Solder Mask Sliver: Preparation, Placement and Process Control

Solder mask is a coating, not a structural material. Where it has to sit between two closely spaced pads or between a pad and a trace, it forms a narrow strip that has to survive printing, placement, reflow and often cleaning. If that strip is too thin, it lifts, and the copper it was protecting becomes a bridge.

The rules that govern those strips are among the least intuitive in a layout rule set, because they are driven by the registration tolerance of the fabrication process rather than by the electrical design. Understanding where the numbers come from makes them easier to apply sensibly.

What a Sliver Actually Is

A solder mask sliver is any strip of mask that is narrower than the process can hold reliably. It appears wherever two openings in the mask come close together, which on a fine pitch device means between every pair of adjacent pads.

The sliver does not fail immediately. It is mechanically weak, and the combination of thermal expansion, paste adhesion and handling will eventually lift it, at which point solder can flow between the pads it was separating. The defect appears after assembly, when the cause is no longer visible.

Mask Dam Between Pads

A mask dam is the deliberate strip of mask between two pads that is intended to prevent bridging. On a coarse pitch device it is a robust feature; on a fine pitch part it becomes thinner than the process can produce, and the fabricator has no choice but to remove it.

Removing the dam is not a defect. It is a decision, and it changes the assembly process: without a dam, the paste deposits on the two pads are separated only by the aperture geometry and the surface tension of the solder, so the stencil design carries more of the bridging risk.

Registration and Its Tolerances

Registration is the alignment of the mask artwork with the copper beneath it. Every process step contributes to the error: the artwork itself, the exposure, the development and the lamination. The total is a few tens of microns on a capable line, and it is the number the design rules are built from.

A dam that is nominally wide enough but sits at the edge of the registration window will be too thin on one side and too wide on the other, and the thin side is where the defect appears. This is why the rule is expressed as a minimum width that accounts for the tolerance rather than as the drawn width.

Mask Defined and Copper Defined Pads

Where the mask opening is smaller than the copper pad, the pad is said to be mask defined. The mask overlaps the copper all around, which gives a defined soldering area and protects the pad edge, at the cost of a smaller opening and a greater sensitivity to registration.

Where the opening is larger than the copper, the pad is copper defined. The solder wets the whole pad and the mask sits away from the joint, which is more tolerant of registration error and produces a slightly wider deposit. The two styles should not be mixed on the same device, because the paste volume and the stand off height would differ across the part.

Tenting and Its Uses

Tenting is the practice of covering a via with solder mask rather than leaving it open. It protects the barrel from flux and from solder, and it prevents a via in a pad from wicking paste away during reflow. A fully tented via has mask on both sides; a partially tented one has it on one side only.

A tent over a via is itself a sliver if the mask annulus is narrow, and a large via is difficult to cover completely, because the mask sags into the hole. Small vias tent reliably, larger ones do not, and a via that must remain sealed should be filled and capped rather than tented.

Writing a Workable Design Rule

A workable design rule states the minimum mask width the process can hold, the registration tolerance that it assumes, and the consequence when the geometry cannot meet it. The last part is what most rule sets omit, and it is the part the designer needs.

The practical consequence of a dam that is too narrow is usually that the mask is removed and the pads become copper defined, so the fallback should be decided in advance. Where that fallback is acceptable, the geometry is not a violation; where it is not, the pad spacing has to change, which is a layout decision with a real cost.

Common Layout Mistakes

The most common mistake is a mask opening that follows the copper exactly, leaving no margin at all. The second is a thermal relief or a ground pour that comes so close to a pad that the mask between them is a sliver, and the third is a via placed close enough to a pad to pinch the mask around it.

Each of these is visible in a design rule check if the rule set contains the mask constraints, and invisible if it does not. Checking the mask layer as a geometry in its own right, rather than as a by product of the copper layer, is what catches them before the data is released; the same discipline is described in the review of pad design standards.

Fabrication and Assembly Consequences

A lifted sliver can bridge two pads during reflow, hold paste that would otherwise have printed cleanly, or leave a fragment on the board that is found later during inspection. The cost of the defect is much higher than the cost of the spacing that would have prevented it.

Where the pitch genuinely cannot provide a dam, the answer is to design the assembly for a dam free process: a thinner stencil, better paste release, and an inspection method that can see the bridges that form. Those are the same measures described in the guidance for manufacturable design, applied where the mask can no longer help.

Process Control and Verification

On a design of this kind, design rule is the item that decides how the rest of the board is arranged. A short note on the drawing about handling, storage or packaging is often worth more than an extra decimal place on a tolerance. Where a value sits close to a process limit, the drawing should say so, since the shop can then open the process window rather than working to a nominal figure that carries no tolerance.

Related reading: our fabrication notes, board quality and design release notes cover the same ground.

Inspection of the Mask

The mask layer is inspected after development and again after the finish has been applied. A sliver that survives development can still lift later, so the check is a combination of the geometry on the panel and a judgement about how much mask is actually present between the openings.

Where a product has a history of lifted mask, the fabricator can reduce the opening slightly or adjust the exposure to strengthen the strip. That adjustment is worth making with the quality of the panel as the reference rather than as a general setting.

FAQ

Is a removed mask dam a defect? No, provided it was removed deliberately and the assembly process was designed for it. It becomes a defect only when the paste and the stencil still assume the dam is present.

How wide should a mask dam be? Wide enough to survive registration on the fabricator’s line. A typical minimum is a little over a tenth of a millimetre for a capable process, and the fabricator’s capability sheet is the authority.

Does tenting affect the assembly? Yes. A tented via does not wick paste, and an untented via in a pad does, so the choice changes the paste volume calculation.