Solder Joint Void Acceptance Criteria

Voids are cavities inside a solder joint, and they are one of the few joint features that can be seen without cutting the board open. Because X-ray makes them visible, they attract attention out of proportion to their effect, and specifications are often written as a single percentage that has little to do with how the joint performs. A workable acceptance criterion depends on where the void sits and what the joint has to do.

Where Voids Come From

Voids form from several mechanisms. Flux volatiles that cannot escape through the molten solder become trapped as bubbles, particularly under a large area such as a thermal pad, where the outgassing path is long. Entrapped air from a paste deposit that was not fully consolidated, shrinkage as the alloy solidifies and moisture released from the board or the component all add to the population. The relative contribution of each depends on the paste, the profile and the geometry.

Because the mechanisms differ, the remedies differ. A profile with a longer soak or a slower ramp lets volatiles escape before the alloy solidifies; a stencil with a better aperture pattern reduces entrapped air; a dry board and dry components reduce moisture. Where a single void percentage is specified without understanding the source, the process engineer is left to guess which lever to pull. Profile guidance is usually the first place to look for the volatile-driven cases.

Thermal and Mechanical Significance

Under a thermal pad, a void reduces the cross-sectional area available for heat flow and raises the junction temperature. The effect is roughly proportional to the void area for small voids, but the position matters: a void directly under the die raises the local thermal resistance more than the same area concentrated at the edge of the pad. That is why a percentage measured over the whole pad can mislead.

Mechanically, a void reduces the load-bearing area and creates a stress concentration. In a joint that sees thermal cycling, the void edge is where a fatigue crack begins, and a void that is large and sharp-edged is more damaging than a rounded one of the same area. In a joint that carries no thermal or mechanical duty, such as a small signal pad, the void is largely a cosmetic matter. Thermal cycling is the test that reveals which of these applies.

How X-Ray Measures Voids

X-ray inspection produces a two-dimensional projection, so the void percentage it reports is calculated from the projected area rather than the volume. Two joints with the same projected percentage can have very different void distributions through the thickness, and a void near the top surface may be invisible entirely. The measurement is also affected by the grey-level threshold used to define a void, which is a software setting rather than a physical fact.

That means the measurement must be defined as carefully as the criterion. The threshold, the magnification, the viewing angle and the software version should be recorded, and the same settings should be used for comparison between lots. Where the joint is critical, a cross section provides the true distribution and should be used to validate the X-ray interpretation. X-ray inspection practice describes the settings that most affect the result.

Writing a Percentage Criterion

Where a percentage criterion is used, it should come from the function of the joint rather than from a round number. A thermal pad under a power device may have a limit derived from the thermal resistance budget; a structural joint may have a limit derived from the mechanical load; a signal joint may have no meaningful limit at all. Deriving the number, even roughly, makes the criterion defensible and gives the process a target.

The criterion should also state the measurement method, because the same joint measures differently under different settings. A specification that says no more than twenty-five percent void without saying how it is measured is not a specification. Where the customer provides the criterion, the same question should be asked, both to ensure the process can meet it and to ensure that the inspection is measuring what the customer intended.

Position-Based Criteria

A position-based criterion is usually more useful than a single number. Voids lying outside the die shadow on a thermal pad, or within a defined distance of the pad edge, can be allowed a larger area because they contribute less to the thermal path. Voids directly under the die, or touching the pad boundary in a way that interrupts the fillet, can be limited much more tightly.

Defining those zones requires a drawing that shows the pad and the die outline, which is available for most components. The benefit is a criterion that reflects the physics and that avoids rejecting acceptable joints for a percentage that has no functional meaning. It also gives the process engineer a target for improvement that is worth pursuing, rather than a number that must be met for its own sake. Thermal via design and the pad geometry determine how much margin exists.

Process Routes to Fewer Voids

Vacuum reflow is the most direct route: applying a vacuum while the solder is molten draws the voids out through the liquid, and the reduction can be substantial. Its cost is equipment and cycle time, and it is usually reserved for products where the thermal performance is critical. Profile changes are cheaper and often effective, particularly extending the soak and slowing the ramp into liquidus.

Paste selection also matters. A paste formulated to reduce voiding is available for many applications, and the powder size and flux chemistry affect the volatile load. Stencil design contributes through the aperture pattern: splitting a large pad into a grid of deposits allows air to escape between them, which often reduces voiding more than any profile change. Paste volume control determines whether the grid is filled consistently.

Verification and Records

Whatever criterion is chosen, the verification should be repeatable and recorded. A sample of joints measured at defined intervals, using fixed inspection settings, gives a trend that shows whether the process is stable and whether a change has helped. Where a criterion is tight, the sample size should be large enough to detect a real shift rather than reporting the variation of a handful of joints.

Records should include the criterion version, the measurement settings and the results, linked to the build. When a customer later questions a joint, that record is what supports the acceptance decision, and when a process change is made, the same record shows whether it worked. Without it, the discussion returns to a single image and an opinion.

Working with the Customer Specification

Many customers specify void limits without a clear functional basis, often inherited from a previous product. Where the process cannot meet the limit, the useful response is to measure the thermal or mechanical consequence rather than to argue about the number. Thermal testing or a cross section that shows the voids are outside the critical area is evidence that the limit is not the right one for this design.

Where the limit is justified, the conversation moves to the process route: a paste change, a stencil change or a vacuum step, each with a cost. Presenting the options with their costs and their expected effect turns a specification dispute into an engineering decision, and it usually produces a criterion that both sides can meet and defend over the life of the product.

FAQ

Is a void always a defect? No. Its significance depends on position and on what the joint has to do. A small void at the edge of a signal joint is not equivalent to a void under a power die.

Is X-ray void measurement exact? It measures a projected area with a software threshold, so it is comparable rather than absolute. Fix the settings when comparing lots.

What is the fastest way to reduce voiding? Often a stencil change that splits a large pad into a grid, followed by a profile change. Vacuum reflow is the most effective but the most expensive.

Should the void criterion be a single percentage? Only where the joint has a uniform function. Position-based criteria usually reflect the physics better.