Lower-energy curing

A coating consumes energy twice, and the two amounts are not comparable. The first is the energy of making it, which is spent once, in the plant that manufactures it. The second is the energy of drying or curing it, which is spent again for every square metre printed or coated, in the lamps, ovens and dryers of the converter, the printer or the coater, for as long as the line runs. Where a coating is applied over large areas on a line that runs continuously, the second figure is usually the larger of the two, and it is the customer who pays it. That puts the useful reductions in the formulation: the cure window designed into a product is the energy bill the press or the coating line meets. We treat it as a specification property, alongside adhesion, rub and gloss, rather than as a virtue to be claimed afterwards.

LED-UV against mercury-arc

A medium-pressure mercury-arc lamp radiates across a broad spectrum, from short-wave UVC through UVB and UVA into the visible, with strong lines in the near-UV and visible near 365, 405 and 436 nm, and a large infrared component alongside. A UV-LED array radiates one narrow band, most commonly centred near 395 nm, with 385 nm and 405 nm also used. What follows is a consequence of that difference in spectrum. The arc lamp’s short-wavelength output is absorbed in the first few micrometres of the film and is what sets the surface, while its infrared output reaches the substrate as heat. An LED array emits almost no infrared, and its own waste heat leaves through the cooling circuit rather than through the web, so the heat load on the substrate is much lower. That is the point for thin heat-sensitive film, for shrink sleeve and for anything that distorts, tunnels or loses register when it is warmed. Narrow-web UV and LED-UV work sits with our Arcray range, and the sleeve and label side with Arcweb.

• No warm-up period and no idle burn: the array emits only while the web is running, although the drivers and the cooling circuit still draw power. Arc systems are often left burning through make-ready and shuttered.
• Instant on and off, which removes the shutter mechanism and the losses around it.
• Substantially less infrared at the substrate, and therefore a cooler web and less chill-roll duty.
• No ozone formed at the lamp, because the short wavelengths that form it are absent. The ozone extraction duty goes with them; extraction for heat and for any volatiles from the job does not.
• Longer service life between replacements, and no mercury lamp to handle as waste at end of life.
• A narrow emission band, which is the source of both the efficiency and the limitations below.

The narrow band forces a different photoinitiator package. Initiators chosen for an arc lamp are often selected for absorption in the UVB and UVC, and much of that absorption sits where an LED array emits nothing at all. LED systems therefore rely on initiators that absorb in the near-UV and at the edge of the visible. Acylphosphine oxides are the usual backbone: they cleave directly on absorption and need no synergist. Thioxanthone types are often blended in to pick up the longer wavelengths, and those do need an amine synergist, which also helps at the surface. The surface is where LED is short of energy, because there is no strongly absorbed short-wavelength output to raise the radical flux that overwhelms oxygen inhibition, so amine synergist, higher initiator loading, higher irradiance or inerting are used in its place. None of this is a drop-in substitution. A formulation moved from arc to LED is a reformulation and should be requalified rather than assumed: solvent rub resistance by ASTM D5402, hardness by ISO 1522 or ASTM D3363, on the customer’s own substrate. Those are methods, not pass marks, so the number of rubs or the pendulum count that will be treated as cured has to be agreed before the trial.

The limitation is worth stating plainly, because it decides jobs. The two ends of the film are constrained differently. At the surface, LED has less to work with against oxygen inhibition, as above. Through the film, the single band has to reach the bottom of the layer while the pigment competes for the same light: titanium dioxide scatters and absorbs hard just below 400 nm, carbon black absorbs everything, and an arc lamp brings both a wider set of wavelengths and, on many installations, more total power to the same problem. Opaque whites, dense blacks and deep blues, high-build tactile and matt layers, and thick decorative effects are therefore the jobs to check first. The practical answers are higher irradiance, dual-wavelength heads, splitting a heavy layer into two lighter ones, reworking the initiator package, or leaving one problem station on an arc lamp while the rest of the deck runs LED. Anyone who tells a press crew that LED cures everything an arc cures, at the same film weight and the same speed, will be found out on the opaque white. High-build work of the kind in our Arctouch and Arcvista families is exactly where this has to be measured rather than assumed.

Low-temperature and ultra-low-cure powder

Powder coating is an adjacent field, and the same argument appears there in its clearest form. Conventional thermosetting powder is typically cured in the region of 180 to 200 degrees Celsius held for ten to twenty minutes, figures quoted here as the industry’s usual window rather than as anyone’s measurement. The schedule is set by the temperature of the part, not of the oven air, so the ramp to temperature has to be added to the dwell. Two things follow. The oven has to bring a large mass to that temperature and hold it there, and the substrate has to survive it, which is why powder has historically been a metal technology. Low-temperature systems bring the schedule down towards 140 to 160 degrees. UV-cured powders lower it again, to a region around 110 to 140 degrees where the powder is melted and flowed thermally, usually by infrared, and then crosslinked in a UV step of seconds rather than minutes. What that unlocks matters more than the energy saved in the oven. Medium-density fibreboard becomes a candidate, though not an easy one: the board has to be conditioned and preheated so that it carries enough surface conductivity for electrostatic application, and its moisture and outgassing are managed rather than eliminated, since gas escaping through a melting film is the usual source of pinholes. Assembled parts become candidates, so items carrying seals, bearings, plastics or electronics need not be coated as separate components, or masked while they are coated. Thinner sections distort less. The oven saving is real but secondary; the gain is a part that could not previously be powder coated at all.

Water-based systems, and why removing water is not free

Water-based chemistry is often described as though the energy question answers itself once the solvent is gone. It does not. At its boiling point the latent heat of vaporisation of water is about 2,260 kJ/kg, against about 370 kJ/kg for ethyl acetate, so removing a kilogram of water takes roughly six times the energy of removing a kilogram of that solvent. The comparison has to be made solvent by solvent rather than in the round. Esters and aromatics sit in the same region as ethyl acetate, but the alcohols used widely in flexo and gravure, ethanol and n-propanol, are nearer 800 to 850 kJ/kg, so against those the gap is closer to three times. Water also has a high specific heat, evaporates more slowly as ambient humidity rises, and holds a surface tension near 72 mN/m that affects wetting and levelling on the way. The consequence on a real line is that the dryer, not the print unit, sets the speed.

Latent heat is not the whole of the plant’s bill either. A solvent dryer has to heat and exhaust a large volume of air to stay well below the lower explosive limit, and most of that heat goes into the air rather than into the solvent, which is why a solvent line’s energy position is worse than the latent heat figures alone suggest. The gains from water-based systems are real and they lie elsewhere than in evaporation: no solvent recovery or thermal oxidation plant to run, a much lower VOC load, though not a zero one, since co-solvents and neutralising amines remain, and a much smaller inventory of flammable liquid in the building. The energy case has to be made with solids content, applied film weight and available dryer length in the argument, because a higher-solids water-based ink that puts less water on the substrate is often most of the answer. This is the balance we work on in the Arcoflex and Arcfibre families, where board and corrugated absorb part of the water and change the calculation in both directions, since the absorption that helps the dryer is also what causes fluting and warp.

Cure on demand

The remaining direction is to stop supplying energy continuously and supply it only when cure is wanted. That covers latent catalysts and blocked crosslinkers that stay dormant until a defined trigger releases them, photo-latent systems where light generates the catalyst rather than the radical, and dual-cure designs that combine a radiation step with a slower moisture or thermal mechanism so that shadowed areas, edges and the underside of a formed part still cure fully. The trigger is not automatically cheap. Many blocked crosslinkers deblock thermally, and where the deblocking temperature is high the energy has been moved rather than removed, so the deblocking temperature, not the principle, is the question to ask. Where the trigger is light or moisture the attraction is clearer: a long, stable working life in the pot or on the press, then a short event that fixes the film, with no need to hold a whole oven at temperature to reach the small part that needs it. Each of these systems trades something, usually pot life against cure temperature, or through-cure against speed. This work is unfinished. Our position across Arclear, Arcray and the water-based families is that a lower cure schedule counts for nothing if adhesion, block resistance or through-cure is quietly given up to reach it, so every low-energy proposal is qualified against the same property set as the system it replaces, on the customer’s substrate and at the customer’s line speed, before it is offered.