Solder Joint Thermal Fatigue and Its Drivers

A solder joint fails in thermal fatigue when the strain from a temperature change accumulates faster than the alloy can accommodate it. The drivers are the geometry, the materials and the cycle itself, and the four are usually examined together because the effect of one depends on the others.

The Strain That Drives It

The board and the component expand by different amounts, and the joint is the item that accommodates the difference. The strain is largest at the joint that is furthest from the neutral point of the package, which is why the corner joints fail first.

The distance from the neutral point is a design figure, so the fatigue life is partly decided by the package size and the board thickness before any process choice is made. Our thermal cycling notes describe how the life is measured.

Cycle Count and Dwell

The number of cycles is the obvious driver and the dwell is the one that is often left out. A long dwell allows the joint to creep, which dissipates part of the strain and changes the failure mode.

That is why a test with a short dwell and one with a long dwell give different results for the same number of cycles. The test has to represent the service rather than the chamber.

Alloy and Microstructure

The alloy sets the creep resistance and the melting point, and the microstructure sets how the strain is distributed between the grains. A fine structure with a well distributed second phase resists fatigue better than a coarse one with a continuous network at the boundaries.

The microstructure comes from the cooling rate, which is a profile parameter rather than a material one. Our joint criteria notes define the appearance and not the structure, which is the reason the structure is measured separately.

The joint height, the pad size and the stand off all change the strain. A taller joint accommodates more displacement for the same strain, which is the reason a ball grid array with larger spheres survives longer than one with smaller spheres.

That is a design lever that is available before the process is chosen. It is often constrained by the package rather than by the board, which is why the package is selected with the life in mind.

The board expansion coefficient, the package expansion coefficient and the underfill modulus all participate. A board with a lower expansion reduces the displacement, and an underfill distributes the strain across the joint area.

Each of those is a material choice with a cost, and each is measured on a cycling test rather than assumed from a data sheet. Our inspection notes describe how the result is verified on the hardware.

The measurement is the number of cycles to a defined resistance rise on a daisy chain, and the second measurement is a section that shows where the crack started. The two together identify the driver.

A crack at the intermetallic points at the interface and a crack in the bulk alloy points at the strain and the microstructure. Our board quality notes describe how the sections are controlled.

Acceptance and Its Evidence

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.

Where the process window is narrow, the measurement resolution has to be better than the window, or the data cannot distinguish a good part from a marginal one. Where two operations share a tolerance, the allocation between them should be explicit rather than left to whichever is measured first.

The narrowest feature on the board usually sets the process window for the whole product, so it deserves the closest attention at review. 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.

Sampling is a compromise between cost and confidence, and the sample size should follow from the failure rate that has to be detected. Where a decision is made by judgement, a boundary sample makes the judgement repeatable between operators and between shifts.

Where a process is at the edge of its capability, the margin should be bought deliberately rather than discovered during production. A record that identifies the operator, the date and the settings is worth more than a record that identifies only the result.

Checks Before Release

The acceptance criteria should be written before the work starts, so that the decision is made by the specification rather than by the person inspecting. Where the supplier and the user both measure the same property, they should agree on the method before the first delivery.

Handling between operations is part of the process, and the damage it causes is often attributed to the operation that preceded it. Documentation exists so that a person who was not present can reproduce the work and reach the same conclusion.

Verification and Records

A measurement taken at the wrong point of the process describes the wrong thing, however carefully it is made.

Does a stronger alloy always last longer? A stronger alloy is usually less ductile, so the benefit depends on the failure mode rather than on the strength alone.

Is the corner joint always the first to fail? It is where the strain is highest, so it is where the first crack appears unless a defect starts one elsewhere.

What does gopcba provide for joint fatigue? We provide a cycling test designed for the service rather than the chamber, dwell set from the board measurement, cooling rate controlled to give the intended microstructure, geometry and material effects measured on a daisy chain, and sections that identify where the crack started.

Points to Confirm at First Article

The checks that matter are the ones performed on the product rather than on a sample kept for the purpose, because a coupon that travels with the panel is the only evidence about that panel. A parameter that is set once and never re verified drifts, and the drift is usually discovered by a defect rather than by the record.

The tooling, the material and the profile form one system, and a change to any of them should be assessed against the other two before it is released.