A foreign matter detection method of a display panel

Abstract

The present disclosure provides a foreign matter detection method of a display panel, specifically comprising the following steps: laser irradiation is performed on a surface area of a to-be-tested sample to obtain a first characteristic spectrum; the first characteristic spectrum is compared with a preset specimen library; if the first characteristic spectrum is consistent with any spectrum in the preset specimen library, the to-be-tested sample is labeled as a normal sample; if the first characteristic spectrum is inconsistent with all the spectra in the preset specimen library, the to-be-tested sample is labeled as an abnormal sample; a roughening treatment is performed on a polyimide surface of the abnormal sample to obtain a rough surface; surface-enhanced Raman scattering detection is performed on the rough surface to obtain a Raman spectrum of the abnormal sample; the Raman spectrum is compared and analyzed with a spectrum library to determine the composition of the foreign matter; the present disclosure solves the problem that foreign matters cannot be traced due to the fact that the diameters of the foreign matters do not reach the detection standard or trace substances are left.

Description

Technical Field

[0001] This disclosure relates to the field of foreign object detection technology, and in particular to a method for detecting foreign objects in a display panel. Background Technology

[0002] In modern electronics manufacturing, especially in display panel (also known as display screen) production, product quality stability and reliability are paramount. As displays rapidly evolve towards higher resolution, thinner profiles, and greater multifunctionality, the requirements for impurity control during production have reached unprecedented levels of stringency. The presence of foreign matter is a key factor affecting display screen quality. Foreign matter is frequently encountered at various processing stages. Among many conventional countermeasures, machine matching based on location is a common method. Based on meticulous production process records and advanced tracking systems, the location of a foreign matter can be quickly linked to the corresponding production machine. Another commonly used method is to trace the source by analyzing the composition of foreign matter through sampling. Using specialized microanalytical instruments, such as scanning electron microscopes paired with energy dispersive spectroscopy (EDS) or Fourier transform infrared spectroscopy (FTIR), the collected foreign matter samples are thoroughly analyzed. By accurately determining the elemental composition and chemical bond characteristics of the foreign matter, the analysis can trace back to the raw material supplier, specific process steps in the production process, and even down to a specific batch of raw materials or a processing operation within a specific time period, thus enabling targeted improvement measures.

[0003] However, the problem of foreign objects in real-world production scenarios is complex, and existing traceability methods have many limitations. On the one hand, current analytical techniques only support effective analysis of foreign object samples with a diameter ≥25μm. When the diameter of the foreign object is at the submicron or even nanometer level, conventional electron microscopes struggle to focus and capture it clearly, and the signals obtained from energy dispersive spectroscopy (EDS) are weak and easily drowned out by background noise, failing to provide reliable evidence for composition determination, making it impossible to trace the origin of many minute foreign objects. Furthermore, when liquid crystal display units or organic light-emitting diode display units reach the display panel manufacturing stage, the internal adhesive encapsulates the foreign object, forming a tight composite structure. During subsequent sectioning for in-depth analysis, the adhesive not only severely hinders direct observation of the foreign object, resulting in blurred images, but also, when using methods such as EDS, a large number of elemental signals from the adhesive itself mix with the foreign object signals, interfering with the analysis results after sectioning, ultimately rendering the traceability results meaningless.

[0004] Furthermore, trace amounts of substances frequently occur throughout the entire display screen manufacturing process, from the adsorption of minute impurities on the surface of raw materials to the residue of chemical reagents used in processing. Trace substances refer to substances present in extremely low amounts in chemical and biological environments, and therefore cannot be detected and analyzed by conventional detection techniques.

[0005] In summary, due to the limitations of current detection technology, existing detection methods cannot comprehensively and effectively cover all potential quality risks when faced with foreign objects or trace substances that do not meet the standard diameter. Therefore, there is an urgent need to develop new detection methods that are more sensitive and accurate. Summary of the Invention

[0006] In view of this, the purpose of this disclosure is to propose a foreign object detection method for display panels to solve the problem of foreign objects whose diameter does not meet the detection standard or whose trace substances remain, making it impossible to trace the source of foreign objects.

[0007] For the purposes described above, this disclosure discloses a method for detecting foreign objects in a display panel, specifically including the following steps:

[0008] The surface area of ​​the sample to be tested is irradiated with a laser to obtain the first characteristic spectrum;

[0009] Compare the first characteristic spectrum with a preset sample library:

[0010] If the first characteristic spectrum matches any spectrum in the preset sample library, then the sample to be tested is calibrated as a normal sample.

[0011] If the first characteristic spectrum is inconsistent with all spectra in the preset sample library, the sample to be tested will be labeled as an abnormal sample.

[0012] The polyimide surface of the abnormal sample was roughened to obtain a rough surface;

[0013] Surface-enhanced Raman scattering detection was performed on the rough surface to obtain the Raman spectrum of the abnormal sample;

[0014] The Raman spectrum is compared and analyzed with the spectral library, including...

[0015] The Raman spectrum is preprocessed to obtain an optimized Raman spectrum;

[0016] Extract key feature data from the optimized Raman spectrum;

[0017] The composition of the foreign substance is determined by matching the key feature data with known spectra in the spectral library.

[0018] Furthermore, it also includes the following steps:

[0019] The roughening process includes primary roughening and secondary roughening.

[0020] Furthermore, the initial roughening process specifically includes:

[0021] A first rough surface is obtained by spraying a mixture of 4,4-diaminodiphenylmethane and 4,4-bisphthalic dianhydride onto the polyimide surface of the abnormal sample.

[0022] Furthermore, the concentration of the mixture is 4-6%;

[0023] Preferably, the concentration of the mixture is 5%.

[0024] Furthermore, the mixture covers the polyimide surface of the abnormal sample for 12-18 minutes;

[0025] Preferably, the mixture covers the polyimide surface of the abnormal sample for 15 minutes.

[0026] Furthermore, the secondary roughening process specifically includes:

[0027] Based on the first rough surface, a mixture of carbon dioxide and acetylene gas is ionized under high pressure in an argon environment to obtain an ion beam. The ion beam is then used to roughen the first rough surface of the abnormal sample a second time to obtain a second rough surface.

[0028] Furthermore, the high voltage is 650-900V.

[0029] Preferably, the high voltage is 700V.

[0030] Furthermore, it also includes:

[0031] After the second roughening treatment, the morphology of the second roughened surface is inspected using an atomic force microscope. If the root mean square value of the surface roughness does not reach the range of 50-80 nm, the roughening treatment steps are repeated until the requirements are met.

[0032] Furthermore, surface-enhanced Raman scattering (SERS) is performed on the rough surface to obtain the Raman spectrum of the anomalous sample, specifically including:

[0033] The laser parameters are optimized, and the laser beam is focused to irradiate the second rough surface. A high-precision spectral acquisition device captures the Raman spectrum generated by the abnormal sample in real time.

[0034] Compared with the prior art, this application has the following technical effects:

[0035] The foreign matter detection method proposed in this application roughens the surface of the sample to be tested (such as the surface of polyimide), making the microstructure of the polyimide surface more complex, increasing its surface area, significantly enhancing its surface Raman effect, effectively capturing minute foreign matter on the polyimide surface, improving the accuracy of energy dispersive spectroscopy analysis, realizing the accurate detection and traceability of minute foreign matter or trace substance residues, thereby accurately determining key information such as the elemental composition and chemical bond characteristics of foreign matter, and tracing back to the source.

[0036] The surface of the sample to be tested (such as the surface of polyimide) undergoes an initial roughening treatment, followed by a secondary roughening treatment to further refine the microstructure and form a multi-layered uneven surface. This significantly increases the surface area and Raman signal intensity, ensuring effective capture of nanoscale foreign matter and significantly improving detection sensitivity and traceability accuracy. Simultaneously, the speed and effectiveness of the secondary roughening treatment are greatly improved, making the overall detection process more efficient and significantly shortening the traceability cycle.

[0037] By collecting spectra of various items, a rich and comprehensive spectral library is constructed. During SERS detection, this library allows for the precise location of characteristic peaks corresponding to target substances. The extensive spectral reserve enables the detection system to familiarize itself with various possible peak shapes, positions, and intensity combinations in advance. This allows for the rapid identification of target peaks from mixed signals, even in the presence of weak interference, reducing false positives and significantly improving the accuracy and reliability of the detection. Attached Figure Description

[0038] To more clearly illustrate the technical solutions in this disclosure or related technologies, the accompanying drawings used in the description of the embodiments or related technologies will be briefly introduced below. Obviously, the accompanying drawings described below are only embodiments of this disclosure. For those skilled in the art, other drawings can be obtained based on these drawings without creative effort.

[0039] Figure 1 This is a schematic diagram of a foreign object detection method for a display panel in one embodiment of the present invention;

[0040] Figure 2 This is a schematic diagram of the test results of relevant foreign object detection methods in the prior art;

[0041] Figure 3 This is a schematic diagram showing the comparison and analysis of Raman spectra and spectral libraries for a foreign object detection method for a display panel in one embodiment of the present invention. Detailed Implementation

[0042] To make the objectives, technical solutions, and advantages of this disclosure clearer, the following detailed description is provided in conjunction with specific embodiments and the accompanying drawings.

[0043] It should be noted that, unless otherwise defined, the technical or scientific terms used in the embodiments of this disclosure should have the ordinary meaning understood by one of ordinary skill in the art to which this disclosure pertains. The terms "first," "second," and similar terms used in the embodiments of this disclosure do not indicate any order, quantity, or importance, but are merely used to distinguish different components. Terms such as "comprising" or "including" mean that the element or object preceding the word encompasses the elements or objects listed following the word and their equivalents, without excluding other elements or objects. Terms such as "connected" or "linked" are not limited to physical or mechanical connections, but can include electrical connections, whether direct or indirect. Terms such as "upper," "lower," "left," and "right" are used only to indicate relative positional relationships; when the absolute position of the described object changes, the relative positional relationship may also change accordingly.

[0044] In modern electronics manufacturing, especially in display screen production, product quality stability and reliability are paramount. As displays rapidly evolve towards higher resolution, thinner profiles, and greater multifunctionality, the requirements for impurity control during production have reached unprecedented levels of stringency. The presence of foreign objects is a key factor affecting display screen quality. Foreign objects are frequently encountered at various processing stages of the production process. Among many conventional countermeasures, machine matching based on location is a common method. Based on detailed production process records and advanced tracking systems, once the location of a foreign object is detected, it can be quickly linked to the corresponding production machine. For example, when a foreign object is detected in a specific area of ​​a semi-finished display screen, by tracing back through the various points the semi-finished product passed through on the production line, the machine that may have introduced the foreign object can be precisely located. This allows for the investigation of the machine's mechanical wear, the integrity of the lubrication system, and the cleanliness of the surrounding environment to pinpoint the source of the foreign object. Another commonly used method is to sample and analyze the composition of the foreign object to achieve source tracing. With the help of specialized microanalytical instruments, such as scanning electron microscopes paired with energy dispersive spectroscopy (EDS) or Fourier transform infrared (FTIR) spectrometers, the collected foreign matter samples are thoroughly analyzed. By accurately determining key information such as the elemental composition and chemical bond characteristics of the foreign matter, the analysis can trace back to the raw material supplier, specific process steps in the production process, and even down to a specific batch of raw materials or a specific time period, thereby enabling targeted improvement measures to be implemented.

[0045] The inventors, through long-term practical work, discovered the following problems in the relevant technology: Current analytical techniques only support effective analysis of foreign matter samples with a diameter ≥25μm. When the diameter of the foreign matter is at the submicron or even nanometer level, conventional electron microscopes struggle to focus and capture it clearly. The signals obtained from energy dispersive spectroscopy (EDS) are weak and easily drowned out by background noise, failing to provide reliable evidence for composition determination and making it impossible to trace the origin of many minute foreign matter. Moreover, when liquid crystal display units or organic light-emitting diode display units reach the display panel manufacturing stage, the internal adhesive encapsulates the foreign matter, forming a tight composite structure. During subsequent sectioning for in-depth analysis, the adhesive not only severely hinders direct observation of the foreign matter, resulting in blurred images, but also, when using methods such as EDS, a large number of elemental signals from the adhesive itself are mixed with the foreign matter signals, interfering with the analysis results after sectioning, ultimately rendering the traceability results meaningless.

[0046] For example, one related foreign object detection method directly performs infrared detection on a substrate containing the foreign object, with the foreign object approximately 10 μm from the substrate surface and the substrate thickness 4-5 mm. There are two detection points: one on the substrate and one on the foreign object. FTIR attenuated total reflectance detection is used, and the results are shown in [Figure number missing]. Figure 1 As can be seen, the spectra of the base part and the foreign body part are basically the same, so the spectrum of the foreign body cannot be obtained, and therefore the composition of the foreign body cannot be analyzed for tracing its source.

[0047] For the reasons mentioned above, this application provides a method for detecting foreign objects in a display panel, such as... Figure 2 As shown, the specific steps include:

[0048] S1: Irradiate the surface area of ​​the sample to be tested with a laser to obtain the first characteristic spectrum;

[0049] The first characteristic spectrum is the spectrum obtained after laser irradiation of the surface area of ​​the sample under test. It reflects the spectral characteristics of a single sample and is a specific spectral data point. The first characteristic spectrum originates from a specific sample under test and is obtained by laser irradiation and spectral acquisition under specific conditions. The selection of samples should cover different batches and different process stages to ensure representativeness and that the selected samples can comprehensively reflect the diverse characteristics of similar substances under normal conditions. The preparation of the sample must strictly follow standardized procedures to ensure a clean and uncontaminated surface. In this embodiment, high-precision slicing technology is used to peel off the display panel sample and cut the layers to a preset thickness, ensuring a smooth and flat cut surface and avoiding the introduction of additional interference.

[0050] To obtain the exact same laser irradiation parameters and spectral acquisition conditions as in the pre-established sample library, the corresponding surface area of ​​the sample to be tested is subjected to stable and uniform laser irradiation. Then, a high-precision spectroscopic analysis instrument is used again to obtain the accurate first characteristic spectrum. This ensures that the acquisition process of the first characteristic spectrum is highly consistent with the spectral acquisition in the pre-established sample library, eliminating the influence of environmental factors and equipment differences, and ensuring data comparability.

[0051] S2: Compare the first characteristic spectrum with a pre-defined sample library:

[0052] S3: If the first characteristic spectrum matches any spectrum in the preset sample library, then the sample to be tested is calibrated as a normal sample.

[0053] S4: If the first characteristic spectrum is inconsistent with all spectra in the preset sample library, the sample to be tested will be labeled as an abnormal sample.

[0054] A laser beam is used to scan the surface of qualified samples point by point, accurately capturing the spectral information at each point to obtain a pre-defined sample library. The pre-defined sample library employs a laser beam whose wavelength, energy, and pulse mode have been precisely calibrated and adapted to the material characteristics of the samples. It is precisely focused on the surface detection area of ​​the qualified samples. A spectral acquisition device with ultra-high sensitivity and precise wavelength resolution is used to simultaneously collect the scattered light generated by the interaction between the laser and the sample. After complex data processing algorithms to remove noise interference and calibrate wavelength deviations, highly accurate and repeatable spectra are obtained. The pre-defined sample library is a database composed of the spectra of multiple qualified samples, representing the spectral characteristic range and standard of a batch of samples under normal conditions. It serves as a reference spectral data set. The data in the pre-defined sample library comes from numerous screened qualified samples that underwent rigorous selection and testing during the establishment of the sample library to ensure they represent a normal state. As a reference standard, the pre-defined sample library provides a basis for judging the normality of unknown samples. By comparing it with primary characteristic spectra, it determines whether the sample to be tested has any abnormalities, and the degree and direction of the abnormality.

[0055] It should be noted that comparing the first characteristic spectrum with a pre-set sample library can be based on comparing spectral characteristic parameters. This involves identifying parameters such as the position, intensity, and width of characteristic peaks in the first characteristic spectrum and the spectra in the pre-set sample library. Differences are then compared between the characteristic peaks of the sample being tested and their corresponding peaks in the normal sample library, such as whether the characteristic peaks have shifted, whether the intensity has changed abnormally, or whether the width has broadened or narrowed. For example, in infrared spectroscopy, the positions of characteristic absorption peaks of specific functional groups are relatively fixed. If the position of a characteristic peak of a certain functional group in the second characteristic spectrum differs significantly from the standard position in the normal sample library, it may indicate an anomaly in the sample. The correlation coefficient between the first characteristic spectrum and each spectrum in the pre-set sample library can also be calculated; the closer the correlation coefficient is to 1, the more similar the spectra are. By setting an appropriate threshold, when the correlation coefficient is below this threshold, a significant difference is considered between the first characteristic spectrum and the spectra in the pre-set sample library.

[0056] In this embodiment, an artificial neural network model can be constructed, using spectral data from a pre-defined sample library as input for training. The network's weights and thresholds can be adjusted to accurately classify spectra. When the first feature spectrum is input into the trained neural network, the model quickly performs in-depth analysis based on its previously learned knowledge. It automatically uncovers subtle features in the spectrum and compares the differences with known normal spectra. Through the network's complex computation and reasoning process, a clear result is ultimately output to determine whether the sample is normal.

[0057] This series of rigorous comparison processes quickly identifies samples containing foreign matter, i.e., abnormal samples, facilitating subsequent component identification and analysis of abnormal samples and improving the accuracy and efficiency of sample detection.

[0058] S5: Roughen the surface of the abnormal sample (polyimide) to obtain a rough surface;

[0059] There are various techniques for roughening the surface of polyimide, including chemical etching, mechanical polishing, and plasma treatment. The optimal method is selected based on the sample material and the required roughness to ensure a uniform and controllable roughened surface, thereby improving the intensity and stability of subsequent detection signals. In this embodiment, chemical etching is used to refine the polyimide surface.

[0060] It should be noted that roughening treatment includes primary roughening treatment and secondary roughening treatment.

[0061] Specifically, the initial roughening treatment includes spraying a mixture of 4,4-diaminodiphenylmethane and 4,4-bisphthalic dianhydride onto the polyimide surface of the abnormal sample to obtain a first roughened surface. The concentration of the mixture is 4-6%. When the concentration is within this range, it ensures a suitable roughness on the polyimide surface without affecting the subsequent Raman scattering detection effect due to excessively high or low concentrations. Preferably, the concentration of the mixture is 5%; at this preferred concentration, 4,4-diaminodiphenylmethane and 4,4-bisphthalic dianhydride achieve optimal synergistic effects, and upon contact with the polyimide surface, they react precisely with the polyimide surface molecules due to their unique chemical activity. On the one hand, these substances can moderately erode the polyimide surface, creating uniformly distributed pits and protrusions with suitable microscale, forming an ideal roughened structure. On the other hand, these reaction products can also form chemical bonds with the polyimide surface, enhancing the adhesion of the mixture to the surface and providing a stable surface environment for subsequent Raman detection. In this embodiment, the mixture is uniformly covered on the polyimide surface by spraying. Compared to other roughening methods, such as smearing or soaking, spraying is more efficient, completing the treatment of the polyimide surface in a short time, thus improving efficiency. Moreover, spraying can also ensure that the mixture covers the polyimide surface more uniformly. By adjusting the parameters of the nozzle, such as nozzle orifice diameter, spraying pressure, and moving speed, the distribution of the mixture can be controlled to ensure a relatively consistent liquid film thickness across the entire polyimide surface, thereby ensuring the uniformity of the roughening treatment and avoiding local over- or under-treatment. In addition, spraying can precisely control the amount of mixture applied. Based on the surface area of ​​the polyimide and the desired treatment effect, the spraying flow rate and time can be accurately adjusted to achieve the optimal coverage of the mixed solution, ensuring both treatment effectiveness and avoiding material waste.

[0062] The duration of the mixed solution covering the polyimide surface of the abnormal sample is 12-18 minutes; preferably, it is 15 minutes. During the initial 12 minutes, the two key components, 4,4-diaminodiphenylmethane and 4,4-bisphthalic dianhydride, begin to penetrate the polyimide surface molecules through their unique chemical activity, attempting to establish connections with the active sites on the surface and initiating a preliminary chemical reaction. If the time is extended to 18 minutes, although the chemical reaction continues, some negative effects begin to appear. Excessive reaction may lead to uneven surface roughness and excessive corrosion in some areas, disrupting the originally regular microstructure. Simultaneously, excessive accumulation of reaction products may, to some extent, obscure the true information of the polyimide surface, interfering with subsequent Raman detection and reducing the accuracy of the detection results. At 15 minutes, the two components in the mixture bonded more tightly to the polyimide surface molecules, the chemical reaction rate reached a delicate equilibrium, more active sites were successfully activated, the intermolecular cross-linking reaction spread out fully, and the roughness of the polyimide surface further increased, forming uniformly distributed and perfectly positioned micro-pits and protrusions. These pits and protrusions not only provided an ideal scattering interface for subsequent Raman detection but also enriched the vibrational modes of surface molecules, which is beneficial for capturing more accurate Raman signals.

[0063] By initially roughening the surface of polyimide, the microstructure of the polyimide surface becomes more complex, increasing its surface area and significantly enhancing its Raman effect. This effectively captures minute foreign matter on the polyimide surface, improves the accuracy of energy dispersive spectroscopy analysis, and enables precise detection and tracing of minute foreign matter or trace substance residues. This allows for the accurate determination of key information such as the elemental composition and chemical bond characteristics of foreign matter, tracing back to the raw material supplier and specific process steps in the production process, and implementing targeted improvement measures.

[0064] Subsequently, the polyimide surface underwent a secondary roughening treatment. Specifically, the secondary roughening treatment included:

[0065] Based on the first rough surface, a mixture of carbon dioxide and acetylene gas is ionized under high pressure in an argon environment to obtain an ion beam. The ion beam is then used to roughen the first rough surface of the abnormal sample a second time to obtain a second rough surface.

[0066] Argon is an inert gas, and creating an argon atmosphere in this process is primarily to provide a stable environment. It prevents unnecessary chemical reactions between the mixed gas and other gases such as oxygen in the air during ionization, ensuring that the ionization process involves only the carbon dioxide and acetylene mixture. This improves the purity and stability of the ion beam composition, which is beneficial for subsequent stable and repeatable secondary roughening treatment of the polyimide surface. The carbon dioxide and acetylene mixture is precisely introduced into this argon atmosphere. The volume percentage of carbon dioxide in the mixture is 60%-70%. This ratio plays an irreplaceable role in optimizing the composition of the ion beam, thereby greatly enhancing the Raman effect. When the volume percentage of carbon dioxide is within this range, the presence of carbon dioxide molecules significantly alters the energy distribution and ion composition of the ion beam during generation. The high stability and specific electronic structure of carbon dioxide molecules allow them to interact with acetylene molecules during ionization, synergistically regulating the characteristics of the generated ion beam. On the one hand, an appropriate amount of carbon dioxide can suppress the generation of some unstable ions, making the ion beam purer and reducing stray ions that may interfere with Raman detection. On the other hand, it can guide the ion beam in a direction that is conducive to Raman scattering, enhancing the intensity and clarity of the Raman signal, and making the final detected Raman spectrum more accurate and reliable.

[0067] The ion beam sweeps across the first rough surface at a precise angle and uniform speed, colliding violently with the surface molecules. On one hand, it can precisely strip away some relatively thin atomic or molecular layers from the surface, further deepening the already rough surface with microscopic landform features such as grooves and protrusions. On the other hand, the impact of the ion beam can also induce chemical reactions such as rearrangement and cross-linking of surface molecules, increasing roughness while optimizing the surface's chemical activity, creating more favorable conditions for subsequent surface-enhanced Raman scattering detection.

[0068] The high voltage range is 650V-900V. In this embodiment, a 700V high voltage is used to ionize the mixed gas. By applying high voltage, the mixture of carbon dioxide and acetylene gains sufficient energy, causing the atoms and molecules in the gas to ionize, forming a plasma state, which is then converted into an ion beam. During this process, factors such as the intensity of the high voltage, the shape and spacing of the electrodes, and the concentration and flow rate of the gas all affect the ionization effect and the properties of the ion beam, such as the energy, velocity, and density of the ions. The converted ion beam has high energy, and when it strikes the polyimide surface, it alters the microstructure of the polyimide surface, creating nanoscale or micrometer-scale roughness. This deep roughening effect increases the surface area of ​​the polyimide surface, and the change in its microstructure enhances the collection and amplification of Raman scattered light. When this roughened polyimide surface is irradiated with a laser, the Raman scattering signal of the adsorbed material on the surface is significantly enhanced due to the surface roughness and the enhancement effect of the local electromagnetic field. This improves the sensitivity and accuracy of foreign object detection, and can more effectively detect trace foreign objects that may exist on the polyimide surface. The type and properties of the foreign objects can be identified by analyzing their Raman spectral characteristics.

[0069] The secondary roughening treatment further refined the microstructure of the polyimide surface, making the grooves deeper and the protrusions more varied, forming highly complex nano- and even micro-scale topography. As a result, when laser light irradiates the substrate surface, the light is repeatedly reflected and scattered within these microstructures, greatly extending the interaction path between the light and heavy metal ions, leading to an unprecedented enhancement of the Raman signal. In actual testing, compared to the untreated polyimide substrate, the treated substrate showed a 10% increase in the Raman signal for heavy metal ions. 3 -105 times, the detection limit is reduced from ppm to ppb, which means that it can accurately capture heavy metal ions in extremely low concentrations. Even in environments with extremely low levels of pollution, it can quickly and accurately detect traces of heavy metal ions, effectively improving detection sensitivity.

[0070] In another embodiment, after the secondary roughening treatment, the morphology of the second roughened surface is inspected using an atomic force microscope. If the root mean square value of the surface roughness does not reach the range of 50-80 nm, the roughening treatment step is repeated until the requirement is met. Only by ensuring that this standard is met can a high-performance second roughened surface be provided for subsequent surface-enhanced Raman scattering detection, maximizing its role in enhancing Raman signals and accurately identifying foreign matter components.

[0071] S6: Perform surface-enhanced Raman scattering detection on the rough surface to obtain the Raman spectrum of the anomalous sample;

[0072] Specifically, the laser parameters are optimized, and the laser beam is used to focus and irradiate the second rough surface. A high-precision spectral acquisition device captures the Raman spectrum generated by the abnormal sample in real time.

[0073] Different wavelengths have drastically different effects on Raman scattering excitation. Choosing the wrong wavelength may result in a weak Raman signal or even ineffective excitation. Therefore, accurately selecting the appropriate laser wavelength for this anomalous sample is crucial. Simultaneously, laser power control is equally critical. Excessive power may damage the fragile sample surface structure, introduce unnecessary thermal noise to interfere with the spectral signal, and potentially induce nonlinear optical effects that mask the true Raman scattering information. Conversely, insufficient power will fail to adequately excite Raman scattering, resulting in insufficient signal intensity and hindering accurate analysis. Through repeated testing and simulation, a precise laser power value was set that ensures effective Raman scattering excitation while maintaining sample stability. Furthermore, parameters such as laser beam spot size and divergence angle also need to be optimized to ensure highly concentrated and uniform laser energy acting on the rough target surface. During acquisition, the acquisition frequency and integration time must be strictly controlled. Excessively high acquisition frequencies may lead to data redundancy and increase system load, while excessively low frequencies may miss crucial spectral changes. The integration time also requires careful consideration; too short a time will result in insufficient Raman signal intensity, while too long a time may introduce excessive ambient light interference or allow subtle variations in the sample itself to affect spectral accuracy. By carefully adjusting these parameters, we can ensure that the dynamic changes in the Raman spectrum of abnormal samples under laser excitation can be captured accurately and completely at every moment, providing a solid data guarantee for subsequent in-depth analysis of foreign matter composition.

[0074] S7: Compare and analyze the Raman spectrum with the spectrum library to determine the composition of the foreign substance.

[0075] Specifically, it includes the following steps:

[0076] S701: Preprocess the Raman spectrum to obtain an optimized Raman spectrum;

[0077] It should be noted that preprocessing includes removing noise interference from the Raman spectrum. Filtering algorithms such as wavelet transform are used to separate high-frequency noise from the original spectral data, making the spectral curve smoother and clearer, improving signal quality, and reducing noise interference. Simultaneously, the spectral data is normalized to unify spectral data from different intensity ranges into the same numerical interval, facilitating subsequent comparative analysis.

[0078] S702: Extracting key feature data from optimized Raman spectra;

[0079] It should be noted that key feature data includes peak positions, intensities, peak areas, and peak half-widths (FWHMs). These features are crucial for identifying the composition of substances, as different substances possess unique combinations of Raman characteristic peaks. For example, specific chemical bonds or functional groups will have specific characteristic peak positions in Raman spectra, and their intensity and shape can reflect the content and environmental information of these chemical bonds or functional groups. These features are quantified and encoded into a digital format that can be recognized and processed by computers, facilitating rapid matching and comparison with data in spectral libraries. For instance, specialized feature extraction algorithms can be used to accurately locate the center position of each characteristic peak and calculate its corresponding intensity value and area, storing these parameters in a specific data structure.

[0080] S703: Determine the composition of foreign matter by matching known spectra in the spectral library with key feature data.

[0081] A spectral library is a database that collects, organizes, and stores characteristic spectra of a large number of different substances. Key feature data from Raman spectra are input into the spectral library retrieval system. The system then uses a pre-defined retrieval algorithm to quickly find the known spectrum that best matches the Raman spectral characteristics of the anomalous sample within the vast library. The spectral library stores a massive amount of Raman spectral information for known substances, covering various fields such as organic compounds, inorganic compounds, and biomolecules. Retrieval algorithms may be based on methods such as matching characteristic peaks, analyzing the overall shape of the spectrum, or machine learning. For example, retrieval algorithms often use similarity-based methods, such as calculating correlation coefficients or Euclidean distances, to measure the similarity between the anomalous sample's spectrum and the spectra in the spectral library. Once a spectrum with high similarity is found, a detailed comparison of the characteristic peaks is conducted, including subtle shifts in position and relative changes in intensity, to ultimately determine the composition of the foreign substance.

[0082] Furthermore, a classification model can be established to match key feature data with known spectra in a spectral library. Machine learning algorithms, such as support vector machines, decision trees, and neural networks, are used to train the spectral data in the library to build a classification model. The Raman spectral feature data of the anomalous sample is input into the trained model, which then predicts the most likely category of the sample or which known spectrum it is most similar to based on the learned features and patterns. In this embodiment, a convolutional neural network model is constructed to match key feature data with known spectra in the spectral library. Key feature data of unknown spectra are input into the model for training. During training, loss functions and optimization algorithms are used to continuously adjust the model parameters to learn the mapping relationship between spectral features and material composition. After training, the model's performance is evaluated and optimized using a test dataset. Finally, the processed key feature data is input into the optimized model, which can then predict the most likely known spectrum and material composition corresponding to the spectrum based on the learned knowledge, thus achieving the matching.

[0083] By extensively collecting spectra of various materials, a rich and comprehensive spectral library has been constructed. During SERS (Surface Enhanced Raman Scattering) detection, when faced with complex spectral signals, the library allows for precise localization of the characteristic peaks corresponding to the target material. Because different materials possess unique Raman spectra, a large spectral reserve enables the detection system to familiarize itself with various possible peak shapes, positions, and intensity combinations in advance. This allows for rapid identification of the target peak from the mixed signal, even in the presence of weak interference, reducing false positives and significantly improving the accuracy and reliability of the detection.

[0084] Even if a spectrum with high similarity to the Raman spectrum of the anomalous sample is found in a spectral library, the composition of the foreign substance cannot be directly identified. Further verification is required. On the one hand, relevant literature should be consulted to verify the variation of Raman spectral characteristics of known substances under different experimental conditions to ensure the reliability of the matching results. On the other hand, other auxiliary analytical methods, such as infrared spectroscopy and mass spectrometry, can be used to cross-verify the composition of the foreign substance. Only after multi-faceted verification can the composition of the foreign substance be accurately determined, providing a scientific basis for subsequent treatment measures.

[0085] It should be noted that the above description describes some embodiments of this disclosure. Other embodiments are within the scope of the appended claims. In some cases, the actions or steps recorded in the claims can be performed in a different order than that shown in the above embodiments and still achieve the desired result. Furthermore, the processes depicted in the drawings do not necessarily require a specific or sequential order to achieve the desired result. In some embodiments, multitasking and parallel processing are also possible or may be advantageous.

[0086] This disclosure is intended to cover all such substitutions, modifications, and variations that fall within the broad scope of the appended claims. Therefore, any omissions, modifications, equivalent substitutions, improvements, etc., made within the spirit and principles of this disclosure should be included within the scope of protection of this disclosure.

Claims

1. A method for detecting foreign objects in a display panel, characterized in that, Specifically, the following steps are included: The surface area of ​​the sample to be tested is irradiated with a laser to obtain the first characteristic spectrum; Compare the first characteristic spectrum with a preset sample library: If the first characteristic spectrum matches any spectrum in the preset sample library, then the sample to be tested is calibrated as a normal sample. If the first characteristic spectrum is inconsistent with all spectra in the preset sample library, the sample to be tested will be labeled as an abnormal sample. The polyimide surface of the abnormal sample was roughened to obtain a rough surface; Surface-enhanced Raman scattering detection was performed on the rough surface to obtain the Raman spectrum of the abnormal sample; The Raman spectrum is compared and analyzed with the spectral library, including: The Raman spectrum is preprocessed to obtain an optimized Raman spectrum; Extract key feature data from the optimized Raman spectrum; The composition of the foreign substance is determined by matching the key feature data with known spectra in the spectral library.

2. The foreign object detection method for a display panel according to claim 1, characterized in that, It also includes the following steps: The roughening process includes primary roughening and secondary roughening.

3. The foreign object detection method for a display panel according to claim 2, characterized in that, The initial roughening process specifically includes: A first rough surface is obtained by spraying a mixture of 4,4-diaminodiphenylmethane and 4,4-bisphthalic dianhydride onto the polyimide surface of the abnormal sample.

4. The foreign object detection method for a display panel according to claim 3, characterized in that, The concentration of the mixture is 4-6%; Preferably, the concentration of the mixture is 5%.

5. The foreign object detection method for a display panel according to claim 3 or 4, characterized in that, The mixture covers the polyimide surface of the abnormal sample for 12-18 minutes; Preferably, the mixture covers the polyimide surface of the abnormal sample for 15 minutes.

6. The foreign object detection method for a display panel according to claim 3, characterized in that, The secondary roughening process specifically includes: Based on the first rough surface, a mixture of carbon dioxide and acetylene gas is ionized under high pressure in an argon environment to obtain an ion beam. The ion beam is then used to roughen the first rough surface of the abnormal sample a second time to obtain a second rough surface.

7. The foreign object detection method for a display panel according to claim 6, characterized in that, The high voltage is 650-900V; Preferably, the high voltage is 700V.

8. The foreign object detection method for a display panel according to claim 6, characterized in that, Also includes: After the second roughening treatment, the morphology of the second roughened surface is inspected using an atomic force microscope. If the root mean square value of the surface roughness does not reach the range of 50-80 nm, the roughening treatment steps are repeated until the requirements are met.

9. The foreign object detection method for a display panel according to claim 1, characterized in that, Surface-enhanced Raman scattering (SERS) of the rough surface to obtain the Raman spectrum of the anomalous sample specifically includes: Optimize the laser parameters, use a laser beam to focus and irradiate the second rough surface, and use a spectral acquisition device to capture the Raman spectrum generated by the abnormal sample in real time.

Images