Crazing in Solder Mask and Laminate: 6 Causes and Controls

Crazing is a network of fine cracks that appears in a coating or in a laminate surface, usually with no single direction, so that the surface looks as though it has been shattered into small cells. Each crack is thin, and individually it looks harmless. A network of them is not, because it opens a path through a material that was there to provide insulation and protection.

The condition appears on solder mask that has been attacked by a chemical, on mask that has been overheated, and on laminate that has been bent or stressed after assembly. Identifying which of those mechanisms is present decides whether the answer is a change of cleaner, a change of cure or a change of handling, and the three have nothing else in common.

What Crazing Is and What It Is Not

Crazing is a cracking pattern within the material itself, not a coating peeling away from the surface and not a scratch. The cracks are usually shallow at first and become deeper with repeated stress, and they tend to start at an edge, a corner or a scratch where the local stress concentration is higher.

It is often confused with the residue left by a chemical attack, with a fine bloom of flux, and with the appearance of a mask that has been over baked. A quick check under magnification, and a look at whether the pattern follows a stress field or the geometry of the panel, is usually enough to tell them apart.

Chemical Causes

Solvent attack is the classic cause on solder mask. A cleaner that is too aggressive, a flux whose activator stays on the surface too long, a solvent used to remove residue that is left to evaporate slowly, and an alcohol used to wipe a board that has not been properly cured can all soften the polymer enough for stress to craze it.

The concentration and the contact time matter as much as the chemistry. A wipe that is normally safe becomes harmful when a puddle is left in a recess, and a cleaning bath that is normally reliable becomes aggressive when its concentration drifts. That is why the chemical controls belong with the cleaning process record rather than with the consumable store, and why the general contamination control routine is the right place to catch the change.

Thermal Causes

Mask and laminate expand when they are heated, and if the coating is cured to a different degree from the layer beneath it, the mismatch produces stress at the interface. A reflow profile that is hotter than the mask was qualified for, a hot air rework operation with the nozzle too close, and a board that is heated twice where the design expected once all raise the risk.

Cure is the other half. A mask that has not been fully cured is softer than the specification assumes, so it crazes at a lower stress, and the solvent it still contains leaves the film progressively as it is heated. Checking the cure against the supplier schedule, and verifying the actual board temperature rather than the oven setting, removes most of the thermal cases.

Mechanical Causes

Depaneling is the most common mechanical source, because a board that is bent along a score line or broken out of a panel sees a local strain that the mask may not survive. Screwing a board down onto an uneven chassis, pressing a probe onto the surface, and handling a panel by its corners all apply stress that concentrates at edges.

The pattern of the crazing often identifies the source. Cracks that follow the edge of a panel or the line of a tab point at depaneling. Cracks that radiate from a mounting hole point at the fastener. Cracks over a large component point at flexing during assembly or at a board support that was missing in reflow.

Where It Hurts

The immediate effect is cosmetic, and cosmetics are the least important part. The real effect is that the mask no longer does its job. A cracked film takes up moisture, holds ionic residue against the surface, and provides a path for electrochemical migration between conductors that were separated by only a few tenths of a millimetre of polymer.

Where a conformal coating is applied over a crazed mask, the coating will often follow the crack pattern and reproduce it on the outer surface, which is why a coating adhesion failure and a crazing problem can look like the same defect. The adhesion test described in the coating adhesion procedure will separate the two.

Detection in Production

Crazing is easier to see under a low angle light or under ultraviolet, because the fluorescence of the mask highlights the crack edges. It should be looked for after depaneling, after any cleaning operation and after any rework, since those are the points where the damage is introduced.

The inspection should be done before the conformal coating is applied, because the coating makes the pattern harder to read and can conceal the origin. Where the assembly is coated, handling after coating is also relevant, and the rules in the post coating handling notes apply to the mask as much as to the coating.

Prevention

Prevention follows the cause. Chemical cases are fixed by specifying the cleaner concentration, contact time and rinse, and by verifying them in the bath rather than on the label. Thermal cases are fixed by curing to the schedule and keeping rework profiles within the qualified range. Mechanical cases are fixed by supporting the panel during depaneling, by controlling the fastener torque and by handling boards by their edges.

The common thread is that each cause produces a small change that is easy to overlook until a network of cracks appears. Where the mask is being qualified for a product, a deliberate stress test on a sample coupon gives the answer long before a production batch does. The mask material, its cure and the way the board is handled are all part of the same solder mask control, and the crack pattern is the evidence that one of them has moved.

Where the same crack pattern appears across a whole batch, the cause is almost always in the process rather than in the handling of one board. A change of cleaner, a cure oven that is running cool, a new depaneling tool and a modified rework profile are all capable of producing a batch wide result, and each of them leaves a record. Comparing the process history of the affected batch with the previous one usually identifies the change within a few minutes.

FAQ

Is crazing in solder mask a cosmetic defect? It starts as an appearance issue and becomes a reliability issue. The cracks allow moisture and ionic material to reach the surface, so a network of them should be treated as a defect rather than as a finish variation.

Does crazing come from the cleaner or from the oven? Both are common, and the crack pattern usually indicates which. A pattern that follows where a liquid pooled points at chemistry, while a pattern that follows a hot area or a panel edge points at temperature.

Can a crazed board be recoated? Recoating over a crazed mask generally reproduces the pattern and does not restore the insulation. Where the cracks are shallow and local, the surface can be cleaned and re-coated, but the cause should be corrected first.