Inkjet printing OLED dry film defect detection method, device and related equipment

Abstract

The application discloses a kind of inkjet printing OLED dry film defect detection method, device and related equipment, belong to photoelectric information engineering field, the method is with the laser of spot less than pixel groove to irradiate OLED panel, with the detector of array multiple subunits receives the reflection, scattering light generated after OLED panel is irradiated;Then the photocurrent intensity of subunit is acquired;The photocurrent intensity of subunit corresponding to reflection center at current time is compared with the photocurrent intensity at last time, when the degree of change exceeds first preset range, according to the photocurrent intensity of each subunit, find out abnormal subunit;According to the number, position distribution and photocurrent intensity of abnormal subunit, exclude defect or determine defect type.The method improves the detection efficiency under industrial production scene, simultaneously does not occupy production time, can cooperate with production rhythm, does not need additional large-scale evaporation or detection equipment supporting, it is good method to improve the overall yield of inkjet printing OLED panel.

Description

Technical Field

[0001] This invention relates to a method, apparatus, and related equipment for detecting defects in inkjet-printed OLED dry films, belonging to the field of optoelectronic information engineering. Background Technology

[0002] The application of OLED devices in new displays and flexible lighting is rapidly expanding. Inkjet printing technology, due to its ability to achieve high-precision, non-contact film formation on flexible substrates, has become a core process for the preparation of LED dry films (such as quantum dot light-emitting layer dry films). The inkjet-printed OLED manufacturing industry is increasingly demanding higher accuracy and efficiency in identifying micro-defects in dry films (3-5μm pinholes, microcracks, and localized thickness unevenness).

[0003] Existing technologies have few detection schemes for dry film defects and film thickness uniformity. Existing detection schemes are mainly divided into full-plate sampling inspection schemes and partial sample sampling inspection schemes. The difference is that full-plate sampling inspection requires printing the entire plate and focuses more on the actual overall dry film formation quality and possible defect problems in production, while partial sample sampling inspection only requires printing small OLED device samples for dry film quality analysis of OLED devices in process testing and other stages.

[0004] The main method for whole-board sampling inspection is white light interferometry. The specific process is as follows: the printed and dried substrate is removed and placed in a vapor deposition equipment to deposit a uniform silver metal film on the dry film surface. Then, white light interferometry is used to detect the height difference between the silver films, thereby indirectly obtaining the film thickness variation and some dry film defects. The shortcomings of this method are obvious. First, additional large vapor deposition equipment is required to deposit silver on the dry film before inspection. Furthermore, this inspection method will directly scrap the substrate, making online inspection impossible. In addition, the vapor deposition process itself, due to factors such as the vapor deposition material, cavity environment, and mask, will also introduce film thickness errors or defects. Directly using the inspection results after vapor deposition to characterize the original dry film quality before vapor deposition will introduce certain errors. This white light interferometry inspection usually requires an "empty pixel slot" as a reference, that is, leaving 10-20 unprinted positions during whole-board printing as a standard reference value for the morphology of the silver film after direct vapor deposition without dry film. This contradicts traditional printing operations, potentially increasing printing time and causing changes or defects in the dry film morphology due to printing time variations, which is also an additional source of error.

[0005] For some sample inspections, laser confocal microscopy (OCT) technology (layer-by-layer scanning) is primarily used. This method utilizes the principle of laser confocal microscopy to scan the dry film thickness layer by layer. While this method also offers high accuracy, its drawbacks remain significant. The biggest disadvantage of laser confocal technology is its slow speed, which is the primary reason it is used for partial sample inspections rather than whole-panel testing. Its detection speed is approximately 3-5 seconds per pixel slot. Therefore, for OLED display panels with millions of pixel slots, online inspection using this method is too time-consuming and cannot meet the needs of industrial production. Furthermore, laser confocal inspection equipment has high environmental requirements. While it may meet the needs of some laboratory-level sample inspections, deploying it in complex industrial production environments will introduce additional errors, compromising its detection accuracy. Summary of the Invention

[0006] To overcome the shortcomings of existing technologies, this invention provides a method, apparatus, and related equipment for detecting defects in inkjet-printed OLED dry films. It utilizes multi-directional laser scattering to detect OLED dry film defects, making it applicable to industrial production scenarios and suitable for online inspection.

[0007] The technical solution adopted by this invention to solve its technical problem is: In a first aspect, this application provides a method for detecting defects in inkjet-printed OLED dry film, comprising: configuring a laser beam that non-perpendicularly illuminates the surface of an OLED panel, wherein the laser spot is smaller than the area of ​​a single pixel slot of the OLED panel; configuring a detector to receive the reflected and scattered light generated after the OLED panel is illuminated, wherein the theoretical reflection center is located in the middle of the detector, and the detector comprises multiple sub-units of an array. The steps include: Obtain the photocurrent intensity of the sub-unit; compare the photocurrent intensity of the sub-unit corresponding to the theoretical reflection center at the current moment with the photocurrent intensity at the previous moment, and when the degree of change exceeds a first preset range, identify abnormal sub-units based on the photocurrent intensity of each sub-unit; and eliminate defects or determine the defect type based on the number, location distribution, and photocurrent intensity of the abnormal sub-units.

[0008] Since the dry film is formed by drying a wet film, assuming no defects such as breakage, the surface height of the dry film should be uniformly varied based on the fluid properties. Therefore, on the detector, this should manifest as a strong central point with weak power at other locations, and the central point remaining constant or changing slowly due to variations in film thickness. Unlike defects, the multi-central point caused by uneven film thickness is slowly varying, and the power density change outward from each center is also relatively gradual. Therefore, by comparing whether the sub-units corresponding to the theoretical reflection center undergo significant changes at different times, a rough judgment of whether the dry film has defects can be made, requiring minimal computation and contributing to efficient detection.

[0009] Of course, to accurately determine the type of defect, the photocurrent intensity of the sub-units outside the theoretical reflection center must be used.

[0010] Furthermore, the size of the light spot is set to be equal to the size of the pixel slot in a first direction, and to be one-tenth to one-fifth of the size of the pixel slot in a second direction; the first direction and the second direction are orthogonal. The pixel slot is typically rectangular, with one of the first and second directions being the length direction of the pixel slot and the other being the width direction of the pixel slot.

[0011] Furthermore, the gain configuration of the detector is as follows: The laser is used to irradiate the plane mirror at the same angle as the laser used to irradiate the OLED panel, and the gain of the sub-unit located at the center of the reflection is set to be below 80% of the saturation power, which is the center gain; The theoretical received power ratio of the sub-unit located at the reflection center to each of the other sub-units at the target defect scale is calculated based on the Mie scattering formula. The gain of each sub-unit, excluding the reflection center, is set according to the product of the center gain and the theoretical received power ratio.

[0012] Furthermore, the step of identifying abnormal sub-units based on the photocurrent intensity of each sub-unit includes: Calculate the mean and standard deviation of the photocurrent intensity of all the sub-units; The absolute Z-score is obtained by dividing the absolute difference between the photocurrent intensity of a certain sub-unit and the mean by the standard deviation. The sub-unit whose absolute Z-score exceeds a preset deviation threshold is the abnormal sub-unit.

[0013] Furthermore, the step of eliminating defects or determining the defect type based on the number, location distribution, and photocurrent intensity of the abnormal sub-units includes: The abnormal sub-units are calculated as the ratio of the number of abnormal sub-units to the total number of sub-units. When the abnormal ratio is less than or equal to a first preset ratio, the defect is eliminated. When the proportion of the number of abnormal sub-units to the total number of sub-units is greater than a first preset proportion, it is determined whether there is an abnormal cluster based on the location distribution of the abnormal sub-units, and the global photocurrent skewness of all sub-units is calculated. The case where there are no abnormal clusters and the proportion of abnormal clusters is greater than the second preset proportion is judged as a slow-change defect; The case where the anomalous clusters exist and the global photocurrent skewness is higher than a preset global threshold is judged as a crack defect; The case where the abnormal percentage is greater than the first preset percentage and less than or equal to the second preset percentage, and the global photocurrent bias is not higher than the preset global threshold, is judged as a scattered defect.

[0014] Furthermore, the step of determining whether an anomalous settlement exists based on the location distribution of the anomalous sub-units includes: Construct a detection matrix using the photocurrent intensity of all the sub-units; The detection matrix is ​​traversed using a sliding window; Gaussian weights are assigned within the window. The aggregation density is obtained by summing the product of the Gaussian weights and the photocurrent intensity within the window. The sub-units corresponding to the center of windows with a cluster density greater than a preset density are marked as high-density center points; Several adjacent points marked as the high-density center points are divided into the anomalous settlements.

[0015] Furthermore, the target defect size is 3μm~5μm.

[0016] Secondly, this application provides an inkjet-printed OLED dry film defect detection device, equipped with a laser beam that non-perpendicularly illuminates the surface of an OLED panel, the laser spot being smaller than the area of ​​a single pixel slot on the OLED panel; equipped with a detector to receive reflected and scattered light generated after the OLED panel is illuminated, the theoretical reflection center being located in the middle of the detector, the detector comprising multiple sub-units of an array; the device further includes: The acquisition module is used to acquire the photocurrent intensity of the sub-unit; The comparison module is used to compare the photocurrent intensity of the sub-unit corresponding to the theoretical reflection center at the current moment with the photocurrent intensity at the previous moment. When the degree of change exceeds a first preset range, the abnormal sub-unit is identified based on the photocurrent intensity of each sub-unit. The judgment module is used to eliminate defects or determine the type of defects based on the number, location distribution and photocurrent intensity of the abnormal sub-units.

[0017] Thirdly, this application provides an electronic device including a processor and a memory, the memory storing computer-readable instructions that, when executed by the processor, perform the steps of the method described in the first aspect.

[0018] Fourthly, this application provides a storage medium having a computer program stored thereon, which, when executed by a processor, performs the steps of the method described in the first aspect.

[0019] The beneficial effects of this invention are: This invention makes it possible to conduct online dry film defect detection of OLED inkjet printed panels, which was originally difficult to apply in the industrial production field. It solves the problem that the product yield cannot be accurately controlled because the dry film defect information cannot be obtained quickly during the production of the entire OLED inkjet printed panel. It improves the detection efficiency in industrial production scenarios, and does not occupy production time. It can be matched with the production cycle and does not require additional large-scale evaporation or detection equipment. It is a good method to improve the overall yield of inkjet printed OLED panels.

[0020] Other features and advantages of this application will be set forth in the following description and will be apparent in part from the description or may be learned by practicing the application. The objectives and other advantages of this application may be realized and obtained by means of the structures particularly pointed out in the written description and the accompanying drawings. Attached Figure Description

[0021] Figure 1 It is a visual representation of the process of setting the gain of each sub-unit.

[0022] Figure 2 This is a visualization of the detector response results for a normal dry film.

[0023] Figure 3 This is a visualization of the detector response results of a slowly varying defect dry film.

[0024] Figure 4 This is a visualization of the detector response results for a dry film with crack defects.

[0025] Figure 5 This is a visualization of the detector response results for a dry film with scattered defects. Detailed Implementation

[0026] This application provides a method for detecting defects in inkjet-printed OLED dry film. A laser beam is configured to non-perpendicularly irradiate the surface of the OLED panel, and the laser spot is smaller than the area of ​​a single pixel slot in the OLED panel. A detector is configured to receive the reflected and scattered light generated after the OLED panel is irradiated. The theoretical reflection center (which can be known by considering the OLED panel under test as a mirror) is located in the middle of the detector (e.g., the very center). The detector includes multiple sub-units of an array. The steps include: Obtain the photocurrent intensity of the sub-unit; compare the photocurrent intensity of the sub-unit corresponding to the theoretical reflection center at the current moment with the photocurrent intensity at the previous moment, and when the degree of change exceeds the first preset range, identify abnormal sub-units based on the photocurrent intensity of each sub-unit; based on the number, location distribution and photocurrent intensity of abnormal sub-units, eliminate defects or determine the defect type.

[0027] The first preset range can be set to 2%~10%, such as 3%, 5%, 8%, etc., which mainly affects sensitivity.

[0028] This invention utilizes a highly coherent laser to illuminate the surface of the printed dry film at an oblique angle, and uses an array detection device at the reflecting end to collect reflected and scattered light from different angles. If the OLED dry film surface is flat, it should exhibit a weak scattering trend overall. If some film surfaces are smooth but not flat, they should exhibit scattering with a stable change in angle, and the range of change is usually small. However, if some film surfaces have defects such as cracks, uneven thickness, or pinholes, obvious and irregular strong scattering signals will appear at the defect locations. By capturing these strong scattering signals from different directions, we can obtain the specific location of the defects and related information.

[0029] Accordingly, this application provides an inkjet-printed OLED dry film defect detection device, equipped with a laser beam that non-perpendicularly illuminates the surface of the OLED panel, the laser spot being smaller than the area of ​​a single pixel slot on the OLED panel; equipped with a detector to receive the reflected and scattered light generated after the OLED panel is illuminated, the theoretical reflection center being located in the middle of the detector, and the detector comprising multiple sub-units of an array; the device further includes: The acquisition module is used to acquire the photocurrent intensity of the sub-unit; The comparison module is used to compare the photocurrent intensity of the sub-unit corresponding to the theoretical reflection center at the current moment with the photocurrent intensity at the previous moment. When the degree of change exceeds the first preset range, the abnormal sub-unit is identified based on the photocurrent intensity of each sub-unit. The judgment module is used to eliminate defects or determine the type of defects based on the number, location distribution, and photocurrent intensity of abnormal sub-units.

[0030] To ensure the laser spot is smaller than the area of ​​a single pixel slot in the OLED panel, we need pixel slot information. This information can be confirmed in various ways, such as consulting the substrate design manual or taking photographic measurements. The main parameters to confirm are pixel slot bank depth, pixel slot width, pixel slot length, and pixel slot spacing. The pixel slot bank depth is primarily used to match the incident angle and detector array angle; the pixel slot width and height are used to match the spot size; and the pixel slot spacing is used to match the laser pulse width. After obtaining the pixel slot information, the spot size can be adjusted in various ways, such as using adjustable zoom lenses, piezoelectric lens controllers, adjustable gratings, or digital micromirror arrays (DMDs). While the methods differ, the goal is to adjust the laser spot size or angle. The laser spot size is typically set to be the same as the pixel slot in one direction, and an ellipse with the other direction being 1 / 5 or 1 / 10 of its width. For example, the major axis of the elliptical laser spot is the same length as or slightly smaller than the pixel slot, and the minor axis is equal to 1 / 5 or 1 / 10 of the pixel slot width.

[0031] By comparing the photocurrent intensity of the sub-unit corresponding to the theoretical reflection center at the current moment with the photocurrent intensity at the previous moment, we can roughly determine whether there is a defect. However, this only uses the center of the detector. To determine the specific type of defect, we need to use other sub-units.

[0032] In industrial production, pits, central voids, microcracks, and fractures can all be considered defects or flaws (die). While their manifestations exhibit strong randomness depending on the specific circumstances, for detector arrays, they uniformly appear as: dispersed and rapidly changing multiple centers with irregular peak power variations at the center. By capturing this irregular, strong scattering signal, the location of the defective dry film can be quickly determined. Compared to white light interferometry and laser confocal microscopy, this detection method is faster, eliminating the need for point-by-point analysis of the dry film's morphology. It only requires rapid screening and identification of defects to complete defect detection, which is clearly more efficient in the industrial production of OLED printed panels. The other two traditional methods are more suitable for small-batch process trials or laboratory device testing.

[0033] Specifically, we can use the Mie scattering principle and the positional relationship between the central sub-unit of the detector and other sub-units, as well as the photocurrent intensity of each sub-unit, to determine the specific defect type.

[0034] Previously, the detector needed to have its gain (the degree of response of photocurrent to light signal) set to generate a photocurrent signal; otherwise, no matter how strong the light, there would only be zero photocurrent. If the gain was set too low, the sub-unit would be insensitive and would not respond to weak light; if the gain was set too high, the detector would easily burn out. Furthermore, the intensity difference between reflected and scattered light generated by laser irradiation of the OLED panel is extremely large, and surface defects in the OLED can cause the reflection point to shift. Therefore, ensuring the detector has sufficient sensitivity without burning out is a key challenge.

[0035] This application requires the detector to consist of multiple sub-units, each of which can be independently gain-adjusted. The shape is not specified, and the positional distribution of the multiple sub-units can be a rectangular array, a circular array, a hexagonal array, etc., which can be combined according to the detection requirements.

[0036] Reference Figure 1 At point a, we will use an 8*8 rectangular array composed of 64 sub-units as an example to explain how to set the gain of the sub-units.

[0037] When the gain of all sub-units is set to the same, such as Figure 1 At point b, the detector array response to the Gaussian spot is shown. The Gaussian spot and specular reflection cause uneven energy distribution on the detector. It should be noted that... Figure 1For the purpose of visualization, warmer colors indicate stronger photocurrent and cooler colors indicate weaker photocurrent. This is only for the purpose of explaining the embodiments of this application. In fact, the acquisition module and the comparison module both use photocurrent intensity values, not images, and there is no need to generate visualization images.

[0038] This non-uniform spot detector response affects the detection intensity of the edge detectors during subsequent detection. Therefore, it is necessary to independently adjust the gain of each sub-detector (sub-unit) to achieve balance, i.e., reduce the gain voltage of the intermediate detectors and increase the gain voltage of the edge detectors. The goal is to configure the detector gain so that when illuminating defect-free pixels, the optical power (photocurrent intensity) of all units is equal, such as... Figure 1 At point c in the middle.

[0039] In some embodiments, we can use a standard, defect-free pixel slot to adjust the detector gain so that all cells have equal optical power when illuminating a defect-free pixel. However, this requires the preparation of standard samples, which is inefficient.

[0040] Preferably, we can adjust the gain using the Mie scattering formula: R1: Irradiate the plane mirror with a laser at the same angle as the OLED panel, and set the gain of the sub-unit located at the center of the reflection to be less than 80% (preferably 50%) of the saturation power. This gain is the center gain.

[0041] R2: The theoretical received power ratio of the sub-unit located at the reflection center to each sub-unit at other locations is calculated based on the Mie scattering formula at the target defect scale.

[0042] R3: Set the gain of each sub-unit except the reflection center according to the quotient of the center gain divided by the theoretical received power ratio.

[0043] Step R1 first places the plane mirror at the same height as the pixel slot. At this point, the mirror reflection is received by one or more sub-units in the detector array. Typically, this mirror reflection power is high, so the reflection gain needs to be reduced or the laser power adjusted so that the sub-unit corresponding to the center of the mirror reflection measures 80% of the saturation power. This is because, regardless of the dry film surface shape or defects, the optical power of this molecular detector cannot exceed the reflected light power of an ideal plane mirror; therefore, the 80% setting is safe. Then, specific gain quantization allocation is performed on the detector array based on the dry film defect characteristics.

[0044] The basis for introducing the Mie scattering model in step R2 is as follows: The typical physical size of the dry film micro-defects (such as microcracks and pinholes) targeted in this invention is mainly concentrated in the range of 3-5 μm. Since this characteristic size is on the same order of magnitude as the laser wavelength λ used for detection, its light scattering behavior is no longer applicable to geometric optics or Rayleigh scattering models, but fully conforms to the applicable scope of Mie scattering. The Mie scattering formula is common knowledge and its calculation is quite lengthy, so it will not be elaborated upon in this invention.

[0045] After obtaining the theoretical Mie scattering spatial light field distribution according to the Mie scattering formula, it needs to be mapped onto the actual physical detector. Since the sub-units used in this invention have a large photosensitive surface (e.g., a diameter of approximately 8 mm), they cannot be considered ideal points. Therefore, it is necessary to perform surface integration over the spatial solid angle (Ω) of each sub-detector, for example, according to the following Equation 1: Formula 1.

[0046] In the formula, I(θ,φ) is the total power of the scattered light theoretically received by the i-th sub-unit; I(θ,φ) is the intensity density of the scattered light in the direction of the spatial deflection angle θ and azimuth angle Φ, calculated by the Mie scattering formula; Ω is the solid angle occupied by the effective photosensitive surface of the i-th sub-detector in space (determined by the detector size of 8mm and the installation distance).

[0047] By using the area integral described above, the theoretically required scattered light power that each sub-detector in the array should receive under a given micro-defect size (with 3 μm as the lower limit of calibration) can be calculated, thereby obtaining multiple ratios of step R2.

[0048] In step R3, the scattering power Pi at large angles at the edge is much smaller than the central power P. center To ensure that the edge sub-units have sufficient sensitivity to weak defect-scattered light, the actual downlink gain G of the i-th sub-unit is... i It is set to be inversely proportional to the theoretical power (or a normalized mapping is used). Its mapping formula can be, for example, Equation 2: Equation 2.

[0049] In the formula, G0 is the center gain obtained in step R1; P is the ratio obtained in step R2. i This represents the theoretically expected scattered light power that each sub-detector in the array should receive. Use a smooth scaling function based on the hardware range, such as a logarithmic function or a limit-linear function, to ensure that the calculated gain does not exceed the maximum physical gain limit allowed by the hardware.

[0050] For the edge sub-detectors, which are theoretically located in the large-angle scattering region (dominated by defect signals), their gain voltage is increased to enhance the sensitivity to capture signals from micro-surface undulations and edge fractures due to their lower theoretical scattering power. Conversely, for the center sub-detectors, located in the low-angle region (where reflection interference from the normal surface of the dry film is significant), a lower gain is maintained. Through this preset gain adjustment mechanism guided by Mie scattering theory, the detector array establishes a high dynamic response range for 3-5 μm dry film defects even before formal scanning begins. This maximizes the capture capability of highly specific scattering signals from micro-defects while ensuring detector safety.

[0051] The target defect scale in step R2 is set to the minimum common size of the aforementioned dry film micro-surface defects, i.e., 3 μm. Of course, it can also be designed to be smaller. If a smaller size is substituted into the Mie scattering formula, it means that the large-angle scattering signal gradually increases, and the corresponding gain at large angles decreases. This may result in a weakening of the detection sensitivity to large-angle scattered light, which is not conducive to the detection of certain micro-surface undulation signals. However, it has a better protection effect on detectors with strong edge signals such as dry film fractures or dry film voids.

[0052] After specific gain adjustment based on Mie scattering, we artificially lowered the high-light gain at the center and increased the low-light gain at the edges, so that the reference response of all detectors was "flattened" when scanning a defect-free standard dry film. On the "flattened" reference surface, if a defect causes abnormally large-angle scattering of light, the signal captured by the edge detectors will suddenly spike.

[0053] The method of this invention is applicable to online dry film inspection scenarios in industrial OLED panel production. It needs to be coordinated with production equipment, and the inspection speed should be designed according to the production cycle. OLED panels are typically composed of repeating pixel slot units. One pixel slot unit consists of a pixel slot and a bank. Since we do not need to focus on the bank, and the strong scattering generated when the laser hits the bank signal may interfere with or even damage the detector array, the laser should be turned off when passing through the bank position. Therefore, the laser signal is a pulse signal, the pulse frequency is proportional to the scanning speed, the pulse width is proportional to the length (or width) of the pixel slot, and the duty cycle is the same as or slightly lower than the ratio of the pixel slot to the bank (the duty cycle can be slightly reduced to protect the detector from edge signal interference).

[0054] The detectors used in this method can be optoelectronic devices such as avalanche diode (APD) arrays and photomultiplier tube (PMT) arrays. These detectors have relatively large individual sub-unit sizes, ranging from 1 to 5 mm, corresponding to a spatial angular resolution of approximately 1.6 to 8 mrad ≈ 0.08° to 0.4°. Therefore, they exhibit poor detail in controlling edge contours, but perform excellently in sampling rates, typically exceeding 100 kHz, with some products reaching over 10 MHz. Thus, they are more suitable for defect detection scenarios where the specific morphology of the defect itself does not need to be described, but rapid capture of the defect and its location is required. Furthermore, APDs and PMTs have stronger weak signal detection capabilities than general CMOS, SCMOS, and EMCCD chip units, making them more suitable for weak signal defect detection scenarios, but they are also more susceptible to interference.

[0055] Since the detector array detects a small number of two-dimensional signals and has higher sensitivity than traditional CMOS cameras, it requires almost no additional signal processing or only simple noise filtering before it can be used for result feedback. Therefore, synchronous result feedback can be achieved. When the above-mentioned changes exceed the first preset range, the photocurrent intensity and the current position data or pixel slot coordinates can be directly sent to the host computer or client. Users can design their own methods for processing or analyzing different types of signals. If it is only for the detection of dry film defects in OLED inkjet printed panels in industrial production, these signals and pixel slot positions can be used directly without processing. Alternatively, after recording the pixel slot positions, more precise detection instruments, such as profilometers, atomic force microscopes, high-resolution microscopes, and other laboratory scientific detection equipment, can be used to further detect the specific defect size distribution and surface fine morphology.

[0056] At this point, we have completed the setting of various parameters before the test. Next, we will continue to explain how to determine the defect type.

[0057] In addition to the visualization of normal dry film test results, Figure 1 Besides point C, it is also possible that... Figure 2 The dry film surface has slight undulations, which are considered to be defect-free and good products. Only some detectors have numerical fluctuations, which are irregular and the fluctuation range is small. Such slight undulations will not cause the change in photocurrent intensity of the sub-unit corresponding to the theoretical reflection center at the current moment to exceed the first preset range compared with the photocurrent intensity at the previous moment.

[0058] When the degree of change exceeds the first preset range, the following calculation process is triggered: First, calculate the global mean U and global standard deviation σ of the photocurrent intensity of the entire detector array (e.g., an 8x8 array). For any sub-unit i, use the absolute value of the Z-score (a statistical concept) for filtering: Formula 3.

[0059] Among them, I i Let be the photocurrent intensity of the i-th subunit.

[0060] If Zi > Th, the sub-unit is determined to be an abnormal sub-unit. Th can be 2.5 to 3, or a preset deviation threshold calibrated based on a priori defect-free standard. Regarding the "priori defect-free standard," this invention requires at least one preparation, and the prior data accumulates with each preparation, becoming historical data. In contrast, white light interferometry requires sample preparation for each iteration, and the test results from a single sample preparation are only valid for that specific sample because the vapor deposition machine is also a variable, and the film thickness deposited each time may differ. Therefore, even if priori defect-free standards may be needed in some embodiments, this invention is still superior to the prior art.

[0061] Next, the proportion of abnormal sub-units to the total number of sub-units is calculated as the abnormality ratio R. g : Formula 4.

[0062] Where N a N represents the number of abnormal units. tatol The total number of sub-units (by Figure 1 The corresponding embodiment is 64), when R g If the percentage is equal to or less than a set lower limit (e.g., 5%), it can be considered a defect-free positive sample, and the subsequent calculation is terminated, continuing to focus only on the photocurrent intensity change at the theoretical reflection center. Otherwise, proceed to the next step to perform a dual density assessment of the detector array, both global and local, determine whether there are anomalous clusters based on the location distribution of anomalous sub-units, and calculate the global photocurrent skewness of all sub-units.

[0063] The situation where there are no abnormal settlements and the proportion of abnormal settlements is greater than the second preset proportion (e.g., 30%) is judged as a slow-change defect; The presence of abnormal clusters and a global photocurrent bias higher than a preset global threshold are identified as crack defects. The case where the abnormal percentage is greater than the first preset percentage and less than or equal to the second preset percentage, and the global photocurrent deviation is not higher than the preset global threshold, is judged as a scattered defect.

[0064] The global photocurrent skewness refers to the skewness of the photocurrent intensity across the entire detector (e.g., 64 sub-units). The formula for calculating the skewness is common statistical knowledge and will not be elaborated upon in this invention. Anomalous clusters can be segmented using methods such as Connected Component Labeling (CCL) and DBSCAN clustering. Preferred segmentation methods are as follows: Use a sliding window to traverse the probe matrix. Assign Gaussian weights within the window. For example, use an n*n sliding window (where n is a positive integer less than the number of sub-units raised to the power of 0.5) to traverse the probe matrix (specifically, a 3*3 sliding window). Assign Gaussian weights within the window, with the largest weight at the center (e.g., 0.5) and decreasing weights at the edges.

[0065] The aggregation density is obtained by summing the product of the Gaussian weights and the photocurrent intensity within the window: Equation 5.

[0066] Where the coordinates (x, y) are an abnormal sub-unit, then It is 1 if it is true, otherwise it is 0. Cluster density; Let (1, 1) be the Gaussian weights; (1, 1) be the first cell of the sliding window, for example, the bottom left corner of the sliding window; and (n, n) be the last cell of the sliding window. Accumulate to This means summing the products (the product of Gaussian weights and photocurrent intensity) corresponding to each cell within the sliding window.

[0067] When the sub-unit corresponding to the center of a window with a cluster density greater than a preset density is marked as a high-density center point, several adjacent high-density center points are classified as anomalous settlements.

[0068] The basis for summarizing various defects is as follows: The first type of defect is the gradually varying defect. These defects are characterized by a uniform increase or decrease in film thickness without any breakage. In detector arrays, they manifest as a high global anomaly rate without significant high-density clusters, and a low global photocurrent bias (gradual light intensity changes), often accompanied by a continuous gradient vector in a single direction covering multiple pixel units. When the dry film thickness distribution is severely uneven due to these defects, the detector array's response is visualized as follows: Figure 3 The gradient direction of the detector array response change is the direction of the film thickness change, such as... Figure 3 In the figure, 'a' and 'b' represent changes in two different directions. These changes are not necessarily unidirectional as shown in the figure; there may also be changes in two or more directions simultaneously. The figure is only shown as a reference for two of these cases.

[0069] The second type of defect is the crack-type defect, specifically a fissure or fracture. The physical characteristic of this type of defect is that when a laser beam strikes a fissure, the scattered light entering the fissure is significantly weakened due to multiple reflections, while the laser beam striking the sharp edges of the fissure undergoes strong, irregular, large-angle scattering. In the detector array, this manifests as a significant high-density anomalous cluster with extremely high global photocurrent skewness. The cluster exhibits a "bimodal or annular" distribution (i.e., the photocurrent of the detector unit at the center of the cluster is below the mean U, while the photocurrent of adjacent detector units on the periphery generates a very large peak value; this situation is not shown in the figure), and the gradient vector shows a steep radial deviation from the center outwards. The response of the detector array when the light spot passes through the microcrack is visualized as follows: Figure 4 The reflected light from its original position is scattered in the surrounding direction. This phenomenon also exists in ordinary cameras but is not obvious. However, by adjusting the gain, the edge detector has a higher gain, so it has a stronger detection capability for this scattering and its performance is more obvious.

[0070] The third type of defect is scattered defects. These defects are characterized by their dispersed, uncertain, and sporadic small size and scale. In detector arrays, they exhibit a low global anomaly rate (compared to the second preset ratio) and low aggregation density (scattered distribution), while also displaying isolated abrupt peaks. The response of detector arrays with this type of defect is visualized as follows: Figure 5 As shown, defects such as pits and bumps are similar to microcracks, all exhibiting a weakening of the central signal accompanied by an enhancement of the surrounding signal; the weakened area is the defect location. The degree of weakening and enhancement is related to the sharpness of the defect edge; the sharper the edge, the more pronounced this effect. For these types of defects, it is not necessary to determine the specific defect type for a single pixel slot; it is only necessary to determine whether the pixel slot is defective (die) or good (pass).

[0071] This application provides an electronic device, including: a processor and a memory, which are interconnected and communicate with each other via a communication bus and / or other forms of connection mechanism. The memory stores a computer program executable by the processor. When the computing device is running, the processor executes the computer program to perform the method in any optional implementation of the above embodiments to achieve the following functions: acquiring the photocurrent intensity of a sub-unit; comparing the photocurrent intensity of the sub-unit corresponding to the theoretical reflection center at the current moment with the photocurrent intensity at the previous moment, and when the degree of change exceeds a first preset range, identifying abnormal sub-units based on the photocurrent intensity of each sub-unit; and eliminating defects or determining the defect type based on the number, location distribution, and photocurrent intensity of the abnormal sub-units.

[0072] This application provides a storage medium storing a computer program. When the computer program is executed by a processor, it executes the method in any optional implementation of the above embodiments to achieve the following functions: obtaining the photocurrent intensity of a sub-unit; comparing the photocurrent intensity of the sub-unit corresponding to the theoretical reflection center at the current moment with the photocurrent intensity at the previous moment, and when the degree of change exceeds a first preset range, identifying abnormal sub-units based on the photocurrent intensity of each sub-unit; and eliminating defects or determining the defect type based on the number, location distribution, and photocurrent intensity of the abnormal sub-units. The storage medium can be implemented by any type of volatile or non-volatile storage device or a combination thereof, such as Static Random Access Memory (SRAM), Electrically Erasable Programmable Read-Only Memory (EEPROM), Erasable Programmable Read Only Memory (EPROM), Programmable Red-Only Memory (PROM), Read-Only Memory (ROM), magnetic storage, flash memory, magnetic disk, or optical disk.

[0073] In the description of this specification, the references to terms such as "one embodiment," "some embodiments," "illustrative embodiment," "example," "specific example," or "some examples," etc., indicate that a specific feature, material, or characteristic described in connection with the embodiment or example is included in at least one embodiment or example of the invention. In this specification, the illustrative expressions of the above terms do not necessarily refer to the same embodiment or example. Furthermore, the specific features, materials, or characteristics described may be combined in any suitable manner in one or more embodiments or examples.

[0074] The above description represents the preferred embodiments of the present invention. It should be noted that those skilled in the art can make various improvements and modifications without departing from the principles of the present invention, and these improvements and modifications are also considered to be within the scope of protection of the present invention.

Claims

1. A method for detecting defects in inkjet-printed OLED dry film, characterized in that, A laser beam is configured to non-perpendicularly illuminate the surface of an OLED panel, the laser spot being smaller than the area of ​​a single pixel slot in the OLED panel; a detector is configured to receive the reflected and scattered light generated after the OLED panel is illuminated, the theoretical reflection center being located in the middle of the detector, and the detector comprising multiple sub-units of an array; The steps include: Obtain the photocurrent intensity of the sub-unit; compare the photocurrent intensity of the sub-unit corresponding to the theoretical reflection center at the current moment with the photocurrent intensity at the previous moment; when the degree of change exceeds a first preset range, identify abnormal sub-units based on the photocurrent intensity of each sub-unit; and eliminate defects or determine the defect type based on the number, location distribution, and photocurrent intensity of the abnormal sub-units. The gain configuration of the detector is as follows: The laser is used to irradiate the plane mirror at the same angle as the laser used to irradiate the OLED panel, and the gain of the sub-unit located at the center of the reflection is set to be below 80% of the saturation power, which is the center gain; The theoretical received power ratio of the sub-unit located at the reflection center to each of the other sub-units at the target defect scale is calculated based on the Mie scattering formula. The gain of each sub-unit, excluding the reflection center, is set according to the product of the center gain and the theoretical received power ratio. The step of identifying abnormal sub-units based on the photocurrent intensity of each sub-unit includes: Calculate the mean and standard deviation of the photocurrent intensity of all the sub-units; The absolute Z-score is obtained by dividing the absolute difference between the photocurrent intensity of a certain sub-unit and the mean by the standard deviation. The sub-unit whose absolute Z-score exceeds a preset deviation threshold is the abnormal sub-unit; The step of eliminating defects or determining the type of defects based on the number, location distribution, and photocurrent intensity of the abnormal sub-units includes: The abnormal sub-units are calculated as the ratio of the number of abnormal sub-units to the total number of sub-units. When the abnormal ratio is less than or equal to a first preset ratio, the defect is eliminated. When the proportion of the number of abnormal sub-units to the total number of sub-units is greater than a first preset proportion, it is determined whether there is an abnormal cluster based on the location distribution of the abnormal sub-units, and the global photocurrent skewness of all sub-units is calculated. The case where there are no abnormal clusters and the proportion of abnormal clusters is greater than the second preset proportion is judged as a slow-change defect; The case where the anomalous clusters exist and the global photocurrent skewness is higher than a preset global threshold is judged as a crack defect; The case where the abnormal percentage is greater than the first preset percentage and less than or equal to the second preset percentage, and the global photocurrent bias is not higher than the preset global threshold, is judged as a scattered defect.

2. The method for detecting defects in inkjet-printed OLED dry film according to claim 1, characterized in that, The size of the light spot is set to be equal to the size of the pixel slot in the first direction and to be one-tenth to one-fifth of the size of the pixel slot in the second direction; the first direction and the second direction are orthogonal.

3. The method for detecting defects in inkjet-printed OLED dry film according to claim 1, characterized in that, The step of determining whether an anomalous settlement exists based on the location distribution of the anomalous sub-units includes: Construct a detection matrix using the photocurrent intensity of all the sub-units; The detection matrix is ​​traversed using a sliding window; Gaussian weights are assigned within the window. The aggregation density is obtained by summing the product of the Gaussian weights and the photocurrent intensity within the window. The sub-units corresponding to the center of windows with a cluster density greater than a preset density are marked as high-density center points; Several adjacent points marked as the high-density center points are divided into the anomalous settlements.

4. The method for detecting defects in inkjet-printed OLED dry film according to claim 1, characterized in that, The target defect size is 3μm~5μm.

5. A defect detection device for inkjet-printed OLED dry film, characterized in that, The device is equipped with a laser beam that non-perpendicularly illuminates the surface of an OLED panel, the laser spot being smaller than the area of ​​a single pixel slot in the OLED panel; and a detector is configured to receive the reflected and scattered light generated after the OLED panel is illuminated, with the theoretical reflection center located in the middle of the detector, which includes multiple sub-units of an array. The gain configuration of the detector is as follows: The laser is used to irradiate the plane mirror at the same angle as the laser used to irradiate the OLED panel, and the gain of the sub-unit located at the center of the reflection is set to be below 80% of the saturation power, which is the center gain; The theoretical received power ratio of the sub-unit located at the reflection center to each of the other sub-units at the target defect scale is calculated based on the Mie scattering formula. The gain of each sub-unit, excluding the reflection center, is set according to the product of the center gain and the theoretical received power ratio. The device further includes: The acquisition module is used to acquire the photocurrent intensity of the sub-unit; The comparison module is used to compare the photocurrent intensity of the sub-unit corresponding to the theoretical reflection center at the current moment with the photocurrent intensity at the previous moment. When the degree of change exceeds a first preset range, abnormal sub-units are identified based on the photocurrent intensity of each sub-unit. The step of identifying abnormal sub-units based on the photocurrent intensity of each sub-unit includes: calculating the mean and standard deviation of the photocurrent intensity of all sub-units; calculating the absolute difference between the photocurrent intensity of a certain sub-unit and the mean by dividing it by the standard deviation, which is the absolute Z-score; the sub-unit whose absolute Z-score exceeds a preset deviation threshold is the abnormal sub-unit. The judgment module is used to eliminate defects or determine the defect type based on the number, location distribution, and photocurrent intensity of the abnormal sub-units: calculating the ratio of the number of abnormal sub-units to the total number of sub-units as the abnormality ratio; when the abnormality ratio is less than or equal to a first preset ratio, the defect is eliminated; when the ratio of the number of abnormal sub-units to the total number of sub-units is greater than the first preset ratio, determining whether there is an abnormal cluster based on the location distribution of the abnormal sub-units, and calculating the global photocurrent skewness of all the sub-units; judging the case where there is no abnormal cluster and the abnormality ratio is greater than a second preset ratio as a gradually varying defect; judging the case where there is an abnormal cluster and the global photocurrent skewness is higher than a preset global threshold as a crack defect; judging the case where the abnormality ratio is greater than the first preset ratio and less than or equal to the second preset ratio, and the global photocurrent skewness is not higher than the preset global threshold as a scattered defect.

6. An electronic device, characterized in that, It includes a processor and a memory, the memory storing computer-readable instructions that, when executed by the processor, perform the steps of the method as described in any one of claims 1-4.

7. A storage medium having a computer program stored thereon, characterized in that, When the computer program is executed by a processor, it performs the steps of the method as described in any one of claims 1-4.

Images