Thin film defect detection method based on electrochemical corrosion

Abstract

The invention discloses a thin film defect detection method based on electrochemical corrosion, which comprises the following steps: carrying out electrochemical corrosion treatment on a thin film to be detected, so that original small-size defects on the thin film are expanded to defects in a size range capable of being observed by an optical microscope; and acquiring a thin film surface image after corrosion treatment through an optical microscope, identifying and counting the defect form, size and distribution through image analysis, and calculating the original defect information of the thin film. According to the method, the thin film is corroded through electrochemical corrosion, then an image is obtained through a low-cost and high-efficiency optical microscope, and the form, size and distribution of defects are identified and counted through image analysis, so that the original defect density, position and type of the thin film are indirectly deduced; the use of high-cost and high-complexity equipment such as a traditional transmission electron microscope is effectively avoided, and a low-cost and high-efficiency evaluation scheme is provided for film quality detection.

Description

TECHNICAL FIELD

[0001] The present application relates to the technical field of thin film quality detection, and particularly relates to a thin film defect detection method based on electrochemical corrosion. BACKGROUND

[0002] Thin film materials are widely used in key fields such as electronic information, optical devices, aerospace, semiconductor manufacturing, etc. For example, molybdenum (Mo) thin films prepared by electron beam evaporation are often used as electrode layers and interconnection layers of semiconductor chips, aluminum oxide (Al2O3) thin films are used as antireflection films or protective layers of optical elements, and titanium (Ti) thin films are used as corrosion-resistant coatings of metal-based materials. The quality of such thin films directly determines the performance and service life of the terminal device, and the "defect problem" is the core bottleneck restricting the quality of the thin film: during the preparation of the thin film, pinholes are easily formed due to splashing of the film material, and impurities such as oil, water vapor and polishing powder left on the silicon wafer substrate weaken the adhesion of the film layer, and the internal stress of the film layer is unbalanced, resulting in micro-cracks. These original defects will gradually develop into the starting point of corrosion and peeling in the service environment (especially in a humid environment), and ultimately lead to device failure (such as semiconductor chip leakage, optical film light transmittance reduction, and corrosion-resistant coating bulging).

[0003] Accurate detection of the density, position and type of original defects of the thin film is a key prerequisite for optimizing the film coating process and ensuring the reliability of the device. The current mainstream thin film defect detection methods are as follows:

[0004] The first type is high-resolution microscopic detection, represented by transmission electron microscopy (TEM) and scanning electron microscopy (SEM). Although this type of method can achieve nanoscale defect observation, it has high equipment cost, complex operation and extremely low detection efficiency, and is only suitable for laboratory-level microscopic analysis, and cannot meet the rapid detection needs of industrial batch thin films; although the ordinary optical microscope has low cost and convenient operation, it is limited by the resolution (the minimum resolvable size is about 0.5 μm), and cannot directly observe the original micro defects (such as <0.5 μm pinholes and micro-cracks) of the thin film.

[0005] The second type is physical property detection, including ultrasonic detection, eddy current detection, infrared thermal imaging, etc. Among them, ultrasonic detection has insufficient penetration of the thin film, and it is difficult to distinguish between micro defects and film thickness fluctuations; eddy current detection is only suitable for conductive thin films, and has low sensitivity to non-metallic defects (such as oxide impurity introduction points); although infrared thermal imaging can achieve large-area detection, defect recognition relies on temperature differences, and the detection rate (about 30%) of micro defects without obvious thermal effects (such as closed micro-cracks) is insufficient, and cannot meet the high-precision detection needs.

[0006] In summary, the current thin film defect detection field faces multiple pain points such as high cost of high-resolution detection, insufficient precision of low-cost detection, and low efficiency, and there is an urgent need for a defect detection method that takes into account "low cost, high efficiency, high controllability, and high analysis precision". SUMMARY

[0007] The present application aims at the deficiencies in the prior art, and provides a thin film defect detection method based on electrochemical corrosion, which adopts electrochemical corrosion to corrode the thin film, then obtains images through a low-cost and high-efficiency optical microscope, and identifies and counts the morphology, size and distribution of defects through image analysis, so as to indirectly infer the original defect density, position and type of the thin film; the use of high-cost and high-complexity devices such as traditional transmission electron microscopes is effectively avoided, and a low-cost and high-efficiency evaluation scheme is provided for thin film quality detection.

[0008] To solve the above technical problems, the present application provides a thin film defect detection method based on electrochemical corrosion, comprising the following steps:

[0009] S1, electrochemically corroding the thin film to be detected, so that the original small-size defects on the thin film are expanded to defects in the observable size range of the optical microscope;

[0010] S2, obtaining the surface image of the thin film after the corrosion treatment through the optical microscope, identifying and counting the morphology, size and distribution of defects through image analysis, and inferring the original defect information of the thin film.

[0011] The present application adopts chemical corrosion development, utilizes the chemical activity difference between the defect area and the non-defect area, and makes the defects "enlarged" to the visible morphology of the optical microscope through the corrosion effect. Electrochemical corrosion as a controllable chemical action has been preliminarily applied in the detection of metal material defects, such as the pitting corrosion detection of steel materials, and is based on the ultra-thin characteristics of the thin film (the thickness is much smaller than that of the block material), but has not been applied in the thin film defect development field. The main reasons are as follows: on the one hand, the existing electrochemical corrosion method is difficult to apply to the thin film - excessive corrosion strength will cause the whole film layer to peel off, and insufficient strength cannot effectively enlarge the original defects; on the other hand, the defect analysis after development lacks precise technical means, and cannot quantitatively infer the density and size of the original defects, nor can it realize the crack path recovery and defect density characterization, resulting in that the detection result can only be qualitatively described, and cannot provide quantitative basis for the optimization of the plating process.

[0012] Further, in S1, the corrosion treatment is specifically: placing the thin film to be detected in a corrosion environment with a humidity of 85%-95% for electrochemical corrosion.

[0013] Further, the corrosion environment is subjected to strengthening treatment to improve the corrosion rate, and the strengthening treatment is one or more of heating, salt water spraying and applying an electric field.

[0014] Further, the temperature of the heating is 70-80℃; the salt water is a sodium chloride aqueous solution with a concentration of 8-10wt%; and the voltage of the electric field is 0.5-2V.

[0015] The existing corrosion development technology mainly depends on natural environment corrosion (such as exposure to atmospheric humid environment), but the process is extremely low in efficiency, and experimental data show that it takes more than 14 days for the original defects of the electron beam Mo film to expand into visible cavities in the natural humid environment, which cannot meet the needs of "real-time detection, rapid feedback process" in industrial production. The application shortens the defect corrosion time to about 24-72 hours by at least one of heating, salt water spraying and applying an electric field, so that the original small defects expand into visible cavities / cracks.

[0016] Further, the material of the to-be-detected film is molybdenum, titanium or aluminum oxide.

[0017] Further, the thickness of the to-be-detected film is 1-500 nm.

[0018] Further, when the thickness of the to-be-detected film is less than 15 nm, the defects after corrosion treatment are cavities, and the original defect information of the film is calculated by image analysis and identification of the defect morphology, size and distribution, including:

[0019] The adaptive threshold segmentation algorithm and morphological operation are used to outline the cavity profile in the image, and the defect distribution thermodynamic diagram is generated based on the calibration multiple relationship of the optical microscope.

[0020] Further, the density of the original defects is obtained through the defect distribution thermodynamic diagram.

[0021] Further, when the thickness of the to-be-detected film is greater than or equal to 15 nm, the defects after corrosion treatment are cracks, and the original defect information of the film is calculated by image analysis and identification of the defect morphology, size and distribution, including:

[0022] The directional filtering and skeleton extraction method are used to track the crack path, and the total length, average width and orientation distribution of the crack are quantified based on the calibration multiple relationship of the optical microscope.

[0023] Further, the position distribution of the original defects is obtained through the orientation distribution of the cracks.

[0024] In the electrochemical corrosion film defect development, the electrochemical corrosion preferentially acts on the defects, because of the difference in film layer characteristics between the defect area and the non-defect area: on the one hand, the film layer at the defect (whether it is a point defect or a line defect) has a continuous damage, and the exposed silicon substrate and the film (such as Mo plating film) form a galvanic couple, the substrate acts as a cathode, and the Mo plating film acts as an anode, accelerating the oxidation and dissolution of the anode Mo (anode reaction: Mo-6e - → Mo 6+ , Mo 6+ Further, in combination with OH- to generate soluble MoO4 2-), while cathode O2 participates in the water film to generate OH- (cathode reaction: O2+2H2O+4e - On the other hand, stress concentration exists at the defects, the stress improves the atomic activity, reduces the activation energy of Mo oxidation reaction, and further strengthens the priority of corrosion.

[0025] For the point defects (original defects are pinholes, impurity introduction points, etc.) corresponding to the thin layer film (<15 nm), the reason why they can be developed by electrochemical corrosion is that: as the local discontinuous area of the film layer, the above-mentioned electrochemical corrosion reaction occurs preferentially at the point defects, and the Mo film layer is continuously dissolved at the defects, and the density of the point defects directly affects the size of the cavity after corrosion. When the point defects are relatively dense, the corrosion and dissolution areas of adjacent point defects will gradually connect as the reaction proceeds, forming a larger visible cavity. Then, through the adaptive threshold segmentation algorithm, the cavity and the background are distinguished, and the morphological operation fills the gap of the cavity edge and accurately outlines the cavity profile, that is, based on the dialog AI driven image analysis, the cavity profile is accurately outlined by identifying the sharp color contrast between the film area (dark blue) and the cavity area (light purple) in the optical image. According to the scale (i.e. the magnification of the optical microscope) in the optical microscope image, the pixel area of the outlined cavity is converted into the actual physical area, the total number of cavities in the detection area and the actual area of the detection area are counted, and finally the defect density (the number of cavities per unit area) is calculated, so as to infer the density of the original point defects.

[0026] For the line defects (original defects are micro-cracks caused by internal stress) corresponding to the thick layer film (>15 nm), the reason why they can be developed by electrochemical corrosion is that: this kind of line defect is a linear area of stress concentration in the film layer, and the electrochemical corrosion preferentially occurs along the linear path of the micro-crack. During the corrosion process, the Mo film layer on both sides of the crack is continuously oxidized and dissolved, so that the original fine line defect is gradually expanded into a visible crack. Then, the direction filter algorithm is used to enhance the linear feature of the crack and suppress the background noise, that is, the crack and the film area are identified by color contrast, the defect is restored to the center according to the crack shape, and the skeleton extraction algorithm simplifies the crack into a single-pixel-width skeleton structure. Since the crack is formed by expanding along the original line defect, the position of the crack skeleton is highly consistent with the position of the original line defect. By counting the total length, branch number and distribution of the crack skeleton in the detection area, and converting the actual spatial range of the detection area according to the scale of the optical microscope, the position of the original line defect can be restored, and the density of the original line defect (the number or total length of the line defects per unit area) can be calculated.

[0027] The beneficial effects of the present application are:

[0028] The film is subjected to corrosion treatment by optimized electrochemical corrosion, then an image is acquired by low-cost and high-efficiency optical microscope, and the morphology, size and distribution of defects are identified and counted by image analysis, so as to indirectly infer the original defect density, position and type of the film; the use of high-cost and high-complexity equipment such as traditional transmission electron microscope is effectively avoided, and a low-cost and high-efficiency evaluation scheme is provided for film quality detection.

[0029] The electrochemical corrosion parameters of the application are controllable, without complex professional skills, the natural corrosion of 14 days is shortened to within 3 days, the detection efficiency is improved, and the film is ensured without falling risk. BRIEF DESCRIPTION OF DRAWINGS

[0030] In order to more clearly illustrate the technical solutions of the present application, the drawings needed in the embodiments will be briefly introduced below. Obviously, the drawings described below are only some embodiments of the present application, and other drawings can also be obtained by those skilled in the art without creative labor.

[0031] Figure 1 is the electrochemical corrosion process schematic diagram of the film to be detected of the present application;

[0032] Figure 2 is the optical microscope image of 10nm film before and after corrosion in example 2 of the present application;

[0033] Figure 3 is the SEM image of 10nm / 35nm thick film after corrosion in example 2 of the present application;

[0034] Figure 4 is the optical microscope image of 10nm / 35nm thick film after corrosion in example 3 of the present application;

[0035] Figure 5 is the 10nm / 35nm thick film defect distribution diagram in example 8 of the present application. DETAILED DESCRIPTION

[0036] The technical solutions of the present application will be described below in conjunction with specific embodiments of the present application. Obviously, the described embodiments are only some of the embodiments of the present application, not all. Based on the embodiments in the present application, all other embodiments obtained by those skilled in the art without creative labor are within the scope of protection of the present application.

[0037] This embodiment provides a film defect detection method based on electrochemical corrosion, comprising the following steps:

[0038] S1, electrochemically etching the film to be detected to expand the original small size defects on the film to the range of observable size defects by optical microscope; wherein the electrochemical etching process is shown in the following figure Figure 1 ;

[0039] S2, obtaining the surface image of the film after etching by optical microscope, identifying and counting the defect morphology, size and distribution by image analysis, and calculating the original defect information of the film.

[0040] As a preferred embodiment, in S1, the etching process is specifically: placing the film to be detected in an etching environment with humidity of 85%-95% for electrochemical etching. The etching environment is subjected to strengthening treatment to increase the etching rate, and the strengthening treatment is one or more of heating, salt water spraying and applying electric field. The temperature of the heating is 70-80℃; the salt water is a sodium chloride aqueous solution with a concentration of 8-10wt%; and the voltage of the electric field is 0.5-2V.

[0041] As a preferred embodiment, the material of the film to be detected is molybdenum, titanium or aluminum oxide; and the thickness of the film to be detected is 1-500nm.

[0042] As a preferred embodiment, when the thickness of the film to be detected is <15nm, the defects after etching are voids, and the identification and counting of the defect morphology, size and distribution by image analysis and the calculation of the original defect information of the film include: using adaptive threshold segmentation algorithm and morphological operation to outline the void profile in the image, i.e. accurately outlining the void profile by identifying the distinct color contrast between the film area (dark blue) and the void area (light purple of the silicon wafer) in the optical image, generating a defect distribution heat map based on the magnification relationship of the optical microscope, and obtaining the density of the original defects through the defect distribution heat map.

[0043] As a preferred embodiment, when the thickness of the film to be detected is ≥15nm, the defects after etching are cracks, and the identification and counting of the defect morphology, size and distribution by image analysis and the calculation of the original defect information of the film include: using directional filtering and skeleton extraction method to track the crack path, i.e. identifying the crack and the film area by color contrast, restoring the defect to the center according to the crack shape (skeleton extraction algorithm simplifies the crack to a single-pixel-width skeleton structure, and since the crack is formed by expanding the original linear defect, the position of the crack skeleton is highly consistent with the position of the original linear defect), quantifying the total length, average width and orientation distribution of the crack based on the magnification relationship of the optical microscope, and obtaining the position distribution of the original defect through the orientation distribution of the crack.

[0044] Example 1

[0045] This example relates to a method for preparing a Mo film to be detected, which includes two stages of silicon substrate pretreatment and electron beam evaporation to prepare Mo film.

[0046] The silicon substrate pretreatment stage, first take the whole piece of single-sided silicon oxide wafer on the qualitative filter paper, with a diamond pen and tweezers along the preset mark cutting into 10x10mm 2 ~20x20mm 2 The sample; then the cut silicon wafer is sequentially placed into acetone solution, anhydrous ethanol and deionized water for ultrasonic treatment, and the ultrasonic parameters are all 25KHz, 300W. The ultrasonic treatment in acetone solution for 5min is to remove the oil and organic contaminants on the surface of the substrate, the ultrasonic treatment in anhydrous ethanol for 5min is to remove the residual acetone, and the ultrasonic treatment in deionized water for 5min is to remove the water-soluble impurities. Finally, the surface moisture of the silicon wafer is blown dry by a nitrogen gun (flow rate 5-10L / min) at an angle of 45°, and then the silicon wafer is placed on a heating platform at 100℃ for 10min to completely remove the residual solvent and water vapor, thereby obtaining a clean substrate for standby use.

[0047] The electron beam evaporation preparation of Mo thin film stage, first, a tungsten crucible (resistant to high temperature ≥3000℃) is used to carry Mo flake, and 1mm thick Mo flake is uniformly laid on the bottom of the crucible to ensure uniform heating area; then the vacuum pump of the coating instrument is started, and the pressure in the vacuum chamber is reduced to 1x10 -4 Pa (to avoid oxidation of the film material), the electron beam acceleration voltage is set to 5-10kV, the beam density is adjusted to local melting of the Mo flake (the melting state is confirmed by observation window), and the film thickness is monitored in real time by a crystal oscillator sensor, and the evaporation rate is stabilized at 0.02-0.03nm / s (to avoid rate fluctuations leading to uneven film layers); then the clean substrate is fixed on the sample holder and moved into the vacuum chamber, and the electron beam evaporation is started to prepare Mo thin films with thicknesses of 10nm, 20nm, 25nm and 35nm respectively; after the coating is completed, the electron beam is turned off, and the vacuum chamber is naturally cooled to room temperature (about 2h), the thin film sample is taken out and placed in a desiccator for storage to avoid moisture.

[0048] Example 2

[0049] This example relates to the electrochemical corrosion (room temperature high humidity environment) of the Mo thin film sample obtained in Example 1:

[0050] One piece of Mo thin film sample with a thickness of 10nm and 35nm respectively is placed in a room temperature, 90%RH environment without applying any acceleration measures; macroscopic observation is carried out every 24 hours by using an optical microscope, and the color change of the sample surface (such as whether yellowing or spot appearing) and the time and position of the first appearance of the corrosion point are recorded. After 14 days of continuous observation, observable black holes appear on the surface of the 10nm thin film, referring to the film image before corrosion in Figure 2 , and Figure 2Image of 10 nm thick Mo film after 14 days of corrosion; fine and scattered micro-cracks appeared on the surface of the film, verifying the mechanism of "electrochemical corrosion preferentially occurring at defects", and obtaining optical microscope images. At the same time, after 14 days, defects were identified by SEM: first, the sample after 14 days of corrosion was fixed on the SEM sample stage and gold was sprayed (thickness 5 nm) to enhance conductivity; the SEM vacuum was set to 5 x 10 -5 Pa, and the acceleration voltage was 15 kV; during verification, first, low magnification (100x-500x) was used for positioning, and the corresponding area was quickly found according to the defect position of the optical microscope image, then high magnification (1000x-5000x) was used for observation to confirm the consistency of the defect morphology (cavity / crack), size and optical microscope image, and to reveal microscopic details (such as cavity bottom substrate exposure, crack branching); the verification results are referenced Figure 3 Image of 10 nm thick Mo film after 14 days of corrosion (cavity defect), and Figure 3 Image of 35 nm thick Mo film after 14 days of corrosion (crack defect), the defect matching degree of SEM and optical microscope image is extremely high, proving the reliability of the optical microscope image, and no TEM is needed to realize defect identification.

[0051] Example 3

[0052] This example relates to the electrochemical corrosion (high temperature and high humidity environment) of the Mo film sample obtained in Example 1:

[0053] A high temperature and high humidity corrosion environment was used, with the core parameters being humidity 90% RH and temperature 75°C; during operation, the 10 nm and 35 nm thick Mo film samples were maintained at the set temperature and humidity for electrochemical corrosion, macroscopic observation was performed using an optical microscope (50x magnification), the color change of the sample surface (such as yellowing, appearance of spots) and the time and position of the first appearance of corrosion points were recorded, the development time was 72 h, optical microscope images were obtained, and reference was made to Figure 4 Image of 10 nm thick Mo film after 14 days of corrosion and Figure 4 Image of 35 nm thick Mo film after 14 days of corrosion.

[0054] Example 4

[0055] This example relates to the electrochemical corrosion (salt water spray environment) of the Mo film sample obtained in Example 1:

[0056] A salt water spray corrosion environment was used, with the core parameters of the salt water spray being 10 wt% NaCl solution and the solution being replaced every 12 hours; during operation, 10 wt% NaCl solution was first prepared and sprayed at room temperature in a 90% RH environment, and the 10 nm thick Mo film sample was placed in the spray for electrochemical corrosion; after the corrosion was completed, the sample was rinsed with deionized water for 10-30 seconds and dried with nitrogen, the development time was 48 h, and optical microscope images were obtained.

[0057] Example 5

[0058] This embodiment relates to the electrochemical corrosion (applied electric field environment) of the Mo thin film sample obtained in Example 1:

[0059] A weak electric field was applied to place the 10 nm film in a 90% RH environment at room temperature. A Keithley 2634B power supply was used, and copper electrodes were attached to both ends of the sample (to avoid damaging the film). A 1V DC voltage was applied to the power supply, and the development time was 24 hours to obtain optical microscope images.

[0060] Example 6

[0061] This embodiment involves the electrochemical corrosion of the Mo thin film sample obtained in Example 1 (high temperature and high humidity + salt spray environment):

[0062] The experiment employed a high-temperature, high-humidity environment with a brine spray. The core parameters were 90% RH humidity, 75℃ temperature, and 8wt% NaCl brine concentration. The sample was preheated for 1 hour before being continuously etched by brine spray. This resulted in clear network cracks appearing on the 35nm thick film within 48 hours. The development time was shortened by 24 hours compared to high-humidity heating alone, and optical microscope images were obtained.

[0063] Example 7

[0064] This embodiment relates to the electrochemical corrosion of the Mo thin film sample obtained in Example 1 (applied electric field + salt spray environment):

[0065] Using an environment of applied electric field and salt spray, with core parameters of room temperature and humidity of 90% RH, salt concentration of 8 wt% NaCl, and DC voltage of 0.8V, a 10nm thick film showed uniformly distributed voids within 24 hours. The development time was shortened by 24 hours compared to salt spray alone, and optical microscope images were obtained.

[0066] Example 8

[0067] This embodiment relates to the optical microscope image analysis and inference of the thin film after corrosion in Example 2:

[0068] (1) In Example 3, the 10nm thick film has mainly "point defects" such as pinholes and impurity introduction points. After electrochemical etching, these defects expand into voids visible under an optical microscope. The derivation logic based on the optical microscope images is as follows:

[0069] First, complete the optical microscope calibration: capture images at the same magnification (50×) as the test film, count the number of pixels corresponding to the feature size of the standard calibration film, and calculate the "actual physical area (S, unit: μm) corresponding to a single pixel". 2 " / pixel"), which converts the optical mirror pixel data into actual physical parameters;

[0070] Subsequently, optical microscopy images were acquired based on the etched thin film. An adaptive threshold segmentation algorithm was used to distinguish between voids and the film background (identifying distinct colors in the image: film areas (dark blue) and void areas (pale purple of the silicon wafer)). Morphological operations were then performed to fill in noise gaps at the void edges, outlining the void contours and generating a defect distribution heatmap. Figure 5 In the middle a, count its pixel count (N) i Based on the calibrated S, calculate the actual physical area (A) of all developing voids within the region. i =N i ×S), while simultaneously counting the total number of pixels in the detection area (N) total ), calculate the actual area of ​​the detection area (A) total =N total ×S), and then the density of the defects after development (ρ=A) is obtained. i / A total ).

[0071] (2) In Example 3, the 35nm thick film's original defects were mainly "line defects" such as microcracks. After electrochemical etching, these defects expanded along the stress path to become cracks visible under an optical microscope. The derivation logic based on the optical microscope images needs to be adjusted to match the characteristics of the line defects.

[0072] Similarly, after completing the optical microscope calibration, optical microscope images of the etched thin film are acquired. Color contrast is used to identify cracks and thin film regions. Directional filtering enhances the linear characteristics of the cracks and suppresses background interference such as film surface textures. Then, skeletonization extraction simplifies wide cracks into a single-pixel-width "crack skeleton." Based on the crack shape, the defect is reconstructed towards the center (the skeletonization extraction algorithm simplifies wide cracks into a single-pixel-width "crack skeleton"), generating a crack defect distribution map. (Refer to...) Figure 5 In the middle b, the total number of pixels in the skeleton (N) is counted. s ), and the converted actual physical area (A) s =N s ×S Count of total pixels in the detection area (N) total ), calculate the actual area of ​​the detection area (A) total =N total ×S), then the defect density ρ=A s / A total .

[0073] The present invention has been described in detail above with reference to specific embodiments and exemplary examples; however, these descriptions should not be construed as limiting the present invention. Those skilled in the art will understand that various equivalent substitutions, modifications, or improvements can be made to the technical solutions and embodiments of the present invention without departing from the spirit and scope of the invention, and all such modifications and improvements fall within the scope of the present invention. The scope of protection of the present invention is defined by the appended claims.

Claims

1. A method for detecting thin film defects based on electrochemical corrosion, characterized in that, Includes the following steps: S1. The thin film to be tested is subjected to electrochemical etching treatment so that the original small-sized defects on the thin film are expanded to defects within the size range that can be observed by an optical microscope. S2. Obtain images of the film surface after corrosion treatment using an optical microscope, identify and statistically analyze the defect morphology, size and distribution through image analysis, and deduce the original defect information of the film.

2. The thin film defect detection method based on electrochemical corrosion as described in claim 1, characterized in that, In S1, the corrosion treatment specifically involves placing the film to be tested in a corrosive environment with a humidity of 85%-95% for electrochemical corrosion.

3. The thin film defect detection method based on electrochemical corrosion as described in claim 2, characterized in that, The corrosive environment is subjected to enhanced treatment to increase the corrosion rate. The enhanced treatment is one or more of heating, salt spraying, and applying an electric field.

4. The thin film defect detection method based on electrochemical corrosion as described in claim 3, characterized in that, The heating temperature is 70-80℃; the brine is an aqueous sodium chloride solution with a concentration of 8-10wt%; and the electric field voltage is 0.5-2V.

5. The thin film defect detection method based on electrochemical corrosion as described in claim 1, characterized in that, The material of the film to be tested is molybdenum, titanium, or aluminum oxide.

6. The thin film defect detection method based on electrochemical corrosion as described in claim 1, characterized in that, The thickness of the film to be tested is 1-500 nm.

7. The thin film defect detection method based on electrochemical corrosion as described in claim 1, characterized in that, When the thickness of the film to be inspected is <15nm, the defects after etching are voids. Image analysis is used to identify and statistically analyze the defect morphology, size, and distribution, and to deduce the original defect information of the film, including: An adaptive threshold segmentation algorithm and morphological operations are used to delineate the contours of each cavity in the image. Based on the magnification relationship of the optical microscope, a heat map of defect distribution is generated.

8. The thin film defect detection method based on electrochemical corrosion as described in claim 7, characterized in that, The density of the original defects is obtained by using a defect distribution heatmap.

9. The thin film defect detection method based on electrochemical corrosion as described in claim 1, characterized in that, When the thickness of the film to be inspected is ≥15nm, the defect after etching is a crack. The image analysis identifies and statistically analyzes the defect morphology, size, and distribution, and deduces the original defect information of the film, including: Directional filtering and skeleton extraction methods are used to track crack paths, and the total crack length, average width and orientation distribution are quantified based on the calibration magnification relationship of optical microscopes.

10. The thin film defect detection method based on electrochemical corrosion as described in claim 9, characterized in that, The original defect location distribution is obtained by the orientation distribution of the cracks.

Images