MOSFET void inspection can help manufacturers and quality teams assess imperfections in the solder layer beneath a power semiconductor die without dismantling or damaging the device. This matters because voids reduce the effective soldered area and can influence how heat moves from the silicon junction to the heat spreader, as well as the joint’s electrical and mechanical behavior. Micro-computed tomography (micro-CT), paired with image processing, offers a way to examine this internal structure and quantify defects.
In a study of a power MOSFET, researchers used scanning electron microscopy with energy-dispersive spectroscopy (SEM-EDS) to examine solder-layer composition and structure, then used micro-CT as a non-destructive inspection method. Image-processing software was applied to identify voids and calculate their area, size category and location. The approach illustrates how internal inspection can support semiconductor manufacturing quality control and device characterization.
Why power MOSFET solder voids matter
The die-attach solder layer provides a connection between the silicon die and the structure that carries heat away from it. Cavities in this joint interrupt the material and reduce the area available for heat transfer. Depending on their size and position, voids may therefore affect a device’s thermal performance. Solder joints also contribute to electrical and mechanical connections, so defects can be relevant to more than thermal management alone.
Voids can form when gas remains trapped in molten solder during assembly. The source study notes that both insufficient and excessive solder can create different assembly problems: too little can contribute to voided or open joints, while too much can lead to solder balling or short circuits. The shape and distribution of voids matter as well as their total area. A void near a chip corner may have a different effect from one near the center, while defects distributed across a grid can present another pattern.
The paper cites a 25% maximum allowable voiding area in the context of IPC-7095. This figure is a reference point discussed by the authors, not a substitute for the applicable acceptance criteria for a particular product, assembly process or quality plan.
Micro-CT for non-destructive MOSFET void inspection
Conventional X-ray inspection can reveal some solder-joint defects, including voids, but image quality and the limitations of a two-dimensional view can make analysis challenging. Micro-CT captures a series of X-ray projections as the sample is rotated. Reconstruction software combines those projections into cross-sectional images and a three-dimensional representation, enabling inspection of internal regions that are not directly visible from the device exterior.
Because the method does not require opening the package, the inspected component can remain intact. This is useful when the purpose is to evaluate a finished device or study its internal construction without destroying the sample. In the reported setup, the operators first selected suitable scan conditions and positioning to obtain adequate contrast. They then reconstructed the acquired data into images that showed the solder layer and its internal voids.
SEM-EDS served a different, complementary purpose in the investigation. It was used to examine the exposed device structure and assess physical and chemical characteristics of the solder layer. Since this preparation involved unpacking the device, it was not the non-destructive inspection step. The distinction is important: microscopy can support detailed material characterization, while micro-CT can examine internal features without the same destructive preparation.
Turning scan images into measurable quality data
Tomographic images need careful processing before void measurements can be trusted. Distortion, contrast variation and the choice of segmentation threshold can affect which regions software classifies as defects. The study used a LabVIEW-based image-processing program to calibrate images against a scale, correct image alignment and distortion, filter the image, and automatically identify void regions.
After segmentation, the software extracted measurements and summaries including:
- Void area and perimeter: measurements used to describe the size and boundary of identified defects.
- Void count and size groups: identification of macrovoids and microvoids for separate analysis.
- Void location: mass-center coordinates that show where defects occur within the solder layer.
- Area statistics: percentage voiding, average measurements and a histogram showing the distribution of void areas.
Calibration is more than a presentation step. The authors reported slightly different void-percentage results when calculations used pixel values versus measurements from a calibrated image: approximately 2.195% and 2.175%, respectively. The difference demonstrates why converting image dimensions to a real scale can matter when reporting quantitative results.
Interpreting the reported inspection results
For the STP20N95K5 device examined, the measured void area was approximately 2.2% of the solder region. The analysis indicated that microvoids contributed 0.3% of the defects, while macrovoids accounted for 99.68% of the void percentage. The device passed the quality-control comparison described in the study, which used the cited 25% limit.
These figures describe the tested component and the study’s inspection and analysis process; they should not be taken as typical results for all power MOSFETs. The broader value of the method is its ability to produce more than a single pass-or-fail percentage. By quantifying defect size and position, engineers can review how voids are distributed and use that information to investigate assembly quality or characterize device construction.
Practical considerations for inspection programs
Micro-CT can provide detailed internal data, but the study also identifies inspection time as a limitation for potential in-line use. Production teams considering the method should account for scan and analysis time alongside the value of the information it provides. Consistent image calibration, threshold selection and segmentation are also important for making results interpretable and comparable.
For quality engineers and system integrators, the practical workflow is to define the inspection region, acquire images with suitable contrast, reconstruct the volume, calibrate the image, and apply a documented segmentation method. Reviewing both total void area and spatial distribution helps avoid reducing an internal defect assessment to one number alone.
Conclusion
Micro-CT-based MOSFET void inspection offers a non-destructive way to view and measure solder-layer defects inside power devices. When combined with calibrated image processing, it can report void area, size categories and location, supporting quality control and manufacturing analysis. The reported device showed approximately 2.2% voiding, but the key takeaway is the inspection method: reliable evaluation depends on sound image acquisition, calibration and interpretation against the acceptance criteria relevant to the application.