A print that scratches off the substrate can be two completely different faults. One is an incomplete cure inside the ink layer; the other is a failure at the interface. They look similar on the shop floor and they have opposite fixes - which is why adding cure energy is often the wrong first move.

This guide shows how to separate under-cured UV ink from adhesion failure: what each fault looks like, which field tests distinguish them, the variables to inspect before touching any setting, how to run a controlled one-variable diagnosis, and the rework and release rules that stop the same failure from returning next week.

Two different failures that look identical

Under-cure is a chemistry problem inside the ink; adhesion failure is a bonding problem at the substrate. The visible symptom - ink coming away - can be identical, while the correct response is not.

A UV ink film is cured by energy that starts a reaction in the photoinitiators and cross-links the material. If the energy delivered is insufficient, or the layer is too thick for the energy to penetrate, the film is soft in the middle. If the ink cures fully but never bonded to the surface, the film is hard and the bond is weak. Printers and operators usually describe both as "the ink doesn't stick", which is exactly the description that leads to the wrong adjustment.

The distinction matters commercially as well as technically. A cure problem is usually solved inside the machine: power, speed, pass count, lamp condition, ink thickness. An adhesion problem is usually solved at the substrate: identity, coating, contamination, preparation, primer. One is an hour of settings work; the other is a supply and process investigation that may reach the material supplier.

What does under-cured UV ink look like?

It feels tacky, smears under pressure and abrades.

Under-cured ink fails inside the layer: the print feels tacky or soft, smears under light pressure, and an abrasion test removes ink without revealing clean substrate.

The failure is typically uniform across the printed area, including the parts where ink sits on ink rather than on the substrate. Stickiness can appear after the part has been stacked or wrapped, because the film was still reacting when it was packed. A fingernail or a coin dragged across the surface will usually deform the ink rather than lift it cleanly, and the edges of the mark are smeared rather than sharp.

Under-cure also has a thickness signature. Thick builds - raised effects, multiple white passes, heavy varnish - cure less completely than thin ones under the same energy, so a fault that appears only on the high-build areas of a job points to energy penetration rather than to surface contamination. Where lamps have been running for a long time, or a lamp has been replaced without re-checking the profile, the same symptom can appear across every job rather than on one product.

What does adhesion failure look like?

The film is hard but lifts to expose bare material.

Adhesion failure occurs at the interface: the ink film feels hard and dry, but it lifts to reveal bare material, and the defect clusters where the surface was contaminated or masked.

The signature is the clean edge. Where a tape test is used - the cross-cut method described in ASTM D3359 is a widely used reference - the ink comes away cleanly and the substrate beneath is visible and undamaged. Failures often follow the geometry of the contamination: fingerprints along a handling edge, a band where a protective film was peeled, a patch where a previous part rested. On coated or film-protected materials, the ink may lift the coating with it, which is a substrate problem rather than a printing problem.

Adhesion failure also tends to be worse in the areas with the least ink. Thin areas have less material to hold onto, so the defect shows first where the print is light. That is the opposite of the under-cure pattern, where the thick areas fail first - which is one of the quickest ways to tell the two apart on a real part.

Quick field tests and what each one rules out

Three cheap checks separate the two faults in most cases: a thumb test, an abrasion test and a tape test. Run them on the same part, in the same order, and record what each one did.

The thumb test - firm pressure with a gloved thumb, held for a few seconds - reveals surface tack. If the ink is sticky or transfers to the glove, treat it as suspected under-cure. The abrasion test - a controlled rub with a defined material and a counted stroke - shows whether the film deforms or releases. Deformation suggests an incomplete cure; a clean release to bare substrate suggests an interface failure. The tape test then distinguishes them at the boundary: ink lifting with the tape and clean material underneath is an adhesion result.

Run the tests on an area that represents the job, not on a convenient corner. Where the part was printed with several layers, test both a high-build area and a thin area and compare: the difference is diagnostic. Record the results in words that can be compared later, because "felt sticky" from two different operators means two different things.

Variables to inspect before touching cure power

Inspect the substrate, the surface and the consumables before changing a curing setting. Each of the variables below can produce a fault that looks like under-cure, and adjusting the machine first destroys the evidence.

Start with material identity and lot, then storage and surface condition. Check the preparation step that was actually performed rather than the one that was specified, because a missed cleaning step or a shortened drying time is a frequent cause of localised failure. Confirm primer presence and age; a primer that has been open too long may no longer perform as recorded. Check the ink batch and whether it matches the recipe, and confirm the layer order against the approved file.

Then look at the machine condition. Head alignment and nozzle condition affect how much ink is laid down, and where a head has been replaced, drop placement changes. The technical documentation published by the component maker, such as Ricoh's industrial inkjet head documentation, explains how head geometry and channel configuration affect the printed result, which is why a head change belongs on the list before any cure adjustment is made. Where the investigation turns to the ink chemistry itself, published print research from bodies such as Fogra is a useful reference for how process conditions affect measurable output.

How is a one-variable diagnosis run?

Change one variable, print a fixed piece, judge once.

Change one variable at a time, print a fixed test piece, and judge it against the same criteria each round. A diagnosis that changes two settings cannot be attributed.

Start with the test that separates the faults rather than with a setting. If the tape test shows clean material underneath, you have an interface problem and the cure settings are not the first place to look. If the thumb and abrasion tests point to a soft film, then work through energy: confirm the lamp condition, check pass count, check the ink thickness, and only then adjust the profile. Print the same artwork and layer structure in each round so the comparison is meaningful.

Keep a written record of each round: what changed, what was printed, what the tests showed. Two or three rounds usually identify the cause; if the same variable keeps appearing, that is the signal to standardise it in the recipe instead of solving it again next month. Where a batch decision depends on counting defects rather than describing them, the acceptance-sampling approach in the NIST engineering handbook provides a documented basis for how many parts to examine.

When the problem is the material rather than the recipe

If the fault follows the material rather than the job - a new supplier, a new coating, a delivery that behaved differently - the recipe is not the cause. The prescription is a substrate decision, not a settings decision.

Look for the pattern. Failures concentrated on one delivery, one supplier or one coating point to the material. Failures that appear across several materials in the same week point to the process or the machine. Failures that follow handling suggest contamination or film residue rather than either. Where the material carries a supplier declaration covering restricted substances, keep it with the qualification record; the requirements summarised by the European Commission under REACH are part of what that declaration has to support in the European market.

When the material is the cause, decide between three options: change the material back to the qualified source, change the preparation to suit the new one, or re-qualify the process for the new material and record it. What should not happen is a permanent adjustment to the cure settings to compensate for a substrate change, because that adjustment will then be wrong for the material you started with. Where inks, primers and cleaning agents are handled during the investigation, the material information required by OSHA's hazard communication standard should be available to the people doing the work.

Stop, rework and release rules

Write the rule before the next failure: which defects can be recured, which can be stripped and reprinted, and which must be scrapped. Deciding this while the job is running guarantees an inconsistent answer.

A practical starting rule set: parts that are under-cured but uncontaminated may be recured and re-tested; parts that failed at the interface should be scrapped rather than reprinted over, because reprinting onto a failing layer hides the cause and doubles the labour; parts that failed after stacking are presumed under-cured and should be re-tested after curing rather than packed again immediately. Whatever the rule, record the decision on the job and keep the rejected part until the cause is closed.

Release only against evidence. A part that passes the tape test and the abrasion test, judged against the same criteria as the approved standard, can go back into flow; a part that "looks fine now" cannot. That single discipline is what stops a rework decision from becoming the next customer complaint.

What records prevent the same failure returning?

Recipe, qualification card, batch log and fault record.

Keep five records: the recipe with approved settings, the substrate qualification card, the consumable batch log, the maintenance log and the fault record for this incident.

Write the fault record while the parts are still on the bench: material and lot, ink batch, the tests performed, the observations, the variables changed in the diagnosis, and the outcome. Note the maintenance dates too - lamp and head changes in the preceding weeks explain a surprising number of faults that looked like a supply problem.

Then feed the conclusion back into the recipe or the qualification card, whichever owns the variable. A fault that ends as a note in a message thread will be diagnosed again next quarter; the same fault recorded against the material or the recipe becomes a condition the shop can recognise in minutes.