PCB Thermal Deformation: Tg, CTE and Z-Axis Expansion

A circuit board changes shape when it is heated, and the change is not uniform. The laminate expands more in the thickness direction than in the plane, the copper expands less than the resin, and above a certain temperature the resin softens and the mechanical behaviour changes completely.

Those effects decide whether a plated through hole survives a soldering cycle, whether a large ball grid array cracks after a hundred thermal cycles and whether a board that is flat in storage is still flat at reflow temperature. This article looks at the mechanisms and at the material parameters that describe them.

What Happens When a Board Heats

Below the glass transition temperature the laminate behaves as a stiff, glassy solid, and its expansion follows a relatively low coefficient. Above that temperature the resin becomes rubbery, the expansion coefficient rises sharply and the material loses much of its stiffness.

For a plated hole the consequence is direct. The copper barrel is constrained by the laminate, and when the laminate expands more than the copper in the thickness direction, the barrel is stretched. If the elongation exceeds what the copper can tolerate, the plating cracks.

<img src="https://www.gopcba.com/wp-content/uploads/2026/08/PCB脉冲电镀.png" alt="PCB cross section after thermal cycling test” />

Glass Transition and Decomposition Temperature

The glass transition temperature is the point at which the resin changes from glassy to rubbery. Standard FR4 sits around 130 to 140 degrees C, mid grade materials are above 150 and high performance laminates exceed 170.

The decomposition temperature is a different parameter: it describes the point at which the resin begins to break down chemically, and it matters for lead free assembly and for multiple reflow cycles. A material with a high glass transition temperature but a modest decomposition temperature may still degrade in a lead free process.

Z-Axis Expansion and Via Stress

The expansion in the thickness direction is described by the coefficient of thermal expansion in the Z axis, and it is the parameter that governs the survival of the plated barrel. Below the glass transition it is typically around 50 to 70 parts per million per degree, and above it the value can triple.

The stress in the barrel depends on the expansion, on the board thickness and on the temperature excursion. That is why a thick backplane with a standard laminate is more vulnerable than a thin board with the same material, and why the same design moved to a thicker panel needs a review of the hole plating.

X-Y Expansion and Registration

In the plane of the board the expansion is smaller, typically 12 to 18 parts per million per degree, and its effect is dimensional rather than mechanical. A component with a different expansion coefficient, such as a large ceramic package, will move relative to the board as the assembly heats and cools.

That mismatch is what produces stress on the solder joints of a large package, and it is the reason for the underfill or the compliant lead design that appears on a package intended for a demanding thermal environment. On the board side, the answer is a material with a lower in plane expansion or a thinner package.

Thermal Deformation in Assembly

During reflow the board is at its most compliant and its most expanded. The combination of the temperature gradient across the panel and the difference in expansion between the copper and the laminate produces a temporary curvature that adds to any permanent warpage the board already has.

That curvature is the mechanism behind placement defects on a large panel. The board sags between the conveyor supports, the height of the placement nozzle changes and a small package ends up tilted. The remedy is usually a support fixture or a carrier rather than a change of material.

Material Choice and Measurement

The material parameters that matter are the glass transition temperature, the decomposition temperature, the Z axis expansion and the total expansion through the assembly temperature range. A laminate with a higher glass transition temperature and a filled resin has a lower Z axis expansion, which is why it is preferred for thick boards and for press-fit assemblies.

The parameters are measured with standard methods: differential scanning calorimetry for the glass transition, thermomechanical analysis for the expansion, and thermogravimetric analysis for the decomposition temperature. The values quoted by suppliers come from those methods, and comparing values from different methods is a common source of confusion.

Design Measures

Keep the board as thin as the mechanical requirement allows, choose a laminate whose glass transition temperature is above the peak assembly temperature with margin, reduce the copper weight on the outer layers where the expansion mismatch is greatest, and avoid placing a large rigid package where the board is least supported.

Where the design uses a thick panel, the plated barrel should be specified with enough copper and the hole should be positioned so that the maximum expansion does not coincide with a thin section of plating. Those choices are cheaper than a field failure.

Test Methods and Qualification

The standard qualification for a plated hole is a thermal cycling or a thermal shock test on a coupon with a daisy chained via chain, measured for resistance change after a defined number of cycles. A rise in resistance indicates a crack, often before it becomes an open circuit.

The test result depends on the profile: the extremes, the ramp rate and the dwell time all change the number of cycles to failure. gopcb builds boards with laminates selected for the assembly temperature and the operating environment, and supplies the coupon and material data that a thermal qualification requires.

Deformation Beyond Reflow: Service Life

The assembly excursion is the most severe thermal event a board experiences, but it is not the only one. A product that operates outdoors sees a daily cycle of perhaps 40 degrees, a power module sees a cycle every time it switches on, and a board in a vehicle sees excursions that are both frequent and rapid. Over years those cycles accumulate, and the failure they produce is the same barrel crack that the reflow excursion would produce in a single step.

The number of cycles to failure depends on the amplitude and on the dwell time as much as on the material. A small excursion repeated often can be more damaging than a single large one, because the creep of the solder and the fatigue of the copper both accumulate. That is why a qualification profile is defined from the actual environment of the product rather than from a standard cycle chosen for convenience.

Interactions with the Rest of the Design

Thermal deformation does not act alone. A board that warps during reflow places the ball grid array joints under a shear load before the product is even switched on, and that preload is added to the stress of every later cycle. Similarly, a component that is too stiff for the laminate beneath it transfers the expansion mismatch into the corner joints, which is where the first crack usually appears.

The practical response is to treat the mechanical and thermal design as one problem. Keeping the board thin, matching the expansion of the package to the laminate where possible, and providing enough copper in the barrel to tolerate the strain all come from the same analysis, and each of them is cheaper to apply at the design stage than to correct after a qualification failure.

FAQ

Does a higher glass transition temperature always mean a better board? Not by itself. A high glass transition temperature helps with the assembly excursion, but the Z axis expansion and the decomposition temperature also matter, and a material with a high glass transition temperature can still have a poor expansion figure.

Why do plated holes crack after assembly rather than before? Because the crack develops under thermal cycling. A barrel that is marginal can pass electrical test at room temperature and fail after a few hundred cycles.

Can thermal deformation be prevented? It cannot be eliminated, because every material expands. It can be managed by selecting a laminate with a low Z axis expansion and by keeping the board thin enough that the absolute movement stays within the ductility of the copper.

Related reading: high temperature PCB materials, PCB warpage and stackup balance, lead free versus leaded solder, and multilayer PCB advantages.