CN-121997234-A - Secondary simulation identification method and system for glass wafer surface microwave morphology

CN121997234ACN 121997234 ACN121997234 ACN 121997234ACN-121997234-A

Abstract

The invention discloses a secondary simulation identification method and a system for a glass wafer surface microwave form, which relate to the technical field of wafer simulation and comprise the steps of obtaining two-dimensional matrix data of wafer surface thickness distribution and performing mask matrix processing; the method comprises the steps of carrying out two-dimensional Fourier transform on processed two-dimensional matrix data to obtain surface thickness frequency domain distribution, identifying PSD power spectral density of the surface thickness frequency domain distribution, extracting a characteristic frequency region of the PSD power spectral density, constructing an adaptive filter, carrying out frequency domain separation on the low-pass filter and the high-pass filter, obtaining high-frequency separation data and low-frequency separation data, and constructing a microwave form secondary simulation surface. The invention solves the technical problem that the surface microwave morphology can not be identified in the range of the whole glass wafer in the prior art, and achieves the technical effects of accurately identifying and reconstructing the surface microwave morphology of the glass wafer.

Inventors

FENG XIANGXU
WU MING
XU YUNMING
TANG CHANGYONG
SHEN JIE

Assignees

浙江蓝特光学股份有限公司
浙江蓝创光电科技有限公司

Dates

Publication Date: 20260508
Application Date: 20260128

Claims (10)

1. The secondary simulation identification method for the microwave morphology of the surface of the glass wafer is characterized by comprising the following steps of: acquiring two-dimensional matrix data of thickness distribution of the surface of the wafer, and performing mask matrix processing on the two-dimensional matrix data; Performing two-dimensional Fourier transform on the processed two-dimensional matrix data to obtain surface thickness frequency domain distribution, identifying PSD power spectral density of the surface thickness frequency domain distribution, and extracting a characteristic frequency region of the PSD power spectral density; constructing an adaptive filter, wherein the adaptive filter comprises a low-pass filter and a high-pass filter which are set based on an adaptive cut-off frequency, and the adaptive cut-off frequency is obtained by performing iterative search by taking a defined optimization objective function as a target of minimizing a space autocorrelation length; performing frequency domain separation on the low-pass filter and the high-pass filter to obtain high-frequency separation data and low-frequency separation data; and constructing a microwave form secondary simulation surface according to the high-frequency separation data and the low-frequency separation data.
2. The method of claim 1, wherein a characteristic frequency region of the PSD power spectral density is extracted, the characteristic frequency region comprising at least a low-frequency characteristic region less than a first preset frequency threshold and a high-frequency characteristic region greater than or equal to the first preset frequency threshold; And simultaneously recording the directivity identifications, the frequency indexes and the sampling intervals of the low-frequency characteristic region and the high-frequency characteristic region.
3. The method of claim 1, wherein the adaptive cut-off frequency is obtained by iterative search with a defined optimization objective function targeting minimizing spatial autocorrelation length, the method comprising: defining a cut-off frequency search range and a search step length; Initializing a self-adaptive filter, selecting candidate cut-off frequency in the cut-off frequency search range, and performing frequency domain separation on the basis of a low-pass filter and a high-pass filter of the candidate cut-off frequency to obtain a high-frequency separation data sample and a low-frequency separation data sample; performing inverse two-dimensional Fourier transform on the high-frequency separation data sample and the low-frequency separation data sample to obtain high-frequency space domain data and low-frequency space domain data; Calculating the space autocorrelation length of the high-frequency space domain data; and re-selecting candidate cut-off frequencies according to the calculated space autocorrelation length until the self-adaptive cut-off frequency aiming at minimizing the space autocorrelation length is obtained.
4. A method according to claim 3, wherein the spatial autocorrelation length of the high frequency spatial domain data is calculated, the method comprising: performing mean value removal processing on the high-frequency spatial domain data, and performing two-dimensional autocorrelation calculation on the processed high-frequency spatial domain data to obtain a two-dimensional autocorrelation function; extracting a one-dimensional autocorrelation function along a zero delay direction from the two-dimensional autocorrelation function; A delay distance from the decay of the spool function value to 1/e is calculated as an autocorrelation length based on the one-dimensional autocorrelation function.
5. The method of claim 3, wherein after calculating the spatial autocorrelation length of the high frequency spatial domain data, the method further comprises: Calculating the space autocorrelation length of the low-frequency space domain data; and judging whether a space autocorrelation length verification result passes or not by comparing the space autocorrelation length of the low-frequency space domain data and the space autocorrelation length of the high-frequency space domain data.
6. A method as claimed in claim 3, wherein the candidate cut-off frequencies are re-selected based on the calculated spatial autocorrelation lengths, the method comprising: Simultaneously acquiring at least two different candidate cut-off frequencies based on the cut-off frequency search range; calculating at least two spatial autocorrelation lengths of the at least two different candidate cutoff frequencies, and updating the cutoff frequency search range according to the size of the at least two spatial autocorrelation lengths; Repeating until the cutoff frequency search range is smaller than a convergence threshold value, and extracting a midpoint in an output search range under the convergence threshold value as an adaptive cutoff frequency.
7. The method of claim 1, wherein after acquiring the two-dimensional matrix data of the wafer surface thickness profile, the method further comprises: performing deviation calculation on the two-dimensional matrix data, and marking the matrix data with the deviation larger than a preset deviation threshold value to obtain two-dimensional abnormal matrix data and two-dimensional normal matrix data; Clustering the two-dimensional abnormal matrix data to obtain a plurality of two-dimensional abnormal matrix data corresponding to a plurality of defect types; Respectively configuring a plurality of self-adaptive cut-off frequencies for the plurality of two-dimensional abnormal matrix data to obtain a plurality of groups of high-frequency separation data and a plurality of groups of low-frequency separation data; And constructing a microwave form secondary simulation surface according to the high-frequency separation data and the low-frequency separation data corresponding to the two-dimensional normal matrix data and the multiple groups of high-frequency separation data and the multiple groups of low-frequency separation data corresponding to the two-dimensional abnormal matrix data.
8. The method of claim 7, wherein a plurality of adaptive cut-off frequencies are respectively configured for the plurality of two-dimensional anomaly matrix data, the method comprising: constructing a plurality of constraint features based on the plurality of two-dimensional anomaly matrix data; analyzing dominant frequency distribution of the plurality of two-dimensional anomaly matrix data; And respectively configuring a plurality of self-adaptive cut-off frequencies for the plurality of two-dimensional abnormal matrix data according to the dominant frequency distribution and the plurality of constraint features.
9. The method of claim 8, wherein a plurality of adaptive cut-off frequencies are respectively configured for the plurality of two-dimensional anomaly matrix data with the dominant frequency distribution and the plurality of constraint features, the method further comprising: Acquiring a plurality of two-dimensional neighborhood matrix data of the two-dimensional anomaly matrix data based on a neighborhood window; and carrying out weight configuration on the dominant frequency distribution and the constraint features according to the two-dimensional neighborhood matrix data to obtain a plurality of self-adaptive cut-off frequencies corresponding to the two-dimensional abnormal matrix data.
10. A system for performing a method for performing a secondary simulation of a microwave morphology of a glass wafer according to any one of claims 1-9, the system comprising: the processing module is used for acquiring two-dimensional matrix data of the thickness distribution of the wafer surface and performing mask matrix processing on the two-dimensional matrix data; the identification module is used for carrying out two-dimensional Fourier transform on the processed two-dimensional matrix data to obtain surface thickness frequency domain distribution, identifying PSD power spectral density of the surface thickness frequency domain distribution, and extracting a characteristic frequency region of the PSD power spectral density; The filter construction module is used for constructing an adaptive filter, the adaptive filter comprises a low-pass filter and a high-pass filter which are set based on an adaptive cut-off frequency, and the adaptive cut-off frequency is obtained by iterative search by taking the minimum space autocorrelation length as a target through a defined optimization objective function; The frequency domain separation module is used for carrying out frequency domain separation on the low-pass filter and the high-pass filter to obtain high-frequency separation data and low-frequency separation data; And the surface construction module is used for constructing a microwave form secondary simulation surface according to the high-frequency separation data and the low-frequency separation data.

Description

Secondary simulation identification method and system for glass wafer surface microwave morphology Technical Field The invention relates to the technical field of wafer simulation, in particular to a secondary simulation identification method and system for a glass wafer surface microwave form. Background During processing and quality inspection of glass wafers, surface microwave morphology is typically presented as a microscopic topographical feature superimposed over an overall thickness relief, with amplitudes at the nanometer scale and distributed throughout the wafer. The existing detection and analysis mode is mainly based on the whole thickness distribution data of the wafer or the surface morphology measurement result of a local area, macroscopic fluctuation and warping components in the whole thickness data are dominant, microwave morphology features are difficult to effectively separate from the whole thickness data, the local surface morphology analysis is limited by a measurement range, and the microwave morphology distribution condition of the whole glass wafer is difficult to reflect, so that the surface microwave morphology cannot be effectively identified and simulated and characterized in the whole glass wafer range. Disclosure of Invention The application provides a secondary simulation identification method and system for a glass wafer surface microwave form, which are used for solving the technical problem that the surface microwave form cannot be identified in the whole glass wafer range in the prior art. In view of the above problems, the application provides a method and a system for secondary simulation identification of a glass wafer surface microwave form. The application provides a second simulation identification method of a glass wafer surface microwave morphology, which comprises the following steps: The method comprises the steps of obtaining two-dimensional matrix data of wafer surface thickness distribution, carrying out mask matrix processing on the two-dimensional matrix data, carrying out two-dimensional Fourier transform on the processed two-dimensional matrix data to obtain surface thickness frequency domain distribution, identifying PSD power spectral density of the surface thickness frequency domain distribution, extracting a characteristic frequency region of the PSD power spectral density, constructing an adaptive filter, wherein the adaptive filter comprises a low-pass filter and a high-pass filter which are arranged based on an adaptive cut-off frequency, the adaptive cut-off frequency is obtained by carrying out iterative search by taking a minimum space autocorrelation length as a target through a defined optimized objective function, carrying out frequency domain separation on the low-pass filter and the high-pass filter to obtain high-frequency separation data and low-frequency separation data, and constructing a microwave form secondary simulation surface according to the high-frequency separation data and the low-frequency separation data. In a second aspect of the present application, there is provided a system for secondary simulation identification of a microwave morphology of a glass wafer surface, the system comprising: the device comprises a processing module, a recognition module, a filter construction module and a surface construction module, wherein the processing module is used for acquiring two-dimensional matrix data of the thickness distribution of a wafer surface, performing mask matrix processing on the two-dimensional matrix data, the recognition module is used for performing two-dimensional Fourier transform on the processed two-dimensional matrix data to acquire surface thickness frequency domain distribution, recognizing PSD power spectral density of the surface thickness frequency domain distribution and extracting a characteristic frequency region of the PSD power spectral density, the filter construction module is used for constructing an adaptive filter, the adaptive filter comprises a low-pass filter and a high-pass filter which are arranged based on an adaptive cut-off frequency, the adaptive cut-off frequency is obtained by performing iterative search with a minimum space autocorrelation length as a target through a defined optimization objective function, the frequency domain separation module is used for performing frequency domain separation on the low-pass filter and the high-pass filter to acquire high-frequency separation data and low-frequency separation data, and the surface construction module is used for constructing a microwave form secondary simulation surface according to the high-frequency separation data and the low-frequency separation data. One or more technical schemes provided by the application have at least the following technical effects or advantages: The method comprises the steps of obtaining two-dimensional matrix data of thickness distribution of a wafer surface, performing mask matrix processing on the tw