UV LED vs mercury lamp curing: the honest comparison

Almost everything written about UV LED versus mercury curing is written by someone selling one of them. This is the comparison with the trade-offs left in, including the places where LED is genuinely worse and where the energy-saving claims fall apart.

What curing actually is

UV ink does not dry. Nothing evaporates and nothing soaks in. Light triggers a chemical reaction — free radical photopolymerisation — that turns liquid monomer into solid polymer, in place, in a fraction of a second. A photoinitiator absorbs a photon, fragments into radicals, those radicals attack monomer double bonds, and a chain reaction propagates until it terminates.

Two consequences follow immediately, and they explain most of what goes wrong. The reaction stops when the light stops — so an under-cured print does not improve overnight the way a solvent print does. And the whole applied film stays on the surface, because none of it evaporated.

Oxygen inhibition: why the surface is always the weak point

Radicals react with atmospheric oxygen — R• + O₂ → ROO• — and the peroxy radical formed is far less reactive. RadTech literature describes oxygen inhibition as “a major drawback to photopolymerizations conducted in air.”

It bites hardest exactly where you notice it: “Oxygen is particularly detrimental to chain propagation at the air-film interface since oxygen can continually diffuse into the interface.” The bulk cures. The top few microns fight for it. That is why an under-cured print feels tacky on top while being solid underneath.

The measured effect is dramatic. Same ink, same lamp, different atmosphere:

Doubling the achievable speed by removing oxygen tells you how much of the cure budget is being spent fighting the air.

The wavelength argument — and the real reason LED prints scratch

This is the section worth reading twice, because it is the honest technical case against LED and nobody selling LED makes it.

Mercury lamps emit across a broad spectrum, with a peak around 365 nm and a significant concentration at 254 nm — in the UVC band. UV LEDs emit a narrow band at one of four peaks: 365, 385, 395 or 405 nm, typically ±5 to ±15 nm. A 395 nm LED emits no UVC at all.

Now chain the facts together, all from GEW:

• “shorter UVC wavelengths are absorbed at the surface of the chemistry”
• “longer UVA and UVV wavelengths penetrate deep”
• the surface is exactly where oxygen inhibition is worst

So mercury dumps UVC into the oxygen-poisoned surface layer, which is precisely what that layer needs. An LED at 395 nm cures the bulk beautifully and can leave the skin tacky. GEW says it outright: “UV LEDs can struggle with surface cure when formulations are not optimized for emitting sources that do not emit UVC.”

The fix is formulation, not more power — which is why LED-cure inks are chemically different, not just relabelled.

Irradiance and dose: the two numbers dealers conflate

Irradiance is power arriving at the surface, W/cm². It does not depend on how fast the job moves. Energy density, or dose, is J/cm² — irradiance integrated over time. Two watts per square centimetre for two seconds is four joules per square centimetre.

They are not interchangeable, and the standard analogy is the best one: “If the cake is baked at twice the temperature for half the time, the cake burns. Conversely, if the cake is baked at half the temperature for double the time, it may not be baked in the center.”

And a warning from Phoseon, who sell LED systems and therefore have no reason to say it: “Research reveals that in many cases, excessive irradiance can negatively affect or hamper proper curing” and “beware of marketing claims, such as ‘highest irradiance’.” More is not better.

Lamp life, and the L-number nobody explains

Mercury lamps last on the order of 1,000 hours (GEW), or “in excess of 1,500 hours, when operated in suitable machines under optimum conditions” (Heraeus).

LEDs are quoted very differently. Phoseon specifies “an L80 lifetime (80% of original output) of at least 40,000 hours”, and notes that “many commercial LED lighting systems define failure at 70%… L70”.

This is the buying-guide point: “40,000 hours at L80” and an unlabelled “20,000 hours” are not comparable numbers. A supplier quoting hours without stating L70 or L80 has told you nothing measurable. Ask which.

GEW also warns that not every LED system reaches those figures: “Not all UV LED system suppliers currently offer designs that meet the highest established lifetimes in excess of 20,000 hours” — inferior systems suffer “catastrophic failure.”

And here is why that matters in India specifically: rated LED life is a thermal-design promise, not a property of the diode. Phoseon: “Two major factors that affect the lifetime of LEDs are temperature and current.” GEW rates one of its LED systems for a maximum ambient of 40 °C. Run an LED array in an un-airconditioned shed through a Gujarat summer and the 40,000-hour figure is describing a different machine than yours. Room conditions are covered in setting up the room.

Heat, ozone and thin material

A mercury lamp emits UV, visible light and infrared in roughly equal proportion. Most of what reaches your substrate is heat, which is why thin PVC and some films distort under mercury and are comfortable under LED. That is a genuine LED advantage and the main reason it took over on flexible media.

Ozone forms when atmospheric oxygen meets wavelengths below 240 nm. Mercury lamps emit there; 385 or 395 nm LEDs do not. So LED installations do not need ozone extraction, which removes ducting from the room.

But the counterweight, again from GEW, who sell both: LEDs are “not… cold curing technology”, and “Only 35–50% of the electricity powering a UV LED array is converted to ultraviolet output” — the remainder is heat that has to be removed, usually by water or forced air. LED moves the heat problem; it does not delete it.

Inks are not interchangeable, and the incompatibility runs one way

GEW: “Existing inks, coatings, and adhesives formulated for medium-pressure mercury lamps must be reformulated for UV LED curing systems.”

But the reverse is often fine. Marabu publishes an LED ink that cures under both LED at 385–395 nm and a conventional mercury lamp. So: LED-formulated ink frequently runs under mercury; mercury ink does not run under LED. If you are running a mixed fleet, that asymmetry decides which ink you standardise on.

Is mercury being banned?

You will read constantly that it is. It is not.

We read the instruments directly. Industrial UV curing lamps are not subject to any Minamata Convention or EU mercury phase-out — every lamp entry in both is qualified “for general lighting purposes.” Regulation (EU) 2024/1849, read on EUR-Lex, carries that qualifier on all its Annex II lamp entries and makes no mention of UV curing.

GEW, writing in September 2025: “While Minamata does not presently require a ban of mercury vapour UV curing lamps,” it does require parties to phase out or reduce mercury where possible.

The live regulatory date is RoHS, not Minamata — the relevant exemption runs into early 2027 with reassessment during 2026. For context, a curing lamp contains 10–100 mg of mercury, averaging about 25 mg.

So choose on cure quality, running cost and the material you print — not on a ban that has not been legislated.

What we will not claim

That UV LED saves 80–90% of the energy. Every such figure we found is a vendor claim with no stated test method, duty cycle or dose. GEW’s own number is “over 65%”, equally unsubstantiated.

The defensible argument is structural rather than numerical. A mercury lamp cannot be switched on and off quickly, so it burns full power behind a shutter whenever the machine pauses. LEDs “don’t use energy unless they are operating.” For a stop-start job shop — which is most Indian shops — idle burn probably dominates the bill. That is a real argument. It just is not 90%.

We also publish no typical mJ/cm² dose for UV inkjet. The only sourced dose figures are for coatings under mercury lamps, which is a different application.

Which to buy

Nearly everything in this range is UV LED, and for a job shop that is the right default: no warm-up, no ozone extraction, no shutter, far less heat into thin material, and idle time that costs nothing. The trade-off you accept is a narrower spectrum that demands ink formulated for it, and a cure that is more sensitive to surface conditions.

What to check on a specific machine is in what to inspect at a demo — particularly the point that a machine can move faster than it can cure. Machine-by-machine detail is on the UV flatbed and roll-to-roll pages, and the ink families are compared in the UV, solvent or eco-solvent guide.

Sources

• GEW — spectral bands, UVC surface absorption, LED surface cure, lamp life, LED electrical efficiency, ink reformulation, mercury regulation update (September 2025)
• Phoseon — LED peak wavelengths and tolerances, L80/L70 lifetime definitions, temperature and current as lifetime factors, irradiance warning
• RadTech proceedings — free radical photopolymerisation, oxygen inhibition at the air-film interface, measured cure speed against oxygen concentration
• Heraeus — mercury lamp service life
• UV+EB Technology — irradiance versus energy density
• EUR-Lex, Regulation (EU) 2024/1849 — lamp entries qualified “for general lighting purposes”
• Marabu technical data — LED ink cured under both source types