Intermetallic Growth: How Solder Joints Age

Every solder joint contains a layer of intermetallic compound where the alloy meets the copper, and that layer is the reason the joint bonds at all. It is also the reason joints fail late, because the layer continues to thicken with time and temperature, and the joint that measured perfectly at build can become brittle after years of aging.

What the Intermetallic Layer Is

When molten solder wets copper, the two react and form compounds such as Cu6Sn5 at the interface. The reaction is what creates a metallurgical bond; without it, the solder would merely sit on the surface and the joint would be a mechanical contact. A thin, continuous layer is therefore a requirement rather than a defect.

The initial layer forms within seconds at reflow temperature and is typically one to two microns thick on a copper pad. Its thickness depends on the peak temperature and the time above liquidus, which is why one profile can produce a joint with twice the intermetallic of another while both look identical to the eye.

Growth Kinetics and Temperature

After assembly, the layer grows by diffusion, and the growth is roughly proportional to the square root of time. That relationship means the layer doubles in thickness over four times the period, so the early hours of elevated temperature matter far more than a long life at room temperature. The rate is also strongly temperature dependent, with an activation energy in the region of 0.8 to 1.0 electron volts for the common solder systems.

The practical consequence is that an aging test at elevated temperature can represent years of service in a few weeks. An hour at 150 degrees Celsius does far more to the layer than a month at 40, which is why accelerated aging is used to study reliability and why a hot-running product ages faster than its calendar suggests.

Copper, Nickel and Gold Interactions

The substrate determines which compounds form. Copper pads produce Cu6Sn5 next to the solder and Cu3Sn next to the copper, and the Cu3Sn layer is the one associated with voiding and embrittlement. A nickel barrier, as used under ENIG, forms Ni3Sn4 instead and slows the reaction considerably, which is why nickel is preferred where the joint will run hot.

Gold adds a different problem. Gold dissolves rapidly into molten solder, and above roughly three to four percent by weight in the joint it produces a brittle, gold-rich intermetallic that fails under shock. The control is to keep the gold layer thin, which is one reason immersion gold is specified at a fraction of a micron.

<img src="https://www.gopcba.com/wp-content/uploads/2026/08/Turnkey-PCB-Assembly.jpg-1.webp" alt="Microsection showing intermetallic layer at a solder joint interface” />

Aging Tests and What They Predict

Accelerated aging typically holds the assembly at 125 to 150 degrees Celsius for periods of 100 to 1000 hours, after which joints are sectioned and the layer measured. Correlating the result against a known field condition converts the test into a prediction, and the correlation depends on the activation energy assumed, so the number should be documented rather than implied.

Thermal cycling is a different test with a different purpose. Cycling drives fatigue through differential expansion rather than through diffusion, and a joint can pass one test and fail the other. A reliability program therefore usually includes both, with aging looking at the interface and cycling looking at the bulk and the fillet.

When Growth Becomes a Defect

A layer of a few microns is normal and harmless. Problems appear when the layer becomes a significant fraction of the joint thickness, particularly in a joint that is already thin, such as a small chip component with a narrow solder fillet, or a joint that has been aged at high temperature for years. In extreme cases the entire joint becomes intermetallic and loses its ductility.

The other failure mode is the void that appears at the Cu3Sn and copper interface, known as Kirkendall voiding, which forms because copper diffuses out faster than tin diffuses in and leaves vacancies behind. Those voids concentrate stress and become the initiation site for a crack that propagates through the joint.

Alloy and Finish Choices That Slow Growth

Nickel barriers, thin gold and controlled reflow profiles all slow the reaction. Lowering the peak temperature and shortening the time above liquidus reduces the initial layer without affecting wetting, which is one reason a tight profile is worth maintaining even when the joints look good at a higher temperature.

Alloy choice matters too. Lead-free alloys with added nickel or with a small amount of a third element are formulated to slow intermetallic growth, and the supplier data on that behaviour is a legitimate reason to choose one paste over another for a hot-running application.

Inspecting and Measuring the Layer

Layer thickness is measured on a microsection, and the measurement is only as good as the preparation: polishing that smears the soft solder over the interface produces a layer that looks thicker or thinner than it is. Etching, imaging and measuring at several points along the interface all improve the result.

Scanning electron microscopy resolves the individual Cu6Sn5 and Cu3Sn layers, which is what distinguishes the benign total thickness from the problematic distribution. A light microscope can measure the total intermetallic, but it cannot tell you which compound is dominant, and the two have different implications.

Field Failure and Its Relation to Aging

Late-life joint failures usually appear as an open or an intermittent contact at a joint that has seen sustained heat, and the section shows a crack through or beside the intermetallic layer. Correlating that finding with the product’s thermal history, and with any accelerated aging performed at qualification, is what closes the loop between the laboratory and the field.

Where the product runs hot, the thermal design is part of the reliability solution. Lowering the operating temperature of a joint slows the diffusion that causes the failures, which makes thermal management a joint reliability measure rather than only a component protection measure.

Verification and Records

Where an operation cannot be verified afterwards, it has to be controlled during the operation, and that control has to be visible in the record. Consumables have a life measured in cycles, and the replacement point should come from the measurement rather than from a failure.

The sequence of operations is part of the specification, because a different order produces a different result from the same steps.

Related reading: our fabrication notes, board quality and design release notes cover the same ground.

Points to Confirm at First Article

Where a requirement can be measured, it should be measured at the point of manufacture and recorded against the board or the lot it applies to. The first article confirms that the setup matches the intent, and it is the cheapest point at which a wrong setup can still be corrected.

FAQ

How thick should the intermetallic layer be? One to two microns after reflow on a copper pad is typical and healthy. The layer becomes a concern when it grows to a significant fraction of the joint thickness.

What is Kirkendall voiding? Voids that form at the Cu3Sn and copper interface because copper diffuses out faster than tin diffuses in. They concentrate stress and become crack initiation sites.

Why is the layer growth not linear with time? Diffusion-controlled growth follows roughly the square root of time, so the first hours at elevated temperature matter far more than many years at room temperature.