The Highway Code of Stencils
Article Type: New products From: Soldering & Surface Mount Technology, Volume 21, Issue 2
The technical information and guidance contained within the Highway Code of Stencils extends beyond stencil design and optimisation to cover the causes and solutions for defects commonly seen in the Surface Mount Technology SMT process.
Achieving the volume of solder paste and ultimately the solder fillets required by an increasingly diverse component population, present on today’s electronic assemblies, is the first step in avoiding costly rework at the end of the process.
Although the end result is undoubtedly an electrical or electronic device,the vast majority of the assembly process is in fact mechanical. Using a combination of good quality images together with informative text stimulates the reader to investigate their process and its results more thoroughly.
In most production facilities, these faults are identified or highlighted after the reflow process and Section 7, which is featured in this issue of Soldering & Surface Mount Technology covers both the identification/classification of these faults providing hints and guidance on improving or eliminating them.
Tony Weldon
7.1 Solder balling
Sometimes also known as spattering, this is mostly caused by the explosive evaporation of the solvents within the solder paste. As part of the reflow process, when the solder deposit’s temperature is raised, pressure builds up within the deposit until the solvents are able to migrate through this flux membrane. Since these solvents are contained within the printed deposits, it is important that the reflow profile ramp rate is not excessive or the soak zone temperature too high so that the release of these solvents is more gradual.
A similar but far more spectacular fault can be caused by the solder paste absorbing moisture which upon reflow causes mini steam explosions with the accompanying spattered effects.
7.1.1 Isolated solder balls in close proximity to a solder pad
This generally occurs when either the stencil to substrate alignment was not good during printing or the stencil aperture could not gasket to the copper feature beneath. Common reasons include over-etched PCBs, incorrect aperture reduction or printer registration errors.
7.1.2 De-wetting and solder-balling
Poor wetting and solder balling on the same pad is most likely caused by creating an excessive soak phase in the reflow profile. Firstly, the activated flux will clean the solder but extending the soak time can cause re-oxidation resulting in poor wetting and a large solder ball. This can often become more of a problem with the solder pastes used for finer pitch requirements as the smaller solder spheres have a larger surface area on which oxides can form or re-form. Lower residue pastes can also offer a smaller operating window so it is prudent to consult the technical data sheet for the paste selected.
7.1.3 Mid-chip solder balls or solder beading
It is usually evident as an emerging ball or balls beside discrete chip components. This is usually the result of inappropriate reduction to the stencil aperture. Excessive paste deposits are unable to retract to the component terminations or solder pad before the chip component is attracted to the board upon reflow. Any solder not able to flow to the component terminations becomes trapped and is compressed to form solder balls. Aperture modifications to greatly reduce or eliminate this problem are found in Section 4.4.
7.2 Wicking
Wicking occurs when a component lead or termination attracts or absorbs the meted solder and can often result in an open circuit. Typical reasons include too large a temperature difference between the pad to be soldered and the component leads. If the lead is hotter than the adjacent copper pad then the solder will flow onto the lead away from the pad. One obvious area to explore is to prolong the soak time to ensure there is minimal difference between the temperatures of the components and the substrate.
7.3 Wetting
Poor wetting can be the result of increased or excessive oxidation. As discussed above, oxidation or re-oxidation can be limited by a reduction in the heat impact on the solder paste selected. Reductions can be achieved by shortening the overall heating time or lowering the temperature rise ΔT in the pre-heat and soak zone.
De-wetting is caused by overheating during the reflow process. Peak temperature can be lowered or the dwell time reduced to overcome this problem.
7.4 Tombstoning/component lifting
Tombstoning on chip components can be caused by unequal wetting forces on the two terminations. Surface tension at the end that reflows first can attract the component and make it rise often causing it to stand on its end.
Unequal wetting can be caused by incorrect solder pad design, poor solderability of either the component terminations or solder pad, different volumes of solder paste available at each termination or unequal temperature at the two pads. A reduction in temperature difference is required to solve this problem. The temperature rise ΔT used in the pre-heat phase must be reduced and the soak time extended accordingly.
There are a number of reasons for the appearance of tombstoning with discrete components including:
different pad surface areas, from one side of the component to the other;
variations in the thermal demand of the solder pads attributed to tracks and internal layers;
different volumes of solder paste applied to the two pads;
solder paste printed beneath the body of the component;
poor solderability of one or both of the solderable component terminations;
solder resist thickness. Excessive thickness may cause the component to rock; and
inadequate pick and placement pressure causing poor initial adhesion to the paste deposits.
With discrete components getting ever smaller, the surface area of the solder pads influences the surface tension experienced by the components. Tombstoning can be more prevalent with smaller component sizes.
Incorrect pad design
Solder pads should be designed to accommodate the component terminations approximately central to the pad centroids to ensure the forces acting on the two ends of the component are balanced. When the component is out of position,relative to the centroids, the forces at each end will be different causing the possibility of lifting the component. Once movement has started, the momentum continues until the component has no solder termination at one end.
Variations in thermal demand of solder pads
When designing the layout of the PCB, it is important to achieve, as near as possible, equal solder pad and track interconnections at each end of the component. Why?
Since, during the reflow process, the component terminations do not reflow at precisely the same moment.
Differences in the surface area of copper connected to the component terminations, including connecting tracks, can affect the wetting speed and lead to an imbalance of forces on one side of the component compared to the other increasing the incidence of component lifting.
Different volumes of solder paste
There may be occasions, even when the copper pads are equal in size, that the solder paste volumes might be different. Poor registration of the stencil to the PCB can cause just such a problem. Smaller paste deposits may reflow before the larger deposits causing wetting to the termination faster than the other side of the component. This logical process is also evident with variations in small paste deposits used for μBGAs where the smaller paste deposits reflow before the larger ones.
Paste under the component body
Slumping of the solder paste deposits caused either by squashing during placement or cold slump due to environmental conditions are both undesirable as this can cause component lift during reflow. Paste should always be contained within the boundary of the copper pad limits.
Solderability
Solderability issues can exist when the plating finish itself is suspect or subsequent environmental reactions, for example, from incorrect storage have degraded the finish to cause large variations in the wetting attraction forces evident on the two pads. Where the solderability of the terminations is worse on one end of the component the forces acting on the component would pull the component to the better wetting pad before reflow had been effected at the other pad. It may not cause a tombstone but it is important to remember that even a small lift can cause an inadequate solder fillet or an open circuit.
Solder resist thickness
Thicker solder resist and resist variability between adjacent solder pads can cause the component to seesaw rather than seat correctly. This effect can also be seen when interconnecting copper tracking exists between the two pads.
The effect of pick and placement pressure
This can cause the component to simply rest on the surface of the paste deposit as opposed to being held by the tackiness of the flux. In such cases,when the solderability of the component is in question, wetting of the solder may be limited. Care has to be taken with pick and placement pressure, however,as excessive force can lead to both increases in mid-chip solder balls and shorting.
7.5 Component skew
Skewing is also caused by an unequal wetting at the two terminations on a discrete component. The surface tension, at the end to reflow first, often acts to realign the component out of its original placed or desired position. With chip components, the solution is as for tombstoning above. Pad design can also affect the surface tension at one end of a chip component (Figure 3).
Placing two chips on shared pads or having differing width connection tracks or via holes within the pad can influence the realignment of the components concerned.
Some integrated circuits, such as QFNs have comparatively large grounding planes on their underside between the terminations. If solder paste is applied to the corresponding central copper pad without reductions, the volume of solder available causes the component to skew out of position.
7.6 Cold joints
Cold joints will be caused by low-peak temperatures not fully reflowing the solder mass (Figure 4). To achieve the required solder fillet appearance, it is essential that minimum peak temperatures are achieved and also the time above liquidus (TAL) is sufficient.
7.7 Voiding
Voiding or non-metallic pockets within a solder joint are largely caused by out-gassing of solvent materials. Pockets of flux can also become trapped within a solder joint.
Minimisation of this problem relies on reducing the soldering time or the pre-heat ramp-up rate.
7.8 Intermetallic growths
Excessive intermetallic growths are caused by exposure to heat over time where resultant absorption of metallic elements form layers within the solder joint and also at the interface with the surface mount pads and component terminations.
Thicker intermetallic layers make a solder joint harden and become brittle. Care should always be taken to ensure excessive peak temperatures and TAL are avoided.
Grain size: the appearance of large grain sizes in the solder joint is indicative of a slower cooling rate. This can be avoided by ensuring the cooling rate is set between 3 and 4°C/s.
7.9 Cracks
Component cracks can develop due to the thermal stress inside of the component (Figure 5). Care should be taken when heating or cooling to ensure both the PCB and the component population do not undergo rapid temperature changes.
Solder joint cracks are usually the result of mechanical stress at the component terminations. This can be caused by substrate flex which in some cases is exacerbated by the use of inappropriate solder pad designs which creates too much solder on component terminations.
7.10 Fillet lifting
Pad lifting, fillet lifting and fillet tearing (Figure 6) are effects that result in part from the differences in thermal coefficients of expansion between the substrate material and the copper barrels and tracks on the PCB.
There is a relatively large expansion of the laminate material in the Z-axis throughout contact with the molten solder in the wave or selective soldering processes. This expansion causes conical deformation of the copper pads.
When solder joints start to solidify the board material cools down and returns to its planar shape. This movement can create stress on the surface of the solder joint, which at this stage has not gained its full strength. Such stress may cause pad lifting or – if adhesion between the copper pad and substrate is at that point stronger than the solder – will cause cracks in the solder fillet surface known as fillet tearing.
7.11 Solder paste printing defects
Many individual company workmanship standards exist alongside those published by the SMART group and other bodies, but generally they all refer to qualifying the printing process in terms of:
registration of the printed deposits to the copper pad pattern;
the effective solder paste coverage of the pads;
the appearance of the printed deposits and their definition and consistency;
paste thickness achieved; and
appearance of obvious defects such as slumping, bridging and spikes.
7.11.1 Good print
Paste is printed and contained within the boundary of the copper pads.
7.11.2 Mis-aligned
Poor registration of the stencil image to the corresponding PCB features.
Incorrect stencil alignment provides a bleed path, for flux and solder balls,contaminating both the gap between pads and also the underside of the stencil.
7.11.3 Bridged
Bleeding or bridging: the result of incorrect gasketting between the stencil aperture and the copper pad or feature on the PCB. Never assume that the PCB or substrate features match the original data sizes exactly.
Dependent on the thickness of the copper on the PCB, it is possible to lose up to 100 μms from the target size required.
Bridging is often a problem with finer pitched devices and can be caused by a number of factors including:
inadequate gasketting through incorrect stencil aperture reductions,over-etched PCBs, or inaccurate stencil to substrate alignment;
component co-planarity errors or excessive pick and placement pressure;
paste slump caused by too rapid a temperature gradient in the pre-heat zone;and
dirt and particulate contamination, for example a hair or fibre filament across two adjacent pads causes an attraction and flow path between them often resulting in shorts.
When printing fine pitch requirements the stencil must gasket onto the copper pad in order to confine the printed deposit to the pad itself.
7.11.4 Slumping
Paste slump can be caused by using solder paste outside of its usual environmental/temperature/humidity range. The same effect can occur using out-of-date solder paste.
7.11.5 Scavenging
Scooping or scavenging: loss of the solder paste volume, for larger stencil apertures, such as base pads of D-PAK devices. The answer here is to brace the aperture to make several smaller apertures within the original pad border.
7.11.6 Ripped print
Ripped print can be caused by movement at the end of the print stroke or the use of cantilever, not straight lift, separation of the substrate from the stencil postprinting.
If using a simple cantilever separation make use of a two-axis or table up/down control to separate by a minimum of 1.5× stencil thickness.
7.11.7 Insufficients
Too little solder paste printed usually the result of incorrect aperture sizes or excessive printing speeds.
7.12 Screening can requirements
Paste migration can lead to depleted solder fillets and voids.
Co-planarity issues are not limited to semiconductors; RFI shielding cans play an increasing role in the quality of finished assemblies today.
Often “pressed” (or stressed) cans will not sit flat on the PCB itself. Several methods are currently used to overcome such problems, including preheating to anneal the cans and placing weights on the cans during reflow but this can affect the reflow characteristics of the assembly and is also likely to lead to problems with joint integrity during cooling and use.
Some screening cans do not exhibit severe co-planarity problems as they are manufactured using photo chemical machining which does not induce stress into the material. Once plated and assembled the seating faces on all four sides largely fulfill the flatness tolerances required.
The use of selective thickness printed solder deposits can also help to overcome any co-planarity problems. Printing increased height paste deposits for the screening can, on the same printing pass, eliminates problems associated with this non-conformance and provides stronger solder fillets which increase mechanical security and prevent problems associated with voided fillets.
Often, paste deposits are printed in isolated blocks (Figure 7). The solder does not, however, reflow sufficiently and creates voids and blowholes in the sealing fillet, which may require reworking. One answer is to ensure the paste deposits do not have these large interrupts that cause the effects in the first place. It is possible to produce integral deposits, which assist in reflowing without voids. Replacement of the solid metal tags with suitable mesh aperture patterns that join the isolated deposit apertures permits the deposition of sufficient solder material and flux in the paste. This encourages the necessary conditions to provide surface tension ensuring solder flows over all areas giving an even fillet all around the can.
Since both the PCB and screening can initially possess good solderable finishes, the addition of printed solder paste combined with the heat applied during reflow often encourages the migration of solder up the can walls away from the PCB. The result is sometimes less than pleasing in terms of appearance and may not provide effective shielding or the mechanical strength required in the solder fillet (Figure 8).
A unique solution to this problem of solder fillet migration is the “reflow plimsoll line”. The interruption in the plating finish on the can wall allows the solder deposit only minimal migration and ensures fillet strength and volume are maximised. It also avoids the problems associated with the formation of blowholes, gaps and the untidy appearance of the can wall finish.
7.12.1 Stencil design considerations for screening can applications
A traditional approach, including webs between adjacent apertures as in Figure 9, often results in isolated paste deposits that do not always deliver integral solder fillets and may need rework.
The enhanced stencil design in Figure 10 including metal mesh ensures paste deposits on whole of can wall and the presence of flux and paste encourages the reflow process. Multi-level stencils can also ensure appropriate paste deposits.
