Quality control / PCB cross-section analysis
A finished PCB hides most of the features that determine interconnection reliability: copper inside plated holes, inner-layer registration, laminate interfaces, via fill and the connection between an internal pad and the hole wall. Microsection analysis makes those features measurable by cutting a representative sample, mounting it in resin, grinding and polishing to the target plane, and inspecting the cross-section under magnification.
The practical rule: A microsection is strong evidence about the sampled location and the process represented by that sample. It is not proof that every hole on every delivered board is identical. The specification must define what is sampled, how it is prepared, which features are measured, the acceptance criteria and how results are linked to the production lot.
1. A microsection turns hidden construction into measurable evidence
IPC-A-600 includes cross-sectional acceptance topics for plated-through holes, internal annular ring, lifted lands, foil and barrel cracks, plating thickness, voids, wicking, inner-layer separation, filled vias and cap plating. The IPC P-MICRO checklist also illustrates many phenomena that may appear in a plated-hole cross-section, including smear, nail heading, hole-wall pull-away, nodules, cracks, resin recession and laminate defects.
| Feature | What can be measured or observed | Typical process question |
|---|---|---|
| Plated-through hole or via | Copper thickness, continuity, nodules, voids, cracks, roughness and hole-wall separation | Did drilling, desmear and metallization create a sound barrel? |
| Inner-layer connection | Registration, annular ring, breakout, smear and separation at the copper interface | Did the drill hit the internal land and did plating connect cleanly? |
| Multilayer stackup | Finished core/prepreg thickness, layer spacing, copper thickness and overall construction | Does the finished board match the approved stackup? |
| Laminate | Voids, delamination, cracks, measling, resin recession and glass-fibre conditions | Did lamination or thermal stress damage the dielectric structure? |
| Filled or capped via | Fill condition, voiding, copper cap thickness and interface integrity | Is the via-in-pad structure complete and suitable for the next process? |
| External features | Finished conductor thickness, solder mask coverage and some surface-finish layers | Are the specified coatings present at the sampled location? |
Microsectioning complements other controls. Electrical test can detect opens and shorts but does not directly quantify plating thickness or show an incipient crack. X-ray can inspect internal geometry without cutting the board, but it normally does not provide the same direct metallographic view of an interface. A cross-section provides local structural evidence; electrical, dimensional and process controls provide broader coverage.
2. The sample must represent the question being asked
Microsectioning is destructive, so the sample is normally a quality-control coupon designed into the production panel. The coupon should reproduce the relevant board construction, copper plating cycle, hole types and critical geometry. A coupon from the same panel is much stronger evidence than an unrelated reference coupon or a sample from another job.
A coupon cannot represent a feature it does not contain. If the product uses blind microvias, stacked vias, high-aspect-ratio PTHs, filled and capped via-in-pad or an unusual layer transition, the test vehicle and sampling plan must include those structures. For a failure investigation, cutting the actual suspect board may be necessary because the goal is no longer routine process verification but locating a product-specific defect.
| Sample source | Best use | Limitation |
|---|---|---|
| Panel coupon | Routine lot or process verification using controlled geometry | Represents only features and panel conditions designed into the coupon. |
| Finished product board | First-article validation or failure analysis at a specific location | Destroys the sampled board and may still be local evidence only. |
| Process test panel | Characterizing plating, lamination or thermal-stress capability | Must be demonstrably representative of the production process. |
Define sampling by production lot, panel, shift, plating batch or another meaningful process boundary. “Provide a microsection report” is incomplete unless it also states the represented lot and the disposition rule when the sample fails.
3. Preparation quality determines measurement quality
A typical workflow selects and cuts the coupon, embeds it in resin, grinds toward the target plane, polishes the face and may etch the surface to distinguish copper structures. Eurocircuits describes the same basic sequence for routine PCB cross-sections. Each step can introduce artefacts if uncontrolled.
- Wrong section plane: polishing away from the hole centre can make the apparent wall thickness, hole diameter or annular ring misleading.
- Edge rounding: aggressive polishing can round copper edges and distort thin-layer measurements.
- Smearing or pull-out: poor cutting or grinding may create damage that resembles a manufacturing defect.
- Over-etching: excessive chemical etching can attack copper and exaggerate interfaces.
- Uncalibrated imaging: a microscope photograph without scale, magnification and calibration is not a dimensional result.
The report should identify orientation and section plane. A transverse section through the hole centre answers different questions from an axial section along the barrel. When acceptance depends on a minimum value, the measurement location and method should be consistent with the referenced standard and customer specification.
4. Read plated holes as a connected system
Do not reduce PTH evaluation to one copper-thickness number. Reliability depends on the relationship between drilling, hole-wall preparation, metallization, internal pads and laminate condition.
Plating thickness and distribution. Measure at the locations defined by the inspection plan. An average can hide a local thin area, while one isolated point may not represent the barrel. Record the individual values, minimum and measurement locations rather than reporting only “pass”.
Voids, nodules and roughness. A void interrupts the plated wall; a nodule or rough region may reduce finished hole diameter even when copper is present. Acceptance depends on size, location, circumferential extent and the applicable class/specification—not merely on whether a feature is visible.
Cracks and separation. Barrel, corner and foil cracks have different locations and implications. Thermal stress or reflow simulation may reveal weakness that is not visible in the as-produced sample. Hole-wall pull-away, resin recession and inner-layer separation should be distinguished rather than grouped under a generic “delamination” label.
Desmear and inner-layer connection. Resin smear remaining on an internal copper interface can interfere with metallurgical connection. Excessive etchback can also change the geometry. The cross-section should show whether the internal copper is properly exposed and joined to continuous plating.
5. Use the same section to verify stackup and special structures
A multilayer cross-section can compare finished dielectric spacing, core and prepreg construction, copper thickness and total board thickness with the approved stackup. It can also reveal layer-registration error, laminate voids, cracks or delamination after thermal conditioning. This is particularly useful when impedance, thin dielectrics or sequential lamination make the internal construction critical.
For filled and capped vias, inspect more than whether the hole appears closed. The plan may need to evaluate fill voids, fill-to-wall interface, cap copper thickness, surface depression or protrusion and any crack after thermal stress. Blind and stacked microvias may require additional planes or samples because one section may miss the critical interface.
Microsectioning also has limits for very thin surface finishes. Eurocircuits notes that a section can measure the nickel layer in ENIG, while the immersion-gold layer is too thin for their routine microsection method and is measured by X-ray. Select the measurement technology for the layer and resolution required.
6. A useful report connects images, measurements and disposition
| Report item | Useful content |
|---|---|
| Identification | Part number, revision, PO/work order, lot, panel/coupon ID and sample date. |
| Applicable criteria | Customer drawing, IPC document and class/revision, plus any approved exceptions. |
| Sample definition | Coupon or product location, hole/via type, diameter, aspect ratio and section orientation. |
| Preparation | As-produced or thermally stressed condition and relevant preparation method. |
| Images | Overview and detailed micrographs with scale, magnification and labelled features. |
| Measurements | Individual readings, units, locations, required limits and calibrated equipment reference. |
| Conclusion | Pass/fail by characteristic, reviewer, nonconformance reference and lot disposition. |
Data to send with an RFQ: identify the PCB revision, applicable acceptance standard and class, critical hole and via structures, requested thermal conditioning, sampling frequency, report format and whether destructive product-board sampling is permitted. If the requirement supports qualification or a regulated record, state the traceability and retention expectations before quotation.
Technical references
- IPC-A-600 — Acceptability of Printed Boards (table of contents).
- IPC P-MICRO — Phenomena in Cross Section of Plated Through Holes.
- Eurocircuits — Microsections.
- Eurocircuits — Microsection of a PTH.
- Würth Elektronik — PCB Test Methods.
Need microsection evidence for a PCB build?
JEXPCB can review the stackup, hole structures, acceptance criteria and requested documentation before fabrication and coordinate project-specific quality records.

