Substrate welding line detection method and device based on thermal-acoustic coupling excitation

Abstract

The invention discloses a substrate welding line detection method based on thermal-acoustic coupling excitation. The method comprises the following steps: applying synchronous thermal excitation and ultrasonic phased array scanning to a substrate; synchronously acquiring a thermal response sequence image of the surface of the substrate and full-matrix echo data of the ultrasonic phased array; processing the thermal response sequence image to obtain a thermal feature image; processing the full-matrix echo data to obtain acoustic features; and fusing the thermal characteristics and the acoustic characteristics to judge the state of the welding lines in the substrate. The method disclosed by the invention can be used for efficiently and accurately detecting and quantitatively evaluating the anisotropy of the welding lines of the aluminum diamond substrate on line.

Description

TECHNICAL FIELD

[0001] The present application relates to the technical field of material nondestructive testing, in particular to a substrate fusion line detection method and device based on thermal-acoustic coupling excitation. BACKGROUND

[0002] Aluminum diamond substrate has become a key heat dissipation material for a new generation of high-power IGBT modules due to its extremely high thermal conductivity and matching thermal expansion coefficient of silicon chips. However, internal fusion lines are easily formed during the preparation of the substrate. Although such fusion lines are not macroscopic cracks, they can cause anisotropy in the thermal and mechanical properties of the substrate, that is, there are differences in physical parameters such as thermal conductivity and sound speed in different directions, thereby affecting the uniformity of heat dissipation and long-term reliability of the chips thereon. Therefore, it is crucial to screen the aluminum diamond substrate for fusion lines before module packaging.

[0003] Existing nondestructive testing methods such as single ultrasonic testing or infrared thermal imaging testing have limitations. Ultrasonic testing is sensitive to acoustic impedance changes and is good at finding defects such as delamination and pores, but for fusion lines showing gradient changes, the image contrast is low and it is difficult to quantify the anisotropy caused thereby. Infrared thermal imaging is sensitive to thermal conductivity differences, but is easily disturbed by uneven surface emissivity and environmental thermal noise, and has a high false positive rate when used alone. SUMMARY

[0004] The present application relates to the technical field of material nondestructive testing, in particular to a substrate fusion line detection method and device based on thermal-acoustic coupling excitation. The second object is to provide a substrate fusion line detection device based on thermal-acoustic coupling excitation for realizing the above detection method.

[0005] Technical solution: The substrate fusion line detection method based on thermal-acoustic coupling excitation disclosed by the present application comprises the following steps: applying synchronous thermal excitation and ultrasonic phased array scanning to the substrate; synchronously collecting thermal response sequence images of the substrate surface and full-matrix echo data of the ultrasonic phased array; processing the thermal response sequence images to obtain thermal characteristics; processing the full-matrix echo data to obtain acoustic characteristics; fusing the thermal characteristics and the acoustic characteristics to determine the internal fusion line state of the substrate.

[0006] Further, the thermal excitation is periodic thermal excitation.

[0007] Further, when processing the thermal response sequence images, phase-locked thermal processing is adopted, and the thermal characteristics are phase images representing internal thermal resistance abnormalities.

[0008] Further, processing the full matrix echo data includes reconstructing an internal structure image using a full focusing algorithm, analyzing and calculating the sound velocity distribution and acoustic anisotropy coefficient by analyzing the sound wave signals in different propagation directions as acoustic characteristics.

[0009] Further, the fusion of thermal characteristics and acoustic characteristics determines the internal fusion line state of the substrate, including the following steps: inputting the thermal characteristics and acoustic characteristics into a pre-trained machine learning classifier using feature-level fusion, and the classifier outputs a determination result.

[0010] Further, the output of the classifier includes the severity level of the fusion line.

[0011] To achieve the above detection method, the application also provides a substrate fusion line detection device based on thermal-acoustic coupling excitation, comprising: a clamp for carrying and fixing the substrate; a thermal excitation unit for applying controllable thermal excitation to the substrate; an infrared thermal imager for collecting thermal response sequence images of the substrate surface; an ultrasonic phased array probe for scanning the substrate and collecting full matrix echo data; a control and processing unit for executing the substrate fusion line detection method based on thermal-acoustic coupling excitation of claim 1, the control and processing unit being in communication connection with the thermal excitation unit, the infrared thermal imager and the ultrasonic phased array probe.

[0012] Further, the thermal excitation unit is an infrared lamp array, and the ultrasonic phased array probe is a multi-chip linear array probe.

[0013] Further, it further comprises a detection chamber, a conveying track is arranged through the detection chamber, the clamp is a pneumatic clamp arranged at intervals on the conveying track, and the clamp is transported to a detection position by the conveying track, the thermal excitation unit is located at the bottom of the clamp corresponding to the detection position, is fixed by lifting equipment, and the substrate is lifted to a detection height by rising, the infrared thermal imager and the ultrasonic phased array probe are arranged in the detection chamber corresponding to the detection position, and the control and processing unit controls the whole equipment and performs data collection and processing.

[0014] Advantages: Compared with the prior art, the application has the following obvious advantages: Through the coupling excitation and information fusion of the two physical modalities of heat and sound, the advantages are complementary; thermal imaging effectively captures the abnormal thermal conductivity caused by the fusion line, acoustic imaging accurately depicts the structure and mechanical anisotropy of the fusion line, and machine learning algorithm is used for comprehensive interpretation, which significantly improves the accuracy and reliability of the detection; The method is non-invasive, fast, and can realize 100% incoming inspection of aluminum-diamond substrates during mass production. Attached Figure Description

[0015] Figure 1 This is a flowchart of the detection method of the present invention; Figure 2 This is a diagram of the detection device of the present invention. Detailed Implementation

[0016] The technical solution of the present invention will be further described below with reference to the accompanying drawings and embodiments.

[0017] Combination Figure 1 and Figure 2 Weld lines were detected on aluminum-diamond substrates.

[0018] Figure 2 The detection device designed to realize the substrate weld line detection method based on thermo-acoustic coupling excitation of the present invention includes a detection chamber 6, a through-passage conveyor track 7 is provided in the detection chamber 6, the clamps 2 are pneumatic clamps spaced apart on the conveyor track 7 and are transported to the detection position by the conveyor track 7, the thermal excitation unit 3 is located at the bottom of the clamp 2 corresponding to the detection position and is fixed by a lifting device, and the substrate 1 is lifted to the detection height by lifting. The detection chamber 6 is equipped with an infrared thermal imager 4 and an ultrasonic phased array probe 5 corresponding to the detection position.

[0019] In the device: clamp 2 is used to support and fix the substrate 1.

[0020] The thermal excitation unit 3 is an infrared lamp array used to apply controllable thermal excitation to the substrate 1.

[0021] Infrared thermal imager 4 is used to acquire sequential images of the thermal response on the surface of substrate 1.

[0022] The ultrasonic phased array probe 5 is a multi-chip linear array probe used to scan the substrate 1 and acquire full matrix echo data.

[0023] The system employs an industrial computer control and processing unit 8 that integrates a data acquisition card, a signal generator, and a running detection algorithm. The control and processing unit 8 is communicatively connected to the thermal excitation unit 3, the infrared thermal imager 4, and the ultrasonic phased array probe 5 to control the overall equipment and perform data acquisition and processing.

[0024] When conducting the test, combine Figure 1 and Figure 2 This includes the following steps: The aluminum diamond substrate to be tested is fixed in the fixture 2 and transferred to the detection position, so that it is relative to the infrared thermal imager and ultrasonic phased array probe of the thermal excitation unit. Synchronous periodic thermal excitation (such as 1Hz modulated infrared lamp array) and ultrasonic phased array fan-shaped scanning are applied to the substrate 1. Synchronously acquire thermal response sequence images of the substrate surface and full-matrix echo data of the ultrasonic phased array; The phase image representing internal thermal resistance anomaly is extracted by phase-locked heat treatment on the thermal response sequence images as the thermal phase anomaly feature; the internal structure image is reconstructed by using the full-focusing algorithm, and the sound velocity distribution map and acoustic anisotropy coefficient are obtained by analyzing and calculating the sound wave signals in different propagation directions as the acoustic anisotropy feature; The thermal phase anomaly feature and the acoustic anisotropy feature are input into the pre-trained feature-level fusion machine learning classifier, and the output of the classifier includes the severity level of the fusion line, and the output is the conclusion of "no fusion line", "slight anisotropy" or "severe anisotropy".

[0025] The thermal excitation unit 3 is lowered, and the substrate 1 falls to the clamp 2, and is transmitted out of the detection position by the conveying track 7. At the same time, the next clamp 2 drives the next substrate 1 to be transmitted to the detection position. After positioning, the pneumatic clamp is deflated, and the thermal excitation unit 3 is lifted for detection operation.

Claims

1. A method for detecting weld lines on a substrate based on thermo-acoustic coupling excitation, characterized in that, Includes the following steps: Synchronous thermal excitation and ultrasonic phased array scanning are applied to the substrate; Simultaneously acquire thermal response sequence images of the substrate surface and full matrix echo data of the ultrasonic phased array; Process thermal response sequence images to obtain thermal characteristics; Process the full matrix echo data to obtain acoustic characteristics; The state of weld lines inside the substrate is determined by combining thermal and acoustic characteristics.

2. The substrate weld line detection method based on thermo-acoustic coupling excitation according to claim 1, characterized in that: The thermal excitation is periodic thermal excitation.

3. The substrate weld line detection method based on thermo-acoustic coupling excitation according to claim 1, characterized in that: Phase-locked thermal processing is used when processing thermal response sequence images, and the thermal characteristics are phase maps that characterize internal thermal resistance anomalies.

4. The substrate weld line detection method based on thermo-acoustic coupling excitation according to claim 1, characterized in that: Processing full matrix echo data includes reconstructing the internal structure image using a full focusing algorithm, and analyzing and calculating the sound velocity distribution map and acoustic anisotropy coefficients from acoustic signals in different propagation directions as acoustic features.

5. The substrate weld line detection method based on thermo-acoustic coupling excitation according to claim 1, characterized in that: The determination of the weld line state inside the substrate by integrating thermal and acoustic features includes the following steps: inputting the thermal and acoustic features into a pre-trained machine learning classifier that uses feature-level fusion, and the classifier outputs the determination result.

6. The substrate weld line detection method based on thermo-acoustic coupling excitation according to claim 5, characterized in that: The classifier's output also includes the severity level of the weld lines.

7. A substrate weld line detection device based on thermo-acoustic coupling excitation, characterized in that, include: A clamp (2) for supporting and fixing the substrate (1); Thermal excitation unit (3) for applying controllable thermal excitation to substrate (1); Infrared thermal imager (4) used to acquire thermal response sequence images of the substrate (1) surface; Ultrasonic phased array probe (5) used to scan the substrate (1) and acquire full matrix echo data; The control and processing unit (8) is used to perform the substrate weld line detection method based on thermo-acoustic coupling excitation as described in claim 1. The control and processing unit (8) is communicatively connected to the thermal excitation unit (3), the infrared thermal imager (4), and the ultrasonic phased array probe (5).

8. The substrate weld line detection device based on thermo-acoustic coupling excitation according to claim 7, characterized in that: The thermal excitation unit (3) is an infrared lamp array; the ultrasonic phased array probe (5) is a multi-chip linear array probe.

9. The substrate weld line detection device based on thermo-acoustic coupling excitation according to claim 7, characterized in that, It also includes a detection chamber (6), in which a through-passage conveyor track (7) is provided. The clamp (2) is a pneumatic clamp spaced at intervals on the conveyor track (7) and is transported to the detection position by the conveyor track (7). The thermal excitation unit (3) is located at the bottom of the clamp (2) corresponding to the detection position and is fixed by a lifting device. The substrate (1) is lifted to the detection height by the lifting device. The detection chamber (6) is equipped with an infrared thermal imager (4) and an ultrasonic phased array probe (5) corresponding to the detection position. The control and processing unit (8) controls the overall equipment and performs data acquisition and processing.

Images