UV Ink Adhesion on Glass, Metal and Coated Board: Primers, Pre-Treatment and the Tape Test

UV ink adhesion on glass and metal substrates is the failure mode that costs the most and shows up the latest. The print looks perfect leaving the machine. It survives handling, packing and shipping. Then six weeks later a customer runs a fingernail along the edge of a printed bottle, or peels a label off a printed panel, and the ink lifts in a sheet.

This is not usually an ink problem or a machine problem. It is a surface problem, a cure problem, or a testing problem — and all three are inside your control if you know what to specify. What follows is what actually governs the bond, what to do to each substrate family before it goes on the bed, and the test protocol that catches a weak bond before you invoice the job rather than after.

Why UV ink sticks, and why it stops

A UV-curable ink is a liquid mix of acrylate monomers and oligomers plus a photoinitiator. Under UV energy the photoinitiator fragments, free radicals form, and the monomers cross-link into a solid polymer network in a fraction of a second. That is the whole appeal — instant cure, no drying time, the printed part is handling-ready as it leaves the machine.

It also creates the problem. Cross-linking pulls molecules closer together, so the film shrinks as it cures. That shrinkage is locked in as internal stress in a film that is now bonded to a rigid substrate. The bond has to hold that stress for the life of the part.

Adhesion itself comes from two mechanisms working together:

• Mechanical keying — the liquid ink flows into surface texture, then hardens around it. Sanded acrylic, etched glass and abraded metal all key well. Polished glass and smooth anodised aluminium key poorly.
• Chemical and polar interaction — the ink has to wet the surface first. Wetting is governed by surface energy: as a working rule, the substrate's surface energy needs to sit roughly 8–10 mN/m above the ink's surface tension before the ink spreads instead of beading.

Everything that follows is about raising surface energy, adding texture, or bridging the two materials chemically with a primer.

What each substrate family actually needs

Longrun lists UV-curable ink bonding to acrylic, glass, metal, wood, PVC, leather and board across its machines — a substrate range that is genuinely broad, and precisely why the pre-treatment question cannot be answered once for the whole list. Each family fails differently.

Three details in that table are worth pulling out, because they are the ones that catch experienced print buyers.

Corona treatment decays. Treating polyolefin parts raises surface energy, but the effect falls off over hours to days depending on the polymer, additive package and storage temperature. A part treated by your supplier three weeks before it reaches your press may have reverted most of the way. If you print polyolefins, dyne-test at the press, not at goods-in.

Anodising is not one thing. Sealed anodised aluminium — hydrated to close the pores — is a harder surface to bond to than the unsealed layer, because the sealing step removes exactly the porosity the ink would otherwise key into. If you are specifying anodised parts for printing, say so to the anodiser and discuss the seal.

Silicone is unrecoverable. Slip additives in board coatings, mould release on plastic parts, and silicone spray anywhere in the building will produce a low-energy surface that no cleaning routine fully rescues. On coated board this is a lot-to-lot variable, which is why board adhesion has to be retested per lot rather than approved once.

The cross-hatch tape test, done properly

The cross-hatch test is the standard method and it is also the most frequently mis-run test in the industry. It is worth doing correctly because a correctly-run test is a contractual number and a casually-run one is an opinion.

The two governing methods are ASTM D3359 (tape test, Method B is the cross-cut lattice) and ISO 2409 (cross-cut test). Both cut a lattice through the ink film to the substrate, apply tape over it, remove the tape, and grade what came off.

The detail that trips people up more than any other: the two scales run in opposite directions. ASTM D3359 grades from 5B (nothing removed) down to 0B (more than 65% removed). ISO 2409 grades from Class 0 (nothing removed) up to Class 5 (worst). So "5B" and "Class 0" both mean a perfect result, and a report that just says "class 5" is ambiguous unless it names the standard. Write the standard and the classification into your acceptance criteria together.

The other details that decide whether a result is repeatable:

• Cut spacing depends on film thickness. The methods specify spacing bands by coating thickness — thin films get the closest spacing. A digital ink film is thin; use the spacing the method assigns to it, not the widest blade you own.
• The tape is specified. The methods call for a defined pressure-sensitive tape with a stated adhesion strength. Office tape gives results that vary between rolls and between brands, which makes cross-supplier comparison meaningless.
• The dwell and pull are specified. Smooth the tape down, leave it the stated dwell, then pull it back rapidly over itself at close to 180°. A slow peel at 45° passes prints that a correct pull would fail.
• Cut through to the substrate. A cut that stops inside the ink film tests the ink against itself, not against the surface.
• The white underbase is a second interface. On a CMYK-over-white job there are two bonds — white to substrate, and colour to white. A cross-hatch failure that leaves white on the part and lifts the colour is an interlayer problem, not a substrate problem, and it usually points at cure or ink limit rather than pre-treatment. The layer-order and white-underbase article covers how that stack is built.

Run one more test alongside it. A solvent double-rub check — rubbing a solvent-soaked cloth back and forth over the print under light pressure and counting cycles to breakthrough — tells you about cure state rather than adhesion. Under-cured ink fails it early. Since under-cure is the single most common root cause behind adhesion complaints, running the two tests together tells you which lever to pull.

Why a sample that passes on day one fails at 30 days

This is the pattern that ruins jobs, and there are five separate mechanisms behind it. Any of them can produce a print that grades 5B on the day it is made and 1B a month later.

1. Post-cure. Free-radical polymerisation does not stop when the part leaves the lamp. Cross-linking continues in the dark for hours to days, and the film keeps shrinking. Stress rises after the test you already passed. Testing a print an hour after it comes off the machine measures the film at its most compliant.
2. Water at the interface. On glass and other high-energy inorganic surfaces, an apparently strong bond can be attacked by moisture that migrates to the interface and hydrolyses it. Ambient humidity is enough given time. This is exactly what a silane-type promoter is for — it forms a covalent bridge that water does not displace as readily.
3. Plasticiser migration. Flexible PVC and coated leather contain plasticisers that migrate into the cured ink film over weeks, softening it and undermining cohesion. A rigid ink on a flexible substrate can look perfect at 24 hours and crack or lift at a month.
4. Thermal cycling. Metal and glass expand and contract at rates different from the cured acrylate film. A part that lives in a warehouse, a vehicle or a window sees daily cycles. Each one flexes the bond.
5. Contamination that arrives later. Anti-fingerprint coatings, packaging plasticisers, and handling oils all reach the part after printing. A part printed straight off the delivery pallet and a part printed after three weeks in your rack are not the same substrate.

The practical response is a conditioned test protocol rather than a single test. Print your sample set, then:

• Hold 24–72 hours before the first cross-hatch. Let post-cure finish. Record the hold time on the report.
• Immerse a duplicate in water for 24 hours, dry it, and cross-hatch it. This is the fastest proxy for the slow moisture attack that shows up at week four.
• Thermal-cycle a third duplicate through the range the part will actually see, then cross-hatch.
• Retain a fourth sample untouched and retest it at 30 days. Diary it. This is the sample that would have told you.
• Repeat on a second material lot. Board coatings, anodising batches and recycled liner vary; a single-lot approval is a single-lot approval.

Turning this into something you can put in a contract

"Good adhesion" is not a specification. A test result is. Write acceptance criteria that name the method, the classification, the conditioning and the substrate lot — for example: cross-hatch to ASTM D3359 Method B, classification 4B or better, tested after a 48-hour hold and again after 24-hour water immersion, on buyer-supplied substrate from the lot identified on the packing list. That sentence is enforceable. The alternative is not.

Then use the sample stage properly. Longrun's stated build sequence includes a test-print step where the buyer's own files and substrates are run before shipment, and standard models can be sampled the same day, with print samples and technical Q&A available through ChinaMakersHub before purchase. That is the window in which adhesion gets proved, and payment terms for this class of equipment commonly run T/T with a deposit and the balance before shipment — which means your negotiating position on adhesion exists exactly once. Guidance on structuring that stage generally is in the sample order process guide and the pre-shipment inspection guide.

What to ask the supplier next

Take these to the builder in writing. Several of them are not answerable from any public specification sheet, which is the point of asking.

1. Which primer, if any, do you recommend for each of my substrates — and is it supplied with the machine, sourced locally, or my problem entirely?
2. What ink chemistry families can this machine run — is there a flexible-grade option for leather and soft PVC as well as a rigid grade, and can both be run on one machine or does it require a flush?
3. What cure settings are exposed to the operator? Can lamp power and pass count be tuned per substrate and saved as a preset, or is cure fixed?
4. What curing technology is fitted, at what output rating? Lamp type is not something to assume from a category description.
5. Do you run any adhesion test at the test-print stage, and will you report the method and result with my samples rather than just sending prints?
6. Send me two lots. Ship your own material — two different production lots of the same substrate — and require both printed and returned, so lot variation shows up before purchase rather than after.
7. What data exists on lightfastness, abrasion or chemical resistance for printed output in my application? If none, say none; a stated gap is more useful than an assurance.
8. How do I re-establish a known-good state? After a head change or an ink change, what has to be re-run before adhesion results are comparable to the ones I approved?

Longrun Printing Machinery builds UV flatbed, corrugated-carton and cylindrical machines in Shenzhen and lists that substrate range on its profile — the Longrun Printing Machinery factory page is the place to send the substrate question, and if the parts you print are thick, raised or three-dimensional, the companion piece on high-clearance UV printing covers the geometry side of the same problem. For the enquiry itself, the product specification sheet guide shows how to phrase requirements so three quotes come back comparable.