Ultrasonic inspection device and ultrasonic inspection method

What is AI technical title?
AI technical title is built by Patsnap AI team. It summarizes the technical point description of the patent document.
The ultrasonic inspection device addresses the challenge of detecting defects at multiple interfaces by correcting and grouping reflected signals and using an image quality conversion table, resulting in high sensitivity and efficient defect detection within a single probe scan.

JP7655806B2Active Publication Date: 2025-04-02HIATACHI POWER SOLUTIONS CO LTD

View PDF 7 Cites 0 Cited by

Patent Information

Application Number: JP2021112338
Authority / Receiving Office: JP · JP
Patent Type: Patents
Current Assignee / Owner
Filing Date: 2021-07-06
Publication Date: 2025-04-02
Estimated Expiration: 2041-07-06

AI Technical Summary

Technical Problem

Existing ultrasonic inspection methods struggle to achieve high sensitivity and efficiency in detecting defects at multiple inspection interfaces within a single probe scan, particularly in objects with complex multilayer structures like electronic components.

Method used

The ultrasonic inspection device employs a processing unit that corrects and groups reflected signals from multiple measurement points, estimates reference reflection intensities, and generates an image quality conversion table to enhance image quality and detect defects effectively.

Benefits of technology

This approach enables highly sensitive detection of defects at each inspection interface from ultrasonic reflected waves acquired in a single probe scan, improving both inspection speed and sensitivity.

✦ Generated by Eureka AI based on patent content.

Smart Images

Figure 0007655806000001
Figure 0007655806000002
Figure 0007655806000003

Patent Text Reader

Abstract

To provide an ultrasonic inspection device capable of performing highly sensitive detection of defects at each inspection interface from ultrasonic reflected waves that are batch acquired by one probe scan on an object to be inspected having a multi-layer structure with multiple inspection interfaces in a non-destructive inspection by ultrasonic waves.SOLUTION: The ultrasonic inspection device performs the steps of: grouping extracted reflected signals into multiple groups (S107); estimating a reference reflection intensity which is a reflection intensity of a normal part as a reference and a reflection intensity range (S108); calculating an image quality conversion table that non-linearly converts the reflection intensity based on the estimated reflection intensity range (S109); converting the reflected signal within each group according to the image quality conversion table (S110); and extracting defects by comparing the converted reflection intensity to the converted reference reflection intensity (S111).SELECTED DRAWING: Figure 5

Need to check novelty before this filing date? Find Prior Art

Description

[Technical field]

[0001] The present invention relates to a non-destructive inspection device, and more particularly to an ultrasonic inspection device and method for determining the presence or absence of defects such as peeling that exist inside an object to be inspected, such as an electronic component, and for visualizing the internal state, using ultrasonic waves. [Background technology]

[0002] Non-destructive inspection methods for detecting defects from images of an object to be inspected include a method in which ultrasonic waves are irradiated onto the object to be inspected, the reflected waves are detected, an ultrasonic image is generated, and defects are identified, and a method in which X-rays are irradiated onto the object to be inspected, the X-rays that pass through the sample are detected, and an X-ray image is generated and defects are identified.

[0003] In general, to detect defects in a multi-layered structured object to be inspected using ultrasonic waves, the reflection characteristics due to differences in acoustic impedance are utilized. Ultrasonic waves propagate through liquids and solid materials, and reflected waves (echoes) are generated at the boundaries and gaps between materials with different acoustic impedances. Here, the reflected waves from defects such as peeling and voids have higher intensity than those reflected from areas without defects, so an image can be obtained in which defects present in the object to be inspected are made visible by imaging the reflection intensity at the bonding surfaces of each layer of the object to be inspected. In order to emphasize the defective areas on the image and enable highly sensitive inspection, it is necessary to irradiate ultrasonic waves to the object to be inspected, and adjust the height of the ultrasonic probe that receives the reflected waves relative to the object to adjust the focus position so that the ultrasonic beam is most focused at the bonding interface to be inspected (hereinafter referred to as the inspection interface) (hereinafter the focused position will be referred to as the focal position).

[0004] On the other hand, in recent years, electronic components, which are representative of the objects to be inspected, have been stacked vertically, and the number of places to be inspected, such as the joints between chips, has increased. For this reason, in order to perform highly sensitive inspections, as mentioned above, it is necessary to perform the tasks of adjusting the focus position, scanning the probe to obtain reflected waves, and generating images for the number of inspection interfaces, which increases the inspection time, and this has become an issue. For this reason, it is often the case that the processing time is reduced by focusing between the inspection interfaces and acquiring images of both the upper and lower inspection interfaces in one probe scan, but this raises concerns about a decrease in defect detection sensitivity. In addition, in order to achieve both inspection throughput and inspection sensitivity, it is necessary to convert the image quality of the image obtained at the out-of-focus position, generate an image with high visibility, and improve inspection sensitivity.

[0005] A method described in Patent Document 1 is a conventional technique that addresses this issue. For an ultrasound image generated by ultrasound, an evaluation weight is set for each small region, and the histogram obtained for the small region is weighted based on the evaluation weight, and the gain and dynamic range for adjusting the image quality of the entire image are obtained from the integrated histogram. Then, the image quality is improved by adjusting the gain and dynamic range. [Prior art documents] [Patent documents]

[0006] [Patent Document 1] JP 2007-117168 A Summary of the Invention [Problem to be solved by the invention]

[0007] The method described in Patent Document 1 targets medical ultrasound diagnostic images, and since the region of interest is known in advance, it is possible to set evaluation weights in advance, but it is not possible to set the region of interest in cases where it is not possible to specify it in advance, such as in electronic components. Also, since image quality is adjusted based on the luminance histogram of the generated image, there is a possibility that it will not be possible to emphasize minute signals that are buried in noise at the image conversion stage due to the weak reflected strength of ultrasound.

[0008] Therefore, an object of the present invention is to provide an ultrasonic inspection device and an ultrasonic inspection method that enable highly sensitive detection of defects at each inspection interface from ultrasonic reflected waves acquired collectively by a single probe scan for an object to be inspected having a multilayer structure with multiple inspection interfaces in ultrasonic non-destructive inspection. [Means for solving the problem]

[0009] In order to solve the above problems, the ultrasonic inspection device of the present invention is an ultrasonic inspection device that scans a surface of an object to be inspected with an ultrasonic probe, emits ultrasonic waves from the ultrasonic probe toward the object to be inspected, receives reflected waves returning from the object to be inspected, and inspects an internal state of the object to be inspected based on the received reflected waves, and a processing unit of the ultrasonic inspection device receives a first gate which is a first reception time region and a second gate which is a second reception time region set by a user, scans the ultrasonic probe and corrects reflected signals obtained from a plurality of measurement points on the object to be inspected in the reception time axis direction based on the first gate, and The present invention is characterized in that a reflected signal from the inspection interface is extracted based on the second gate from the signal, the extracted reflected signals are grouped into a plurality of groups, the reflection intensity of a normal part is estimated for each group as a reference reflection intensity, an image quality conversion table for nonlinearly converting the reflection intensity based on the estimated reference reflection intensity is calculated, the reflected signal in each group is converted according to the image quality conversion table, the converted reflection intensity is compared with the converted reference reflection intensity to extract defects, an image of the inspection interface is generated from the converted reflection intensity in each group, and a defect map is output in which the extracted defects are displayed on the image of the inspection interface. Other aspects of the present invention will be described in the embodiments described later. Effect of the Invention

[0010] According to the present invention, in ultrasonic nondestructive testing, defects at each test interface of an object to be inspected having a multilayer structure with multiple test interfaces can be detected with high sensitivity from ultrasonic reflected waves acquired all at once by a single probe scan. [Brief description of the drawings]

[0011] [Figure 1] 1 is a block diagram showing the configuration of an ultrasonic inspection apparatus according to a first embodiment. [Diagram 2] 2 is a diagram showing the configuration of a detection unit of the ultrasonic inspection device according to the first embodiment. FIG. [Diagram 3] 1 is a schematic diagram showing a vertical structure of a semiconductor package having a multilayer structure, which is an example of an object to be inspected according to the first embodiment. FIG. [Figure 4] 5A to 5C are diagrams showing examples of reflected signals obtained due to differences in focus positions from an object to be inspected having a plurality of bonding interfaces, which is an inspection target according to the first embodiment. [Diagram 5] 4 is a flowchart showing a processing procedure of an ultrasonic inspection method according to the first embodiment. [Figure 6] 5A to 5C are diagrams illustrating an example of behavior of a reflected signal depending on the difference in defect occurrence depth and focus position according to the first embodiment. [Figure 7] 5A to 5C are diagrams illustrating an example of correction of a reception time shift of a reflected signal from an interface according to the first embodiment. [Figure 8] 5 is a diagram showing an example of grouping of reflected waves and setting of normal ranges of intensity according to the first embodiment; FIG. [Figure 9] 5A to 5C are diagrams illustrating an example of an overview of image quality conversion and defect determination processing within a group according to the first embodiment. [Figure 10] 10 is a flowchart showing a processing procedure of an ultrasonic inspection method according to a second embodiment. [Figure 11A] FIG. 11 is a diagram showing the relationship between an image and a small window according to the second embodiment. [Figure 11B] FIG. 13 is a diagram showing a histogram of luminance within a small window. [Figure 11C] FIG. 13 is a diagram showing an example of an image quality conversion table. [Figure 12] FIG. 2 is a diagram illustrating a hardware configuration of a processing unit and the like of the ultrasonic inspection device. [Figure 13] 5A to 5C are diagrams showing examples of output of defect detection results and a defect map according to the embodiment; DETAILED DESCRIPTION OF THE PREFERRED EMBODIMENTS

[0012] The present embodiment is directed to an electronic component having a multi-layer structure such as an IC chip as a main inspection target. The present embodiment relates to an ultrasonic inspection device, an ultrasonic inspection method, and a program that irradiate ultrasonic waves by a single probe scan to an inspection target having multiple bonding interfaces of different depths, including fine mechanical components and circuit patterns inside, generate an inspection interface image with good visibility over the entire area from ultrasonic reflected waves from each bonding interface acquired at once, and enable detection of minute defects at each interface.

[0013] That is, in this embodiment, the reception times of all ultrasonic reflected waves obtained from the measurement area of the object to be inspected are corrected and aligned in the time direction. After alignment in the time direction, the reflected waves are grouped into groups of adjacent reflected waves in both the reception time direction and the azimuth direction, and an image quality conversion table (see, for example, FIG. 11C) is generated for each group, and an image within the group is generated from the reflection intensity within the group based on the image quality conversion table. Furthermore, the images within each group are integrated to generate an image of the inspection interface. This embodiment is particularly effective for high-speed, high-sensitivity nondestructive inspection using ultrasonic waves of objects to be inspected that have fine and multi-layered structures.

[0014] DETAILED DESCRIPTION OF THE PREFERRED EMBODIMENTS An embodiment for carrying out the present invention will be described in detail with reference to the accompanying drawings. For the sake of simplicity, the following description will be given taking as an example an object to be inspected having a multi-layer structure formed by stacking a plurality of electronic devices, such as a 2.5-dimensional or 3-dimensional semiconductor packaging product. However, the present invention should not be interpreted as being limited to the description of the embodiment shown below. It will be easily understood by those skilled in the art that the specific configuration can be changed without departing from the concept or spirit of the present invention.

[0015] First, when ultrasonic waves are irradiated toward the surface of the object to be inspected, the ultrasonic waves propagate inside the object to be inspected, and if there is an interface where the material properties (acoustic impedance) change, part of the ultrasonic waves is reflected. In particular, if there is a gap, most of the ultrasonic waves are reflected. For this reason, by capturing the reflected waves from the desired heterogeneous bonding interface (hereinafter referred to as the inspection interface) and imaging its intensity, an ultrasonic inspection image can be generated in which defects such as voids and peeling are evident. In the following, the "boundary surface" may be referred to as the "interface." Furthermore, the "position" may be referred to as the "location." In the following, the detection target will be described as defects at the heterogeneous bonding interface of a multilayer structure product.

[0016] <<First embodiment>> 1 is a block diagram showing the configuration of an ultrasonic inspection device 100 according to the first embodiment. The ultrasonic inspection device 100 includes a detection unit 1, an A / D converter 6, a processing unit 20 (a signal processing unit 7, an overall control unit 8), a scanner 13, a mechanical controller 16, a user interface unit 17, a storage unit 19, etc.

[0017] The detection unit 1 is configured with an ultrasonic probe 2 (ultrasonic probe) and a flaw detector 3. The flaw detector 3 drives the ultrasonic probe 2 by providing a pulse signal 31 to the ultrasonic probe 2. The ultrasonic probe 2 driven by the flaw detector 3 generates ultrasonic waves U1 and transmits them to the object to be inspected (sample 5) via water. When the transmitted ultrasonic waves U1 are incident on the sample 5 having a multi-layer structure, a reflected wave U2 is generated from the surface of the sample 5 or the heterogeneous junction interface. In this specification, the object to be inspected and the sample have the same meaning. For convenience of explanation, the ultrasonic waves generated by the ultrasonic probe 2 are referred to as "transmitted waves" and the ultrasonic waves received by the ultrasonic probe 2 are referred to as "reflected waves". The reflected waves may also be referred to as "ultrasonic reflected waves" as appropriate.

[0018] As described above, the flaw detector 3 sends a pulse signal 31 to the ultrasonic probe 2, which converts the pulse signal 31 into ultrasonic waves and transmits them to the sample 5. The ultrasonic probe 2 receives the reflected wave U2 from the sample 5 and transmits it to the flaw detector 3. The flaw detector 3 converts the reflected wave 32 into an RF (Radio Frequency) signal 33 and transmits it to the processing unit 20 (control unit) via the A / D converter 6.

[0019] The processing unit 20 realizes mechanical control by sending a control signal to the mechanical controller 16 in order to scan an appropriate portion of the sample 5 using the ultrasonic probe 2. Automatic control (scanning) of the ultrasonic probe 2 is performed through the sequence of the processing unit 20 → mechanical controller 16 → scanner 13 (see FIG. 2) → ultrasonic probe 2 → flaw detector 3.

[0020] As described above, the reflected wave U2 is received by the ultrasonic probe 2, and the flaw detector 3 receives it as a reflected wave 32, which is subjected to necessary processing and converted into a reflection intensity signal. Next, this reflection intensity signal is converted into digital waveform data by the A / D converter 6 and input to the signal processing unit 7. This transmission and reception of ultrasonic waves is performed by sequentially scanning the inspection area on the sample 5.

[0021] The signal processing unit 7 is a processing unit that processes the reflection intensity signal A / D converted by the A / D converter 6 to detect internal defects of the sample 5. The signal processing unit 7 is configured to include an image generating unit 71, a defect detecting unit 72, a data output unit 73, a parameter setting unit 74, etc.

[0022] The image generating unit 71 performs signal processing, described later, on the waveform data input from the A / D converter 6 to the signal processing unit 7. Through this processing, the image generating unit 71 generates an inspection image (inspection interface image) of each bonding interface of the sample 5 from the digital waveform data. The defect detecting unit 72 performs processing, described later, on the image of the inspection interface generated by the image generating unit 71 to detect defects such as peeling and voids. Furthermore, the data output unit 73 generates data to be output as inspection results, such as information on each defect detected by the defect detecting unit 72 and an image for observing a cross section, and outputs the data to the overall control unit 8.

[0023] FIG. 2 is an explanatory diagram showing the configuration of the detection unit 1 of the ultrasonic inspection device 100. In FIG. 2, a coordinate system 10 shows a coordinate system of three orthogonal axes, XYZ. The detection unit 1 has a scanner table 11 and a water tank 12 provided on the scanner table 11. The detection unit 1 also has a scanner 13 that is provided on the scanner table 11 so as to straddle the water tank 12 and is movable in the XYZ directions. The scanner table 11 is a base that is installed almost horizontally (on a surface parallel to the XY plane). The Z axis is an axis along the direction of gravity. Water 14 is poured into the water tank 12 up to the height shown by the solid line, and a sample 5 is placed at the bottom of the water tank 12 (in the water). As described above, the sample 5 is an electronic component that includes a multilayer structure inside, mainly a semiconductor package product. The water 14 is a medium necessary for efficiently propagating the ultrasonic waves emitted from the ultrasonic probe 2 into the inside of the sample 5.

[0024] The ultrasonic probe 2 transmits ultrasonic waves from the ultrasonic emission part at the lower end to the sample 5 and receives the reflected waves returning from the sample 5. The ultrasonic probe 2 is attached to a holder 15 and can be freely moved in the XYZ directions by a scanner 13 driven by a mechanical controller 16. As a result, the ultrasonic probe 2 receives reflected waves from one or more bonding interfaces of the sample 5 that have been accepted in advance by a user (or selected by the signal processing unit 7) while moving in the X and Y directions, and can obtain a two-dimensional image of the bonding interface in the measurement area (XY plane) and detect defects. The ultrasonic probe 2 is connected to a flaw detector 3 that converts the reflected waves into a reflection intensity signal via a cable 22. The two-dimensional image obtained by the ultrasonic inspection device 100 is a cross-sectional image parallel to the XY plane at depth Z. In the following description, a "cross-section along the aaa plane" may be abbreviated as cross-section [aaa]. For example, a cross-section along the XY plane is called a "cross-section [XY]".

[0025] The contents of the signal processing unit 7 will be further described with reference to FIG. 1 and FIG. 2. The image generating unit 71 generates an image from the digital data obtained from the A / D converter 6. The digital data is the result of A / D conversion by the A / D converter 6 of the reflected waves received by the ultrasonic probe 2 after returning from the surface and each bonding interface in the measurement area [XY] of the sample 5 received from the user. The defect detection unit 72 processes the image generated by the image generating unit 71 to make the internal defects visible or detect them. The data output unit 73 outputs the inspection result in which the internal defects are made visible or detected by the defect detection unit 72. The parameter setting unit 74 receives parameters such as measurement conditions input from the outside (for example, a user operating the user interface unit 17) and sets them in the image generating unit 71 and the defect detection unit 72. In the signal processing unit 7, for example, the setting information of the parameter setting unit 74 is stored in the storage unit 19.

[0026] The overall control unit 8 receives parameters (corresponding to conditions described later) and the like from the user. The overall control unit 8 is also connected to a user interface unit 17 and a memory unit 19. The user interface unit 17 highlights defective locations detected by the signal processing unit 7 on the inspection interface image, and displays information such as the number of defects, and the coordinates and dimensions of each defect. The memory unit 19 stores features and images of the defects detected by the signal processing unit 7. The mechanical controller 16 drives the scanner 13 based on control commands from the overall control unit 8. The signal processing unit 7, the flaw detector 3, and the like are also driven by commands from the overall control unit 8.

[0027] The hardware configurations of the signal processing unit 7, the overall control unit 8, and the mechanical controller 16 will be described later with reference to Fig. 12. The signal processing unit 7, the overall control unit 8, and the mechanical controller 16 may be separate pieces of hardware as shown in Fig. 1, or may be integrated into common hardware. Alternatively, the signal processing unit 7 and the overall control unit 8 may be integrated into common hardware without integrating the mechanical controller 16. In the following description, regardless of whether they are integrated, hardware that includes at least one of the signal processing unit 7, the overall control unit 8, and the mechanical controller 16, or that is integrated, may be simply referred to as a "controller."

[0028] <Sample> 3 is a schematic diagram showing the vertical structure of a semiconductor package having a multilayer structure, which is an example of an object to be inspected according to the first embodiment. FIG. 3 shows an object to be inspected 400, which is an example of a sample 5. The object to be inspected 400 is an electronic component having a multilayer structure that is the main object to be inspected, and this figure shows a schematic diagram of the vertical structure of the electronic component. A coordinate system 401 shows a coordinate system of three orthogonal axes, XYZ. The coordinate system 401 is the same as the coordinate system 10 in FIG. 3.

[0029] The test object 400 is a semiconductor device 42 bonded to the printed wiring board 40 at the bottom layer via solder balls 41. The semiconductor device 42 has multiple chips (two chips 43 and 45 in this example) stacked inside it, and the chips are bonded to each other via a bonding portion 44, which is a TSV (Through Silicon Via). The chip and the interposer substrate 46 are bonded to each other via a bump layer 47, which is a microbump. The periphery of the microbump is sealed with a liquid sealing material (underfill, the black part in the figure), and overmolded to seal the whole with resin 48 (the shaded part in the figure), so that it is protected from the outside. If there are defects such as peeling, voids, or cracks at each bonding portion, the device will not operate normally, so high-sensitivity inspection of the bonding portions 44, bump layer 47, etc. is required. When ultrasonic waves 49 are incident from the front side (upper part in the figure) of the test object 400, the ultrasonic waves 49 propagate into the inside of the test object 400. The ultrasonic wave 49 is reflected by the surface of the resin 48, and by the joints 44 with the chips 43, 45, the bump layer 47, and other locations with different acoustic impedances, and these are received by the ultrasonic probe 2 as one reflected wave.

[0030] Fig. 4 is a diagram showing an example of a reflected signal obtained from an object having a plurality of bonding interfaces, which is an inspection target according to the first embodiment, depending on the difference in focus position. Referring to Fig. 4, a problem of ultrasonic inspection of an object having a multi-layer structure and a plurality of bonding interfaces that require high-sensitivity inspection will be described. Note that, hereinafter, when the subject is omitted, the signal processing unit 7 is the processing subject.

[0031] Bonded interfaces 50 and 51 in Fig. 4 are schematic diagrams of the bonded portion 44 and the bump layer 47 of the inspection object 400 shown in Fig. 3. Ultrasonic beams 52 and 53 indicate ultrasonic beams emitted from the ultrasonic probe 2 (see Fig. 1) toward the inspection object.

[0032] Ultrasonic beam 52 is in a state where the focus position is adjusted to the bonded interface 50, that is, where the Z position of the ultrasonic probe 2 is adjusted so that the ultrasonic beam is most focused (→← in the left diagram of FIG. 4) at the position of the bonded interface 50. Ultrasonic beam 53 is in a state where the ultrasonic probe 2 is lowered in the Z direction (probe movement) and the focus position is adjusted between the bonded interface 50 and the bonded interface 51 (→← in the right diagram of FIG. 4).

[0033] Reflected waves 54 and 55 are ultrasonic beams 52 and 53 at their respective focus positions, and are reflected from the bonding interface 50. The reflection intensity of reflected wave 55 is smaller than that of reflected wave 54, which indicates that the image of the inspection interface generated based on the reflection intensities also has a low contrast.

[0034] Therefore, in order to obtain clear inspection images of each bonding interface and perform highly sensitive inspection, it is necessary to perform the operation of adjusting the focus position to each bonding interface, scanning the probe to generate images of the inspection interface, and repeating this operation for each inspection interface, which increases the time required for inspection. In recent years, multi-layer structures have progressed further, and achieving both high sensitivity and high speed inspection of each inspection interface has become a challenge.

[0035] In the method of this embodiment described below, for an object to be inspected having multiple inspection interfaces, multiple inspection interface images can be generated simultaneously using a small number of probe scans, and minute defects can be detected with high sensitivity from each of the generated inspection interface images, thereby achieving both a reduction in image acquisition time and high-sensitivity inspection.

[0036] <Processing of this embodiment> 5 is a flowchart showing the procedure of the ultrasonic inspection method according to the first embodiment, with reference to FIG.

[0037] First, the processing unit 20 acquires a first reference wave, which is a reflected wave of the ultrasonic wave irradiated by the detection unit 1 to the object to be inspected (sample 5) (S101). The first reference wave is acquired from an arbitrary position in the measurement area. The first reference wave may be acquired from at least one position in the XY plane in the measurement area. The acquired first reference wave is displayed on the user interface unit 17 (display unit). The processing unit 20 accepts setting conditions for generating an inspection interface image described later based on the first reference wave (S120). The acceptance is performed by a user who visually recognizes the first reference wave inputting the setting conditions into the user interface unit 17 (S121). The setting conditions include, for example, an S gate (time range, intensity threshold) which is a first gate, an F gate (time range) which is a second gate for generating an inspection interface image described later, the number of bonding interfaces included therein, and the polarity of the reflected wave for generating the image. The setting conditions are an example of the above-mentioned parameters.

[0038] FIG. 6 is a diagram showing an example of behavior of a reflected signal depending on the depth of defect occurrence and the focus position according to the first embodiment. An example of the conditions accepted in S120 will be described with reference to FIG. 6. Reflected signal waveforms 60 to 65 are reflected waves obtained from a sample 5 consisting of two bonded interfaces. Among the diagrams shown side by side in the upper, middle, and lower parts of FIG. 6, the reflected signal waveforms 60 and 61 in the upper part are reflected waves from defects obtained by adjusting the focus position to the upper bonded interface (interface F1), the reflected signal waveforms 64 and 65 in the lower part are reflected waves from defects obtained by adjusting the focus position to the lower bonded interface (interface F2), and the reflected signal waveforms 62 and 63 in the middle part are reflected waves from defects obtained by adjusting the focus position to the middle of the two bonded interfaces (interfaces F1 and F2). The reflected signal waveforms 60, 62, and 64 in the left column are defects existing at the interface F1, and the reflected signal waveforms 61, 63, and 65 in the right column are defects existing at the interface F2. In addition, the solid-line rectangle 67a in the reflected signal waveform 60 indicates the reception time of the reflected wave from the defect at interface F1, and the dashed-line rectangle 68a indicates the reception time of the reflected wave from the defect at interface F2. When the focus is adjusted to interface F1, the defect at interface F1 can obtain the largest reflected wave, and the intensity decreases as the focus position deviates, making detection more difficult. Conversely, the defect at interface F2 becomes easier to detect as the focus position approaches interface F2.

[0039] In this embodiment, the focus position is set to a state like the reflected signal waveforms 62 and 63 in the middle, the first reference wave is acquired and displayed, and the setting conditions from the user are accepted. The gate 66b in the reflected signal waveform 62 or the reflected signal waveform 63 is an S gate for detecting the reflected wave from the surface, and the gate 67b is an F gate for detecting the reflected wave from the bonded interface and generating an image of the inspection interface. In this example, the number of bonded interfaces is set to "2" and the polarity is set to "-" by the user. In this example, one F gate with a wide time width is set so as to cover the reception time of the reflected wave from both the interface F1 and the interface F2, but it is also possible to set an F gate for each bonded interface, like the rectangles 67a and 68a.

[0040] Returning to FIG. 5, based on the conditions received in S120, the processing unit 20 acquires ultrasonic reflected waves from each measurement point while scanning the measurement area (S102), and detects the reflected wave from the surface (hereinafter referred to as the surface echo) that first exceeds the height of the S gate within the S gate (S103).

[0041] Next, the processing unit 20 adjusts the time of the reflected waves based on the reception time of the detected surface echo (S104). That is, the processing unit 20 aligns each reflected wave so that the surface echo of each reflected wave coincides with the surface echo of the first reference wave on the time axis, and extracts a reflected wave (interface echo) of the time width set by the F gate (S105).

[0042] Fig. 7 is a diagram showing an example of correction of the reception time shift of a reflected signal from an interface according to the first embodiment. Fig. 7 shows an example of the reflected signal waveform. Reflected signal waveform 70 is an interface echo (a cut-out of the reflected wave within the F gate) of the reflected wave obtained from four measurement points, and has variation in the time direction. Reflected signal waveform 70A is an interface echo extracted from the reflected wave adjusted based on the reception time shift of the surface echo. The reception time is adjusted based on the surface echo, but it is also possible to adjust the time based on a specific local peak within the F gate.

[0043] Returning to FIG. 5, the processing unit 20 determines whether or not the processing of all reflected waves within the measurement area has been completed (S106), and if not (S106→No), returns to S102, and if completed (S106→Yes), proceeds to S107.

[0044] In S107, the processing unit 20 divides the obtained interface echoes in the entire measurement region into groups according to the time domain and the space domain (grouping).

[0045] FIG. 8 is a diagram showing an example of grouping of reflected waves and setting of normal range of intensity according to the first embodiment. An example of the grouping is shown in FIG. 8. A structure 80 shows an outline of the structure of a sample 5. There are two bonded interfaces 80a and 80b with different depths (Z direction) inside. Also, it is shown that there are two types of structures XY1 and XY2 in the azimuth space (XY plane). In this embodiment, the interface echo is grouped into two groups, the bonded interfaces 80a and 80b, in the time domain, and into structures XY1 and XY2 and the other three groups in the space domain. A reflected signal waveform 81 shows an example of grouping the region of the structure XY1 into time domains Z1 and Z2 (groups 81a and 81b), and similarly, a reflected signal waveform 82 shows an example of grouping the region of the structure XY2 into time domains Z1 and Z2 (groups 82a and 82b). The number of divisions of the time domain corresponds to the number of bonded interfaces of the condition accepted in S120 (see FIG. 5).

[0046] Returning to FIG. 5, next, the processing unit 20 estimates the reflected waves from normal parts (not reflected waves from defective parts) and normal variation for each group (S108). Since the reflected waves within a group may include reflected waves from defects, the normal part may be determined by using the average or median of the reflected waves within the group, or by using the maximum frequency value from the maximum intensity histogram. The variation is calculated based on the standard deviation of the maximum intensity histogram. α1, α2, β1, and β2 in FIG. 8 indicate the calculated variations in normal parts.

[0047] Then, the processing unit 20 calculates an image quality conversion table for improving image quality and making defects visible based on α1, α2, β1, and β2 (S109). The image quality conversion table is a table for nonlinearly converting reflection intensity.

[0048] Fig. 9 is a diagram showing an example of an overview of image quality conversion and defect determination processing within a group according to the first embodiment. In Fig. 9, the coefficients set in the image quality conversion table of each of groups 81a, 81b, 82a, and 82b are indicated by up and down arrows (↑↓). Here, an example is shown in which a coefficient that extends the reflection intensity within each group by gain is set, but a coefficient having logarithmic or exponential characteristics may also be used.

[0049] Returning to Fig. 5, next, the processing unit 20 performs image quality conversion by converting the total reflection intensity value using the coefficients set in the image quality conversion table (S110), compares the reflected wave at each measurement point after image quality conversion with a normal part reflected wave (estimated value) that has been similarly image quality converted, and extracts as defects any points where the difference is greater than a preset threshold value (S111). The reflected signal waveforms 91 and 92 in Fig. 9 described above are the differences between the reflected waves of the reflected signal waveforms 81 and 82 and the normal part reflected waves after image quality conversion, respectively. It is shown that a uniform threshold value is set and defects are detected from interfaces F1 and F2.

[0050] The processor 20 also extracts a representative value from each reflected wave to generate an image with improved image quality through image quality conversion (S112). The representative value may be the maximum intensity value within the F gate of the reflected wave at each measurement point, or the maximum intensity of each of the number of interfaces within the F gate (here, two, time domains Z1 and Z2) may be used.

[0051] The processing unit 20 judges whether or not all groups have been processed (S113), and if not (S113→No), the process returns to S108, and if completed (S113→Yes), the process proceeds to S114. In S114, the processing unit 20 reconstructs an image of the entire measurement region that has been image-converted using the extracted representative value (S114). Here, if one representative value is used in S112, one image is generated in which the interfaces F1 and F2 are integrated, and if a representative value is generated from each time domain, an image is generated for each interface.

[0052] Finally, the processing unit 20 generates a defect map in which the defects detected in S111 are highlighted (colored, etc.) on the reconstructed image of the measurement area after image quality conversion (S115). Here too, a representative value is set to one in S112, and finally, the defects detected in S111 are highlighted (colored, etc.) on the reconstructed image after image quality conversion. When there is one reconstructed image, the defects are displayed in different colors for each detected group (depth). When reconstructing into an image for each interface in the F gate, the defects are highlighted on each reconstructed image.

[0053] 13 is a diagram showing an example of the output of a defect detection result and a defect map according to the embodiment. Image 1300 shows the result of comparing the reflection intensity of the inspection interface before image conversion with the estimated reflection intensity of the normal part, and image 1301 shows the result of comparing them after image conversion. Defects of different depths exist in the area surrounded by circles, and the defects become apparent by image conversion. Image 1302 is a defect map in which the detected defects are superimposed on the inspection interface image after image conversion, with hatch patterns changed according to the reception time (i.e., depth). It becomes possible to check the defects at a glance.

[0054] As described above, in this embodiment, an example has been shown in which grouping according to the spatial domain of reflection intensity is performed based on differences in structures in the azimuth space (XY plane) and an image quality conversion table is generated, but the image quality conversion table may also be generated dynamically.

[0055] <<Second embodiment>> Fig. 10 is a flowchart showing the processing procedure of the ultrasonic inspection method according to the second embodiment. Fig. 10 is the same as the processing of extracting interface echoes from reflected waves in all measurement regions by F gates in S101, S120, and S102 to S105 in Fig. 5. Explanation of the same processing will be omitted.

[0056] In the second embodiment, the processing unit 20 extracts an interface echo (S105), and then extracts a representative value (S206) in the same manner as in S112. The processing unit 20 determines whether or not the processing of the total reflected waves within the measurement region has been completed (S207), and if not (S207→No), the process returns to S102, and if completed (S207→Yes), the process proceeds to S208.

[0057] In S208, the processing unit 20 generates an inspection interface image with the representative value as the brightness value of the image. It may be simply left as an array in the memory. This is, of course, an interface inspection image (data array) based on the representative value of the reflected intensity of ultrasonic waves in the F gate before image quality conversion. The processing unit 20 sets a small window on the generated inspection image (S209), and dynamically calculates an image quality conversion table based on the distribution of brightness values in the small window (S210). Next, the processing unit 20 converts the brightness value of each pixel in the small window based on the calculated image quality conversion table (S211). Then, the processing unit 20 moves the small window on the inspection image (S212), and performs defect detection processing on the image quality converted by setting a threshold value (S213).

[0058] The processing unit 20 determines whether or not the processing of all images has been completed (S214), and if the processing of all images has not been completed (S214→No), the processing returns to S209, and if the processing of all images has been completed (S214→Yes), the processing proceeds to S114.

[0059] In S114, the processing unit 20 generates an image of the entire inspection interface from the images converted in each small window (S114). Finally, the processing unit 20 generates a defect map in which the defects detected in S213 are highlighted (by color, etc.) on the reconstructed image of the measurement area after image quality conversion (S115). The defects are highlighted on the image of the inspection interface before image quality conversion generated in S208.

[0060] So far, we have described examples of highlighting defects by overlaying them on either the image before image conversion or the image after image conversion. However, it is also possible to overlay the image before image conversion using another method, such as deconvolution processing, which removes factors that cause image degradation due to the spread of ultrasonic beams, and output it.

[0061] Fig. 11A is a diagram showing the relationship between an image and a small window according to the second embodiment. Fig. 11B is a diagram showing a histogram of luminance in a small window. Fig. 11C is a diagram showing an example of an image quality conversion table. An example of a method of calculating an image quality conversion coefficient according to the second embodiment will be described using Figs. 11A to 11C.

[0062] The image 1100 shown in FIG. 11A is an image of the inspection interface generated in S208. The small window 1101 is an example of a small window set in the image 1100 of the inspection interface. The histogram 1102 shown in FIG. 11B is a histogram of the luminance in the small window of the image 1100, with the horizontal axis being the luminance value and the vertical axis being the number of pixels of each luminance value in the small window. Here, the luminance range W including the luminance value M having the peak (maximum frequency) of the histogram is determined from the standard deviation. The image quality conversion table 1103 shown in FIG. 11C is an example of a calculated image quality conversion table. The horizontal axis is the luminance value before conversion, and the vertical axis is the luminance value after conversion. In this embodiment, it is shown that linear conversion is performed outside the selected luminance range W, and conversions 1104 and 1105 are performed within the luminance range W. In other words, the image quality conversion table is a table that performs nonlinear conversion. It is possible to generate and output two image quality conversion images using two types of conversion coefficients, or to integrate them into one image and output them.

[0063] The brightness range (gradation width) of the input and output of the image quality conversion table described above may be the same or different. For example, it is possible to convert a value in an 8-bit range to an 8-bit range so as to emphasize a specific gradation, or it is possible to compress the gradation range of a reflected wave with high brightness resolution having a larger gradation range of 10 bits to 16 bits while maintaining the resolution. Also, the small window set in the image may be the same as the entire measurement area.

[0064] On the other hand, grouping may be performed using design data (layout data) without setting a small window. Here, the number of image quality conversion tables is not limited to two, and multiple types may be used.

[0065] <Hardware configuration> FIG. 12 is a diagram showing a hardware configuration of a processing unit and the like of the ultrasonic inspection device. The computer 1200 shown in Fig. 12 is one embodiment of the processing unit 20 (signal processing unit 7, overall control unit 8), user interface unit 17, and storage unit 19 shown in Fig. 1. Each unit may be realized by multiple computers 1200. For example, parallel calculation may be performed by multiple computers 1200. Also, a tablet computer equipped with the user interface unit 17 may be included.

[0066] The computer 1200 includes a memory 1201, a processor 1202, a storage device 1203 such as a hard disk (HD), a communication unit 1204 such as a network interface card (NIC), and a user interface unit 1205. Note that, while a central processing unit (CPU) or a graphics processing unit (GPU) can be considered as an example of the processor, other semiconductor devices may be used as long as they are capable of executing predetermined processing.

[0067] Then, a program stored in the storage device 1203 is loaded into the memory 1201, and the loaded program is executed by the processor 1202. This embodies the functions of the image generating unit 71, the defect detecting unit 72, the data output unit 73, the parameter setting unit 74, and the overall control unit 8 shown in Fig. 1. The storage device 1203 can also correspond to the storage unit 19 in Fig. 1. Although not shown in the figure, the computer 1200 may have a display, a touch panel, a mouse, and a keyboard as the user interface unit 17.

[0068] The ultrasonic inspection device 100 of this embodiment has a means for dividing the ultrasonic reflected waves of each measurement point obtained from an object to be inspected having a complex and multi-layered structure in the reception time direction and the azimuth direction, adaptively calculating an image quality conversion table from the ultrasonic reflection intensity contained in each same reception time domain and same azimuth space domain, and converting the ultrasonic reflection intensity based on the image quality conversion table to generate an image. It also has a means for identifying defects for each reception time domain and azimuth space domain, and displaying the defects in the generated image by color coding for each reception time domain. As a result, an inspection image with high contrast over the entire inspection area is generated from a signal with a wide dynamic range with various reflection intensities of ultrasonic waves obtained to a narrower dynamic range, making it possible to detect minute defects with weak reflection intensity.

[0069] Conventionally, in ultrasonic inspection of semiconductor electronic components, the number of images of the inspection interface has increased due to the lamination of the internal structure. Although high-sensitivity inspection can be achieved by repeatedly adjusting the focus position for each inspection interface and scanning the inspection area with a probe, the problem is that the data acquisition time increases. According to the ultrasonic inspection device 100 of this embodiment, defects at each inspection interface can be detected with high sensitivity from ultrasonic reflected waves acquired collectively by a single probe scan for an object to be inspected having a multi-layer structure with multiple inspection interfaces. This makes it possible to achieve both high-speed inspection processing and high sensitivity inspection.

[0070] The ultrasonic inspection apparatus of the present embodiment described above has the following features. The ultrasonic inspection device 100 scans the surface of the object 400 under inspection with an ultrasonic probe (e.g., ultrasonic probe 2), emits ultrasonic waves from the ultrasonic probe toward the object 400 under inspection, receives reflected waves returning from the object 400 under inspection, and inspects the internal state of the object 400 under inspection based on the received reflected waves. The processing unit 20 of the ultrasonic inspection device 100 receives a first gate (e.g., S gate) which is a first reception time domain and a second gate (e.g., F gate) which is a second reception time domain set by the user (S120 in FIG. 5), corrects the reflected signals obtained from a plurality of measurement points on the object under inspection by scanning the ultrasonic probe in the reception time axis direction based on the first gate (S104 in FIG. 5), and extracts a reflected signal from the inspection interface from the reflected signal after the reception time correction based on the second gate (S105).

[0071] The processing unit 20 groups the extracted reflected signals into a plurality of groups (S107), estimates the reflection intensity of normal parts for each group as a reference reflection intensity (S108), calculates an image quality conversion table that nonlinearly converts the reflection intensity based on the estimated reference reflection intensity (S109), converts the reflected signals in each group in accordance with the image quality conversion table (S110), and compares the converted reflection intensity with the converted reference reflection intensity to extract defects (S111).

[0072] An image of the inspection interface is generated from the reflection intensity converted within each group (S114), and a defect map in which the extracted defects are displayed on the image of the inspection interface is output (S115). According to the ultrasonic inspection device of this embodiment, for an object to be inspected having a multi-layer structure with multiple inspection interfaces, it is possible to detect defects at each inspection interface with high sensitivity from ultrasonic reflected waves acquired collectively by a single probe scan.

[0073] The grouping of the reflected signals can be based on either or both of the received time domain and the azimuth spatial domain (see Figs. 8 and 9).

[0074] The second gate is a time range that includes the reflected waves from multiple interfaces (see Figure 6).

[0075] The inspection image and defect map generated from the second gate are one sheet, and the defect map is color-coded according to the grouping based on the reception time domain. In Fig. 6, one setting of an F gate with a wide time width is accepted, and the display method of the defects detected in the F gate is changed according to the depth of occurrence as shown in Fig. 13, and the defects are displayed superimposed on one defect map, which enables the user to reduce the time required for setting the F gate.

[0076] For example, in the case of two bonded interfaces, in the conventional method, the user needs to estimate the generation time of the reflected wave from each interface from the first reference wave, and set the start time and time width of the F gate so that the reflected wave from the interface is included but the reflected wave from another interface is not included. Then, the user needs to visually check the images of interfaces F1 and F2 generated based on the two set gates, and if the desired image is not obtained, the user needs to readjust the F gate, and repeat this process.

[0077] In contrast, in this embodiment, a gate is simply set roughly to cover the reflected waves from multiple interfaces. In addition, since there is only one defect map, the occurrence status of defects in the time domain and the space domain is visualized simultaneously, so that the time required for the user to check and analyze the defect detection results can be reduced.

[0078] A defect map of the detected defects is highlighted and superimposed on an image of the inspection interface before intensity conversion, an image of the inspection interface after intensity conversion, and an image of the inspection interface that has been image-quality converted using a different method (see Figure 13).

[0079] In the ultrasonic inspection apparatus of the second embodiment, the processing unit 20 groups the extracted reflected signals into groups of multiple reflected waves obtained from measurement points within a certain distance in the azimuth direction (S209 in FIG. 10), estimates the reflection intensity of normal parts for each group as a reference reflection intensity (see brightness value M in FIG. 11), calculates an image quality conversion table that nonlinearly converts the reflection intensity based on the estimated reference reflection intensity (S210), converts the reflected signals in each group in accordance with the image quality conversion table (S211), compares the converted reflection intensity with the converted reference reflection intensity or a preset threshold value to extract defects (S213), generates an image of the inspection interface from the converted reflection intensity in each group (S114), and outputs a defect map in which the extracted defects are displayed on the image of the inspection interface (S115).

[0080] The image quality conversion table for the reflected signal is of one or more types, and the image and defect map obtained after the intensity conversion are one. According to this embodiment, since there is one defect map, the occurrence status of defects in the time domain and the space domain can be visualized simultaneously, and therefore, the time required for the user to check and analyze the defect detection results can be reduced. [Explanation of symbols]

[0081] 1. Detection section 2 Ultrasound probe 3 Flaw detector 4 Reflected wave 5. Sample 6 A / D Converter 7 Signal Processing Section 8 Overall control section 11 Scanner stand 12 Aquarium 13. Scanner 15 Holder 16 Mechanical Controller 17 User Interface Section 19 Memory section 20 Processing section 44 Joint 47 Bump Layer 50,51 Joint interface 52,53 Ultrasonic beam 66b Gate (First Gate, S Gate) 67b Gate (Second Gate, F Gate) 70,70A reflected signal waveform 71 Image Generation Unit 72 Defect detection section 73 Data output section 100 Ultrasonic inspection equipment 400 Test subject 1100 images 1101 Small Window 1102 Histogram 1103 Image quality conversion table 1300 images (images of the inspection interface before intensity conversion) 1301 Image (image of inspection interface after intensity conversion) 1302 Image (Defect Map) F1,F2 interface M Luminance value W Brightness Range

Claims

1. 1. An ultrasonic inspection device that scans a surface of an object to be inspected with an ultrasonic probe, emits ultrasonic waves from the ultrasonic probe toward the object to be inspected, receives reflected waves returning from the object to be inspected, and inspects an internal state of the object to be inspected based on the received reflected waves, The processing unit of the ultrasonic inspection device includes: a first gate which is a first reception time region and a second gate which is a second reception time region set by a user are received, the ultrasonic probe is scanned to correct reflected signals obtained from a plurality of measurement points on an object under test in a reception time axis direction based on the first gate, and a reflected signal from the inspection interface is extracted from the reflected signal after the reception time correction based on the second gate; The extracted reflected signals are grouped into multiple groups. for each group, estimating a reflection intensity of a normal portion as a reference reflection intensity, calculating an image quality conversion table for nonlinearly converting the reflection intensity based on the estimated reference reflection intensity, converting the reflection signals in each group according to the image quality conversion table, and comparing the converted reflection intensity with the converted reference reflection intensity to extract defects; An image of the inspection interface is generated from the converted reflection intensity within each group, and a defect map is output in which the extracted defects are displayed on the image of the inspection interface.