Solder mask selection decided by ink family, thickness over copper, dam width and CTI rather than datasheet headline numbers, with a drawing checklist.

PCB Solder Mask Selection: Ink, Thickness and CTI Checks

Solder mask cross section showing coverage over copper, a land opening, the dam between lands and a tented via with thickness labels

Most solder mask selection decisions are made before the panel reaches the imaging line, in a conversation that lasts about ninety seconds and covers only the ink brand and the surface finish. Solder mask selection deserves more than that, because the same ink that performs perfectly on a 1.6 mm four-layer control board will blister on a 0.8 mm HDI build with 0.5 mm pitch lands, and the drawing will not have said anything different in either case.

This guide treats the mask as three separate specifications that have to agree: the ink family, the thickness and geometry it has to hold, and the electrical job it does when creepage distances get tight. Each section gives the number that matters and the way it is checked on the floor. If you are still deciding the dielectric system underneath, the laminate and material guide covers that layer of the decision.

What the Solder Mask Actually Has to Do

The mask is a permanent polymer coating that stays on the board for its entire life, so it is judged on four jobs at once. It has to insulate between conductors that are closer than the laminate alone would allow, act as a dam that keeps molten solder where the pad is, protect copper from oxidation and handling damage, and survive assembly without cracking, blistering or losing adhesion.
Those jobs pull in different directions. A thicker coat insulates and survives handling better, but it is harder to develop cleanly at fine geometry and more likely to trap solvent. A thinner coat images better at 4 mil dams, but coverage over a 2 oz copper step becomes marginal. Everything in the rest of this article is a consequence of that trade-off.
One clarification that saves arguments later: the mask is not a substitute for the fabrication drawing. If the drawing does not state the ink type, the thickness target, the via treatment and the acceptance criteria, the fabricator will apply its house standard, which is optimised for yield rather than for your assembly process.

Solder Mask Selection Starts With the Ink Family

Four families cover almost everything built today, and the choice is driven by geometry and volume rather than by preference. Photoimageable liquid ink applied by screen printing is the workhorse for multilayer and HDI work. The same chemistry applied by spray or curtain coating suits thin panels and high-mix shops. Dry film mask is laminated as a sheet and imaged, which gives the most uniform thickness. Thermal-cure screen ink is the low-cost option for simple boards, patterned by the screen itself rather than by exposure.
Ink familyHow it is appliedGeometry it holdsBest fitWhere it hurts
Photoimageable liquid, screen printedScreen coat, pre-cure, expose, develop4–5 mil dam on 1 oz copperMultilayer and HDI volume workMesh marks and thickness variation across a large panel
Photoimageable liquid, spray or curtain coatedSpray or curtain coat, expose, developFiner dams on thinner filmsThin panels, dense fine-pitch boards, high mixCoverage over tall copper steps; needs flat, clean panels
Photoimageable dry filmVacuum lamination, expose, developUniform thickness, fine openingsFine-pitch HDI where film uniformity mattersConformability around thick copper and heavy via fill
Thermal-cure screen inkScreen print, oven cure, openings set by the screenCoarse openings, wide damsSimple double-sided and low-density boardsRegistration on fine pitch, limited tenting of small vias
Two practical rules follow from the table. First, never compare inks by datasheet numbers alone, because the datasheet describes a 25 micron film on smooth copper, not the 15 micron film you will get over a trace edge. Second, if the board mixes heavy copper with fine pitch, no single coat does both well and the conversation should move to which region can accept a compromise.

Thickness Targets Over Copper, Edges and Pads

Thickness is specified on the finished board over the conductor, not on the wet film. A common floor is 0.4 mil (10 µm) over conductor edges, with Class 3 work frequently calling for 0.8 mil (20 µm) and above. Confirm the figure in your own acceptance standard rather than assuming the house number, since the standard you cite in the purchase order is what an inspector will measure against.
Location on the boardTypical targetWhy it is specified thereHow it is verified
Over a conductor on the outer layer0.4 mil (10 µm) minimum; 0.8 mil (20 µm) for Class 3 workInsulation and resistance to handling damageMicrosection on a coupon, or an optical thickness gauge
At a land openingNo ink on the land; opening 2–4 mil larger than the land per sideSets the wetted area and the solder fillet shapeFirst-article visual against the drawing, then AOI
Between two lands (dam)4 mil (0.1 mm) minimum, and pitch-limitedPrevents solder bridging between adjacent padsAOI with dam width measurement at the tightest pitch
Over a tented viaFull coverage with 0.5 mil over the rimStops flux and contamination entering the barrelVisual and cross-section on the qualification coupon
At the board edge and rout5–10 mil clearance from the rout pathStops the mask flaking along a machined edgeVisual at the array frame
The number people forget is the one over a copper step. On a heavy copper board the mask film thins dramatically as it drapes over the trace shoulder, so a nominal 25 micron coat can measure 8 microns at the edge. If the drawing asks for coverage over 2 oz copper, say so explicitly and let the fabricator choose a coat weight that survives the step.

Resolution, Registration and Dam Width

Dam width is the single most useful capability question, because it combines ink resolution, imaging exposure and registration in one number. Ask what the minimum dam is at your copper weight and pitch, not what the ink datasheet claims. A shop that holds 4 mil dams reliably on 0.5 mm pitch is telling you something concrete about its exposure and development control.
Registration is the other half. The mask opening has to sit concentrically on the land, and on a 0.4 mm pitch array there is very little room for error before the opening covers part of the neighbouring land and creates a bridge risk. Registration is limited by the lamination and imaging alignment, not by the ink, which is why the same ink behaves differently on panels of different size.
Below about 4 mil, dams start to fail in a characteristic way: a sliver of ink lifts during development, the opening merges with its neighbour, and the board passes visual inspection in one orientation while bridging at reflow. If a design needs dams under that width, the correct answer is usually a different pad geometry or a change of pitch, not a thinner ink.

Adhesion, Cure Window and the Two Failure Modes

Almost every field failure I have seen on mask traces back to cure, in one direction or the other. Under-cure leaves residual solvent and unreacted photoinitiator, which shows up as ionic contamination, patchy adhesion and occasionally as a dull, uneven surface after reflow. Over-cure makes the film brittle, so it cracks at via rims and along rout edges when the panel is depanelised.
The objective evidence is boring and effective: a cure schedule held inside a documented window, a tape peel or cross-hatch on the first article, pencil hardness on a coupon, and an ionic contamination measurement on the finished panel. Ask for the cure window in the process control plan and then ask what the operator does when a panel is double-passed through the oven, because that is where the window usually gets lost.
Adhesion also depends on what is underneath. Oxide treatment, micro-etch depth and surface cleanliness before coating determine whether the film keys to copper or merely sits on it. A board that passes every measurement on a smooth test coupon can still fail adhesion over a brushed, contaminated production panel.
Three solder mask defect signatures: a skip at a land edge, a blister after reflow and a crack at a tented via, each with cause and control

Surface Finish Compatibility

The mask sees the finish chemistry and the assembly thermal profile, so the pair has to be specified together. This is where a technically adequate ink becomes an inadequate choice.
Surface finishWhat it does to the maskWhat to specify
Lead-free HASLMolten solder near 260–270 °C plus a hot-air wipe; the harshest thermal event in common useInk rated for the thermal excursion, and a thickness that survives the wipe without thinning at the land edge
ENIGAlkaline and acid baths at 80–90 °C over several transfersChemical resistance, plus adhesion checked after the full plating cycle rather than before it
Immersion silver or immersion tinMildly acidic, short dwell, low thermal loadStandard photoimageable liquid ink; watch for staining at mask edges
OSPAcidic, low temperature, no thermal shockConfirm the ink is compatible with the acid cleaner, then treat it as a low-risk pairing
Hard gold on edge fingersPlating, taping and tape removal over a masked surfaceInk that resists tape peel; fingers are usually masked off and plated separately
ENIG deserves a note because the failure is so often blamed on the wrong layer. A board can show poor solder wetting that looks like a mask issue when the real cause is the nickel bath, which is covered in this ENIG solderability and black pad analysis. Check which layer failed before changing the ink.

CTI, Creepage and the Electrical Job

On mains-connected and industrial boards the mask stops being cosmetic and becomes part of the insulation coordination. Comparative tracking index describes how a material resists forming a conductive carbon path across a wet, polluted surface, and the mask film sits directly in that path between two conductors.
The important subtlety is that the ink does not upgrade the laminate. If the base material is rated for one tracking class and the ink for a better one, the assembly is limited by the weaker of the two along any path that crosses bare laminate, for example in a mask opening or at a board edge. Where creepage distances are tight, specify a tracking-improved ink and confirm the laminate class at the same time, following the spacing logic in this CTI ratings and creepage guide.
For products above roughly 300 V working voltage, write the requirement into the drawing as a material property with the standard cited, for example the spacing rules of IEC 60664-1, and ask the fabricator for the ink’s tracking data. A generic claim of good insulation is not something an inspector can check.

Tenting, Plugging and What Happens at the Via

Via treatment is a separate decision from mask type and is the most commonly unstated item on a fabrication drawing. Tenting covers the hole with mask during the normal coat. Plugging fills the barrel with resin or ink before or after coating. Leaving the via open is the third option, and it is legitimate when the via is used as a test point or when outgassing during assembly must be avoided.
Tenting works well down to about 0.3 mm and starts to fail below that, because the film cannot bridge the hole without support and sags into the barrel during cure. Dense via fields on a thin board are the worst case: the film stretches over many holes at once and cracks at the rims. The comparison between the two treatments is set out in this tenting versus plugging comparison, and the resin-versus-ink choice for plugged vias is covered in the via plugging materials guide.
For via-in-pad the decision is already made for you: the via must be filled and planarised, because a tented via under a component pad will cause voiding or an open joint. That process belongs with the plating and fill step, not with the mask.

Incoming Checks Before the Ink Reaches the Line

The mask is a purchased chemical, and it deserves the same incoming discipline as laminate. Four checks catch nearly all of the problems that would otherwise appear as random line defects.
  • Lot certificate matching the ink part number, batch and date, with a shelf life that has not expired.
  • Viscosity measured against the process window on receipt, not only when a defect appears.
  • Storage and mixing conditions recorded, including the pot life clock once the hardener is added.
  • First-article verification on a production coupon: cross-hatch adhesion, thickness, dam width and cure hardness.
Then close the loop with the acceptance criteria in the same document. The workmanship requirements for mask integrity, coverage and marking should reference a published standard such as the IPC criteria at Global Electronics Association, while the flammability and material ratings come from the supplier’s UL record. Two references, two different questions.

Solder Mask Selection Checklist for the Drawing

The drawing is the only place where all of this survives. Put the six items below on it and most of the arguments disappear before the first panel is coated.
Drawing itemWhat to writeWhy it matters
Ink familyPhotoimageable liquid, applied by screen or spraySets resolution, thickness range and cost level
Thickness0.4 mil minimum over conductor edges; 0.8 mil for Class 3 areasGives the inspector a measurable target
Dam widthState the minimum acceptable dam at the tightest pitchConverts geometry into a pass or fail decision
Via treatmentTented, open or filled and planarised, by via sizeRemoves the most common silent default
Tracking requirementTracking index or class, plus the reference standardMakes an electrical property verifiable
Acceptance standardClass 2 or Class 3 per the workmanship standard you citeAligns the visual criteria with the rest of the drawing
If only one item can be added, make it the via treatment. It is the field that most often changes without anyone noticing, and the one whose absence produces the most expensive late discovery.

Frequently Asked Questions

What is the minimum solder mask thickness over a trace?

0.4 mil, or 10 microns, over the conductor edges is the common floor, and Class 3 work is often specified at 0.8 mil or 20 microns. Measure on the finished board rather than the wet film. Over a heavy copper step the film thins sharply at the trace shoulder, so a nominal coat can fall below the minimum right where it matters.

Four mil, roughly 0.1 mm, is the practical minimum for a stable dam, and it depends on both copper weight and pitch. Below that, slivers lift during development and adjacent openings can merge, which shows up as solder bridging at reflow rather than at inspection. If the design needs less, change the pad geometry instead of the ink.

LPI stands for liquid photoimageable, so yes in everyday usage, but the application method still matters. The same chemistry screen printed and spray coated behaves differently over copper steps and on thin panels. When a supplier says LPI, ask how it is applied before accepting the capability claim.

It can, on any path where the tracking route crosses the mask film, but it cannot improve the laminate underneath. If the base material is in a lower tracking class, a better ink does not raise the assembly’s rating along paths crossing exposed laminate. Specify both properties and confirm which one governs each gap.

Moisture is the usual cause: solvent left in an under-cured film, or water absorbed on the copper before coating, expands at reflow temperature and lifts the film. Over wide copper areas the effect is strongest because there is more trapped volume. Bake panels before coating and hold the documented cure window.

No. A tented via under a land will trap air and flux during assembly, which produces voids or an open joint. Vias inside pads must be filled and planarised before the mask step, and the fill belongs to the plating process rather than to the coating line.

Four: a dam width measurement at the tightest pitch, cross-hatch adhesion on a production coupon, thickness over a conductor edge, and a cross-section through a tented via. Those four cover the geometry, the chemistry and the via treatment, which is where nearly all field failures originate.

Put the Mask Requirement on the Drawing, Not in an Email

Send the stackup with the tightest pitch, the copper weight and the assembly thermal profile, and the engineering team will come back with the ink family, the thickness target and the via treatment that fit your build rather than the shop default.

The manufacturing services overview shows how coating, imaging and finish steps are sequenced, which is useful when a mask change has to be coordinated with the finish specification.

Share your love

Newsletter Updates

Enter your email address below and subscribe to our newsletter