The Polycarbonate Greenhouse That Went Cloudy in Year Three: A UV Post-Mortem

Post-mortem examination of a twin-wall polycarbonate (PC) greenhouse sheet removed after its third year of service shows that the visible clouding is concentrated in the outer 50–150 µm of the sky-facing surface rather than in the bulk sheet or interior rib structure. At 20× magnification, the affected zone exhibits a fine, diffuse craze pattern that is not removed by wiping with distilled water or isopropanol. The interior surface remains transparent, and the webs between the upper and lower walls are free of liquid filling. Measurements on the failed surface using ASTM D1003-13 show total haze in the thin-wall region above 15%, against a typical as-extruded value below 2.0%. This loss of transparency is accompanied by an increase in yellowness index to more than 8 under ASTM E313-20 and by a reduction in 60° specular gloss under ASTM D2457-21 to the 20–40 GU range. Published multi-year field data for this exact sheet configuration remain limited, but the post-mortem signature is consistent with accelerated weathering data for unstabilized or partially stabilized polycarbonate under ISO 4892-2:2013.

Why Does the Polycarbonate Haze Threshold Emerge in the Third Service Year?

Although the third-year haze threshold is often discussed as a material limit, it is not an intrinsic degradation clock for polycarbonate. It is the point at which photochemical chain scission, UV absorber depletion, and thermal stress combine to create a light-scattering surface layer. Polycarbonate absorbs strongly in the 290–330 nm ultraviolet band because the bisphenol A carbonate repeat unit contains photosensitive aromatic carbonate linkages. The dominant reaction pathway is the photo-Fries rearrangement, which converts carbonate groups into hydroxybenzophenone, aryl aldehyde, and conjugated chromophore species. These products absorb in the violet-blue visible region and produce the characteristic yellow-to-brown tint seen in weathered PC. Simultaneously, radical-induced chain scission lowers surface molecular weight, reduces craze resistance, and allows microvoid formation when the sheet is subjected to thermal expansion and contraction.

During peak irradiance, greenhouse roof surface temperatures commonly rise 30–50°C above ambient air. The coefficient of linear thermal expansion for polycarbonate is approximately 6.5 × 10⁻⁵ K⁻¹ per ISO 11359-2:2021, and diurnal temperature swings of 40°C or more generate repeated surface strain. Under these conditions, the surface craze network grows through a combination of photo-oxidative embrittlement and mechanical fatigue. Laboratory xenon-arc testing under ISO 4892-2:2013 with 0.51 W/m²/nm at 340 nm and black standard temperature 65°C reproduces the surface-restricted haze in unstabilized PC within 1500–3000 h, depending on cycle humidity.

In failed PC greenhouse sheets, yellowing and haze appear together but do not progress at the same rate. Yellowing is a chemical chromophore phenomenon measured by yellowness index; haze is a physical light-scattering phenomenon caused by refraction and reflection at microcracks and pits. A sheet can retain a yellowness index below 5 while haze already exceeds 10% if the surface craze density is high and the chromophore concentration remains moderate. Conversely, a thermally over-processed sheet may yellow without initial haze. Consequently, replacement decisions based only on yellowness can underestimate the structural damage to the surface.

Because the melt volume-flow rate increase shown in the table is a direct indicator of molecular weight loss, it provides a rapid post-mortem screening method. Polycarbonate glazing grades are typically produced with a relatively high melt viscosity to retain toughness; a shift from 5–7 cm³/10 min to more than 10 cm³/10 min under ISO 1133-1:2022 at 300°C/1.2 kg corresponds to a substantial reduction in average molecular weight and a loss of ductility. The notched impact value may drop by more than two-thirds before the sheet exhibits gross cracking. A panel can therefore remain intact for several seasons after first clouding but fail suddenly under hail or wind load.

UV Cap Layer Erosion, Yellowing Indices, and the Microcrazing Sequence

Commercial greenhouse polycarbonate is normally protected by a coextruded UV-stabilized cap layer that is 25–100 µm thick on the sky-facing side. The cap layer is not a topical coating; it is a PC surface layer into which a benzotriazole or triazine UV absorber has been compounded at a loading typically between 0.2 wt% and 0.5 wt%. The absorber operates by converting incident UV into heat before the radiation reaches the unprotected bulk. When the cap layer is intact, the underlying PC remains transparent and tough because the damaging 290–330 nm wavelengths are attenuated within the first 25–50 µm of the surface.

When the cap layer fails, the process is primarily a thickness-loss and absorber-depletion problem. UV absorbers can migrate, volatilise, or undergo photo-degradation over time. Surface ablation from windborne dust, hail, and cleaning removes the absorber-rich polymer and shifts the UV absorption front deeper into the sheet. Once the cap layer is breached or depleted below a critical screening concentration, the underlying PC begins to photo-oxidize. Cross-sectional microscopy of failed greenhouse panels typically shows a high craze density in the outer 10–30 µm, a transition zone at 30–80 µm, and an unaffected core.

Under processing conditions, phosphite antioxidants and phenolic antioxidants protect the melt during coextrusion, but they are not long-term UV stabilizers. Benzotriazole or triazine UV absorbers provide surface screening but do not fully protect the polymer from radical attack if the cap layer is too thin. Hindered amine light stabilizers are generally avoided in polycarbonate because amine functionality can promote carbonate hydrolysis under thermohumid conditions. A greenhouse-grade PC sheet therefore typically combines a low-volatility UV absorber in the cap layer with a minimal non-amine antioxidant package in the core. This formulation boundary is one reason a general-purpose PC sheet without the coextruded cap layer can show early clouding even when it meets an initial haze specification.

Within the polycarbonate greenhouse roof, night-time condensation forms on the underside of the upper wall and drains toward the gutter. The condensate is rarely pure water. It can contain dissolved carbon dioxide, nitrate and ammonium from fertilizer decomposition, organic acids from plant transpiration, and traces of crop protection chemicals. In a closed greenhouse, condensate pH can fall below 5.5 or rise above 8.5 depending on the crop and climate-control strategy. Polycarbonate is most stable in neutral-to-slightly acidic water, but alkaline condensate is especially damaging once the UV cap layer is breached because hydroxide attacks the carbonate linkage. The damage appears as surface etching, microcavitation, and loss of gloss rather than bulk fracture.

A second field variable is mechanical abrasion. Polycarbonate has relatively low surface hardness, and even routine cleaning with dry cloths or brushes can remove the UV-protected skin. In high-wind regions, dust and sand act as a continuous abrasive on the sky-facing surface. Post-mortem examinations of greenhouse panels from arid and coastal sites frequently show gloss loss on the upper surface that is absent on the interior face. A 60° gloss value below 40 GU under ASTM D2457-21 is a useful field threshold for suspecting cap layer damage. Hail impacts also act as stress concentrators; surface crazes often initiate at hail dimples and then propagate across the roof panel.

Roof orientation and slope influence the third-year clouding pattern. A south-facing roof in the northern hemisphere receives a larger UV dose and also experiences longer periods of standing condensate if the slope is below 5°. In those low-slope installations, condensation accumulates near the eave and can wick into the flutes, holding moisture against the surface for hours after sunrise. The combination of UV, alkaline or acidic moisture, and thermal cycling accelerates cap-layer degradation. This explains why post-mortem panels often show more severe haze near the eave and around roof clamps than in the centre of the roof, even though the centre receives similar UV irradiance.

If the Replacement Sheet Lacks a Verified Sky-Facing UV Cap Layer and Moisture Control

If a single-sided product is installed upside down, the unprotected surface is placed at the source of UV, and field observations in high-UV regions show visible yellowing and haze in less than 18 months. The protective layer is usually marked by the manufacturer on the masking film. If the film is removed and the orientation becomes uncertain, a solvent wipe test or attenuated total reflectance infrared spectroscopy can identify the cap layer side by its higher UV absorber concentration. Installing the sheet with the wrong orientation voids the manufacturer’s weathering warranty and converts a third-year clouding failure into a first-year failure.

Moisture control during storage and installation is also a processing boundary. Polycarbonate must be dried to below 0.02% moisture before thermoforming or any thermal cutting operation. Residual moisture above this level at processing temperatures of 280–300°C causes hydrolysis, molecular weight loss, and visible splay or bubbles in the finished sheet. Drying is typically performed in a desiccant dryer at 120°C for 4 h with a dew point of -40°C or lower. The dried granules are then processed on a vented twin-screw extruder with an L/D ratio between 32:1 and 36:1, using a flat sheet die and a polishing stack to set surface gloss. The cap-layer extruder is typically 24:1 to 30:1 L/D, and its melt temperature is kept within the same narrow window to avoid degrading the UV absorber.

In high-UV sites, a sheet that passes the laboratory xenon-arc requirement under ISO 4892-2:2013 for 2000 h can still exhibit haze in year three if the UV cap layer is underspecified or mechanically damaged. The laboratory test provides a controlled UV and moisture exposure but does not fully reproduce the greenhouse roof environment of fertilizer-derived condensate, abrasive dust, and thermal cycling. The operational boundary for polycarbonate greenhouse glazing therefore includes a verified UV-stabilized cap layer, a gentle wet cleaning protocol, adequate roof slope for drainage, and exclusion of abrasive cleaners. Where published field data for a particular product are absent, the post-mortem protocol should record UV dose, roof orientation, sheet age, cap-layer thickness, haze, yellowness index, and gloss to permit a data-based replacement decision.