Microsection Analysis of PCB Quality

What a Microsection Shows

A microsection is a cross-section of a printed circuit board that has been mounted in resin, ground and polished back to a plane, and examined under a microscope. It turns the board’s internal structure into something that can be measured: copper thickness on the surface and in the holes, plating quality, dielectric spacing, and the interfaces between layers.

It is the reference method for verifying what fabrication actually produced. Thickness measurements taken by other means are indirect, but a microsection shows the material itself, including the defects that indirect methods miss. That is why it is used for qualification, for process troubleshooting, and for investigating failures.

Because it is destructive, it is applied to coupons, to test panels, or to samples from a lot rather than to production boards. The result describes the sample, and the sampling plan determines how much of the lot the result can be said to represent.

Sample Preparation

Preparation is where most of the skill lies. The sample is cut from the board, mounted in a resin that supports the edges, and then ground and polished in stages. Each stage removes the damage from the previous one, and the final polish produces a surface that reflects light without smearing the soft copper.

Poor preparation produces artifacts that look like defects. Smearing can push copper across a dielectric boundary and make a void appear filled. Rounded edges can hide a thin plating or make a measurement read low. Pull-out of copper during grinding can leave a gap that looks like an open.

The cut plane must be chosen to include the feature of interest. A section through a via shows the plating in the barrel, while a section across a trace shows the copper thickness and the dielectric spacing. Where a specific defect is being investigated, the section should be positioned through it rather than near it.

What to Measure

The most common measurements are the average and minimum copper thickness around a plated hole, the surface copper thickness, the dielectric thickness between layers, and the plating quality of the barrel. The minimum barrel thickness is the number that matters for reliability, because the failure usually starts at the thinnest point.

Plating quality is assessed visually as well as numerically. A good barrel shows a continuous, uniform copper layer with no voids, cracks, or thin spots, and the interface between the copper and the dielectric is clean. Where the plating has pulled away from the barrel wall, or where there is a gap between the copper and the laminate, the measurement may pass while the joint is unreliable.

Other measurable features include the solder mask thickness, the surface finish layers, the annular ring width, and the presence of intermetallic layers. Each of those has an acceptance criterion, and the measurements should be compared against the specification for the product rather than against a general rule.

Common Findings

Voiding in the barrel plating is the most frequent result, and it usually points to a plating chemistry or agitation problem. Thin plating at the middle of the barrel, caused by poor throwing power in a deep hole, is another common finding and is the reason minimum thickness is specified rather than average. Resin smear or drill debris in the barrel indicates a desmear problem.

Delamination between layers appears as a separation along the interface and points to lamination parameters or to moisture in the materials. Copper cracking, especially at the corner of a pad or a via, is a mechanical finding that follows thermal or flexural stress and is often the cause of a field failure.

Measurement variation between samples is itself a signal. A lot with a wide range of barrel thicknesses has a process control problem even if every sample is within specification, because the tail of the distribution is where failures occur. Trending the minimum thickness rather than only the average makes that visible.

Coupons and Sampling Plans

Most fabrication shops include a test coupon on the production panel, positioned so that it experiences the same plating and lamination conditions as the boards. The coupon carries the same hole sizes, layer count, and copper weights as the product, which makes its microsection representative of the panel.

The sampling plan defines how often a microsection is performed. A common approach is one section per panel lot at the start of production, with additional sections whenever a process change, a new material, or an excursion occurs. For critical products, the plan may require a section from each panel or a defined portion of the lot.

Coupons that do not match the product are a frequent source of false confidence. A coupon with a different hole size or a different stackup will produce different plating results, so the coupon design should be reviewed when the product changes.

Interpreting the Results

A microsection result should be treated as a process indicator rather than as an isolated number. A minimum barrel thickness that passes but is close to the limit on every sample means the process has no margin, and the next small variation will produce a reject. The measurement should be compared against the trend, not only against the specification.

Where a finding is outside the specification, the section should be repeated on additional samples to see whether it is isolated or systematic. A single thin spot may be a local artifact, while a consistent thin region across several samples points to a process problem that requires a change to the plating or drilling parameters.

The interpretation should also account for preparation artifacts. A result that conflicts with other evidence, such as an electrical test failure that does not correspond to the section, should be investigated before the section is accepted as the explanation. Repeating the section in a different location is often the fastest way to resolve the discrepancy.

FAQ

Why is a microsection destructive? The board is cut, mounted, ground, and polished, so the sample cannot be used afterwards. That is why sections are taken from coupons or from a sample of the lot.

What is the most important measurement? The minimum copper thickness in the plated barrel, because reliability depends on the thinnest point rather than the average.

Can the same sample be used for several checks? One section can be measured for several features if the cut plane includes them, but each feature needs its own cut plane. A section through a via does not show a trace spacing.

How often should microsections be performed? At least once per lot at the start of production, and again after any process change, material change, or excursion. Critical products may require more frequent sampling.

What if the section looks good but the board fails electrically? Repeat the section at the failure location. Preparation artifacts and unrepresentative cut planes can both produce a section that does not show the actual defect.

Conclusion

Microsection analysis is the reference method for verifying the internal quality of a fabricated board, and its value depends on careful preparation and on the right sampling. Measuring the minimum barrel plating, the dielectric spacing, and the interface quality, and comparing the results against the trend rather than only the specification, turns a destructive test into real process control. For related topics, see our notes on PCB manufacturing, quality management, PCB capabilities, and PCBA testing for how fabrication quality is verified in 2026.