Barrel Cracks in Plated Holes: 6 Causes After Thermal Cycling

A barrel crack is a fracture in the copper plating inside a plated through hole, usually running around the circumference or along the length of the barrel. It appears after thermal cycling rather than at manufacture, which is what makes it a reliability defect rather than a yield defect.

The crack breaks the electrical path between the layers the hole connects, and it does so intermittently at first, when the two faces are still in contact. That behaviour is why a cracked barrel often passes electrical test and fails later in the field under vibration or further thermal load. That delayed behaviour is what makes the defect expensive, because the manufacturing record is long closed by the time it appears.

<img src="https://www.gopcba.com/wp-content/uploads/2026/08/3.png" alt="Microsection showing a barrel crack in the copper plating of a plated hole” />

What a Barrel Crack Is

In a section, the crack appears as a break in the copper wall, often with a dark line where the plating has separated. It may extend around part of the circumference or the whole wall, and it frequently begins at the knee where the barrel meets the surface pad or at a corner in the barrel.

Cracks are classified by where they occur and how far they extend. A crack confined to one side of the barrel with no measurable resistance change is treated differently from one that runs the full circumference, and the acceptance criteria should reflect the application rather than a single general rule, since a crack that leaves contact intact behaves differently from one that has separated.

Why Thermal Cycling Opens a Barrel

Copper expands far more than the laminate in the plane of the board, but the opposite is true in the thickness direction. The laminate expands much more than copper in the Z axis, so every thermal cycle stretches the barrel along its length while the board around it pushes outward.

If the plating is thin or has low ductility, it cannot accommodate that strain and it fractures. The higher the thermal cycling range, the more strain per cycle, which is why lead-free assembly temperatures push the requirement harder than the older tin-lead processes did, and why a design qualified years ago may not be qualified for a new assembly process.

Plating Thickness and Ductility

Thickness is the first variable. A thin barrel reaches its strain limit sooner, and a barrel that is thin in the middle relative to the ends is weaker than one of uniform wall thickness. The specified minimum applies to the thinnest point in the barrel rather than to an average.

Ductility matters as much. Copper that has been plated with too much organic additive, or that contains codeposited impurities, becomes brittle and cracks at strains a ductile deposit would survive. This is why plating thickness data alone cannot predict reliability, and why elongation testing belongs with thickness measurement.

Laminate Expansion and the Z Axis

The coefficient of thermal expansion through the thickness of the laminate drives the strain the barrel must absorb. A high expansion material, or one used above its glass transition temperature, expands more and stresses the barrel harder, especially in a thick board where the total expansion is larger.

Design choices change this. More glass, a higher glass transition temperature and a thicker board all interact, and the combination that produces the highest strain is a thick board with small holes in a high expansion material. Reviewing the design against the intended thermal environment is worth doing before the layout is released.

Drill Quality and the Knee of the Barrel

The knee, where the barrel turns into the surface pad, is where stress concentrates. A rough drill finish, a smear that was plated over or a nailhead from a damaged inner layer connection all create a geometric discontinuity that behaves as a stress raiser.

Drill quality therefore feeds directly into crack resistance. Heat damaged resin around the barrel, rough walls and excessive fibre protrusion all reduce the mechanical support the laminate gives the plating, which is the same set of conditions that produces poor hole wall adhesion.

Resin and Glass Effects

The resin system sets how much the laminate moves and how well it grips the copper. A resin that absorbs moisture expands more during a reflow excursion, and absorbed water also softens the material, so a board that has been stored damp behaves worse in thermal cycling than a dry one.

Glass style and resin content affect the local expansion around a hole. Where the drill passes close to a glass bundle, the material around the barrel is stiffer in one direction than another, and the resulting asymmetry concentrates strain on one side of the hole, which is one of the reasons cracks often start locally rather than evenly.

Detection: Microsection, Resistance and IST

A microsection after thermal cycling shows whether a crack exists and where it started. It is the reference method, and it depends on preparation quality, because a poorly polished barrel can show an artefact that resembles a crack.

In-situ monitoring measures the resistance of a daisy chained via during the thermal cycle, so a crack is detected at the moment it opens rather than by a section afterwards. Interconnect stress testing does the same thing with a controlled current load, and both methods give a distribution of failures across a sample rather than a single pass or fail.

Preventing Cracks Through Design

Design decides much of the margin. Thicker plating on a small hole, a larger annular ring at the knee, a reduced board thickness where possible, and a laminate with a higher glass transition temperature all raise the strain the barrel can absorb before it cracks.

The count of thermal cycles the product will see also matters. A board that experiences three reflow passes before it is shipped has already consumed part of its life, so the qualification test should include the assembly cycles as well as the service cycles rather than starting from a pristine sample.

Correcting the Process After a Failure

Where a failure occurs, the sequence should start with the section, then thickness and ductility data, then the drill record, then the laminate. Changing the plating first is common and often wrong, because a crack caused by drill damage will return even with thicker copper.

Compare the failed sample with a control from the same lot that passed. That comparison localises the difference to a process step rather than to the design as a whole, and it prevents a broad change that fixes the symptom while leaving the cause in place. Where a crack is accompanied by CAF growth along a fibre path, the investigation has to cover the laminate and the drilling as well as the plating.

FAQ

How thick should barrel plating be to resist cracking? The requirement depends on the board thickness, the hole size and the thermal range, so the figure has to be set per design. What is constant is that the minimum applies at the thinnest point of the barrel, not at the average.

Can a cracked barrel be detected without a section? Resistance monitoring during thermal cycling detects a crack as it opens, which is more informative than a section taken afterwards. A section then confirms the location and the cause.

Does a higher glass transition temperature solve the problem? It reduces expansion at the temperatures that matter during assembly, which helps significantly. It does not compensate for thin or brittle plating, so it should be treated as one factor in the margin rather than as the answer.