CTE Measurement Methods for PCB Laminate Selection

Copper expands about five times as much as the laminate around it, so a plated hole in a printed circuit board is under stress every time the board changes temperature. The coefficient of thermal expansion describes how much a material grows for each degree, and it is one of the few material properties that can be measured directly, interpreted simply and then used to make a real design decision. CTE measurement is therefore a routine part of laminate evaluation rather than a research activity.

What CTE Means for a PCB

The coefficient of thermal expansion is the fractional change in length per degree of temperature change, usually expressed in parts per million. A value of fifty means that a metre of material grows by fifty micrometres for each degree, which is enough to matter when a board is heated from room temperature to reflow temperature, because a twenty centimetre board then grows by a tenth of a millimetre over a two hundred degree rise.

What makes the property important in a circuit board is that it is not a single number. The material behaves differently in the plane of the board and through its thickness, and it behaves differently below and above its glass transition temperature, so the measurement has to be interpreted with all of those conditions in mind.

In Plane and Out of Plane Expansion

In the plane of the board the glass fabric carries most of the load and restrains the resin, so the in plane expansion is relatively low. Through the thickness the fabric contributes less and the resin dominates, so the out of plane expansion is much larger, often two or three times the in plane value. That out of plane figure is the z axis expansion, and it is the value that appears in specifications for via reliability.

The z axis expansion is the one that stresses the plated barrel, because the copper in the hole is restrained in that direction by the laminate around it, and the barrel has nowhere to relieve the resulting strain. The larger the difference between the expansion of the copper and the expansion of the laminate, the more the barrel is stretched as the board heats.

<img src="https://www.gopcba.com/wp-content/uploads/2026/06/SMT车间一角.jpg" alt="Dilatometer measuring thermal expansion of a laminate sample” />

Measurement by Dilatometer

A dilatometer measures a small sample as it is heated, recording the change in length against the temperature. The sample is held in a fixture and the movement is measured with a probe or with an optical system, and the result is plotted as a curve of expansion against temperature rather than as a single number, which is what allows the transition to be located.

The curve is what makes the method useful, because it shows the region where the material changes behaviour. The slope below the transition, the slope above it and the temperature at which the change occurs are all read from the same test, which is why the technique is used for both CTE and glass transition measurement.

Sample Preparation and Conditioning

The sample has to be representative of the board and it has to be prepared in a consistent way. A cube cut from a laminate, a section through the thickness of a finished board and a thin coupon each give a different result, and the method should state which one was used.

Conditioning matters as much as preparation. A sample that has absorbed moisture behaves differently from a dried one, and the first heating cycle often evaporates water and produces an apparent expansion that is not a material property at all. The usual practice is to run a first cycle to dry the sample and to take the measurement from a second cycle, which is why the number of cycles is part of the reported method and why two laboratories can report different figures for the same material.

Temperature Range and Glass Transition

The measurement is normally reported over a range that brackets the glass transition temperature, because that is where the behaviour changes. Below the transition the material is relatively stiff and expands slowly; above it the resin softens and the expansion rate increases sharply, which is exactly the condition the board experiences during reflow.

A material with a high transition temperature keeps its low expansion behaviour to a higher temperature, which is the main reason high transition laminates are chosen for boards that see a demanding thermal cycle and for boards with many sequential lamination steps. The value quoted for the transition itself is measured from the same curve, and the two figures should always be reported together.

Interpreting the Results

The raw curve has to be interpreted before it can be used. The slope below the transition gives the in plane or out of plane CTE depending on the sample orientation, and the slope above it shows how much the material changes once the resin softens. The difference between the two is what creates the stress concentration at the via.

The values should also be compared with the copper, since it is the mismatch rather than the absolute value that causes the problem, and the strain on the barrel is proportional to that difference. Copper expands at roughly seventeen parts per million in all directions, so a laminate whose out of plane value approaches that figure will produce a much lower stress in the barrel than one whose value is several times higher.

Effect on Via Reliability

During a thermal cycle the laminate expands more than the copper in the barrel, so the barrel is stretched along its axis. If the copper cannot accommodate that stretch, it cracks, usually in the middle of the barrel or at the corner where the barrel meets the surface pad.

The ability of the barrel to survive depends on the copper thickness, the plating quality and the aspect ratio of the hole. A thick, ductile deposit tolerates more strain than a thin brittle one, and a hole with a high aspect ratio has less copper to distribute the strain. Those interactions are described in the guide to hole copper.

Material Selection Trade Offs

The material with the lowest expansion is not automatically the right choice. Low expansion usually comes with a higher transition temperature, a more difficult process and a higher price, and the benefit only appears if the design actually experiences the thermal stress that the material is chosen to resist.

The selection should therefore start from the application: the number of thermal cycles, the maximum temperature, the thickness of the board and the aspect ratio of the holes. Those inputs determine how much expansion the design can tolerate, and the material is then chosen to meet that requirement rather than to be the best available. The comparison between material families is described in the guide to laminate material properties.

Specification and Records

The specification should state the value required, the direction in which it applies and the method by which it is measured. A drawing that quotes a single CTE figure without saying which axis or which temperature range will be interpreted differently by every supplier who quotes against it.

The records should include the sample orientation, the conditioning, the number of heating cycles and the transition temperature measured in the same test. That combination allows a later failure to be compared against the material that was actually supplied. The geometry that interacts with the expansion is covered in the guide to aspect ratio.

FAQ

What CTE values are typical for a PCB laminate? In plane values are usually in the range of ten to twenty parts per million, while out of plane values below the glass transition are commonly in the range of forty to seventy. The figures depend on the resin, the glass content and the direction of measurement.

Why is the out of plane value the one that matters most? Because it is the direction in which the laminate expands against the copper in a plated hole. The mismatch stretches the barrel along its axis, and that strain is what produces barrel cracks during thermal cycling.

Can CTE be measured on a finished board? Yes, by cutting a section and measuring through the thickness, and the result describes the material as it exists in the product. The measurement is destructive, so it is normally done on a coupon or on a sample panel rather than on production boards.