Roll Applied Coil and Can Enamel Flowout in Three Second Oven Dwell

Roll application of coil and can enamels transfers a thin, shear-thinned liquid film from rotating applicator rolls to a continuously moving substrate, and the imposed shear history determines the initial surface topography. On a typical three-roll coater, the pickup roll transfers coating from the pan to a metering nip, the metering roll controls wet film thickness through gap and speed ratio, and the applicator roll transfers the metered film to the strip or sheet. The applicator nip generates shear rates often above 10,000 s⁻¹, reducing viscosity under high shear; after the film leaves the nip, shear stress falls to near zero and the film attempts to level. However, the film immediately begins to lose solvent at ambient temperature, and in a high-speed coil line the coated strip enters a forced-convection oven within 0.2–0.8 seconds. In a 3-second oven, the wet film has only a narrow interval at low enough viscosity for capillary leveling before evaporation and early crosslinking raise viscosity beyond the point where surface-tension-driven flow can remove roping, ribbing, and roll-coater pattern. The required dry film thickness for coil enamels is commonly 15–25 µm, while can enamels applied to tinplate or ECCS are often 4–12 µm, so the solvent load and thermal mass differ substantially. The flowout problem is therefore not generic mixing or oven residence, but the interaction of roll-induced ribbing wavelength, solvent evaporation rate, viscoelastic recovery, substrate heat capacity, and peak metal temperature.

Production-scale equipment behaviour makes this worse. Ribbing wavelength from roll application is typically tied to roll diameter and gap, with amplitudes that can remain visible if the film does not reach a low-shear leveling condition. When coil lines run above 120 m/min, the oven residence time of 3 seconds corresponds to an oven length of only 6 m at 120 m/min, and many convection oven sections are physically longer only because line speeds are limited. If line speed increases to 150 m/min, the same oven yields 2.4 seconds dwell, requiring changes in zone air temperatures or coating solvent blend. The film surface may reach the evaporation plateau early while the substrate side remains below the true wet-bulb temperature, creating a vertical viscosity gradient. Flowout at the surface can be prematurely arrested by skinning, while solvent retained near the substrate later volatilizes and produces solvent popping or pinholes. This is the central process conflict: any solvent adjustment that improves leveling generally increases retained solvent near the metal interface, and any cure acceleration that reduces popping generally shortens the leveling window. Published data for this specific configuration is limited to supplier technical bulletins and line trial reports because the exact balance depends on coater geometry, oven nozzle design, and solvent package.

What limits flowout in a 3-second coil oven dwell?

The first limitation is capillary leveling kinetics. The rate at which a sinusoidal surface perturbation decays is proportional to surface tension and amplitude divided by wavelength and viscosity, and the driving force collapses as amplitude decreases. In practice, a ribbed wet film with surface tension measured by ASTM D1331 at 28 mN/m and low-shear viscosity measured by ISO 2884-1 at 100 mPa·s may level within a few seconds only if the rib wavelength remains below roughly 200–400 µm. Roll coaters often produce longer wavelength ribbing if gaps are too large or speed ratios are mismatched, and those longer waves require disproportionately longer leveling times. High-shear viscosity measured by ASTM D4287 at 10,000 s⁻¹ must remain low enough for clean transfer from the applicator roll, but a formulation optimized only for transfer can produce low-shear viscosity too high for flowout. The ratio between high-shear and low-shear viscosities is therefore a critical predictive parameter.

The second limitation is solvent selection and the distillation curve. In a 3-second dwell the solvent blend must flash quickly enough to avoid retained solvent defects such as popping, but not so quickly that the surface becomes immobile before ribbing has decayed. Formulators use solvent blends with a progressive evaporation range measured by ASTM D3539, where low-boiling components provide initial flash and a slower mid-boiling component maintains mobility. If the mid-boiling solvent is too slow, it remains in the film as the crosslinker begins to react, and solvent popping occurs when the film skins over. If the mid-boiling solvent is too fast, the surface skins within the first 1.0–1.5 seconds, freezing the roll pattern. The optimum is often a narrow ester or glycol ether plateau, and this optimum moves with line speed because the heat transfer rate changes the evaporation profile. Solvent retention is determined by ASTM D2369 volatile content and by oven temperature profile logging.

The third limitation is early crosslinking. Polyester-melamine coil enamels and epoxy-acrylate or epoxy-phenolic can enamels are designed to react rapidly at peak metal temperature, but the catalyst or crosslinker must not advance during the warm-up phase. Blocked acid catalysts such as dinonylnaphthalene disulfonic acid and amine-blocked sulfonic acids release at specific temperatures, and the release temperature is adjusted to overlap the end of the leveling window. If the catalyst releases too early, viscosity rises and ribbing is frozen. If it releases too late, the film remains soft at the quench, causing blocking, tackiness, or poor MEK resistance measured by ASTM D5402. In a 3-second oven, the catalyst release must occur within the final 1.0–1.5 seconds of dwell. This creates a processing window that can be as narrow as ±5°C in peak metal temperature for high-reactivity coil enamels.

In an oven profiled with a trailing thermocouple, the substrate temperature response in a 3-second dwell typically shows an initial lag, a rapid ramp, and a short plateau. For a 0.25 mm aluminium strip, the peak metal temperature can be driven from ambient to 200–220°C in 2–3 seconds because of the low thermal mass, while a 0.6 mm steel strip at the same line speed may not reach the same peak metal temperature unless the air temperature is increased or the dwell extended. This gauge sensitivity is one of the least understood production bottlenecks: a coil line running a fixed oven recipe cannot maintain identical flowout and cure when the strip gauge changes, and the line must either adjust speed or zone setpoints. In can coating, sheet-fed roll applicators and wicket ovens have a similar gauge effect with tinplate and black plate; thinner sheet reaches temperature faster, and the edges can overshoot by 8–15°C because of higher surface-to-mass ratio.

Roll-applied can enamels used on food and beverage containers have dry film weights of 4–12 g/m² and are frequently cured at peak metal temperature values near 190–205°C for internal epoxy-phenolic systems, while exterior polyester-melamine enamels may require 210–240°C. The 3-second dwell in a wicket oven is shorter than many coil oven configurations, but the oven air temperature is often higher because sheet pass lines and air impingement permit faster heat transfer. Flowout in these systems is complicated by the very low dry film thickness: the wet film may be only 15–35 µm depending on solids, leaving a smaller reservoir for capillary leveling and less tolerance for surface tension gradients. At such thicknesses, substrate roughness and roll-coater skips become visible as pinholes or telegraphing if the film does not flow out before viscosity increases.

Waterborne can enamels and coil primers introduce an additional constraint: at relative humidity above 60%, the evaporation rate of water drops, and the film remains sufficiently viscous that flowout may be inadequate unless pre-drying is applied. In such conditions, the use of a low-solids water-reducible enamel on a high-speed coil line requires careful zone humidity control and an extended predry at ambient or low heat before the main oven. If not controlled, the film can maintain a surface skin and still contain water at the substrate interface, causing microlamella popping.

Rheological and surface tension thresholds in polyester-melamine and PVC organosol enamels

In polyester-melamine coil enamels, the low-shear viscosity is usually set by resin molecular weight, solvent blend, and flow-control additives, while the crosslinker content and catalyst govern the cure window. Hexamethoxymethylmelamine levels of 5–10 wt% on total binder are common, with hydroxyl-functional saturated polyesters having OH values in the range 10–30 mg KOH/g and acid values below 3 mg KOH/g. The melamine self-condensation reaction is strongly acid-catalyzed, and catalyst choice determines whether viscosity builds early in the oven. Sulfonic acid catalysts blocked with amines or epoxides release at a temperature specific to the blocking chemistry, and the release temperature is selected so that the leveling window closes only after capillary flow has removed ribbing. Dynamic mechanical analysis after cure shows that an over-accelerated system can reach high crosslink density and MEK resistance yet retain visible roll pattern because the film gelled too early. Viscosity during the first 1–2 seconds is therefore monitored by rotational methods at low shear, and the temperature-viscosity profile is compared to the measured solvent loss curve to establish the flowout window.

PVC organosol enamels for can ends and closures behave differently. The dispersed PVC particles do not dissolve until the plasticizer and solvent heat enables gelation, so the initial viscosity is set by the continuous phase and the plasticizer solvency. In the oven, the continuous phase viscosity initially falls as temperature rises, improving flowout, then rises sharply during PVC gelation and fusion. The gelation temperature depends on plasticizer type and PVC particle size, but common dioctyl phthalate or citrated plasticized systems gel in the range 150–180°C, leaving only 1–2 seconds of useful flowout before fusion. The addition of acrylic flow modifiers lowers interparticle friction and reduces low-shear viscosity, but excessive levels may reduce film integrity, food-contact resistance, or overend seam adhesion. The flowout/cure balance is evaluated by ISO 1519 cylindrical bend testing after cure and by ASTM D4145 T-bend testing for coil stock.

Surface tension gradients dominate when flow modifiers migrate or when solvent blends evolve non-uniformly. Acrylic flow modifiers reduce surface tension and accumulate at the air interface, helping to suppress craters and improve leveling, but if their concentration is too high the film can develop surface defects such as crawling or edge retraction. Fluorosurfactants are used at very low levels to reduce surface tension to 22–26 mN/m, but they can destabilize intercoat adhesion if they migrate to the surface and remain unreacted. Surface tension measured by ASTM D1331 before and after baking is therefore paired with intercoat adhesion testing by ISO 2409 cross-cut or ASTM D3359 tape adhesion. Formulators avoid amine-neutralized additives in acid-catalyzed polyester-melamine systems because amine species can block the sulfonic acid catalyst and shift the apparent cure window, causing both undercure and poor MEK resistance. This is a strict incompatibility: systems requiring amine-stabilized waterborne resins must be reformulated with non-amine defoamers and wetting agents.

Surface-tension-driven flow combines with the roll pattern in ways that routine viscometry does not predict. When a solvent blend evaporates non-uniformly, the cooler regions of the film have lower surface tension and pull coating from hotter regions, producing Marangoni flow that can either help level or create orange peel and convection cells. In a 3-second oven the temperature gradient across the width of the strip can be 3–8°C, and the resulting surface-tension gradient is often strongest near the edges, causing edge retraction or fat edges. In roll-applied can enamels, the continuous sheet edge acts as a thermal fin, rising faster and evaporating solvent earlier, so edge flow defects can appear even when the centre of the sheet is smooth. Production lines often use edge masking or edge cooling to manage this, but these measures alter the film thickness at the edges and can create a visible boundary.

High-velocity air impingement nozzles also create local surface disturbances. In the first oven zone, air velocities of 20–35 m/s over the wet film can cause standing waves, wicking, or mottling if the film has not yet developed a stable surface. The film thickness and viscosity at the moment of nozzle impingement determine whether the film will settle back into a smooth layer or freeze the impingement pattern. For this reason, some coil lines use a gentle evaporation section for the first 0.5–1.0 second before entering the high-velocity cure zone. In sheet-fed can lines, the wicket oven air distribution is similarly staged, with lower nozzle pressure at the entrance where the wet film is still mobile. Operators verify the transition using high-speed cameras and compare the visual flowout to gloss measured by ISO 2813 or DOI instruments.

When peak metal temperature overshoots 254°C in a can enamel oven, flowout and fabrication tolerance diverge

For many polyester-melamine coil enamels, peak metal temperature above 250–254°C creates a property cliff edge where crosslink density increases rapidly, but flexibility and impact resistance fall at the same time. The film may pass solvent rub testing because the surface is heavily cured, yet fail ASTM D4145 T-bend at 0T or 1T when the coated panel is fabricated. This divergence is a process conflict: the oven operator may increase temperature to eliminate solvent popping or to compensate for a faster line, but the resulting film becomes hard and brittle, and edge cracking or adhesion loss during roll forming, drawing, or can seaming increases. In can enamels, overbaked epoxy-phenolic films can lose the flexibility required for double seaming and beading, producing fractures at the flange or sidewall beading. The correct response is not always lower oven temperature; sometimes the wet film was applied too thick, and the extra solvent load required more energy to evacuate. The physical measurement that resolves this is a combination of solvent resistance and bend/flex testing after the oven, not a single cure test.

Peak metal temperature overshoot is often non-uniform. On strip lines, the edges of a wide coil can run 8–15°C higher than the centre because of higher heat transfer at the strip edges, and the edge region may show yellowing or embrittlement while the centre remains smooth. On sheet-fed can lines, the corners and edges of rectangular sheets overshoot similarly. The use of infrared pyrometers scanning across the strip can quantify this distribution, but emissivity differences caused by unpainted metal, wet film, and cured enamel make single-point pyrometry unreliable. A trailing thermocouple attached to the sheet or strip under actual line speed remains the most representative measurement because it records the true thermal history. Peak metal temperature tolerance for high-reactivity exterior coil enamels is often quoted as ±5°C, and within that band the flowout/cure balance is maintained. Outside the band, the process fails first at the property cliff edges: underbake causes solvent retention and cohesion failure, while overbake causes flowout to be frozen by premature crosslinking and later embrittlement.

Compliance with food-contact requirements under FDA 21 CFR 175.300 or European frameworks adds a further boundary: post-cure film extractives and migration behaviour can shift with overbake because degradation products or unreacted intermediates change. Overbaked enamel may still meet adhesion but not migration limits, so the oven operator cannot use temperature as an unbounded variable. Similarly, REACH and RoHS restrictions on certain catalyst, plasticizer, and solvent residues constrain the reformulation options available after a peak metal temperature problem is identified.

Troubleshooting solvent popping, ribbing, and pinholing without sacrificing flowout

When a production line exhibits both ribbing and solvent popping, the conventional first response is to raise oven temperature, but this often freezes the rib pattern before leveling and increases the skin formation rate. A systematic approach uses the measured temperature-viscosity profile and the solvent distillation curve to identify whether the film is still mobile when the first high-velocity air impingement occurs. If ribbing is the dominant defect, the coater geometry is examined first: applicator roll speed ratio, metering gap, roll hardness, and roll runout. A speed ratio that is too high or a metering gap that is too small can produce ribbing that no oven can level. If the coater is within specification, the coating is adjusted by reducing low-shear viscosity without changing solids; this is typically done with a small addition of an acrylic flow modifier that does not destabilize the crosslinker. If pinholing is the dominant defect, the solvent retention and film weight are measured because pinholes in roll-applied thin films often originate from microfoam or from retained solvent erupting through a semi-solidified surface.

On multi-roll coaters used for can enamels, shear-induced microfoam occurs when the applicator roll speed exceeds the strip speed by more than the recommended ratio, or when the coating pan level drops and air is drawn into the nip. The resulting microfoam lamellae can survive into the oven, leaving pinholes when the film skins over within 1.5 seconds. The corrective action is not necessarily a defoamer increase; defoamer overload can create craters and intercoat adhesion loss. The preferred adjustment is to reduce roll speed differential, verify roll eccentricity, and ensure that the coating pan level and viscosity are within the ranges specified by the coater manufacturer. If pinholes remain, a vacuum deaeration step or a controlled predry can be applied before the main oven.

Solvent popping is linked to the distillation curve. If the solvent blend contains too much high-boiling component, it will remain in the film when the surface skins, and the vapour pressure at peak metal temperature will exceed the surface layer strength. The formulation response is to reduce the high-boiling fraction or to add a slightly faster mid-boiling solvent so that the film is more depleted before gelation. However, this can reduce flowout because the mobility window shortens. The production response is to reduce wet film thickness or to add a preheat zone that allows evaporation at lower film surface temperature. On thick gauge steel, the metal acts as a heat sink and the film surface can skin before the interface is warm, so raising air temperature makes popping worse while the centre remains underbaked. The correct response is a longer preheat or lower line speed, not higher oven air temperature.

Finally, when a new coil enamel or can enamel is qualified for a 3-second line, the acceptance protocol should include duplicate panels at the specified peak metal temperature, at −5°C, and at +5°C, tested for MEK resistance, pencil hardness, T-bend or cylindrical bend, cross-cut adhesion, and gloss or DOI. The entire set is used to define the operational peak metal temperature window, and any proposed line speed change is evaluated against the resulting dwell time change. If a line speed increase reduces dwell from 3.0 seconds to 2.4 seconds, the same peak metal temperature may still be reached by raising air temperature, but the leveling window is disproportionately shortened because the early evaporation phase is compressed. This is the most common cause of chronic ribbing in lines that otherwise meet cure specifications.