CTE Mismatch PCB Design Guide

CTE mismatch is the difference in thermal expansion between the materials in an assembly. A laminate expands by about fifteen parts per million per degree, a ceramic body by six, and a metal package by twelve. Every temperature change therefore moves the parts by different amounts, and the difference is absorbed by the solder joints, the pads and the laminate underneath them.

What CTE Mismatch Is

The coefficient of thermal expansion is the fractional change in length per degree of temperature. Two materials with different coefficients that are joined together will fight each other when the temperature changes.

The result is a shear strain in the joint between them. A solder joint between a large ceramic package and an FR4 board sees that strain on every thermal cycle, and the strain accumulates into a fatigue crack.

The magnitude of the strain depends on the size of the part, on the temperature swing and on the difference in the coefficients. A large part and a large swing produce the most damage.

The mechanism is the same whether the mismatch is between a component and a board or between two layers of the board itself, and both cases are covered by the thermal management work.

The Materials Involved

The laminate is the largest term in most assemblies. A standard FR4 expands by around fifteen to seventeen parts per million in the plane of the board and more in the thickness direction.

A ceramic package expands by about six, so the mismatch with an FR4 board is a factor of nearly three. A small chip resistor is not affected because the joint is short, while a large ceramic capacitor or a ball grid array package is.

The copper inside the board has its own coefficient, which is much lower than the resin. The board therefore expands differently in a region with heavy copper than in one without, which adds a local mismatch to the global one.

The laminate properties include the expansion figures, and a low expansion material is one of the available measures when the strain is too high.

What Fails First

The solder joint fails first in most cases, because it is the softest element and it takes the strain. A crack starts at the outer edge of the joint, where the strain is highest, and grows across the joint section until the connection opens.

The failure is a fatigue failure, so it depends on the number of cycles and on the temperature swing rather than on a single event. A product that cycles once a day lasts for years, while one that cycles every minute fails in months.

The second failure is the pad and the laminate under it, where the strain can lift a pad or crack the resin. A pad that lifts takes the solder with it and leaves a crater in the board.

The third failure is inside the component, where a large die can crack or delaminate. That failure is not visible from outside, which makes it the most difficult to diagnose.

Design and Material Measures

The most effective measure is to reduce the size of the joint that carries the strain. A component that is broken into several smaller packages has a lower strain per joint, which is why a large ball grid array is often replaced by two smaller ones.

A compliant joint helps as well. A taller solder joint or a column instead of a ball absorbs more of the movement, and a package with a compliant lead reduces the strain at the board.

The material choice is the second lever. A laminate with a lower expansion coefficient reduces the mismatch, and a metal core or a ceramic board removes it altogether for a high power design.

The comparison in the metal core and ceramic guide is the right starting point when the mismatch cannot be managed by geometry alone.

Assembly and Profile Effects

The soldering profile sets the temperature swing that the assembly sees before it ever leaves the factory. A profile with a fast ramp and a fast cool produces a larger swing and a larger strain than a gentle one.

The cooling rate also sets the microstructure of the solder, and a joint that is cooled quickly is stronger and more brittle. The optimum is a compromise, and it belongs to the assembly process rather than to the layout.

A board that is warped during reflow starts its life with a residual stress in the joints, which shortens the fatigue life. The flatness of the board and the balance of the stackup are therefore part of the reliability picture.

The tolerance and reliability work covers the same subject from the assembly side, where the joint geometry and the process window are set.

Test and Verification

The verification is a thermal cycling test with a defined profile, a defined number of cycles and an in situ resistance measurement. The resistance rise across a daisy chained joint is the failure criterion.

The test is expensive and slow, so it is usually run on a coupon or on a simplified assembly rather than on the product. The coupon must have the same joint geometry and the same materials as the product to be meaningful.

A finite element model is used to predict the strain and to compare design options, and the test is then used to confirm the model rather than to explore the space. The model is only as good as the material data, which is why the figures for the solder and the laminate matter.

The results should be recorded with the design, because the next revision will face the same question. A cycle count that is written down saves a second test.

Practical Rules

Keep the largest and the most thermally mismatched components away from the edges of the board, where the constraint is lowest and the movement is highest.

Prefer several smaller packages to one large one where the function allows it, since the strain scales with the distance from the neutral point.

Use a lower expansion laminate where the assembly will cycle frequently, and check that the change does not break the impedance or the stackup.

Measure a coupon when the design is new, and treat the result as a design input rather than a qualification formality.

Additional Considerations for This Build

Practical attention to via pays for itself here, because it decides whether the finished board behaves as the drawing intended. Where the requirement is not stated on the fabrication drawing or in the assembly notes, the shop has to assume a default, and that default is rarely the value the design was simulated with. Stating via explicitly, together with the tolerance that applies, removes the assumption and keeps the result predictable from batch to batch.

Process Control and Verification

On a design of this kind, laminate is the item that decides how the rest of the board is arranged. A short note on the drawing about handling, storage or packaging is often worth more than an extra decimal place on a tolerance. Where a value sits close to a process limit, the drawing should say so, since the shop can then open the process window rather than working to a nominal figure that carries no tolerance.

FAQ

What is a typical CTE for FR4? Around fifteen to seventeen parts per million per degree in the plane of the board, and higher through the thickness. The exact value depends on the resin and the glass.

Why do large components fail first? Because the strain in the joint grows with the distance from the neutral point, which is set by the size of the part.

Can a different solder alloy help? It can, because a more compliant alloy absorbs more strain. The alloy change also affects the process temperature and the joint strength.