Advanced packaging inspection is becoming more challenging as wafer-level packaging (WLP), including fan-out wafer-level packaging (FOWLP), adopts finer circuit features and increasingly complex process flows. Conventional optical inspection can struggle with two different problems: transparent organic residues may be difficult to see, while harmless variations in metal grain texture can be mistaken for defects. A specialized illumination approach described for advanced packaging applications is designed to improve visibility of organic material while reducing interference from metal surface texture.
Why advanced packaging inspection is becoming more demanding
Advanced packaging processes bring technologies and materials associated with front-end manufacturing into back-end applications. In fan-out packaging, redistribution layers (RDLs) form conductive paths, while organic polymer dielectrics provide insulation and planar surfaces between layers. As these structures become more complex, maintaining process control increasingly depends on detecting defects at critical steps.
Feature dimensions add to the challenge. Early fan-out implementations used RDL line widths of around 20 micrometers; the source describes a trajectory toward 2 micrometers, with further reductions anticipated. At smaller dimensions, inspection must distinguish defects that could affect the structure from surface variations that are part of the material itself. Missing a critical defect or generating excessive nuisance detections can both make inspection results less useful for process monitoring.
Two inspection challenges: nuisance signals and transparent residue
Metal grain texture can resemble defects
Metal layers are essential to packaging interconnects, but their surfaces can contain visible grain patterns. In bright-field and dark-field inspection, variations in grain appearance may be identified as anomalies even when they do not represent a process defect. The described grain features can be larger than the RDL lines themselves, making it difficult to separate their signals from those of genuinely important features.
When inspection systems count large numbers of these nuisance signals, they can obscure defects that need attention. Inspection methods therefore need to handle both sensitivity and selectivity: detecting relevant abnormalities without treating every change in surface appearance as a failure.
Organic residues may be hard to see
A separate challenge arises when etching or descumming leaves organic material behind. Such residues can be transparent at visible wavelengths, producing little signal under conventional bright-field or dark-field illumination. Residue on contacts, including bumps and pillars, is a particular concern because it may be located on a feature that needs to remain exposed.
The inspection issue is not simply that the residue is small. Its optical response can make it blend into the surrounding structure, leaving a conventional image that appears acceptable even when material remains. That limits the usefulness of inspection if the relevant defect does not produce enough contrast to be recognized.
How specialized illumination changes the image
The Clearfind illumination technology described in the source uses optical properties associated with certain organic materials. Some organic molecules and polymers respond differently to illumination than metals, silicon, and other inorganic materials commonly found in semiconductor manufacturing. Under suitable illumination, this difference can create a detectable color-contrast signal from organic features against an inorganic background.
The emitted signal is described as anisotropic, meaning it does not respond to surface topography in the same way as ordinary reflected light. This can make the signal less vulnerable to surface features directing reflected light away from the detector. The approach is intended to increase sensitivity to organic residues while reducing interference from surrounding features, including variations associated with metal grains.
The system described uses a laser-based light source rather than the broadband source typically used in white-light inspection. The source notes that this provides greater stability in spectral range and output power. Autofocusing uses a near-infrared laser triangulation system to maintain the distance between the imaging optics and the scanned area. Images are acquired with a high-resolution camera. These elements support image acquisition, while the distinguishing feature is the illumination method and the contrast it can produce.
What the TSV example illustrates
The source presents a 300-millimeter silicon wafer with revealed through-silicon via (TSV) pillars. The TSVs are on the wafer backside, with the front side attached to a carrier. In the example, the TSV nail diameter is about 8 micrometers, and the spacing between TSVs is about 56 micrometers.
Two TSVs illustrate the difference between bright-field and specialized illumination images. One has unetched organic residue remaining on it; the other is fully exposed. In the bright-field images, both appear acceptable, so the residue is not readily distinguishable. In the Clearfind images, the residue-covered TSV produces a bright signal while the exposed TSV appears dark. The comparison shows how changing illumination can reveal a material difference that is not apparent in the conventional image.
A wafer map in the example also shows a concentration of detected defects toward the wafer’s right edge. The map provides a spatial view of the inspection results; it should be interpreted as a pattern to investigate rather than, by itself, an explanation of the cause.
Practical considerations for packaging teams
For process engineers and manufacturing teams, the central lesson is that inspection performance depends on the interaction between the defect material, the surrounding structure, and the illumination used. A process that leaves transparent residue may require an inspection method that generates contrast from that material. Conversely, a method that responds strongly to metal grain texture may create a high nuisance count without improving visibility of the defects of interest.
When evaluating inspection for WLP or FOWLP, teams can consider:
- Defect type: Identify whether the primary concern is organic residue, metal-related variation, or another feature requiring detection.
- Material contrast: Consider whether the target defect is visible under the illumination used in the current inspection flow.
- Inspection selectivity: Review whether nuisance detections from metal texture make it harder to identify actionable findings.
- Process location: Pay particular attention to steps where residues may remain on contacts, bumps, pillars, or other features that need to be exposed.
- Spatial patterns: Use wafer-level maps to locate areas with concentrated detections and guide follow-up process review.
Conclusion
As packaging geometries shrink and process integration becomes more demanding, advanced packaging inspection must reveal relevant defects without being overwhelmed by harmless material variation. The illumination approach described here addresses a specific gap in conventional optical inspection: it can make certain organic residues more visible while reducing sensitivity to signals associated with metal grain texture. For manufacturers developing inspection strategies, matching illumination to the defect and materials of interest is a practical step toward clearer process-monitoring results.