A method and system for characterizing deep hole etching defect topography
By combining scattering and ellipsometric measurements with deep learning, the problem of online non-destructive characterization during deep hole etching was solved, enabling accurate detection of hole diameter consistency and defect types in high aspect ratio etched through holes, and providing a one-stop solution.
Patent Information
- Authority / Receiving Office
- CN · China
- Patent Type
- Patents(China)
- Current Assignee / Owner
- WUHAN UNIV
- Filing Date
- 2023-03-24
- Publication Date
- 2026-05-29
Smart Images
Figure CN116403927B_ABST
Abstract
Description
Technical Field
[0001] This invention belongs to the field of nano-characterization technology, and more specifically, relates to a method and system for characterizing the morphology of deep hole etching defects. Background Technology
[0002] The demand for morphological characterization of deep-hole etching defects has become increasingly prominent recently. For example, in the widely used 3D NAND flash memory technology, the high process uniformity and aspect ratio requirements necessitate significant morphological characterization of deep-hole etching defects. High aspect ratio channel etching is one of the most important and challenging steps in the 3D NAND process, and a key step in uniformly drilling holes on the stacked layers to form memory cell channels. This etching process requires precisely etching vertical vias with consistent top and bottom diameters, thus requiring accurate online monitoring and non-destructive characterization of the deep-hole etching state. Furthermore, as chip sizes continue to shrink, the integration complexity of 3D NAND processes is increasing. Due to the increased number of stacked deposition layers and the greater thickness difference from the wafer center to the edge, the potential for defects is also increasing. For high aspect ratio memory layer mask etching, due to the change in the flux ratio of ions and neutral particles with etching depth, when the aspect ratio is greater than 40, neutral particles cannot reach the bottom, and when the aspect ratio is greater than 50, 50% of the ions cannot reach the bottom. This will result in defects such as inconsistent upper and lower apertures, deviation from the preset trajectory, and curved etch pits. Moreover, for three-dimensional structures, these defects are all-round, and these defects will seriously affect the quality of the product. Therefore, non-destructive characterization of these defects is of great significance. Summary of the Invention
[0003] This invention provides a method and system for characterizing the morphology of deep hole etching defects, thereby solving the problem that existing technologies cannot achieve online, non-destructive characterization during deep hole etching.
[0004] This invention provides a method for characterizing the morphology of deep hole etching defects, comprising the following steps:
[0005] Step 1: Select an incident angle that meets the high sensitivity condition through scattering measurement, and record this incident angle as the first incident angle;
[0006] Step 2: Take the first incident angle as the incident angle for performing generalized ellipticity measurement of the full Mueller matrix, and obtain the first ellipticity measurement information through ellipticity measurement; Based on the first ellipticity measurement information, use deep learning to determine the type of deep hole etching defects.
[0007] Step 3: Increase the angle of incidence, and record this angle of incidence as the second angle of incidence;
[0008] Step 4: Use the second incident angle as the incident angle for generalized ellipticity measurement of the full Mueller matrix, and obtain the second ellipticity measurement information through ellipticity measurement; based on the second ellipticity measurement information, perform mask deformation closed-loop judgment; if the mask is determined to be deformed, replace the mask and return to step 2; if the mask is determined not to be deformed, obtain the deep hole etching defect morphology characterization information based on the deep hole etching defect type determined in step 2.
[0009] Preferably, in step 1, sensitivity information is obtained using a scattering measurement device, the sensitivity information including sensitivity values corresponding to several different incident angles; based on the sensitivity information, an incident angle that satisfies the high sensitivity condition is selected.
[0010] Preferably, in step 2, the first elliptic measurement information corresponding to the test sample is obtained using an elliptic measurement device. The first elliptic measurement information includes the 16-element Mueller matrix of the test sample based on wavelength.
[0011] The 16-element Mueller matrix is decomposed into a dichroic Mueller matrix, a phase-delayed Mueller matrix, and a scattering depolarization Mueller matrix;
[0012] The fitting curves corresponding to the Mueller matrix of the 16 elements, the dichroic Mueller matrix, the phase delay Mueller matrix and the scattering depolarization Mueller matrix are obtained by deep learning, and the above curves are denoted as the measurement fitting curves.
[0013] The measured fitting curve is compared with the database, which contains fitting curves corresponding to various deep hole etching defect types. The type of deep hole etching defect is determined based on the comparison results.
[0014] Preferably, in step 2, the measured fitting curve is compared with the critical point of multiple fitting curves in the database to find the fitting curve with the closest critical point and determine the type of deep hole etching defect.
[0015] The critical point is the good product critical point, including one or more of the following: the upper and lower aperture difference critical point, the etching trajectory tilt angle critical point, and the curvature of the curved etch pit good product critical point.
[0016] Preferably, the test sample is a three-dimensional stacked flash memory structure.
[0017] Preferably, step 2 further includes: performing a closed-loop judgment of the incident angle sensitivity; if the type of deep hole etching defect cannot be determined using deep learning, it is considered that the incident angle sensitivity is too low, and the process returns to step 1 to reselect the incident angle; if the type of deep hole etching defect can be determined using deep learning, the process proceeds to step 3.
[0018] Preferably, in step 4, based on the second elliptic measurement information, the deformation amount of the mask is obtained using deep learning, and the deformation amount is used to determine whether the mask is deformed.
[0019] Preferably, the objective lens in the scattering measurement device is capable of simultaneously measuring omnidirectional angles and multiple incident angles.
[0020] Preferably, the sample stage in the ellipticity measuring device is capable of omnidirectional angular rotation.
[0021] On the other hand, the present invention provides a characterization system for the morphology of deep hole etching defects, comprising:
[0022] A scattering measurement device is used to perform scattering measurements and obtain sensitivity information, the sensitivity information including sensitivity values corresponding to several different incident angles;
[0023] An elliptic measurement device is used to perform elliptic measurement and obtain first elliptic measurement information and second elliptic measurement information.
[0024] The analysis and judgment device is used to determine the type of deep hole etching defect based on the first elliptic measurement information using deep learning; to perform mask deformation closed-loop judgment based on the second elliptic measurement information; and, when it is determined that the mask is not deformed, to obtain deep hole etching defect morphology characterization information based on the determined deep hole etching defect type.
[0025] The deep hole etching defect morphology characterization system is used to implement the steps in the deep hole etching defect morphology characterization method described above.
[0026] One or more technical solutions provided in this invention have at least the following technical effects or advantages:
[0027] This invention first selects an incident angle that meets high sensitivity conditions through scattering measurement. Then, based on this incident angle, it performs a full Mueller matrix generalized ellipsometric measurement to obtain first ellipsometric measurement information. Deep learning is then used to determine the type of deep hole etching defects. Next, to detect whether the mask used in deep hole etching is deformed, the incident angle is increased to accommodate mask detection. Then, based on the increased incident angle, a full Mueller matrix generalized ellipsometric measurement is performed to obtain second ellipsometric measurement information, which is then used for mask deformation closed-loop judgment. If mask deformation is determined, the mask is replaced and deep hole etching defect detection is performed again; if mask non-deformation is determined, the morphological characterization information of the deep hole etching defect can be obtained based on the previously determined defect type, thus arriving at the final defect type conclusion. This invention fully utilizes the characteristics of scattering measurement, ellipsometric measurement, and deep learning technologies to achieve online, non-destructive characterization during deep hole etching, providing a one-stop solution for the detection of high aspect ratio, vertically etched, and uniformly sized etched through-holes. Attached Figure Description
[0028] Figure 1 A flowchart illustrating a method for characterizing the morphology of deep hole etching defects provided in Embodiment 1 of the present invention;
[0029] Figure 2 This is a schematic diagram of an ellipticity measurement device in a characterization system for the morphology of deep hole etching defects provided in Embodiment 2 of the present invention;
[0030] Figure 3 This is a schematic diagram of a scattering measurement device in a characterization system for the morphology of deep hole etching defects provided in Embodiment 2 of the present invention. Detailed Implementation
[0031] To better understand the above technical solutions, the following will provide a detailed explanation of the technical solutions in conjunction with the accompanying drawings and specific implementation methods.
[0032] Example 1:
[0033] Example 1 provides a method for characterizing the morphology of deep hole etching defects, see [link to example]. Figure 1 This includes the following steps:
[0034] Step 1: Select an incident angle that meets the high sensitivity condition through scattering measurement, and record this incident angle as the first incident angle.
[0035] That is, a highly sensitive incident angle is quickly selected through scattering measurement, which is an angle-based scattering measurement.
[0036] Specifically, sensitivity information is obtained using a scattering measurement device, which includes sensitivity values corresponding to several different incident angles; based on the sensitivity information, an incident angle that meets the high sensitivity condition is selected.
[0037] The objective lens in the scattering measurement device can simultaneously measure all omnidirectional angles and multiple incident angles.
[0038] Step 2: Use the first incident angle as the incident angle for generalized ellipticity measurement of the full Mueller matrix, and obtain the first ellipticity measurement information through ellipticity measurement; based on the first ellipticity measurement information, use deep learning to determine the type of deep hole etching defects.
[0039] Specifically, step 2 includes the following sub-steps:
[0040] Step 201: Obtain the Mueller matrix by ellipsometry.
[0041] Using the first incident angle as the incident angle, the first elliptic measurement information corresponding to the test sample is obtained using an elliptic measurement device. The first elliptic measurement information includes the 16-element Mueller matrix of the test sample based on wavelength.
[0042] Step 202: Decompose the Mueller matrix.
[0043] The 16-element Mueller matrix is decomposed into three factors using a polarization decomposition method: the dichroic Mueller matrix, the phase-delay Mueller matrix, and the scattering depolarization Mueller matrix. The dichroic Mueller matrix and the phase-delay Mueller matrix are independent of depolarization.
[0044] Step 203: Fit the wavelength-based Mueller matrix curve.
[0045] The fitting curves corresponding to the Mueller matrix of the 16 elements, the dichroic Mueller matrix, the phase delay Mueller matrix, and the scattering depolarization Mueller matrix are obtained by deep learning, and these curves are denoted as the measurement fitting curves.
[0046] Step 204: Determine the defect type.
[0047] The measured fitting curve is compared with the database, which contains fitting curves corresponding to various deep hole etching defect types. The type of deep hole etching defect is determined based on the comparison results.
[0048] Specifically, the measured fitting curve is compared with the critical points of various fitting curves in the database to find the fitting curve with the closest critical point and determine the type of deep hole etching defect; the critical point is the good product critical point, including but not limited to the critical point of the difference between the upper and lower hole diameters, the critical point of the etching trajectory tilt angle, and the good product critical point of the curvature of the curved etch pit, etc.
[0049] The basis for determining the type of deep hole etching defects is as follows: a database is constructed by comparing the critical points of the fitted curves of various defect types obtained through extensive simulation and experimental characterization. Based on this database, the defect type is determined by comparing the measured fitted curves with the critical points of the curves for comparing various types of defects.
[0050] Step 205: Closed-loop judgment of incident angle sensitivity.
[0051] If the type of deep hole etching defect cannot be determined using deep learning, the incident angle sensitivity is considered too low, and the process returns to step 1 to reselect the incident angle; if the type of deep hole etching defect can be determined using deep learning, the incident angle sensitivity is considered high, and the process proceeds to step 3.
[0052] The test sample in step 2 can be a three-dimensional stacked flash memory structure, or other nanostructures that require deep hole etching.
[0053] The sample stage in the elliptic measurement device described in step 2 can rotate in all directions, enabling omnidirectional angular measurement. Through deep learning, the specific orientation of structural defects such as 3D NAND can be obtained.
[0054] Step 3: Increase the angle of incidence, and denote this angle of incidence as the second angle of incidence.
[0055] Since deep holes require a smaller angle of incidence than masks, the angle of incidence needs to be increased in order to better characterize the size of the mask.
[0056] Step 4: Use the second incident angle as the incident angle for generalized ellipticity measurement of the full Mueller matrix, and obtain the second ellipticity measurement information through ellipticity measurement; based on the second ellipticity measurement information, perform mask deformation closed-loop judgment; if the mask is determined to be deformed, replace the mask and return to step 2; if the mask is determined not to be deformed, obtain the deep hole etching defect morphology characterization information based on the deep hole etching defect type determined in step 2.
[0057] Specifically, based on the second elliptic measurement information, the deformation of the mask is obtained using deep learning, and the deformation is used to determine whether the mask is deformed.
[0058] Example 1 includes dual closed-loop detection: incident angle sensitivity closed-loop judgment and mask deformation closed-loop judgment. Example 1 can obtain accurate defect type conclusions and realize defect morphology characterization.
[0059] In summary, Example 1 provides a non-destructive characterization method for multiple processes, multiple azimuth angles, and multiple defect types. It uses a scattering measurement device to find the sensitivity of each incident angle, and then uses a full Mueller matrix ellipsometric measurement device to obtain the Mueller matrix elements of the sample. The Mueller matrix is decomposed, and the defect type is determined by the 16 Mueller matrix elements and the decomposed matrix. When a defect occurs, the ellipsometric measurement device increases the incident angle to measure whether the mask is deformed, forming a closed-loop characterization and control. Example 1 can achieve online, rapid, and omnidirectional non-destructive characterization in the deep hole etching process of various nanostructures, including 3D NAND.
[0060] Example 2:
[0061] Example 2 provides a characterization system for the morphology of deep hole etching defects, comprising:
[0062] A scattering measurement device is used to perform scattering measurements and obtain sensitivity information, the sensitivity information including sensitivity values corresponding to several different incident angles;
[0063] An elliptic measurement device is used to perform elliptic measurement and obtain first elliptic measurement information and second elliptic measurement information.
[0064] The analysis and judgment device is used to determine the type of deep hole etching defect based on the first elliptic measurement information using deep learning; to perform mask deformation closed-loop judgment based on the second elliptic measurement information; and, when it is determined that the mask is not deformed, to obtain deep hole etching defect morphology characterization information based on the determined deep hole etching defect type.
[0065] The deep hole etching defect morphology characterization system provided in Example 2 is used to implement the steps in the deep hole etching defect morphology characterization method as described in Example 1. The system's functionality is described in Example 1 and will not be repeated here.
[0066] The scattering measurement device and the ellipsometric measurement device will be further described below.
[0067] See the schematic diagram of the elliptic measurement device. Figure 2 The optical path for measuring the generalized ellipsometrics of the full Mueller matrix is as follows: First, the first light source 101 emits a multi-wavelength light source, which then enters the first collimating lens 102. The collimated beam is converted into linearly polarized light by the first polarizer 103. The linearly polarized light then undergoes phase delay by the first rotating compensator 104, and is focused onto the first sample stage 111 by the first beam splitter 105. The first sample stage 111 rotates omnidirectionally, and the light, carrying information about the test sample, is reflected by the test sample and enters the second rotating compensator 106 for a second phase change by the first beam splitter 105. It then passes through the analyzer 108 and the second collimating lens 109, and is collected by the spectrometer 109. The data is then sent to the first PC terminal 110 to calculate and fit the measurement results (including fitting curves corresponding to the 16-element wavelength-based Mueller matrix, dichroic Mueller matrix, phase-delayed Mueller matrix, and scattering depolarized Mueller matrix of the sample). To obtain sufficient spectral information, the rotational speeds of the first rotating compensator 104 and the second rotating compensator 106 are different.
[0068] See the schematic diagram of the scattering measurement device. Figure 3 The scattering measurement optical path is as follows: First, the light emitted by the second light source 201 is converted into single-wavelength light by the monochromator 202, and then becomes linearly polarized light by the second polarizer 203. The light then enters the second beam splitter 204 and is focused into the objective lens 208. The second sample stage 209 adjusts its height and direction according to the objective lens 208 so that the test sample is located at the focal point of the objective lens 208. The light, after reflection, carries the information of the test sample and enters the lens group 205 through the second beam splitter 204. Finally, the light beam carrying the sample information is collected by the CCD module 206, and the data is sent to the second PC terminal 207 to calculate and fit the measurement results (including: sensitivity values corresponding to different incident angles).
[0069] The method and system for characterizing the morphology of deep hole etching defects provided in this invention have at least the following technical advantages:
[0070] The method for characterizing the morphology of deep-hole etching defects provided by this invention can be achieved using common optical instruments and data processing methods, resulting in low system cost and low implementation difficulty. In the nanolithography process, the method provided by this invention can characterize various types of defects in deep-hole etching of various nanostructures, including 3D NAND, in real time online, and can also provide feedback on the measurement results. This invention fully utilizes the characteristics of scattering measurement, ellipsometric measurement, and deep learning technologies, providing a one-stop solution for the detection of etched through-holes with high aspect ratios, vertical etching, and consistent top and bottom diameters.
[0071] Finally, it should be noted that the above specific embodiments are only used to illustrate the technical solutions of the present invention and not to limit it. Although the present invention has been described in detail with reference to examples, those skilled in the art should understand that modifications or equivalent substitutions can be made to the technical solutions of the present invention without departing from the spirit and scope of the technical solutions of the present invention, and all such modifications or substitutions should be covered within the scope of the claims of the present invention.
Claims
1. A method for characterizing the morphology of deep hole etching defects, characterized in that, Includes the following steps: Step 1: Select an incident angle that meets the high sensitivity condition through scattering measurement, and record this incident angle as the first incident angle; Step 2: Take the first incident angle as the incident angle for performing generalized ellipticity measurement of the full Mueller matrix, and obtain the first ellipticity measurement information through ellipticity measurement; Based on the first ellipticity measurement information, use deep learning to determine the type of deep hole etching defects. Step 3: Increase the angle of incidence, and record this angle of incidence as the second angle of incidence; Step 4: Use the second incident angle as the incident angle for performing generalized ellipsometric measurement of the full Mueller matrix, and obtain the second ellipsometric measurement information through ellipsometric measurement; Based on the second elliptic measurement information, a mask deformation closed-loop judgment is made; if the mask is determined to be deformed, the mask is replaced and the process returns to step 2. If the mask is determined to be undeformed, the morphological characterization information of the deep hole etching defect is obtained based on the deep hole etching defect type determined in step 2.
2. The method for characterizing the morphology of deep hole etching defects according to claim 1, characterized in that, In step 1, sensitivity information is obtained using a scattering measurement device. The sensitivity information includes sensitivity values corresponding to several different incident angles. Based on the sensitivity information, an incident angle that meets the high sensitivity condition is selected.
3. The method for characterizing the morphology of deep hole etching defects according to claim 1, characterized in that, In step 2, the first elliptic measurement information corresponding to the test sample is obtained using an elliptic measurement device. The first elliptic measurement information includes the 16-element Mueller matrix of the test sample based on wavelength. The 16-element Mueller matrix is decomposed into a dichroic Mueller matrix, a phase-delayed Mueller matrix, and a scattering depolarization Mueller matrix; The fitting curves corresponding to the Mueller matrix of the 16 elements, the dichroic Mueller matrix, the phase delay Mueller matrix and the scattering depolarization Mueller matrix are obtained by deep learning, and the above curves are denoted as the measurement fitting curves. The measured fitting curve is compared with the database, which contains fitting curves corresponding to various deep hole etching defect types. The type of deep hole etching defect is determined based on the comparison results.
4. The method for characterizing the morphology of deep hole etching defects according to claim 3, characterized in that, In step 2, the measured fitting curve is compared with the critical point of multiple fitting curves in the database to find the fitting curve with the closest critical point and determine the type of deep hole etching defect. The critical point is the good product critical point, including one or more of the following: the upper and lower aperture difference critical point, the etching trajectory tilt angle critical point, and the curvature of the curved etch pit good product critical point.
5. The method for characterizing the morphology of deep hole etching defects according to claim 3, characterized in that, The test sample is a 3D stacked flash memory structure.
6. The method for characterizing the morphology of deep hole etching defects according to claim 1, characterized in that, Step 2 further includes: performing closed-loop judgment of incident angle sensitivity; if the type of deep hole etching defect cannot be determined by deep learning, it is considered that the incident angle sensitivity is too low, and the process returns to step 1 to reselect the incident angle; if the type of deep hole etching defect can be determined by deep learning, the process proceeds to step 3.
7. The method for characterizing the morphology of deep hole etching defects according to claim 1, characterized in that, In step 4, based on the second elliptic measurement information, the deformation of the mask is obtained using deep learning, and the deformation is used to determine whether the mask is deformed.
8. The method for characterizing the morphology of deep hole etching defects according to claim 2, characterized in that, The objective lens in the scattering measurement device can simultaneously measure all omnidirectional angles and multiple incident angles.
9. The method for characterizing the morphology of deep hole etching defects according to claim 3, characterized in that, The sample stage in the elliptic measurement device can rotate in all directions.
10. A characterization system for the morphology of deep hole etching defects, characterized in that, include: A scattering measurement device is used to perform scattering measurements and obtain sensitivity information, the sensitivity information including sensitivity values corresponding to several different incident angles; An elliptic measurement device is used to perform elliptic measurement and obtain first elliptic measurement information and second elliptic measurement information. An analysis and judgment device is used to determine the type of deep hole etching defect based on the first elliptic measurement information using deep learning. Used to determine mask deformation closed-loop based on the second elliptic deviation measurement information; Furthermore, if the mask is determined to be undeformed, the morphological characterization information of the deep hole etching defect is obtained based on the determined type of deep hole etching defect; The characterization system for deep hole etching defect morphology is used to implement the steps in the characterization method for deep hole etching defect morphology as described in any one of claims 1-9.