Polycarbonate (PC) is an amorphous engineering thermoplastic with a typical density of 1.20 g/cm³, tensile yield strength in the range 60–70 MPa when tested according to ISO 527-2:2012, and notched Izod impact values commonly above 600 J/m per ASTM D256. Transparent solid and multiwall PC sheet products rely on a coextruded UV-absorber cap layer, often 30–70 µm thick, to prevent photo-Fries rearrangement at the weathering surface. That cap layer is frequently identified only by a removable masking film; after film removal, the exposed face is not visually distinguishable under white light. The resulting site condition is installing polycarbonate the wrong way up: the UV face nobody marks.

When the unmarked cap face is placed inward, the unprotected core polymer faces solar UV. The cap layer, now on the interior, cannot provide a sacrificial absorption zone at the air-polymer interface. UV-B (280–315 nm) and UV-A (315–400 nm) photons are absorbed in the outer 100–200 µm of unprotected PC. The photochemical pathway includes photo-Fries rearrangement of carbonate linkages to phenyl salicylate and 2,2′-dihydroxybenzophenone chromophores, which absorb in the 350–450 nm region and produce yellow to amber discoloration.

Installing Polycarbonate the Wrong Way Up: The UV Face Nobody Marks

Cut-to-size sheets in glazing, canopy, and machine-guard applications are frequently trimmed by 20–50 mm per edge to fit a rafter or frame profile. Printed edge legends, if present, are removed in that trimming operation. On CNC router lines, the cut sheet is stacked flat after trimming, and the original face orientation is lost because the masking film may have been removed at the supplier or on first handling. The failure is therefore not an inherent material deficiency but an orientation error caused by the absence of a post-masking face identifier.

On construction sites, sheets are often delivered with the masking film partially peeled at corners to inspect clarity; once peeled, the film cannot be reliably re-adhered. The cap face is not distinguishable by touch. In multiwall sheets, the extruded ribs are symmetric and do not code orientation. The only reliable orientation markers on many products are supplier-printed legends on the masking surface, which are removed as soon as the film is stripped. This creates a single-point failure in material control: a canopy installation may not reveal the error until the first outdoor exposure begins.

Coextrusion of PC sheet uses a primary extruder with L/D ratio 32:1–38:1 and a secondary extruder for cap stock, combined through a feedback or multimanifold die. The cap layer thickness is monitored with beta or X-ray transmission gauges and controlled by gravimetric feeders. Benzotriazole or triazine UV absorbers are let down into the cap layer at final concentrations typically 2–5 wt%. Die flow calculations assume a cap viscosity close to that of the core resin; deviations greater than 30 Pa·s at 10⁻¹ s⁻¹ shear rate can produce layer-thickness nonuniformity. When layer thickness is below the specified lower control limit, the UV face may still be present but unable to provide the published protection. The unmarked face problem is compounded by this batch-to-batch variance.

On a three-roll calender line running 6–9 m/min, contact cooling below 140 °C before embossing can cause microcracking if the sheet is handled too early; this surface defect is unrelated to UV orientation but can be confused with later weathering damage. Industrial reports from calendering lines also note that sheets flipped by forklift suction lifters may lose stack identity, increasing the probability of wrong-face installation.

What Changes When Unmarked Cap Layers Are Exposed on the Interior?

The immediate weathering response is an increase in yellowing index measured to ASTM E313-20 on the exposed uncoated face. The surface forms a brittle photodegradation crust. Tensile elongation at break falls because surface flaws propagate under load. For a correctly installed sheet, the cap layer absorbs UV before it reaches the bulk. For a reversed sheet, the protective cap side is hidden, and discoloration is visible through the sheet thickness after sufficient exposure.

Initial optical clarity of the wrong-oriented sheet is unchanged; light transmittance values remain in the range 85–90% for clear PC sheet per ISO 13468-2:2021. The failure is delayed, which reduces installer feedback. After weathering, the uncoated face yellows and microcrazes, but the interior cap face remains visually stable. Because the sheet is transparent or translucent, the yellowing is visible from both sides, and building occupants may see an amber cast before the structural reduction is noticed.

Photo-Fries rearrangement proceeds through a concerted migration of an aryl ester group from oxygen to ortho or para positions, generating 2,2′-dihydroxybenzophenone and phenyl salicylate type structures. The carbonyl-containing photoproducts extend conjugation and increase absorbance in the visible tail. Concurrently, chain scission reduces molecular weight at the surface, which lowers the critical strain for crazing. The cap layer is designed to sacrifice itself rather than the bulk; if installed inward, the uncoated bulk becomes the sacrificial layer, and the cap layer merely protects an interior boundary that does not see UV.

Tensile tests per ISO 527-2 on weathered specimens generally show first a loss of elongation rather than a large loss of tensile strength. A reduction in elongation at break from 110% to 20% may occur before the yield stress changes by more than 10%. The residual tensile strength can mislead a visual assessment; impact tests are a more sensitive indicator of surface embrittlement. Notched Izod impact retention can fall below 40% for the reversed sheet after equivalent exposure.

Cleaning incompatibility is accelerated by the wrong orientation. Aromatic solvents, ketones such as methyl ethyl ketone, and alkaline cleaners are already excluded from PC maintenance because they promote environmental stress cracking; when the uncoated surface has been UV-damaged, the susceptibility to stress cracking under these agents increases. The optical haze value measured per ASTM D1003-13 can rise above 20% on the degraded face, while the interior cap side remains comparatively clear. The visual defect is therefore asymmetric: one face yellows and microcrazes, the other remains relatively unchanged.

Accelerated Weathering, Yellowing Index, and Orientation Risk

Standard xenon arc weathering for PC sheet is commonly specified under ISO 4892-2:2013 Cycle 1 at 0.51 W/m² at 340 nm, black standard temperature 65 °C, and 102 min dry/18 min water spray. Table 1 provides representative orientation-dependent property data observed in industrial weathering programs for solid PC sheet with a 35–50 µm cap layer. The values are comparative, not universal; lamp spectral distribution, sheet thickness, and cap chemistry shift absolute results.

The table shows the failure mode is not a bulk loss of PC mass but a surface boundary condition. The unmarked UV face, when placed inward, produces a sheet that appears intact at installation but crosses the optical and mechanical acceptance limits within the first outdoor exposure period in aggressive climates. The cap layer cannot migrate through the thickness to the exposed surface; UV absorber mobility in polycarbonate is negligible at ambient temperature.

Comparative outdoor exposure in Florida, Arizona, and high-altitude sites has shown that orientation is more important than sheet thickness for UV durability. A thick uncoated sheet fails at the exposed surface before the bulk load-bearing capacity is consumed, because the photodegradation front advances only microns per year. The cap layer confines this damage to a designed sacrificial layer. The orientation error therefore converts a long-lived glazing system into a sheet with surface-limited service life.

Frequent accelerated test cycles include ISO 4892-2:2013 Cycle 1 and ASTM G154 Cycle 1 for fluorescent UV. Xenon arc is preferred for glazing materials because it includes the 300–400 nm region more closely. Water spray is specified to simulate condensation and to remove water-soluble photoproducts. In a reversed sheet, the uncoated face exposed to xenon arc may show microcrazing before yellowing saturates. Gloss retention measured at 60° per ISO 2813 may fall below 50% on the degraded face while the cap side remains above 90%. Published data for this specific configuration is limited because outdoor exposure series are often terminated after 10 years and data are supplier proprietary.

When the Cap Layer Is Installed Inward, What Failure Timing Is Expected?

Failure timing depends on equivalent UV irradiance. In high-UV climates with annual solar UV-A+B irradiance in excess of 500 MJ/m², unprotected PC surfaces may develop visible yellowing within 12–24 months and surface microcrazing within 36 months. In temperate latitudes, the same changes may require 4–7 years. The cap-protected orientation under equivalent exposure commonly resists visible yellowing for 10 years or more in solid sheet when the cap thickness and absorber concentration meet supplier specification.

Thermal movement is unaffected by orientation, but the degraded exposed face loses ductility. PC linear thermal expansion is 65–70 × 10⁻⁶ m/m/°C. A 4,000 mm panel subjected to a 40 °C temperature rise develops a dimensional change of 10.4–11.2 mm. If the sheet is restrained at fasteners after surface embrittlement, stress cracking propagates from drilled holes. Correct orientation does not eliminate this risk but confines the damage to the cap layer rather than the load-bearing core.

Installation torque and clamping also interact with the reversed orientation. PC sheet requires oversized holes and clearance for thermal expansion; the use of screws without expansion slots can cause cracking even in virgin sheet. Once the exposed face has become brittle, cracks propagate from fastener holes at loads well below the design wind load. In multiwall sheets, capillary condensation in flutes can carry UV-degraded surface residues into drainage channels, but this secondary effect is minor compared with the loss of impact resistance.

Where the masking film has been removed and the cap face is unmarked, field verification is not standardized under ISO or ASTM. A handheld 365 nm UV-A lamp in a darkened room can reveal a difference on some coextruded sheets: the cap layer may appear slightly duller or milkier than the uncoated core because the UV absorber quenches fluorescence. This response is formulation-dependent and should not be used as the sole acceptance criterion. Laboratory confirmation can be performed by solvent extraction of a 25 mm × 25 mm surface area with isopropanol, followed by UV-Vis spectrophotometry over 250–400 nm. Benzotriazole absorbers show absorbance near 300–360 nm; the uncoated face lacks that extractable peak. The extraction is mild and does not typically damage the sheet, but it removes surface contaminants and should be followed by a cleaning step if the sheet is to be installed.

Cross-sectional optical microscopy at 200× after microtoming can show a distinct surface layer of 20–70 µm with slightly different birefringence or additive speckle. This is destructive and normally limited to quality audits. Site inspection can also use a portable glossmeter and a 365 nm LED lamp. The cap face often has a slightly different surface texture from the calender embossing roll; however, texture is not a reliable proxy because both faces may be embossed or polished. Non-destructive ultrasonic thickness gauging does not resolve the thin cap layer. Eddy current methods are not applicable to a dielectric polymer. For applications where orientation is critical, the purchase specification can require a post-trimming surface mark or retained masking film on the cut edge.