Thermal Shock Testing Of Plated Through Holes

Thermal shock testing exposes a plated through hole to a rapid change of temperature so that the difference in expansion between copper and laminate loads the plating. It is the standard way to find out whether a barrel will survive the thermal history of a real product, and it is applied both to coupons that travel with a production panel and to finished assemblies.

This article describes how the test is set up, what happens inside the barrel during a cycle, how a crack is found, and how the results are judged against an acceptance criterion.

What The Test Does

A plated through hole is a composite structure. The copper barrel is bonded to the resin of the drill wall, and the resin expands roughly five to seven times as much as the copper for a given rise in temperature, measured in the direction of the board thickness. As the board heats, the laminate tries to grow through the thickness and drags the barrel with it; as it cools, the laminate contracts and the barrel is left in tension. Every cycle adds a small amount of damage at the interfaces where the stress is greatest.

The two critical locations are the corner where the barrel meets the surface pad and the point where the barrel meets an inner layer connection. Both are geometric discontinuities, and both concentrate strain. This is why a thermal shock test is aimed at the plating and the laminate together, not at the copper alone, and why the quality of the hole wall preparation before plating matters as much as the plating itself. The defects that originate at these interfaces are catalogued under copper plating defects prevention.

Test Conditions And Cycles

A thermal shock cycle moves the sample between two chambers held at fixed temperatures, with a transfer time of a few seconds to a minute. A thermal cycling test, by contrast, controls the ramp rate and holds the sample in one chamber, so the sample changes temperature slowly. The distinction matters: a rapid transfer imposes a steeper gradient through the board and puts more strain on the barrel, so the two tests are not interchangeable and the results cannot be compared directly.

Typical conditions for a plated hole are a low extreme between minus forty and minus fifty five degrees Celsius and a high extreme between one hundred and twenty five and one hundred and fifty degrees, with a dwell long enough for the sample to reach the chamber temperature. That dwell is usually ten to fifteen minutes. The number of cycles is set by the product specification and by the class of the board; a hundred cycles screens the process, while a qualification programme may run five hundred or a thousand, sometimes beyond the point of failure to establish a margin.

The Plated Through Hole Under Stress

The strain that opens a barrel crack is driven by the difference in expansion through the board thickness, so a thicker board with a smaller hole is more severely loaded. The plating thickness matters as well, but not in the direction most people expect: a very thin barrel ruptures quickly, while a very thick barrel resists the tensile load but transmits more stress to the interface with the inner layer, where the crack then begins instead. The usual specification window for barrel copper reflects that trade.

Chemistry contributes through the material properties of the deposit. A ductile, fine grained deposit accommodates strain by deforming, while a brittle deposit with high internal stress cracks early. The additives that control grain structure and ductility are described under electroplating additives for PCB, and the same parameters that give good throwing power in a high aspect ratio hole also influence how the deposit behaves under thermal load.

Finding A Barrel Crack

After the cycling sequence the sample is microsectioned at the holes of interest and inspected at magnification. A well prepared microsection shows the plating, the inner layer connections, the resin and the glass bundles in one plane, and a crack appears as a dark line that separates the copper from itself or from the pad. The section must be cut through the centre of the hole, because an off-centre cut shows a misleading barrel thickness and can hide a crack that runs around one side.

Cracks are classified by where they sit and how far they extend. A crack that runs part way around the barrel, and does not cross the full plating thickness, is usually treated as a defect to monitor. A crack that runs around the whole circumference or through the full thickness of the plating is a failure. Where the crack passes through the inner layer connection it is more serious still, because that is also the electrical path. Sectioning practice and the appearance of the layers involved are discussed under the quality characteristics of a board design.

Interconnect Stress Test And Resistance Monitoring

Sectioning shows what happened to the holes that were cut, and it gives no statistical weight. An interconnect stress test fills that gap. A coupon containing a chain of plated holes and the traces between them is cycled, and its resistance is measured continuously. As a barrel cracks, the resistance of the chain rises, first in small steps and then sharply when a link opens. The advantage is that thousands of holes are stressed at once and the failure is detected by an instrument rather than by an operator’s eye.

The result is a distribution: the number of cycles at which each chain crosses a resistance threshold. Comparing the distribution before and after a process change is far more sensitive than comparing a handful of microsections, because a shift in the tail of the distribution appears long before any single hole shows a visible crack. For this reason the resistance method is used for process qualification and the microsection for root cause analysis.

Judging Results And Following Up

The acceptance criterion is normally expressed as a number of cycles with no crack exceeding a defined length, or as a resistance change below a stated percentage after a stated number of cycles. The criterion belongs to the product specification, and it should name the hole size, the board thickness and the plating thickness it applies to, because a result is meaningless without them.

When a test fails, the follow-up follows the interfaces. The drill wall is examined for resin smear and for glass fibre protruding into the barrel, the plating is checked for thickness and ductility, and the lamination is checked for voids and for separation around the hole. A failure that appears only at the inner layer connection points at the etch and oxide treatment before lamination, while a failure that appears at the surface corner points at the plating and the pad geometry. Isolating the cause is the purpose of the test; the pass or fail result is only the trigger. Dimensional behaviour of the laminate itself is covered under PCB dimensional stability and expansion.

FAQ

Is thermal shock the same as thermal cycling? No. Thermal shock transfers the sample rapidly between two fixed chambers and produces a steep gradient. Thermal cycling ramps the sample slowly within one chamber. The shock test is more severe on a plated barrel, and the two produce different failure distributions.

How many cycles are enough? The number comes from the product specification or from the class of the board, not from a universal figure. A hundred cycles is a common screening level, while a qualification programme for a harsh environment may run five hundred to a thousand.

Why microsection instead of relying on resistance alone? The resistance test tells you that a chain has degraded and gives statistical confidence. The microsection tells you where the crack started and what caused it, which is what a process fix needs.