Voiding in BGA assembly is due to a variety of reasons. No doubt the flux used in the solder paste which accounts to approximately 50% of the solder paste’s volume is a key factor.

Therefore choosing a solder paste with low voiding potential is a critical step. The resins and solvents chosen by the solder paste manufacturer will need to be both thermally stable at higher reflow temperatures but the solvents must also vaporize during the preheating cycle.

Solvents in solder paste are used primarily to give it rheological character the printing process. They are also present to dissolve organic activators. Choosing the correct solvent blends to achieve good printing but also avoid entrapment during the reflow cool down stage is an area manufacturers of solder pastes spend considerable energy optimizing.

Optimization of the thermal profile is also an important factor which can contribute to voids. Avoiding lower preheat temperatures and excessive peak temperatures will help reduce voids.

Voiding in BGA assembly can also occur due to the board finish out-gassing or component out-gassing.

A pre-bake can help determine this.

The main cause for voids in BGA solder balls is too low a ramp temperature or to fast a ramp zone. In either case the solder balls do not reach a high enough temperature to boil off the fluxes and they become encapsulated in the solder balls causing voids.

There are arguably many variables that contribute to the increased voiding characteristics of SAC alloy solder joints. Strictly speaking from a materials perspective, though, the problem has to do with the proclivity of SAC materials for volatile formation.

SAC alloys form more gasses and these volatiles cannot escape as easily from a molten SAC alloy as they can from a conventional SnPb alloy. They have to travel a greater distance to escape and, therefore, become trapped inside the solder joint and form voids. When this condition is combined with the unique geometry of the QFN, voiding may become even more prevalent.

Unlike BGAs where there are bumps or a QFP where there are leads, the QFN provides no standoff so there is nothing to absorb stress or allow for volatile escape. What’s more, the pad in the center of the QFN which is primarily used for thermal transfer presents large area soldering challenges and, consequently, issues with voiding.

Because there is such a large surface area and no standoff to allow volatiles to escape, these gasses may become entrapped and cause void formation. Though many would argue that some level of voiding is acceptable, Henkel’s stance has always been that reducing voids as much as possible is the best approach.

Plus, with QFNs, the voids aren’t just a problematic from a mechanical perspective, but can also result in thermal transfer impedance issues as well. This can lead to resistive heating and, if the voids are sizeable, hot spots can develop and may lead to thermal damage of the device.

Resolving the QFN voiding challenge may not be as difficult as it seems, however. Through a two-pronged materials-based and process-based approach, the materials experts at Henkel have successfully reduced the incidence of voiding in QFNs in both laboratory and high-volume production environments.

Henkel’s work has revealed that modifications to the solder paste flux system can significantly reduce void formation. The flux’s solvent concentration and boiling temperature, flux content and flux activator concentration all play a role in volatile formation. By altering the flux system to reduce volatile generation, voiding is lessened significantly. Using this technique, Henkel has developed some innovative, low-voiding solder pastes that will, undoubtedly, enable the QFN to be the powerhouse package it was intended to be.

A low-voiding solder paste in combination with optimized reflow profiles is clearly the best method for ensuring void reduction. I would be remiss if I didn’t also mention the potential impact of varying the print patterns for the QFN’s center pad as another possible void reduction mechanism.

Depending on the size of the device, limited success has been realized through printing a pattern – such as a snowflake or cross – instead of covering the entire pad with paste, which may allow for some area through which gasses can escape.

Pages have been, and still could be written on this subject.

There are basically two types of voids; a) gas /air pocket, and b) entrapped flux.

Voids can be large enough to engulf an entire solder joint, or tiny like a thousand champagne bubbles.

There are many causes and contributors to void production, with some of the more noteworthy being:

Component cleanliness, existing void in component ball, component lead finish quality, cleanliness of the PWB (fabrication), moisture absorption, surface contamination, surface finish selection, surface oxidation, plating thickness, via-in-pads, flux chemistry volatility, solder paste alloy, solder paste powder type, solder paste powder quality, paste deposition, profile type, profile parameters, component placement oven environment, and oven heat source.

Some voiding types can be profiled out through tedious trial and error manipulation of the reflow process parameters. Some will react to an aperture change in your stencil, others may require a flux chemistry change. There will be voiding that may require you to change more than one variable in order to be minimized or eliminated, and there are some voids that you will not be able to impact much at all, and leadless packages such as BGA’s cannot necessarily be treated the same as say an LCC with a large center ground plane.

It's unfortunate that we cannot tell just by looking at an assembly what will void and what will not. Our industry experience can give us a good indication; however, you may find that what may cause your worst headache is no problem at all for another manufacturer.

When considering voiding, the reflow cycle can turn into a fine balancing act at best. You must accommodate all of the requirements of the flux chemistry and alloy selection, achieve proper peak temperatures, and allow a long enough time above liquidus to produce a proper IMC, while allowing entrapped flux to escape from the molten alloy. Breaching any one of these variables either way for just a few seconds can be the difference between a void-free process or a voiding nightmare.