention

[0006] The main purpose of the present disclosure is to propose a method for analyzing position detection signals, aiming to solve the problem that the position detection information analysis method has limitations.

[0007] To achieve the above objectives, the present disclosure proposes a method for analyzing a position detection signal, comprising:

[0008] obtaining signal characteristics of a position detection signal;

[0009] According to the signal characteristics, the time-frequency analysis of the position detection signal is performed using wavelet transform to obtain the time-frequency information;

[0010] According to the time-frequency information, the distortion information and graphic mark information of the position detection signal are analyzed, and then the defects of the graphic mark in the spatial domain are diagnosed, realizing the analysis capability of the time-frequency-spatial domain of the position detection signal.

[0011] Optionally, the signal characteristics include whether the signal contains distortion;

[0012] According to the signal characteristics, the position detection signal is subjected to time-frequency analysis using wavelet transform. The steps of obtaining time-frequency information include:

[0013] When the signal contains distortion, wavelet transform is used to perform time-frequency analysis on the position detection signal to obtain time-frequency information.

[0014] Optionally, after performing time-frequency analysis on the position detection signal using wavelet transform according to signal characteristics to obtain time-frequency information, the following steps may be further performed:

[0015] Perform preliminary frequency domain analysis on the position detection signal through Fourier transform to obtain basic information;

[0016] Based on the basic information and time-frequency information, the comprehensive information of the position detection signal is obtained.

[0017] Optionally, the wavelet transform includes any one of discrete wavelet transform, continuous wavelet transform and wavelet packet.

[0018] Optionally, the wavelet transform is a discrete wavelet transform;

[0019] The steps for time-frequency analysis using wavelet transform include:

[0020] Format the position detection signal and set the sampling frequency and resolution;

[0021] According to the sampling frequency and resolution, set the number of decomposition layers and vanishing moment of discrete wavelet transform;

[0022] According to the number of decomposition layers and the vanishing moment, discrete wavelet transform is used for time-frequency analysis.

[0023] Optionally, the steps of performing time-frequency analysis using discrete wavelet transform include:

[0024] According to discrete wavelet transform, select the wavelet basis type;

[0025] According to the wavelet basis type, obtain the wavelet coefficient sequence and the wavelet decomposition graph.

[0026] Optionally, the wavelet basis type includes any one of a haar wavelet basis, a db wavelet basis, and the like.

[0027] Optionally, the step of performing time-frequency analysis using discrete wavelet transform according to the number of decomposition levels and the vanishing moment includes:

[0028] Different wavelet bases, vanishing moments and decomposition levels are used to perform time-frequency analysis multiple times to improve the accuracy of detection signal analysis and evaluation.

[0029] Optionally, the position detection signal includes a first position detection signal and a second position detection signal with two different directions;

[0030] The directions of the first position detection signal and the second position detection signal are arranged to be perpendicular to each other, and the analysis of the signals can be expanded from one dimension to two dimensions.

[0031] Optionally, based on signal characteristics, the step of performing time-frequency analysis on the position detection signal using wavelet transform to obtain time-frequency information includes:

[0032] Using wavelet transform to perform time-frequency analysis on the first position detection signal and the second position detection signal respectively, comprehensively obtaining time-frequency information, and further obtaining two-dimensional spatial information of the detection signal;

[0033] The wavelet transform types used for the position detection signals in two different directions may be different.

[0034] The technical solution provided by this disclosure employs the mathematical method of wavelet transform, leveraging its multi-resolution characteristics to perform wavelet decomposition on the acquired position detection signal, simultaneously obtaining the signal's frequency and time domain characteristics. Frequency domain analysis reveals the distribution of the position detection signal within a specific frequency range, as well as its frequency characteristics. Time domain analysis reveals the overall variation of the position detection signal throughout the test cycle, as well as the moments at which detection signal distortions caused by external interference, marking defects, and deformations occur, as well as the spatial locations corresponding to the markings. BRIEF DESCRIPTION OF THE DRAWINGS

[0035] Figure 1 This is a schematic diagram of a position detection graphic mark in an embodiment of the method for analyzing position detection signals disclosed herein;

[0036] Figure 2 It is the position detection signal in the time domain;

[0037] Figure 3 It is a frequency domain position detection signal based on Fourier transform;

[0038] Figure 4 For position detection signals with various signal distortions;

[0039] Figure 5 for Figure 4 A local magnified view of the distortion of the mid-position detection signal;

[0040] Figure 6 It is the frequency domain analysis result based on Fourier transform when the signal is distorted;

[0041] Figure 7 The position detection signal analysis result based on discrete wavelet transform in one embodiment provided by the present disclosure;

[0042] Figure 8 for Figure 7 The analysis results of position detection signal based on wavelet packet (low frequency part);

[0043] Figure 9 for Figure 7 The analysis results of position detection signal based on wavelet packet (high frequency part);

[0044] Figure 10 A schematic diagram of two-dimensional scanning detection of a position detection graphic mark in another embodiment provided by the present disclosure. DETAILED DESCRIPTION

[0045] In order to make the objectives, technical solutions and advantages of the present disclosure more clearly understood, the present disclosure is further described in detail below in conjunction with specific embodiments and with reference to the accompanying drawings.

[0046] It should be noted that if directional indications are involved in the embodiments of the present disclosure, the directional indications are only used to explain the relative positional relationship, movement status, etc. between the components under a certain specific posture. If the specific posture changes, the directional indications will also change accordingly.

[0047] In addition, if there are descriptions involving "first", "second", etc. in the embodiments of the present disclosure, the descriptions of "first", "second", etc. are only for descriptive purposes and cannot be understood as indicating or implying their relative importance or implicitly indicating the number of the indicated technical features. Therefore, the features defined as "first" and "second" may explicitly or implicitly include at least one of such features. In addition, the technical solutions between the various embodiments can be combined with each other, but they must be based on the ability of ordinary technicians in this field to implement them. When the combination of technical solutions is contradictory or cannot be implemented, it should be deemed that such a combination of technical solutions does not exist and is not within the scope of protection required by the present disclosure.

[0048] The present disclosure provides a method for analyzing a position detection signal, comprising:

[0049] S10, obtaining signal characteristics of the position detection signal;

[0050] S20. Perform time-frequency analysis on the position detection signal using wavelet transform according to signal characteristics to obtain time-frequency information;

[0051] S30. Analyze the distortion information and graphic mark information of the position detection signal based on the time-frequency information, and then diagnose the defects of the graphic mark in the spatial domain, thereby realizing the frequency-spatial domain analysis capability of the position detection signal.

[0052] In the technical solution provided by the present disclosure, the mathematical method of wavelet transform is adopted, and the characteristics of multi-resolution are utilized to perform wavelet decomposition on the obtained position detection signal, and the time domain and frequency domain characteristics of the signal are obtained at the same time. It is suitable for the analysis of non-stationary signals. For frequency domain analysis, the distribution of the position detection signal at a specific frequency can be obtained, and the frequency characteristics of the position detection signal can be obtained. For time domain analysis, the overall change of the position detection signal throughout the test cycle can be obtained, and the time at which the detection signal is distorted due to external interference and mark defects and deformations, as well as the spatial position corresponding to the mark, can be obtained.

[0053] It should be noted that distortion information includes the frequency, time of occurrence, and spatial location of the distorted signal, which can be used to diagnose the source of the problem. Graphical mark information includes defects and deformations of the mark, as well as changes in position information caused by the mark.

[0054] Please note that Figure 1 In step S10, for Figure 1 The position graphic mark in the image is scanned at a uniform speed in the normal direction of the mark, and then a triangular wave signal of the diffraction light intensity formed by the graphic mark is obtained to obtain a position detection signal.

[0055] Furthermore, the signal characteristics include whether the signal contains distortion; step S20 includes:

[0056] When the signal contains distortion, wavelet transform is used to perform time-frequency analysis on the position detection signal to obtain time-frequency information.

[0057] Perform time-frequency analysis on non-stationary position detection signals to make up for the deficiency that Fourier transform can only perform frequency domain analysis on stationary signals

[0058] On the other hand, after step S20, the method further includes:

[0059] S21. Perform preliminary frequency domain analysis on the position detection signal through Fourier transform to obtain basic information;

[0060] S22. Obtain comprehensive information of the position detection signal based on the basic information and the time-frequency information.

[0061] When there are non-stationary distorted signals, the Fourier analysis method is often not capable of analysis. On the basis of Fourier, wavelet is also used for further analysis.

[0062] See also Figure 2 The triangle wave signal contains multiple diffraction order information, which changes periodically with the spatial shape of the graphic mark; see Figure 3The Fourier method is used to perform frequency domain analysis on the triangle wave signal. The peak values ​​corresponding to each frequency domain correspond to the 1st, 3rd, 5th, 7th and 9th diffraction orders respectively. The signal frequency of each diffraction order increases exponentially, and the amplitude of the signal decreases continuously.

[0063] However, see Figures 4 and 5 , four different types of signal distortion are superimposed on the four different positions of the triangle wave signal of the position detection signal, namely low-frequency signal jitter, random Gaussian noise, instantaneous signal mutation and constant signal value for multiple consecutive sampling periods; please refer to Figure 6 , the Fourier method is used to analyze the distorted signal, and the frequency characteristics of the triangular wave change significantly. However, the frequency intervals corresponding to the several signal distortions cannot be clearly distinguished from the frequency diagram itself, and the time and space intervals where the signal distortion occurs cannot be obtained. The Fourier method cannot effectively analyze the position detection signal with multiple distortion conditions through the frequency characteristics; further, in this embodiment, when there is signal distortion, the wavelet transform method is used to utilize its multi-resolution characteristics to perform frequency domain-space domain analysis on the signal.

[0064] It should be noted that this embodiment is not limited to the above four types of signal distortion.

[0065] Specifically, there are multiple ways to select the type of wavelet transform. In this embodiment, the wavelet transform includes any one of discrete wavelet transform, continuous wavelet transform and wavelet packet.

[0066] In a specific embodiment provided by the present disclosure, the wavelet transform is a discrete wavelet transform;

[0067] The steps for time-frequency analysis using wavelet transform include:

[0068] S201, formatting the position detection signal and setting the sampling frequency and resolution;

[0069] S202, setting the number of decomposition layers and vanishing moment of discrete wavelet transform according to the sampling frequency and resolution;

[0070] S203. Based on the number of decomposition layers and the vanishing moment, discrete wavelet transform is used to perform time-frequency analysis.

[0071] A too large value of the vanishing moment will affect the compact support of the wavelet function. Generally, a compromise is required based on the results of the detection signal analysis. In this embodiment, the vanishing moment is selected to be 5-9.

[0072] Furthermore, in step S203, using discrete wavelet transform to perform time-frequency analysis includes:

[0073] S2031. Select a wavelet basis type according to discrete wavelet transform;

[0074] S2032. Obtain a wavelet coefficient sequence according to the wavelet basis type, and obtain a wavelet decomposition graph.

[0075] It should be noted that the wavelet basis type includes any one of a haar wavelet basis, a db wavelet basis, and the like.

[0076] In the technical solution disclosed herein, the low-frequency graph of the signal after wavelet transform processing reflects the overall overview of the signal, covering a signal frequency range of 0 to F / 2^(L+1); wherein the sampling frequency is F, the number of decomposition layers is L, and the vanishing moment is N; the high-frequency graph of the signal after wavelet transform processing reflects the detailed features of the signal in different frequency bands. When the number of decomposition layers is L, the frequency range corresponding to this layer is F / 2^(L+1) to F / 2^L.

[0077] Among them, the sampling frequency is F, the number of decomposition levels is L, and the vanishing moment is N.

[0078] Since the transformed signal can reflect information in the entire time domain, it can reflect the moment when the distortion occurs in the position detection signal. At the same time, the frequency band in which the distortion is located can be determined by the signal amplitude values ​​of different layers corresponding to that moment. Relative to the graphic mark, the detection signal is obtained by uniformly scanning the graphic mark. The distribution of the detection signal in time is uniform, and it can be considered that the time domain and space domain of the detection signal are one-to-one corresponding. In this embodiment, based on the moment when the distortion signal occurs, the position where the distortion occurs corresponding to the graphic mark can be obtained. Similarly, based on the frequency-space information of the distortion signal, the distortion signal can be diagnosed to further determine the source of the distortion signal, such as external interference, signal noise, instantaneous signal mutation, graphic mark defect, graphic mark deformation, etc.

[0079] In one embodiment provided in the present disclosure, db wavelet is used, the vanishing moment is selected as 5, the decomposition level is 7, and the wavelet decomposition diagram is as follows: Figure 7 shown.

[0080] The detection signal distortion caused by low-frequency signal jitter interference appears in the high-frequency part after decomposition in the 7th layer, and this distortion is not obvious in other layers; the detection signal noise caused by Gaussian random noise is reflected in all layers; the detection signal distortion caused by instantaneous mutation is mainly concentrated in the high-frequency part of the detection signal, and has little effect on the low-frequency part of the detection signal; the signal distortion caused by the signal value being a constant for multiple consecutive sampling periods is mainly concentrated in the low-frequency part of the detection signal, and has little effect on the high-frequency part of the detection signal; through discrete wavelet transform, not only can the detection signal distortion caused by different sources and the frequency range in which they are located be clearly distinguished, but the time range of the distortion can also be located in the time domain.

[0081] In addition, different wavelet transform types can be used for multiple analyses. For example, in order to solve the shortcoming of the discrete wavelet transform's insufficient analysis ability in the high-frequency part of each decomposition layer, wavelet packets can be used for further analysis.

[0082] In the disclosed embodiment, the wavelet packet decomposition diagram is as follows: Figure 8 and Figure 9 As shown in the figure, the detection signal distortion corresponds to the low-frequency part and the high-frequency part of the third-level decomposition respectively; relative to the graphic mark, the detection signal is obtained by uniformly scanning the graphic mark, and the temporal distribution of the detection signal is uniform, and thus it can be considered that the time domain and the spatial domain of the detection signal are one-to-one corresponding; according to the occurrence time of the detection signal distortion obtained by the above analysis, the position area corresponding to the occurrence of the graphic mark can be obtained; according to the frequency-space information of the distorted signal, the detection signal distortion caused by the deformation or defect of the graphic mark can be further estimated.

[0083] Optionally, after step S203, the following steps are included:

[0084] S204. Perform frequency domain analysis multiple times using different wavelet bases, vanishing moments, and decomposition levels. Repeated iterative analysis improves the accuracy of detection signal analysis and evaluation.

[0085] In another embodiment provided by the present disclosure, the analysis of the signal can be extended from one dimension to two dimensions. In the analysis of the two-dimensional position detection signal, the position detection signal includes a first position detection signal and a second position detection signal with two different directions.

[0086] The directions of the first position detection signal and the second position detection signal are arranged perpendicular to each other.

[0087] See also Figure 10 In another embodiment provided by the present disclosure, a uniform scanning motion is performed at a specific angle, and the obtained detection signal is projected into two orthogonal XY directions, so that position detection signals in the two XY directions can be obtained simultaneously.

[0088] Furthermore, step S20 includes:

[0089] S24, performing time-frequency analysis on the first position detection signal and the second position detection signal using wavelet transform, comprehensively obtaining time-frequency information, and further obtaining two-dimensional spatial information of the detection signals;

[0090] The wavelet transform types used for the position detection signals in two different directions may be different.

[0091] Position detection signals in two different directions can be analyzed and diagnosed by using any one of the methods of one-dimensional discrete wavelet transform, one-dimensional continuous wavelet transform or wavelet packet transform.

[0092] In this embodiment, the position detection signal is sampled at fixed intervals to form a grid array. Due to the uniform scanning motion, the entire grid represents a position array relative to the graphical mark. Each element of the grid array corresponds to a specific location (X, Y) on the graphical mark. Using this grid array, combined with the wavelet transform coefficients in the two aforementioned directions, wavelet transform results corresponding to two orthogonal directions can be obtained at a specific location (X, Y), enabling further two-dimensional analysis.

[0093] The above specific embodiments further illustrate the purpose, technical solutions and beneficial effects of the present disclosure in detail. It should be understood that the above are only specific embodiments of the present disclosure and are not intended to limit the present disclosure. Any modifications, equivalent substitutions, improvements, etc. made within the spirit and principles of the present disclosure should be included in the scope of protection of the present disclosure.

Claims

1. A method for analyzing a position detection signal, characterized in that: include: Acquiring signal characteristics of a position detection signal; the position detection signal includes a triangular wave signal of diffracted light intensity formed by the graphic mark; Performing time-frequency analysis on the position detection signal using wavelet transform according to the signal characteristics to obtain time-frequency information; Based on the time-frequency information, distortion information and graphic mark information of the position detection signal are analyzed to diagnose defects of the graphic mark in the spatial domain, thereby realizing the time-frequency-spatial domain analysis capability of the position detection signal; the distortion information includes the frequency, time of occurrence, and spatial position of the distorted signal; the graphic mark information includes defects and deformations of the mark, as well as changes in position information caused by the mark; The signal characteristics include whether the signal contains distortion; The step of performing time-frequency analysis on the position detection signal using wavelet transform according to the signal characteristics to obtain time-frequency information includes: When the signal contains distortion, wavelet transform is used to perform time-frequency analysis on the position detection signal to obtain time-frequency information.

2. The method for analyzing a position detection signal according to claim 1, wherein: According to the signal characteristics, the position detection signal is subjected to time-frequency analysis using wavelet transform to obtain time-frequency information, and the following step further comprises: Performing a preliminary frequency domain analysis on the position detection signal by Fourier transform to obtain basic information; Comprehensive information of the position detection signal is obtained according to the basic information and the time-frequency information.

3. The method for analyzing a position detection signal according to claim 1, wherein: The wavelet transform includes any one of discrete wavelet transform, continuous wavelet transform and wavelet packet.

4. The method for analyzing a position detection signal according to claim 1, wherein: The wavelet transform is a discrete wavelet transform; The steps for time-frequency analysis using wavelet transform include: Formatting the position detection signal and setting the sampling frequency and resolution; Setting the number of decomposition layers and the vanishing moment of the discrete wavelet transform according to the sampling frequency and the resolution; According to the number of decomposition levels and the vanishing moment, discrete wavelet transform is used to perform time-frequency analysis.

5. The method for analyzing a position detection signal according to claim 4, wherein: The steps of using discrete wavelet transform for time-frequency analysis include: Selecting a wavelet basis type according to the discrete wavelet transform; According to the wavelet basis type, a wavelet coefficient sequence is obtained, and a wavelet decomposition graph is obtained.

6. The method for analyzing a position detection signal according to claim 5, wherein: The wavelet basis type includes any one of a haar wavelet basis and a db wavelet basis.

7. The method for analyzing a position detection signal according to claim 4, wherein: The step of performing time-frequency analysis using discrete wavelet transform according to the number of decomposition levels and the vanishing moment includes: Time-frequency analysis was performed multiple times using different wavelet bases, vanishing moments and decomposition levels.

8. The method for analyzing a position detection signal according to claim 1, wherein: The position detection signal includes a first position detection signal and a second position detection signal with two different directions; The directions of the first position detection signal and the second position detection signal are arranged to be perpendicular to each other.

9. The method for analyzing a position detection signal according to claim 8, wherein: The step of performing time-frequency analysis on the position detection signal using wavelet transform according to the signal characteristics to obtain time-frequency information includes: Using wavelet transform to perform time-frequency analysis on the first position detection signal and the second position detection signal respectively, to obtain the time-frequency information, and further obtain the two-dimensional spatial information of the detection signal; The position detection signals in two different directions use different wavelet transform types.

Images