Solder Balling in Reflow: Profile and Paste Drivers

Solder balling is the ring of small spheres that appears on the solder mask around a reflowed joint. Each sphere is a droplet of paste that melted but did not merge with the main fillet, and it was left behind when the bulk of the alloy pulled back to the pad. The defect is usually cosmetic on a large pitch part and a real risk on a fine pitch one, where a ball can bridge two conductors or roll under a component.

The causes divide into three groups: the thermal path the paste experiences, which the reflow profile sets, the volume and shape of the deposit, and the condition of the paste itself. All three are visible in the process records, and the fix is rarely a single parameter change. The rest of this article follows those three groups in the order in which they are usually checked.

What a Solder Ball Is

A ball forms when a fragment of paste is separated from the main deposit by the solder mask or by the edge of the pad. The fragment melts, but the oxide on its surface and the distance to the main body of liquid metal keep it from coalescing. Surface tension then pulls each fragment into a sphere, which is the shape with the least surface area for a given volume, and the sphere stays where it formed unless the flux gases move it.

Two conditions have to be present. There has to be separated paste, and the thermal path has to melt it while the surrounding metal is still separate. That is why balling is associated with a slow or uneven preheat: the bulk of the paste melts and recedes before the fragments have been activated, and the fragments then sit on a surface that is no longer wet by the flux.

Preheat and the Solvent Window

Preheat has two jobs. It brings the assembly to a temperature at which the flux activates, and it drives off the solvent in the paste at a controlled rate. A ramp that is too fast boils the solvent under a skin of dried paste, and the vapour that escapes throws fragments of paste away from the deposit. A ramp that is too slow dries the paste to a powder that no longer coalesces properly.

A useful window for most no-clean pastes is a soak between 150 and 180 °C lasting 60 to 120 seconds, with the ramp from ambient held below about 3 °C per second. The soak should be long enough that the paste surface is dry to the eye before the alloy melts, and short enough that the flux has not begun to be consumed. Extending the soak at the upper end of the window is the usual first corrective action for balling.

Peak Temperature and Time Above Liquidus

The peak and the time above liquidus decide whether the fragments have a chance to merge. A peak of 235 to 245 °C with 45 to 75 seconds above 217 °C for a tin-silver-copper alloy gives the liquid metal time to flow and coalesce. A profile that just touches the melting point and falls away leaves the fragments as spheres, because the flux that was protecting them from oxidation has already been used up.

The time above liquidus has an upper limit as well. Too long at temperature thickens the intermetallic layer at the pad and darkens the flux residue, and the residue that remains is harder to clean. Balling that persists when the peak and the dwell are both inside the window is usually a sign that the problem is upstream of the oven rather than inside it.

Stencil Aperture and Paste Release

Paste smears on the underside of the stencil are the most common source of stray fragments. If the aperture walls are rough, the paste does not release cleanly, and a smear is deposited beside the pad on the next print stroke. Laser cut apertures with a wall angle that narrows towards the board release better than straight walls, and a nano coating on the stencil reduces the adhesion between the paste and the metal.

The area ratio of the aperture sets the release as well, since a small opening with a deep side wall holds the paste by friction. The relationship between aperture size, stencil thickness and release is worked through in the notes on stencil release. A stencil that is due for cleaning and a stencil that is worn both produce the same defect, and both are visible on a print inspection system.

Paste Volume and Slump

A deposit that is too tall slumps sideways under its own weight and leaves a thin skirt of paste around the pad. That skirt is separated from the main deposit by a distance that the liquid metal cannot bridge, and it becomes the ball population. Optimising the stencil thickness against the pad geometry is the standard remedy, with a volume that fills the joint without overprinting onto the mask.

Slump also depends on the paste rheology and on the humidity of the room. A paste at the end of its working life loses its thixotropic structure and slumps more, and a print room held at the wrong humidity changes the water absorption of the flux vehicle. The volume tolerance quoted by a print inspection system is the practical control on this effect, and the paste condition side is covered in the notes on paste shelf life.

Moisture and Outgassing

Moisture in the assembly is the classic cause of violent balling, because water turns to steam at 100 °C and expands by a factor of more than a thousand. The steam escapes through the paste and ejects droplets of metal and flux that land on the mask and form spheres. A board that has been stored in a humid room, or a paste that has been left open on the printer, will show this behaviour even with a correct profile.

The controls are storage and handling rather than the oven. Boards are held in the specified conditions, paste is returned to its container and sealed when the line stops, and the print room is held within its humidity window. Component moisture, which drives the same outgassing from inside a package, is a separate control described in the notes on moisture sensitivity levels.

Reflow Atmosphere and Oxygen

Oxygen accelerates the oxidation that keeps a fragment from coalescing. Reflow in nitrogen reduces the oxygen concentration to the low hundreds of parts per million, and the visible effect is a reduction in balling together with a brighter joint surface and less residue discolouration. The improvement is largest for small apertures and for pastes with an aggressive activator.

Nitrogen is not a substitute for a correct profile. A board that is outgassing moisture or receiving a smeared print will still produce balls in an inert atmosphere, only fewer of them. Where the defect is marginal, the atmosphere can be the difference between a pass and a fail, and the flow rate and purity behind that decision should be recorded rather than left to the gas supplier default.

Inspection and Acceptance

Balling is judged against a written criterion rather than by eye. A common rule allows a limited number of balls smaller than a stated diameter in an area away from the joints, and prohibits any ball that reduces the clearance between two conductors by more than a fraction of the gap or that is loose on the surface. The criterion should state the magnification and the lighting, because balls are much easier to see under low angle light.

Automatic optical inspection can be taught to find balls, and the effort is worthwhile on fine pitch products. The program should be taught on good boards as well as on defective ones, or the false call rate will make the inspection unusable. Verification is by unaided inspection of the flagged area at higher magnification, exactly as for any other feature.

Process Control and Records

The controls that keep balling out of production are the profile with its soak window and time above liquidus, the stencil clean frequency, the paste working life and the print room environment. Each is recorded, and the record should be specific enough that a change can be dated: paste lot, stencil serial number, profile revision and room humidity.

Where the defect appears, the fastest diagnostic sequence is to look at the paste deposit first, then at the profile, and last at the paste itself. Balling is a print and thermal defect far more often than it is a material defect, and a paste lot is usually the last item to change in a stable process. The record of those three variables turns a recurring defect into a solved one.

FAQ

Are solder balls always a defect? Small isolated balls away from a joint usually pass a written criterion. Balls that are loose, that sit between fine pitch leads, or that reduce the conductor clearance are rejected.

Can a longer soak always remove balling? Up to a point. An excessive soak consumes the flux before the alloy melts, which produces a different defect at the same location.

Does lead-free paste ball more than tin-lead? It melts higher and needs a longer time above liquidus, so the process window is narrower. The mechanism is the same.