Cleaning and Coating Polycarbonate Without Killing the UV Layer

Polycarbonate (PC) sheet, lenses, and molded housings intended for outdoor glazing, automotive lighting, and medical windows typically carry a coextruded or coated UV-protective layer containing a benzotriazole, hydroxyphenyl triazine, or benzophenone absorber. The cap layer is not a cosmetic coating. It blocks UV-B and UV-A photons before they reach the bulk PC, preventing photo-Fries rearrangement, chain scission, yellowing, and impact loss. The physical and chemical threat during cleaning and coating is therefore not the PC alone: the process must preserve a UV cap that may be only 10–50 µm thick while removing mold release, fingerprint oils, cutting fluids, and particulate contamination. Production failures typically appear not as immediate fracture but as haze development, yellowing, and coating delamination after 500–1000 h of accelerated weathering.

Because cleaning precedes coating on a UV-capped PC surface, the first boundary is chemical. A neutral aqueous detergent at pH 6–8 and 40–60 °C in a flat-bed spray washer or ultrasonic tank removes water-soluble soils without attacking the carbonate linkages of PC. Alkaline cleaners above pH 10 are rejected because hydroxide ions hydrolyze the surface carbonate linkages and etch the UV cap. Acidic cleaners below pH 4 are not used for the same reason. Isopropanol/water at 30–50 vol% is acceptable for short-contact wiping, but pooled solvent or immersion under load can initiate environmental stress cracking in stressed lenses. Amine-based glass cleaners and anti-static formulations are incompatible with PC and are not used unless the UV cap supplier has published specific compatibility data. The controlled screening method is ASTM D543-20 or ISO 2812-1:2011: expose 50 mm × 50 mm plaques for 24 h at 23 ± 2 °C, then measure mass change, yellowness index per ASTM E313-20, and notched Izod impact per ASTM D256-23. A cleaner is accepted only when yellowness index changes less than 2 and impacted strength retention exceeds 90% relative to the unexposed control.

Which Solvent Classes Extract Benzotriazole Absorbers from the Cap Layer?

On outdoor glazing and automotive lenses, solvent attack on a UV cap layer is not always visible as dissolution. Benzotriazole absorbers with molecular weights of 300–450 g/mol can migrate out of the top 10–50 µm without causing immediate surface tack. Extraction is detected by wipe staining, increased haze by ASTM D1003-21, and reduced UV absorbance by ASTM D2244-23 or UV-Vis spectroscopy over 300–400 nm. Aliphatic hydrocarbons such as heptane have low absorber solubility but leave oily residues that interfere with coating wetting. Isopropanol and ethanol provide polar soil removal with low absorber extraction under short-contact conditions. Denatured ethanol is rejected if it contains methyl isobutyl ketone, methanol, or aromatic denaturants. Propylene glycol methyl ether acetate, butyl acetate, acetone, and methyl ethyl ketone dissolve PC and are not used. If a specific cleaning solvent is required for a high-contamination line, it is qualified only after ISO 22088-3 bent-strip testing on actual UV-capped production parts, followed by 500 h of ASTM G154-23 exposure to verify that the UV cap has not been depleted. Polycarbonate has a reported Hildebrand solubility parameter near 19–21 MPa^1/2; cleaning solvents with total solubility parameters in that range can swell and extract the cap even if they do not dissolve the bulk part. This is why a high-flash-point solvent may be safe for metal cleaning and destructive to UV-capped PC.

Mechanical Cleaning as a Cap-Layer Removal Process

For an injection-molded PC lens with a 25 µm cap, wiping is a material-removal process, not a surface rinse. A single scratch from a 50 µm particle can penetrate to the unmodified PC. Cleanroom wiping therefore uses ionized air blow-off or vacuum removal first, then a pre-wetted knitted polyester or polyester-cellulose cleanroom wipe. Wipe direction is single pass with 1/4 overlap; circular hand wiping re-deposits particles and increases local haze. Dry wiping is avoided because it generates static charge and abrades the cap. Isopropanol-impregnated wipes specified for polycarbonate are used only when saturated; a damp wipe has higher friction and can leave streak defects. Re-used microfiber cloths are not allowed on UV-capped PC because they carry cured coating particles from other lines. After cleaning, the surface is inspected under 1000 lux at 30° for haze and interference color shifts. A Taber abrasion test per ASTM D1044-19 can quantify cap loss: uncoated PC typically shows significant haze after 100 cycles with CS-10F wheels, while a UV-capped PC should remain below 10% haze if cap thickness is not compromised. Published data for specific cap-layer grades is limited; the relevant pass/fail is the change after cleaning, not the absolute value.

When Plasma Activation Replaces Solvent Wiping Before Hardcoat Application

Plasma and corona treatments are used to raise the surface free energy of PC from 34–38 mN/m to 50–60 mN/m for coating wetting, measured by ASTM D2578-23 dyne solutions or ISO 19403-2:2017. Oxidative plasma also attacks the organic UV absorber. Atmospheric corona with ceramic electrodes at 20–80 W·min/m² can be used on flat sheet, but UV-capped grades need the lower end of this range. The treated surface must be coated within 15–30 min because the polar groups re-orient. Vacuum plasma cleaning of molded optical parts uses argon or oxygen; oxygen plasma at powers above 300 W and exposure longer than 60 s can consume more than 5 µm of the cap surface and create low-molecular-weight yellowing precursors. Argon or nitrogen plasma is preferred because it cleans and cross-links the surface with less oxidative attack. The treatment window is verified by UV-Vis transmission through the cap before and after plasma, and by ASTM G154-23 exposure of treated uncoated plaques for 500 h. If the plasma-treated surface yellows faster than an untreated control, the plasma dosage is reduced. Published data for specific coextruded cap-layer formulations under plasma is limited; production lines therefore use gradient trials rather than a single fixed recipe.

Polysiloxane Hardcoat Curing Does Not Have to Sacrifice the UV Cap

Thermal polysiloxane hardcoats are flow-coated, dip-coated, or spray-coated onto UV-capped PC. Typical flow-coat viscosity is 5–15 mPa·s at 25 °C. Thermal cure at 120–130 °C for 30–45 min is below the heat deflection temperature of unreinforced PC at 1.82 MPa per ASTM D648-18, which is commonly reported at 126–130 °C. The process risk is not bulk deformation but localized rack shadowing and air temperature variation. Racks with high thermal mass can keep parts below 120 °C for part of the cycle, producing soft or undercrosslinked coating, while stagnant zones above 130 °C can accelerate migration of low-molecular-weight absorber fractions and create pinholes. Adhesion is tested after 24 h conditioning by cross-cut tape pull per ASTM D3359-17 using 3M 600 tape or by ISO 2409:2020. A minimum of 4B is specified for large flat glazing; 5B is expected for clean plasma-prepared PC coated within 1 h of surface treatment.

UV-curable hardcoats present a direct conflict. Mercury arc lamps emit UV photons that the cap layer is designed to absorb. Benzophenone and alpha-hydroxy ketone photoinitiators absorb primarily below 400 nm, where the cap blocks photon penetration; the result is a surface-cured skin over an undercured interface. The process solution is a bisacylphosphine oxide photoinitiator with absorption in the 400–420 nm tail, and a UVA dose limited to 800–1200 mJ/cm². The cap retains its protective function if the total UV dose does not exceed its saturation absorption capacity. Coated plaques are exposed to ASTM G154-23 cycle 1 UVA-340 for 500 h; yellowness index increase per ASTM E313-20 must remain below 2, and adhesion per ASTM D3359-17 must not fall below 4B.

After cleaning, coating, and curing, residual surfactant from a neutral detergent trapped under a silane primer or hardcoat becomes a discontinuity that passes time-zero adhesion but fails after 500 h of ASTM G154-23 exposure. Flood rinsing with deionized water below 25 µS/cm conductivity and air-knife drying at 0.4–0.6 MPa supply pressure removes cleaner residues without abrading the cap. The final surface is inspected at 1000 lux and a 30° viewing angle for haze, streaks, and interference color shifts that indicate cap-layer thinning. One coated and one uncoated plaque from each rack are subjected to ASTM D3359-17 adhesion before weathering; duplicate plaques are exposed to ASTM G154-23 for 500 h and then tested for yellowness index and adhesion. Coated plaques are rejected if yellowness index increase exceeds 2 per ASTM E313-20 or adhesion falls below 4B after 500 h.