Ultrasonic Detection Method, System and Device for Integrated Circuit Plastic Encapsulation

Abstract

The present invention relates to the field of ultrasonic detection technology, and particularly to an ultrasonic detection method, system and device for integrated circuit plastic packaging. According to the proportion of the suspected defect area and the distance distribution from the module area in the ultrasonic detection image of each sample, the distribution suspected index is obtained by analyzing from the distribution position, and combined with the similarity of the ultrasonic signal fluctuation between pixel points in the area, the pore feature index is obtained by analyzing from the internal distribution similarity, and the possible pore defect index is obtained; the possible pore defect index is corrected by combining the overlapping performance of the suspected defect areas between multiple ultrasonic detection images, and the corrected possible index is obtained for pore detection. The present invention improves the accuracy of distinguishing electromagnetic noise and pore defects through the similarity representation degree and overlapping distribution degree of defect features in the defect areas between ultrasonic detection images of multiple samples, makes defect detection more accurate, and further improves the reliability of subsequent process parameter optimization.

Description

Technical Field

[0001] The present invention relates to the technical field of ultrasonic detection, and particularly to an ultrasonic detection method, system and device for integrated circuit plastic packaging. Background Art

[0002] The encapsulation of integrated circuits is to provide a stable and reliable working environment for chips, avoiding the influence of the external environment on integrated circuits and affecting their service life, such as dust, humidity, static electricity, etc. The plastic packaging process of integrated circuits is the stage most prone to internal defects during the entire integrated circuit plastic packaging process, and optical detection methods can only detect surface defects. Therefore, an ultrasonic detection method is needed to obtain internal defect images of integrated circuit plastic packaging through the packaging material.

[0003] Currently, for ultrasonic detection of integrated circuit plastic packaging, it is necessary to judge the existing defect types based on the detection images, establish the mapping relationship between the defects and the process parameters of the packaging, and thus obtain the optimal process parameters for integrated circuit packaging to improve the usability of integrated circuits.

[0004] Due to the precision of integrated circuit devices, the electromagnetic components inside the ultrasonic detection equipment emit electromagnetic waves during normal operation, and the electromagnetic waves are received by the ultrasonic detection equipment, thus interfering with the normal operation of the ultrasonic equipment and causing electromagnetic noise in the ultrasonic detection images of integrated circuit plastic packaging. Tiny changes in packaging parameters will result in defects during the packaging process. Pore defects are a common type of packaging defect. The internal pores or cavity defects that appear after integrated circuit packaging will cause image manifestations in the ultrasonic images that are similar to electromagnetic noise, affecting the judgment of real pore defects in integrated circuit plastic packaging, making the results of defect detection inaccurate, and further affecting the subsequent corresponding judgment of defects and process parameters, resulting in inaccurate selection of process parameters. Summary of the Invention

[0005] In order to solve the technical problem that the internal pores or cavity defects that appear after integrated circuit packaging will cause image manifestations in the ultrasonic images that are similar to electromagnetic noise, affecting the judgment of real pore defects and making the results of defect detection inaccurate, the purpose of the present invention is to provide an ultrasonic detection method, system and device for integrated circuit plastic packaging, and the specific technical solutions adopted are as follows:

[0006] The present invention provides an ultrasonic detection method for integrated circuit plastic packaging, and the method includes:

[0007] Obtain the ultrasonic detection images of each packaging sample under the same batch of processes, and determine the suspected defect areas and module areas in each ultrasonic detection image through template matching; obtain the ultrasonic signals at the positions of each pixel point in each suspected defect area in the ultrasonic detection image;

[0008] For each ultrasonic detection image, according to the proportion of each suspected defect area and the distance from the module area, obtain the distribution suspicion index of each suspected defect area; according to the similarity of each pixel point in each suspected defect area in terms of distribution position and ultrasonic signal fluctuation with other pixel points, obtain the pore feature index of each pixel point; according to the distribution suspicion index of each suspected defect area and the pore feature index of the pixel points, obtain the possible pore defect index of each suspected defect area;

[0009] For each suspected defect area of each ultrasonic detection image, analyze the overlapping distribution degree of the suspected defect areas with other ultrasonic detection images, and combine the possible pore defect index to obtain the corrected possible index of each suspected defect area; perform pore defect detection based on the corrected possible index.

[0010] Further, the method for obtaining the distribution suspicion index includes:

[0011] For any suspected defect area, obtain the shortest distance between the suspected defect area and the module area, perform negative correlation mapping and normalization processing, and obtain the position distribution index of the suspected defect area;

[0012] Normalize the overlapping area between the suspected defect area and the module area to obtain the distribution proportion index of the suspected defect area;

[0013] Combine the position distribution index and the distribution proportion index of the suspected defect area to obtain the distribution suspicion index of the suspected defect area.

[0014] Further, the method for obtaining the pore feature index includes:

[0015] For a pixel point in any suspected defect area, obtain the time series fluctuation sequence of the pixel point according to the time series fluctuation of the ultrasonic signal of the pixel point based on frequency;

[0016] Successively take each other pixel point in the suspected defect area where the pixel point is located as the analysis similarity point; perform negative correlation mapping and normalization processing on the distance between the pixel point and the analysis pixel point to obtain the distance credibility of the analysis pixel point;

[0017] Perform negative correlation mapping on the DTW value between the time series fluctuation sequences of the pixel point and the analysis pixel point to obtain the amplitude similarity of the analysis pixel point;

[0018] Combine the distance credibility and the amplitude similarity of the analysis pixel point to obtain the similarity index of the analysis pixel point;

[0019] Take the average value of the similarity indexes of all other pixel points of the pixel point as the pore feature index of the pixel point.

[0020] Further, the method for obtaining the possible index of the pore defect includes:

[0021] For any suspected defect area, the ratio of the mean value to the standard deviation of the pore feature indexes of all pixel points in the suspected defect area is used as the suspected pore feature index of the suspected defect area;

[0022] The product of the suspected pore feature index of the suspected defect area and the suspected distribution index is used as the possible index of the pore defect of the suspected defect area.

[0023] Further, the method for obtaining the corrected possible index includes:

[0024] For any ultrasonic detection image, the suspected defect area in the ultrasonic detection image is sequentially used as the target area; each other ultrasonic detection image outside the ultrasonic detection image is sequentially used as the analysis image;

[0025] When there is an overlap between the target area and the suspected defect area in the analysis image, the area ratio of the overlap between the target area and the suspected defect area in the analysis image is used as the defect coincidence degree of the analysis image; the product of the possible index of the pore defect of the suspected defect area overlapping with the target area in the analysis image and the defect coincidence degree is used as the repeated suspected index of the analysis image;

[0026] When there is no overlap between the target area and the suspected defect area in the analysis image, the preset repeated index is used as the repeated suspected index of the analysis image;

[0027] The sum value of the repeated suspected indexes of all analysis images is used as the correction adjustment degree of the target area; combining the possible index of the pore defect of the target area with the correction adjustment degree and performing normalization processing to obtain the corrected possible index of the target area.

[0028] Further, the pore defect detection based on the corrected possible index includes:

[0029] The maximum corrected possible index in each ultrasonic detection image is used as the detection index of each ultrasonic detection image; when the detection index is greater than the preset defect threshold, the encapsulated sample corresponding to the ultrasonic detection image is recorded as having a pore defect.

[0030] Further, the determination of the suspected defect area and the module area in each ultrasonic detection image by template matching includes:

[0031] For any ultrasonic detection image, the ultrasonic detection image is subjected to template matching with the module area in the preset template image to obtain the module area of the ultrasonic detection image;

[0032] The ultrasonic detection image is input into the trained neural network model, and the suspected defect area of the ultrasonic detection image is output.

[0033] Further, the method for obtaining the timing fluctuation sequence includes:

[0034] Perform Fourier transform on the ultrasonic signal of the pixel to obtain the frequency of the ultrasonic signal of the pixel;

[0035] Periodically divide the ultrasonic signal of the pixel according to the frequency to obtain the amplitude of each period; arrange the amplitudes in chronological order to obtain the timing fluctuation sequence of the pixel.

[0036] The present invention also provides an ultrasonic detection system for integrated circuit plastic encapsulation, including:

[0037] A data acquisition module, configured to acquire ultrasonic detection images of each encapsulation sample under the same batch of processes, determine the suspected defect areas and module areas in each ultrasonic detection image through template matching; acquire the ultrasonic signals at the positions of each pixel in each suspected defect area in the ultrasonic detection image;

[0038] A pore defect analysis module, for each ultrasonic detection image, obtain the distribution suspected index of each suspected defect area according to the proportion of each suspected defect area and the distance from the module area; obtain the pore feature index of each pixel according to the similarity of the distribution position and ultrasonic signal fluctuation of each pixel in each suspected defect area to other pixels; obtain the possible pore defect index of each suspected defect area according to the distribution suspected index of each suspected defect area and the pore feature index of the pixel;

[0039] A correction analysis and detection module, for each suspected defect area of each ultrasonic detection image, analyze the overlapping distribution degree of the suspected defect areas with other ultrasonic detection images, and combine the possible pore defect index to obtain the possible correction index of each suspected defect area; perform pore defect detection based on the possible correction index.

[0040] The present invention also provides an ultrasonic detection device for integrated circuit plastic encapsulation, including a memory, a processor, and a computer program stored in the memory and executable on the processor. When the processor executes the computer program, the steps of an ultrasonic detection method for integrated circuit plastic encapsulation as described in any one of the above are implemented.

[0041] The present invention has the following beneficial effects:

[0042] The present invention takes into account the location characteristics of the presence of pore defects and the randomness characteristics of electromagnetic noise. According to the proportion of the suspected defect area and the distance distribution from the module area in the ultrasonic detection image of each sample, the distribution suspected index is first analyzed from the distribution position. Combining with the similarity of the ultrasonic signal fluctuations between pixel points in the area, the pore feature index is analyzed from the internal distribution similarity. The possible index of pore defects is obtained to reflect the possible degree of pore defects, providing an analysis basis for the screening of pore defects. Further, by using the high repeatability of packaging defects under the same process and combining the overlapping performance of the suspected defect areas between multiple ultrasonic detection images, the possible index of pore defects is corrected to obtain a more accurate corrected possible index for pore detection. The present invention improves the accuracy of distinguishing electromagnetic noise and pore defects through the similarity characterization degree and overlapping distribution degree of defect characteristics in the defect areas between ultrasonic detection images of multiple samples, making defect detection more accurate, and further improving the reliability of subsequent process parameter optimization. BRIEF DESCRIPTION OF THE DRAWINGS

[0043] In order to more clearly illustrate the technical solutions and advantages in the embodiments of the present invention or the prior art, the following will briefly introduce the drawings required for the description of the embodiments or the prior art. Obviously, the following drawings are only some embodiments of the present invention. For those of ordinary skill in the art, other drawings can be obtained based on these drawings without creative efforts.

[0044] Figure 1 It is a flowchart of an ultrasonic detection method for integrated circuit plastic packaging provided by an embodiment of the present invention;

[0045] Figure 2 It is a schematic diagram of a partial ultrasonic detection image with pore defects provided by an embodiment of the present invention. DETAILED DESCRIPTION OF THE EMBODIMENTS

[0046] In order to further elaborate on the technical means and effects adopted by the present invention to achieve the predetermined invention purpose, the following, in combination with the drawings and preferred embodiments, details the specific embodiments, structures, features, and effects of an ultrasonic detection method, system, and device for integrated circuit plastic packaging proposed according to the present invention. In the following description, different "one embodiment" or "another embodiment" do not necessarily refer to the same embodiment. In addition, the specific features, structures, or characteristics in one or more embodiments can be combined in any suitable form.

[0047] Unless otherwise defined, all technical and scientific terms used herein have the same meaning as commonly understood by those of ordinary skill in the technical field to which the present invention belongs.

[0048] The following specifically describes the specific solutions of an ultrasonic detection method, system, and device for integrated circuit plastic packaging provided by the present invention in conjunction with the accompanying drawings.

[0049] Example 1:

[0050] Please refer to Figure 1 , which shows a flowchart of an ultrasonic detection method for integrated circuit plastic packaging provided by an embodiment of the present invention. The method includes the following steps:

[0051] S1: Obtain the ultrasonic detection images of each packaging sample under the same batch of processes, and determine the suspected defect areas and module areas in each ultrasonic detection image through template matching; obtain the ultrasonic signals at the positions of each pixel point in each suspected defect area in the ultrasonic detection image.

[0052] In the implementation of the present invention, the plastic packaging process of the integrated circuit includes: selecting a plastic packaging block of appropriate size and being able to wake up the material in a dry environment; after the material waking is completed, transferring the product to be packaged to a preheating table, setting the preheating temperature, and performing preheating; then adjusting the heating devices of the plastic packaging film according to the upper and lower molds to keep the mold temperature at a constant temperature state, and moving the preheated whole mold frame into the plastic packaging mold through the loading and unloading rack to complete the plastic packaging process for the integrated circuit.

[0053] At the same time, record the packaging process parameters of the packaged circuit, such as preheating temperature, mold temperature, or plastic packaging time, etc. Different process parameters will result in different plastic packaging defects. Therefore, usually, more optimal process parameter selections are determined through defect detection under different packaging process parameters. In the embodiment of the present invention, the packaged circuits under the same batch of unified process parameters are sampled to obtain multiple packaging samples under this process parameter, and then these packaging samples are ultrasonically detected to obtain ultrasonic detection images.

[0054] During the ultrasonic detection process of integrated circuit plastic packaging, it is easily interfered by the normal co - operation of electronic components in the detection equipment, resulting in electronic noise in the ultrasonic image. There are a large number of electronic components in the ultrasonic detection equipment, such as amplifiers, digital - to - analog converters, etc. The normal co - operation of these components generates thermal noise. Thermal noise is generated due to the random movement of electrons, causing random voltage fluctuations in the output signal, thereby making the intensity of the ultrasonic wave show random fluctuations, resulting in an image performance similar to a speckle - like pattern on the ultrasonic image. This shape is very similar to the characteristics of pore defects during the packaging process. Pore defects also appear as speckle shapes in the ultrasonic image of the integrated circuit. Tiny pores cause the ultrasonic echo signal to weaken and appear as speckles in the ultrasonic image, thus affecting the judgment of real defects.

[0055] For the convenience of further analyzing the authenticity of the pore defects subsequently, preliminary data analysis of the ultrasonic detection images is performed to obtain them. Through template matching, the suspected defect areas and module areas in each ultrasonic detection image are initially analyzed. In the embodiment of the present invention, for any ultrasonic detection image, the ultrasonic detection image is subjected to template matching with the module area in the preset template image to obtain the module area of the ultrasonic detection image. The preset template image is a real integrated circuit template, including a processor module, a memory module, an I / O module, or a logic control module of the integrated circuit, etc. The corresponding area position of the module in the ultrasonic detection image is marked as the module area through template matching.

[0056] Furthermore, the ultrasonic detection image is input into the trained neural network model, and the suspected defect area of the ultrasonic detection image is output. The trained neural network model is used to perform preliminary detection and judgment on the pore defects, and the suspected defect area is obtained for subsequent authenticity analysis. It should be noted that template matching and neural network model classification are technical means well-known to those skilled in the art and will not be elaborated here.

[0057] Considering that electromagnetic interference shows a random pulse phenomenon in the time domain and the ultrasonic signal has the characteristic of unstable electronic interference, the analysis of the ultrasonic signal is added in the subsequent analysis. In the embodiment of the present invention, the ultrasonic signal at the position of each pixel point in each suspected defect area in the ultrasonic detection image is obtained through an ultrasonic probe, that is, the waveform from the signal emission to reception by the probe.

[0058] S2: For each ultrasonic detection image, according to the proportion of each suspected defect area and the distance from the module area, the distribution suspected index of each suspected defect area is obtained.

[0059] If pore defects are generated during the integrated circuit packaging process, defects such as micro pores or voids are often related to the packaging structure of the integrated circuit. Such defects often appear at the bonding positions between the packaging material and each module of the circuit during the packaging process, such as the pin part of the processor chip. Therefore, pore defects are often located near each module of the circuit. Please refer to Figure 2 , which shows a schematic diagram of a partial ultrasonic detection image with pore defects provided by an embodiment of the present invention.

[0060] Therefore, the closer the position of the defect is to the module position of the integrated circuit, and the higher the overlapping area of the appearance positions, the greater the possibility of abnormality due to packaging factors during the packaging process, and the greater the possibility that the defect area belongs to a real pore defect. Therefore, the analysis is first carried out from the distribution of the suspected defect areas.

[0061] Preferably, in the embodiment of the present invention, the method for obtaining the distribution suspected index includes:

[0062] First, for any suspected defect area, obtain the shortest distance between the suspected defect area and the module area. After performing negative correlation mapping and normalization processing, obtain the position distribution index of the suspected defect area. Through the shortest distance between the suspected defect area and the module area, reflect the authenticity of the defect from the distribution position characteristics between the suspected defect area and the module. In the embodiment of the present invention, by calculating the Euclidean distance between the center points of the suspected defect area and each module area, and taking the minimum Euclidean distance as the shortest distance, determine the position situation between the areas.

[0063] It should be noted that Euclidean distance, negative correlation mapping, and normalization processing are all technical means well-known to those skilled in the art. Negative correlation mapping can adopt forms such as inverse proportion or negative exponential power, and normalization processing can adopt linear normalization or standard normalization, etc., without further limitation and elaboration.

[0064] Further, normalize the overlapping area between the suspected defect area and the module area to obtain the distribution proportion index of the suspected defect area. The larger the overlapping proportion area, the higher the probability of the defect being reflected.

[0065] Finally, combine the position distribution index and the distribution proportion index of the suspected defect area to obtain the distribution suspicion index of the suspected defect area. In the embodiment of the present invention, take the product of the position distribution index and the distribution proportion index as the distribution suspicion index of the suspected defect area. The larger the position distribution index and the distribution proportion index, the closer the suspected defect area is to the module distribution and the greater the overlapping degree, so the probability that the area is a real pore defect is higher.

[0066] S3: Obtain the pore feature index of each pixel point according to the similarity of the distribution position and ultrasonic signal fluctuation between each pixel point in each suspected defect area and other pixel points; obtain the possible pore defect index of each suspected defect area according to the distribution suspicion index of each suspected defect area and the pore feature index of the pixel point.

[0067] However, defects not in the module part of the integrated circuit are not necessarily real defects. Since electromagnetic noise randomly appears in the ultrasonic image, the area that appears in the module of the circuit may still be electromagnetic noise interference.

[0068] Considering that electromagnetic interference shows the characteristic of random pulses in the time domain, after the ultrasonic signal instantaneously has a large amplitude change, it quickly returns to the normal amplitude. This makes the signal exhibit the characteristic of unstable electronic interference. In normal ultrasonic signals, the signal usually shows a stable change in time series. Therefore, the amplitude of the ultrasonic signal shows different performances at different positions.

[0069] For the real encapsulated pore defects, since the positions of the defects are fixed, the passing of the ultrasonic signals through the pore defects will cause the same degree of signal change in the ultrasonic signals, and the distance between the ultrasonic signals and the defect positions remains unchanged. Therefore, the decay periods of the signals are basically the same. Therefore, further analyze from the distribution of different pixel points and the fluctuation of ultrasonic signals to analyze the degree of pore defects represented by each pixel point position.

[0070] Preferably, in the embodiment of the present invention, the method for obtaining the pore characteristic index includes:

[0071] First, for any pixel point in a suspected defect area, according to the time-sequence fluctuation of the ultrasonic signal of the pixel point according to the frequency, obtain the time-sequence fluctuation sequence of the pixel point. In the embodiment of the present invention, perform Fourier transform on the ultrasonic signal of the pixel point to obtain the frequency of the ultrasonic signal of the pixel point, and analyze the intensity of different frequency components in the signal through Fourier transform. Divide the ultrasonic signal of the pixel point periodically according to the frequency to obtain the amplitude of each period, and arrange the amplitudes in chronological order to obtain the time-sequence fluctuation sequence of the pixel point. It should be noted that Fourier transform and period amplitude acquisition are well-known public technical means for those skilled in the art and will not be elaborated here.

[0072] The amplitude change of the ultrasonic signal reflects the internal structure change of the integrated circuit at the monitoring position. If it belongs to the defect caused by normal encapsulation, all pixel points in the defect area jointly reflect the internal pore defects of the integrated circuit encapsulation. Since the pixel points jointly reflect the internal defect characteristics, they have high similarity, that is, the same degree of amplitude appears in similar periods. Therefore, the ultrasonic signals at similar positions monitor the same defect position of the encapsulation pores, and the ultrasonic signals are similar.

[0073] Further, sequentially take each other pixel point in the suspected defect area where the pixel point is located as an analysis similarity point, and analyze all other pixel points in the area. Perform negative correlation mapping and normalization processing on the distance between the pixel point and the analysis pixel point to obtain the distance credibility of the analysis pixel point. If the distribution distance is small, it is more likely to correspond to the same pore defect situation, and at this time, more attention should be paid to the similarity analysis. When the distribution distance is large, the pore defects may not be at the same position at this time, and the low similarity may also be a normal phenomenon. Therefore, adjust through the distance between the distributions.

[0074] Further, perform negative correlation mapping on the DTW value between the time-sequence fluctuation sequences of the pixel point and the analysis pixel point to obtain the amplitude similarity of the analysis pixel point. The DTW value reflects the similarity degree between the sequences. When the DTW value is smaller, it means that the two sequences are more similar. It should be noted that the DTW value is obtained through the DTW algorithm, which is a well-known public technical means for those skilled in the art and will not be limited here.

[0075] Further, by combining the distance credibility and amplitude similarity of the analyzed pixel points, a similarity index of the analyzed pixel points is obtained. In the embodiments of the present invention, the product of the distance credibility and amplitude similarity of the analyzed pixel points is used as the similarity index of the analyzed pixel points. When the distance credibility and amplitude similarity are larger, it indicates that the analyzed pixel point is closer to the pixel point and the ultrasonic signals are more similar, so the similarity index is larger.

[0076] Finally, the mean value of the similarity indices of all other pixel points of the pixel point is used as the pore feature index of the pixel point. Through comprehensive similarity analysis of all other pixel points, the significance level of the pore feature of the pixel point is obtained.

[0077] For each suspected defect area, by combining the distribution position and the pore feature of the similarity points, the possible degree of the micro-pore defect is comprehensively reflected. Therefore, in the embodiments of the present invention, according to the distribution suspicion index of each suspected defect area and the pore feature index of the pixel points, a pore defect possibility index of each suspected defect area is obtained, including:

[0078] For any suspected defect area, the ratio of the mean value to the standard deviation of the pore feature indices of all pixel points in the suspected defect area is used as the pore feature suspicion index of the suspected defect area. When the representation degree of the overall pore feature of the pixel points in the suspected defect area is larger and the fluctuation degree is smaller, it reflects a higher authenticity of the pore defect in the area.

[0079] Finally, the product of the pore feature suspicion index and the distribution suspicion index of the suspected defect area is used as the pore defect possibility index of the suspected defect area. When the pore feature suspicion index and the distribution suspicion index are larger, it indicates a higher possibility that the suspected defect area is a real pore defect.

[0080] S4: For each suspected defect area in each ultrasonic detection image, analyze the overlapping distribution degree of the suspected defect areas with other ultrasonic detection images, and combine the pore defect possibility index to obtain a corrected possibility index for each suspected defect area; perform pore defect detection based on the corrected possibility index.

[0081] Considering that the ultrasound at the module boundary part is also affected by the contours of each module of the integrated circuit, making the amplitude difference reflect the normal changes of the integrated circuit. At this time, when it is further interfered by electromagnetic noise, the noise at this time still has a high similarity with the real packaging air hole defect. Utilizing the feature that the packaging air hole defect of the integrated circuit is caused by inaccurate packaging process parameters, there is a high probability that the integrated circuit can reproduce the packaging air hole defect under the same process parameters. However, if it is manifested as electromagnetic noise, the position, size, etc. of the ultrasonic images detected each time have a high randomness. Therefore, during multiple ultrasonic detection processes, if the characteristics such as the position of the defect have a high repeatability, the greater the possibility of it being a real packaging defect.

[0082] Therefore, by analyzing the overlapping situation of the defect areas with other ultrasonic detection images, the possible indicators of each defect area are further adjusted to improve the detection accuracy. Preferably, in the embodiment of the present invention, the method for obtaining the corrected possible indicators includes:

[0083] For any ultrasonic detection image, the suspected defect areas in this ultrasonic detection image are sequentially used as the target areas, and each suspected defect area is analyzed. Each other ultrasonic detection image outside this ultrasonic detection image is sequentially used as the analysis image, and each other ultrasonic detection image is analyzed in turn.

[0084] When there is an overlap between the target area and the suspected defect area in the analysis image, it indicates that the defect may have reproducibility in this analysis image. The proportion of the overlapping area between the target area and the suspected defect area in the position-related suspected defect area is used as the defect coincidence degree of the analysis image. In the embodiment of the present invention, the area proportion is the ratio of the overlapping part to the area of the target area. The higher the coincidence degree, the higher the similarity of the size or shape of the defect area, and the higher the credibility of the area as a real air hole defect.

[0085] Furthermore, the product of the possible air hole defect indicator of the suspected defect area overlapping with the target area in the analysis image and the defect coincidence degree is used as the repeated suspected indicator of the analysis image. The larger the repeated suspected indicator, the more likely the reproduction situation in the analysis image is the situation of a real air hole defect, reflecting the degree of correction influence of the analysis image on the reproduction situation of the target area.

[0086] When there is no overlap between the target area and the suspected defect area in the analysis image, it indicates that there is no reproduction situation at this time. The preset repeated indicator is used as the repeated suspected indicator of the analysis image. In the embodiment of the present invention, the preset repeated indicator is set to 0, and the specific value can be adjusted by the implementer himself / herself and is not limited here.

[0087] The sum value of the repeated suspected indicators of all analysis diagrams is used as the correction adjustment degree of the target area, reflecting the degree of influence of the overall target area by the recurrence degree. Finally, by combining the possible indicators of the pore defects in the target area with the correction adjustment degree and performing normalization processing, the corrected possible indicators of the target area are obtained. In the embodiment of the present invention, the sum value of the possible indicators of the pore defects in the target area and the correction adjustment degree is normalized to obtain the corrected possible indicators of the target area. The larger the corrected possible indicator, the higher the possibility that the target area has pore defects.

[0088] The corrected possible indicators of the suspected defect areas analyzed in the ultrasonic detection diagrams can be used to judge the defect situation in the integrated circuit. In the embodiment of the present invention, the maximum corrected possible indicator in each ultrasonic detection diagram is used as the detection indicator of each ultrasonic detection diagram, and the defect characterization possibility in the ultrasonic detection diagram is characterized by the maximum and minimum values.

[0089] When the detection indicator is greater than the preset defect threshold, it indicates that the pore defects are significant, and the corresponding packaging sample of the ultrasonic detection diagram is recorded as having pore defects. In the embodiment of the present invention, the preset defect threshold is set to 0.9, and the specific value can be adjusted by the implementer according to the specific implementation scenario and is not limited here.

[0090] So far, the situation of the pore defects in the samples has been completed. Subsequently, based on the defect situations under different process parameters, the mapping analysis between the defects and the process parameters can be carried out to determine the optimal process parameters to improve the encapsulation quality.

[0091] It is worth mentioning that all the data participating in the calculation in this embodiment have been dimensionless processed to avoid the influence of different dimensions on the parameter analysis. The dimensionless processing is a well-known technical means in the art and will not be elaborated and limited here.

[0092] In summary, the present invention considers the position characteristics of the pore defects and the randomness characteristics of the electromagnetic noise. According to the proportion of the suspected defect areas in the ultrasonic detection diagrams of each sample and the distance distribution from the module area, the distribution suspected indicators are first analyzed from the distribution position, and combined with the similarity of the ultrasonic signal fluctuations between the pixel points in the area, the pore characteristic indicators are analyzed from the internal distribution similarity to reflect the possible degree of the pore defects, providing an analysis basis for the screening of the pore defects. Further, by using the repeatability of higher packaging defects under the same process and combining the overlapping performance of the suspected defect areas between multiple ultrasonic detection diagrams, the possible indicators of the pores are corrected to obtain more accurate corrected possible indicators for pore detection. The present invention improves the accuracy of distinguishing the electromagnetic noise and the pore defects through the similarity characterization degree and the overlapping distribution degree of the defect characteristics in the defect areas between the ultrasonic detection diagrams of multiple samples, making the defect detection more accurate, and thus improving the reliability of the subsequent optimization of the process parameters.

[0093] Embodiment 2:

[0094] The present invention also provides an ultrasonic detection system for integrated circuit plastic encapsulation, including:

[0095] A data acquisition module, configured to acquire ultrasonic detection images of each encapsulation sample under the same batch of processes, determine the suspected defect regions and module regions in each ultrasonic detection image through template matching; and acquire the ultrasonic signals at the positions of each pixel point in each suspected defect region in the ultrasonic detection image.

[0096] An air hole defect analysis module, for each ultrasonic detection image, to obtain the distribution suspected index of each suspected defect region according to the proportion of each suspected defect region and the distance from the module region; to obtain the air hole feature index of each pixel point according to the similarity of the distribution position and ultrasonic signal fluctuation of each pixel point in each suspected defect region to other pixel points; and to obtain the possible air hole defect index of each suspected defect region according to the distribution suspected index of each suspected defect region and the air hole feature index of the pixel points.

[0097] A correction analysis and detection module, for each suspected defect region of each ultrasonic detection image, to analyze the overlapping distribution degree of the suspected defect regions with those in other ultrasonic detection images, and combine with the possible air hole defect index to obtain the possible correction index of each suspected defect region; and to perform air hole defect detection based on the possible correction index.

[0098] It should be noted that for the system provided in the above embodiments, only the above division of each functional module is used for illustration. In actual applications, the above functions can be allocated to different functional module units according to needs, that is, the internal structure of the system is divided into different functional module units to complete all or part of the functions described above. Since the specific implementation process of an ultrasonic detection system for integrated circuit plastic encapsulation in this embodiment is the same as that of an ultrasonic detection method for integrated circuit plastic encapsulation described above, it will not be elaborated here too much.

[0099] Embodiment 3:

[0100] The present invention also provides an ultrasonic detection device for integrated circuit plastic encapsulation, including a memory, a processor, and a computer program stored in the memory and executable on the processor. When the processor executes the computer program, the steps of an ultrasonic detection method for integrated circuit plastic encapsulation as described in any one of the above are implemented.

[0101] It should be noted that: the above sequence of the embodiments of the present invention is only for description and does not represent the superiority or inferiority of the embodiments. The processes depicted in the drawings do not necessarily require the specific order or continuous order shown to achieve the desired results. In some embodiments, multitasking and parallel processing are also possible or may be advantageous.

[0102] Each embodiment in this specification is described in a progressive manner. For the same or similar parts among the embodiments, reference can be made to each other, and the differences between each embodiment and other embodiments are emphasized.

Claims

1. An ultrasonic detection method for integrated circuit plastic packaging, characterized in that: The method comprises: Obtain the ultrasonic inspection image of each packaged sample under the same batch process, determine the suspected defect area and module area in each ultrasonic inspection image through template matching; obtain the ultrasonic signal of each pixel position in each suspected defect area in the ultrasonic inspection image; For each ultrasonic detection image, the distribution suspected index of each suspected defect area is obtained according to the proportion of each suspected defect area and the distance between the suspected defect area and the module area; the pore characteristic index of each pixel point is obtained according to the similarity between each pixel point in each suspected defect area and other pixel points in distribution position and ultrasonic signal fluctuation; the possible pore defect index of each suspected defect area is obtained according to the distribution suspected index of each suspected defect area and the pore characteristic index of the pixel point; For each suspected defect area of ​​each ultrasonic test image, the overlapping distribution degree of the suspected defect areas with other ultrasonic test images is analyzed, and the corrected possible index of each suspected defect area is obtained in combination with the possible index of pore defects; and pore defect detection is performed based on the corrected possible index; The method for obtaining the distribution suspected indicator includes: For any suspected defect area, the shortest distance between the suspected defect area and the module area is obtained, and after negative correlation mapping and normalization processing, the position distribution index of the suspected defect area is obtained; Normalizing the overlapping area between the suspected defect area and the module area to obtain a distribution ratio index of the suspected defect area; Combining the position distribution index and the distribution ratio index of the suspected defect area, obtaining the distribution suspected index of the suspected defect area; The method for obtaining the pore characteristic index includes: For any pixel point in a suspected defect area, a time sequence fluctuation sequence of the pixel point is obtained according to the time sequence fluctuation of the ultrasonic signal of the pixel point according to the frequency; Each other pixel in the suspected defect area where the pixel is located is used as an analysis pixel in turn; the distance between the pixel and the analysis pixel is negatively correlated and normalized to obtain the distance credibility of the analysis pixel; Negative correlation mapping is performed between the DTW values ​​of the time series fluctuation sequence between the pixel point and the analysis pixel point to obtain the amplitude similarity of the analysis pixel point; The similarity index of the analyzed pixel points is obtained by combining the distance credibility and amplitude similarity of the analyzed pixel points; The average of the similarity indexes of all other pixels of the pixel is used as the pore characteristic index of the pixel.

2. The ultrasonic detection method for integrated circuit plastic packaging according to claim 1, characterized in that: The method for obtaining the possible indicators of the pore defects includes: For any suspected defect area, the ratio of the mean value and the standard deviation of the pore characteristic index of all pixels in the suspected defect area is used as the suspected pore characteristic index of the suspected defect area; The product of the suspected index of the pore characteristics and the suspected index of the distribution of the suspected defect area is used as the possible index of the pore defect of the suspected defect area.

3. The ultrasonic detection method for integrated circuit plastic packaging according to claim 1, characterized in that: The method for obtaining the correction possible indicator includes: For any ultrasonic test image, the suspected defect area in the ultrasonic test image is taken as the target area in turn; each other ultrasonic test image outside the ultrasonic test image is taken as the analysis image in turn; When the target area overlaps with the suspected defect area in the analysis diagram, the overlapping area ratio of the target area and the suspected defect area in the analysis diagram is used as the defect overlap of the analysis diagram; the product of the possible index of pore defects and the defect overlap of the suspected defect area in the analysis diagram that overlaps with the target area is used as the repeated suspected index of the analysis diagram; When there is no overlap between the target area and the suspected defect area in the analysis diagram, the preset repetition index is used as the repetition suspected index of the analysis diagram; The sum of the repeated suspected indicators of all analysis graphs is used as the correction adjustment degree of the target area; the possible indicators of pore defects in the target area are combined with the correction adjustment degree and normalized to obtain the possible correction indicators of the target area.

4. The ultrasonic detection method for integrated circuit plastic packaging according to claim 1, characterized in that: The method of performing pore defect detection based on the correction of possible indicators includes: The maximum possible correction index in each ultrasonic test image is used as the detection index of each ultrasonic test image; when the detection index is greater than the preset defect threshold, the packaging sample corresponding to the ultrasonic test image is recorded as having a pore defect.

5. The ultrasonic detection method for integrated circuit plastic packaging according to claim 1, characterized in that: The method of determining the suspected defect area and the module area in each ultrasonic detection image by template matching includes: For any ultrasonic detection image, template matching is performed between the ultrasonic detection image and the module area in the preset template image to obtain the module area of ​​the ultrasonic detection image; The ultrasonic detection image is input into the trained neural network model, and the suspected defect area of ​​the ultrasonic detection image is output.

6. The ultrasonic detection method for integrated circuit plastic packaging according to claim 1, characterized in that: The method for acquiring the time series fluctuation sequence includes: Performing Fourier transform on the ultrasonic signal of the pixel point to obtain the frequency of the ultrasonic signal of the pixel point; The ultrasonic signal of the pixel point is divided periodically according to the frequency to obtain the amplitude of each period; the amplitudes are arranged in time sequence to obtain the time sequence fluctuation sequence of the pixel point.

7. An ultrasonic detection system for integrated circuit plastic packaging, used to implement the ultrasonic detection method for integrated circuit plastic packaging as claimed in claim 1, characterized in that: include: The data acquisition module is used to obtain the ultrasonic inspection image of each package sample under the same batch process, determine the suspected defect area and module area in each ultrasonic inspection image through template matching; and obtain the ultrasonic signal of each pixel position in each suspected defect area in the ultrasonic inspection image; The pore defect analysis module is used to obtain the distribution suspected index of each suspected defect area according to the proportion of each suspected defect area and the distance from the module area for each ultrasonic detection image; obtain the pore characteristic index of each pixel point according to the similarity between each pixel point and other pixel points in the distribution position and ultrasonic signal fluctuation in each suspected defect area; obtain the possible pore defect index of each suspected defect area according to the distribution suspected index of each suspected defect area and the pore characteristic index of the pixel point; The correction analysis detection module is used to analyze the overlapping distribution degree of the suspected defect areas between each suspected defect area of ​​each ultrasonic detection image and other suspected defect areas, and obtain the corrected possible indicators of each suspected defect area in combination with the possible indicators of pore defects; and perform pore defect detection based on the corrected possible indicators.

8. An ultrasonic detection device for integrated circuit plastic packaging, comprising a memory, a processor, and a computer program stored in the memory and executable on the processor, characterized in that: When the processor executes the computer program, the steps of the ultrasonic detection method for integrated circuit plastic packaging as described in any one of claims 1 to 6 are implemented.

Images