Solder Mask Cure Verification: Degree of Cure and Final Hardness
Solder mask that looks fully cured may still be under cured, and an under cured mask fails later as poor adhesion, blistering in reflow or ionic leakage between conductors. The solder mask cure state has to be measured, and the measurement has to be made on the same surface the process produced.
What Final Cure Actually Has to Achieve
Solder mask is applied as a photoimagable or thermally cured polymer and then cured to crosslink the resin. Crosslinking converts the soft, soluble film into a hard, chemically resistant coating that will survive plating, hot air solder levelling, reflow and cleaning. The degree of cure is the fraction of the available reactive groups that have reacted, and it is the property that governs hardness, adhesion, solvent resistance and ionic cleanliness of the finished surface.
The consequences of under cure are cumulative and delayed. An under cured mask is more permeable to moisture and to ionic species, so flux residues trapped beneath it can migrate between conductors, the mechanism that ionic contamination testing is designed to detect. It adheres less strongly to copper, so it lifts at panel edges during routing and blisters over a copper plane during reflow. It also releases volatiles in the reflow oven, which can deposit on the oven interior and on other boards. None of these symptoms appears at the mask station, which is why the cure state has to be measured rather than inferred from a visual inspection. Over cure is a problem too, though a smaller one. Excessive thermal exposure embrittles the mask, reduces its elongation and makes it prone to cracking at the corners of large pads and around plated holes.
Thermal Cure Schedules and the Effect of Board Thickness
A typical final cure is 150 C for 60 minutes, but the schedule is a surface temperature and time and not an oven setting. The board takes time to reach temperature, and the time is longer for thick, copper heavy panels because the copper planes conduct heat away and the thermal mass is larger. A 1.6 mm four layer board reaches cure temperature in a few minutes; a 3 mm twelve layer board with heavy copper planes may take ten times as long.
That means the cure schedule has to be qualified on the worst case thickness and copper distribution in the product family, and the soak time has to be counted from the moment the board surface reaches temperature rather than from the oven entry. Confirm the board temperature with a thermocouple attached to the board surface under a patch of mask, and record the trace with the cure specification. Where the mask is curable by ultraviolet light followed by a thermal step, the two stages have different roles. The ultraviolet exposure sets the pattern by crosslinking the exposed areas so the unexposed material can be developed away; the thermal step completes the cure. Reducing the thermal step and relying on the ultraviolet dose produces a mask that is hard at the surface and soft underneath, which passes a pencil hardness test and fails an adhesion test.
Measuring Degree of Cure by FTIR
Fourier transform infrared spectroscopy measures the absorbance of chemical bonds, and the cure reaction consumes specific groups. For an epoxy based mask the epoxide band near 915 wavenumbers is the usual choice; for an acrylate system the carbon to carbon double bond near 810 or 1636 wavenumbers is used. The degree of cure is the reduction of that band relative to a reference band that does not change with cure, such as the aromatic ring absorption near 1510 wavenumbers.
The measurement is a ratio, so it does not require a calibration standard, but it does require a consistent baseline and a consistent sampling depth. Attenuated total reflectance samples only the top micron or two of the film, which is exactly where the surface cure is highest; transmission measurements on a peeled film sample the bulk. For a cure specification, state which technique is used, because two laboratories reporting different numbers may simply be sampling different depths. Report the result as a percentage relative to the fully cured reference, and set the acceptance limit from a sample that has been cured under the production schedule and has passed the performance tests. A typical limit for a photoimagable mask is 90 percent or more of the reference. Note that FTIR measures the whole sample area, so measure at several locations across the panel and on more than one panel per lot.
Differential Scanning Calorimetry and Residual Enthalpy
Differential scanning calorimetry measures the heat absorbed or released as a sample is heated, and an under cured mask shows a residual exotherm as the remaining reactive groups react during the scan. The residual enthalpy divided by the total enthalpy of a fully uncured sample gives the fraction that did not react, and the degree of cure is the complement.
The technique is destructive and needs a few milligrams of material, so it is a laboratory method rather than a production one. It is valuable for qualifying a new mask or a new cure schedule because it measures the bulk rather than the surface, and it catches the case where the surface is cured and the bulk is not. Run the reference on a sample of the same mask that has been fully cured by an extended schedule, so that the total enthalpy is realistic. Where both FTIR and calorimetry are available, run them together during qualification and then use the faster method for routine control. The correlation between the two is what makes the routine method defensible.
Pencil Hardness and Surface Cure
Pencil hardness is the classic shop floor check. A set of pencils of known hardness is drawn across the cured surface at a fixed angle and load, and the hardest pencil that does not cut or scratch the film is recorded. The method is quick and cheap, and it measures the surface rather than the bulk, which makes it a good indicator of the thermal cure step but a poor indicator of through thickness cure.
Standardise the test conditions or the results are meaningless. The load, the angle, the stroke length and the pencil grade all change the outcome, and different operators sharpen pencils differently. Follow the method in the adhesion and cure testing standard, use a fixed pencil holder, and re verify the pencil set periodically against a reference panel. Record the result as the grade that passes, together with the conditions. Where the mask is being cured over a copper plane, check the hardness over copper as well as over bare laminate. Copper conducts heat away, so the mask over a large plane is often less cured than the mask in an open area, and that difference is the one that matters when the board reaches reflow.
Solvent Rub and Chemical Resistance Checks
A solvent rub test applies a defined solvent to the cured film with a defined cloth and rubbing pressure and counts the number of strokes needed to break through. It is a direct measure of crosslink density and it correlates well with the performance the mask has to deliver in cleaning and in flux removal. Common solvents for the test are methyl ethyl ketone, isopropyl alcohol and the cleaning chemistry the product will actually meet.
Set the pass criterion from a reference panel rather than from a generic table. The number of strokes a well cured mask survives depends on the mask type and on the film thickness, and a specification copied from another product will either pass everything or fail everything. Record the solvent, the cloth, the load and the stroke count with the result. Chemical resistance in service is broader than the rub test. A mask that survives methyl ethyl ketone may still be attacked by a high pH aqueous cleaner or by a specific flux activator. Where the product will meet a particular chemistry, test with that chemistry as well, using a coupon that has been through the full assembly process rather than a freshly cured panel.
Adhesion Tests: Tape, Cross Hatch and Pull
Adhesion is the property that the customer experiences as a failure, and it should be measured directly. A tape test applies a specified tape to a cross hatched pattern and pulls it off at a controlled angle; the amount of mask removed is graded against a standard scale. The test is quick and is sensitive to both under cure and surface contamination.
The pull test gives a numerical result. A wire or a pad is soldered to the solder mask or to a test feature and pulled normal to the surface, and the force at failure is recorded. It separates adhesion differences that a tape test grades as a pass, and it is the better choice for comparing two suppliers or two cure schedules. Test after the thermal excursions the board will actually see. A mask that passes a tape test as cured may fail after one reflow cycle, because the thermal expansion mismatch between the mask and the copper loads the interface. Where the product is reflowed twice, run the adhesion test after two cycles, and include a humidity soak before the test if the product will be stored in a humid environment.
When Cure Looks Right and Is Not
Three situations produce a mask that looks cured and is not. The first is a surface cured film over an under cured bulk, produced by an ultraviolet dose that is too high relative to the thermal step. The second is a mask cured over a heavy copper plane where the copper acts as a heat sink, so one region of the panel is fully cured and another is not. The third is a mask cured in an oven that is loaded too densely, so the boards in the centre of the load never reach the set temperature.
Each has a characteristic signature. The first passes hardness and fails FTIR of the bulk or a solvent rub that reaches deeper. The second gives different results in different regions of the same panel. The third gives different results between the edges and the centre of the load. Recording where the sample came from is therefore as important as the measurement result, and a single sample from the centre of a panel tells you nothing about the panel.
Records and Incoming Verification
Record the cure schedule with the measured board temperature trace, the oven identification, the load pattern and the results of the routine checks with the location each sample came from. When a customer report of mask blistering arrives, the record distinguishes a cure problem from a laminate moisture problem or a contamination problem, which have different corrective actions.
Verify incoming printed panels with the same discipline used for any other purchased process. Ask the supplier for the cure schedule and for evidence that it was verified on a board of comparable thickness and copper distribution, and confirm with your own measurements on the first article. Where the supplier’s cure schedule is quoted as an oven setting without a board temperature measurement, treat the panel as unverified and measure it yourself.
FAQ
How is degree of cure measured on solder mask? FTIR is the routine method: it compares the absorbance of a bond consumed by the cure reaction against a reference band and reports the result as a percentage of a fully cured sample. Differential scanning calorimetry measures the residual exotherm and is used for qualification because it samples the bulk rather than the surface.
Is pencil hardness enough to confirm cure? No. Pencil hardness measures the surface and is useful for checking the thermal step, but a mask can be hard at the surface and soft underneath. Combine it with a solvent rub and an adhesion test, and confirm the through thickness cure with FTIR or calorimetry during qualification.
Why is the mask over a copper plane less cured? Copper conducts heat away from the mask, so the mask over a large plane reaches a lower temperature than the mask over bare laminate for the same oven setting. Measure the cure in both regions, and qualify the schedule on the worst case copper distribution in the product family.
