Skip to content
Theme
OnDevice Tools

Tech BusinessChips & Semiconductors

Acoustic micro imaging helps inspect advanced chip packages

Semiconductor Engineering explains how high-frequency ultrasound can reveal hidden defects in advanced packaging, and why the method matters as chip designs grow more complex.

Acoustic micro imaging scanner inspecting a semiconductor package for hidden gaps and delamination

Key takeaways

  • Acoustic micro imaging uses high-frequency ultrasound to reveal gaps and delamination in advanced chip packages.
  • AMI software can count defects and help decide whether parts are accepted or rejected.
  • Nordson Test & Inspection developed SpinSAM, which can inspect four 300mm wafers at once.
  • Machine learning can help AMI software ignore via patterns that resemble delamination.
In this article
  1. What acoustic micro imaging does
  2. How inspectors read the results
  3. Why X-rays and destructive tests still have a role
  4. Machine learning and faster inspection
  5. Why it matters for chip makers

What acoustic micro imaging does

Semiconductor Engineering published a guide on 16 November 2026 about acoustic micro imaging, or AMI, an ultrasound-based inspection method used to find hidden defects in advanced chip packages.

According to Semiconductor Engineering, acoustic micro imaging is a high-frequency ultrasound inspection method used in advanced packaging. The publisher describes it as a way to produce clear pictures of hidden gaps and layer separation that may be difficult to see.

The process starts with a transducer. This device converts a pulse of electrical energy into mechanical energy, creating an ultrasound wave. Ultrasound cannot pass through air. The resulting images can reveal where gaps or separated layers exist inside a package.

Advanced packaging is important for modern chips. As features become smaller and spacing becomes tighter, tiny defects can cause electrical or mechanical failures later. Acoustic micro imaging helps manufacturers inspect these structures without damaging the device.

How inspectors read the results

Illustration: temperature, market trend, news and the labels Acoustic Micro Imaging, Advanced Packaging, Machine Learning

Each transducer is designed for a particular frequency, and inspectors choose it based on the sample being checked. The transducer moves over the sample, and the system creates multiple images. AMI software then analyses those images.

The software helps decide whether a part should be accepted or rejected, and it can count defects and highlight areas that need attention. Some defective samples show obvious air gaps. Others show thinning of the attach material at solder pads, which could lead to delamination if stress is applied.

An uneven solder layer or thermal interface material layer may also point to a problem in the build process. Thermal interface material, or TIM, is another layer that can be uneven. These clues help engineers trace defects back to process issues.

Why X-rays and destructive tests still have a role

Acoustic micro imaging can gather a lot of information without harming the sample. Even so, companies may still use destructive tests during research and development. A destructive test can confirm that a process is fully developed before it enters production.

Semiconductor Engineering gives capacitor inspection as an example. A new capacitor maker may require a destructive test to confirm voids or delaminations found by AMI software. After that proof of concept is complete, the company can add the software to its inspection process.

AMI also differs from transmission scanning, which is more comparable to X-ray. Transmission scanning places one transducer above the sample and another below it. With transmission scanning and X-ray, the result is a combined image. AMI, by contrast, examines one layer at a time.

The two approaches can complement each other. Some tasks suit AMI, while others suit X-ray. For some applications, both may be recommended. For spherical objects, including ball grid arrays, a combined image is often preferred.

Machine learning and faster inspection

The inspection method must keep pace with more complex chip designs. The software may now have to inspect multi-die stacks, not just a single flip chip. This makes the inspection task more complex.

Machine learning is one way to address this complexity. Some wafers have via patterns that look similar to delamination. Humans have traditionally had to distinguish between them. AMI software can now use machine learning to identify via patterns and exclude them from the defect count.

Customer demand for higher resolution and better throughput is also driving development. The guide says Nordson Test & Inspection created SpinSAM, an automated tool. It can inspect four 300mm wafers at once.

Traditional acoustic microscopy systems often use a raster pattern to move the transducer. This pattern is a small grid of individual pixels. It collects data at each point, but the tool must reduce speed when it reaches the end of a line. That reduces throughput.

Some systems leave out lines and estimate missing data points. Others use several transducers and join the resulting images. Both methods have drawbacks.

SpinSAM instead rotates the wafer. This keeps the transducer and wafer moving continuously. It allows inspectors to scan each pixel faster. It aims to improve throughput without the distortion that can come from stitching images together.

Why it matters for chip makers

Inspection is essential for advanced packaging. It supports higher yield and helps maintain device reliability. As packages become more complex, tiny defects may create serious problems.

Acoustic micro imaging gives engineers a non-destructive way to see hidden gaps and delamination. It can support production decisions, guide process development, and work alongside other inspection methods. As chip designs continue to evolve, tools like AMI and machine learning-assisted software will be important for finding defects early.

Sources

  1. What Is Acoustic Micro Imaging? A Quick Guide To This Inspection ProcessSemiconductor Engineering

Tags

Spotted a mistake? See our corrections policy. Read how News by OnDevice works in our editorial policy.