Evaluation and Validation of an Accelerated Weathering Procedure to Characterise the Release of Bisphenol A from Polycarbonate Under Exposure to Simulated Environmental Conditions

Abstract

Bisphenol A (BPA) has been listed as a substance of very high concern (SVHC) due to its endocrine-disrupting properties according to REACH in 2017. European competent authorities have prepared a REACH restriction proposal to reduce BPA levels in the environment. The proposed limit for the concentration of free BPA and other bisphenols in articles is 10 mg kg⁻¹. If exceeded, migration testing can demonstrate that no more than 0.04 mg L⁻¹ is released from the product or material over its lifetime. German authorities are drafting a new restriction proposal after the original was temporarily withdrawn. The residual and migration limits mentioned above were key requirements from the previous restriction proposal. Numerous national and international standards exist for assessing how environmental factors affect the physical and chemical properties of products and materials—such as notch impact strength and tensile strength—but these standards do not cover the release of pollutants. A standardised procedure that covers all aspects of artificial weathering and monitors the subsequent release of pollutants is necessary, especially in the context of the regulation of these substances. An accelerated weathering procedure was established for non-protected samples. This material was not intended for outdoor applications. The testing procedure applied a typical weathering scenario that represents Central European climate conditions. The procedure was validated and applied to samples under distinct quality assurance aspects. Released BPA is quantified via an organic isotope dilution LC-MS/MS method. In parallel, identical samples were weathered outdoors on a weathering rack. Haze and yellowness index are measured to compare outdoor and weathering chamber results.

1. Background

Real weathering refers to the process by which natural impacts, such as global radiation, rain, wind, and temperature fluctuations, gradually deteriorate and change the appearance, texture, and properties of materials over time. Weathering is indeed a time-consuming process, and the extent and rate of weathering can vary depending on several factors, including the type of material, the local climate, the presence of pollutants or contaminants, and the specific environmental conditions. Some materials might show noticeable weathering effects over a relatively short period, while others might take many years or even decades to exhibit significant changes. Moreover, it is challenging to observe and investigate the release of pollutants from materials under real environmental relevant conditions.

In this current study, the determination of the BPA release in ultra-trace amounts from polycarbonate materials is very challenging, as cross-contamination can occur, and many microorganisms can metabolise BPA. One way to study the behaviour of a polymer or plastic in the presence of rain, global radiation, heat, and freezing periods in a shorter time is to use accelerated artificial weathering experiments. Artificial weathering is therefore a suitable solution for the determination of the BPA release. The advantages are an exact quantification of the cumulated release in relation to the surface area of the weathered materials, precise control and recording of the simulated environmental impact factors, as well as economic aspects like time and costs.

Many weathering experiments, investigating the influence of humidity, radiation, or temperature (in all possible combinations), have been made [1,2,3]. Various standards for these aspects of weathering are available [4,5,6]. When materials are exposed to global radiation and air pollutants like corrosive gases, they can break down through photooxidative processes or other degradation mechanisms induced by the applied weathering scenarios.

These approaches tackle only aspects concerning the physicochemical properties of the weathered materials [7,8,9]. A standardised procedure that covers all aspects of artificial weathering and the release of constituents, compounds, or even pollutants from the material is essential, especially for the possible regulation of pollutants or hazardous substances. Most of these artificial weathering procedures only investigate the changes in the material’s behaviour or their optical properties, like haze, toughness, or the yellowness index of the weathered materials. To evaluate these effects, analytical methods like microscopy, spectroscopy, or the analysis of colour changes are applicable.

Due to the unique properties of polycarbonate, such as transparency, strength, and stability, it can be used in many different applications. Polycarbonate consists exclusively of polymerised Bisphenol A, which could be released again under the environmental conditions described below. This would pose a greater risk to the environment and humans than if, for example, other plastics such as polypropylene, polystyrene, or PVC were to be investigated. Therefore, PC was selected as the test material due to its widespread use in outdoor applications (greenhouse glazing, automotive, construction), where these materials are exposed to a variety of environmental impacts. These impacts (Figure 1) include physical influences like temperature and global radiation, or chemical influences such as rain or corrosive gases like ozone. Furthermore, biological impact (microorganisms), and mechanical stress (hail, sand) affect materials outdoors.

All these impacts can damage or age the material. An ageing process summarises the following aspects: “processes that occur in a polymeric material during a specified period, and that usually result in changes in physical and/or chemical structure and the values of the properties of the material” [10]. These changes might induce leaching processes and/or the migration of pollutants/chemicals/substances out of the material into the environment.

The primary goal of this study is to establish a procedure that meets all these criteria, enabling the accurate quantification of substance release from materials during artificial weathering. This weathering process in a weathering chamber is accelerated under precisely defined and documented conditions. Additionally, the release can be related to a defined surface area by the self-designed sample holder.

2. Materials and Methods

2.1. Weathering Chamber for Accelerated and Artificial Weathering

All experiments were carried out in a weathering chamber SunEvent UV/200/20/80 (Figure 2) from Weiss Technik (2020, Reiskirchen, Germany). In this chamber, the samples can be exposed to global radiation, heat, frost, and humidity in various combinations.

The applied weathering scenario simulates Central European average weather conditions in conformity with a standard method from RAL (German National Committee for Delivery and Quality Assurance). The programme shown in Figure 3 includes extreme weather phenomena, such as a high-temperature phase and long wetting periods. One cycle has a duration time of 24 h and consists of six stages with a time frame of four hours each. This circadian is repeated 42 times giving a duration of six weeks. One cycle consists of three raining periods at 23 °C, two high-temperature periods at 70 °C, and one freezing period at −10 °C with constant UV-A irradiation (at about 45 W m⁻²) using UV-A-340 nm fluorescent tube lamps (Figure 3). With this continuous irradiation, an acceleration factor of two is already obtained.

For the simulation of global radiation, 16 UV-A-340 fluorescent tube lamps (Atlas Material Testing Technology GmbH, Linsengericht, Germany) were used. Figure 4 shows the spectra obtained from these UV-A-340 lamps and a global radiation reference spectrum [11]. The reference spectrum shows the standard reference spectral irradiance for hemispherical solar irradiance incident on cloudless atmospheric conditions on a sun-facing, 37° tilted surface. The UV-A-340 lamps radiate in the range from 295 to 440 nm and thus simulate the high-energy small-wavelength range of global radiation (Figure 4). The spectral output of the UV-A-340 lamps is mainly in the UV-A region from 315 to 400 nm with a small part in the UV-B range. The radiation dose is 45 W m⁻², and the spectral maximum at 340 nm has a radiation dose of 0.864 W m⁻². The weathering chamber is equipped with a quartz glass filter that cuts off all wavelengths below 300 nm to avoid energy-rich radiation, which can break chemical bonds. Moreover, the part of the spectrum below 300 nm does not occur in global radiation but could trigger other degradation mechanisms which are not occurring at higher wavelengths. The UV-A radiation has no solar NIR (near-infrared) part in the spectrum, which can heat the surface of the weathered samples.

For quality assurance, the radiation dose was measured three times a week in the weathering chamber including the quartz glass filter using an ALMEMO ^® 2590A, a professional measuring instrument and data logger equipped with a FLA623 UV-A radiation probe for UV-A measuring (ALMEMO^® 2590A with FLA623 radiation probe, Ahlborn, Holzkirchen, Germany). The measurements of irradiance were carried out towards the end of the 24 h cycle at about 20 degrees and without rain.

To avoid blank values of several polymer-associated additives, the chamber is equipped with a stainless-steel rainwater reservoir. All internal tubes were custom-built and contained no polymeric material. A self-designed sample holder (Figure 5) was used for the determination of the Bisphenol A release from polycarbonate materials with a defined surface of 0.307 m². The holder consists of two aluminium plates. The lower plate is six mm thick and measures 636 × 800 mm. In it, 12 cavities with two mm depth are milled for depositing the specimens, and holes with M5 threads are drilled. The upper plate is three mm thick and has 12 windows measuring 218 × 138 mm and holes for the screws with six mm diameter. To protect the sample backsides and the edges from scattered radiation and raining water simulant, the samples were wrapped in blank value-free PTFE sheet and Teflon^® sealing tape (PTFE Band Type FRp, 12 × 0.1 mm, 60 g m⁻², Fermit, Vettelschoß, Germany). Then, they are placed in the sample holder and both plates of the sample holder are screwed together with a torque of 1.4 Nm using a Wiha TorqueVario Screwdriver (Wiha, Schonach, Germany). The constant torque of 1.4 Nm ensures reproducible conditions, and the material is not physically stressed.

2.2. Real Weathering

The samples were weathered outdoors in Horstwalde, Germany, in parallel (52°05′44.6″ N, 13°24′34.1″ E). They were placed in the same sample holder as in the weathering chamber and positioned on a weathering rack at a 45° angle to the south, according to ASTM D1435 (Figure 6) [6]. Weather report with data for temperature, duration of sunshine, and rainfall is provided by the DWD (German Weather Service).

2.3. Sample Materials

The moulded samples are made of polycarbonate with a very low content of UV absorber. No coating or coextrusion layer of a UV absorber polycarbonate material was applied to the samples; therefore, the samples were unprotected. The PC samples were provided by PC/BPA group of Plastics Europe.

2.4. Chemicals

The analytical standards were Bisphenol A (99.8%) from Dr. Ehrenstorfer (Augsburg, Germany), d₁₆-Bisphenol A (99%) were provided by CDN Isotopes (Pointe-Claire, QC, Canada) and ¹³C₁₂ Bisphenol A (99%) in ACN were purchased from CIL (LGC standards Ltd., Teddington, UK). A Custom-made control standard with 50 ng L⁻¹ BPA, d₁₆-BPA, and ¹³C₁₂-BPA was purchased from Neochema GmbH (Bodenheim, Germany). Ultrapure water was generated with an ELGA Purelab^® flex (Veolia Water Technologies, Paris, France) water purifier. ULC/MS grade solvent acetonitrile was obtained from Biosolve B.V. (Valkenswaard, The Netherlands). All the materials and solvents are tested for blank values.

2.5. Validation of the Weathering Chamber

The recovery rates for potential released BPA were determined with a solution of fully deuterated BPA in Milli-Q^®-Water. Four polystyrene cuvettes were fixed with polypropylene cable straps in the chamber. They were filled with 2 mL d₁₆-BPA solution each, and the weathering programme was started. After 24, 48, and 96 h, a sample was taken out of the raining water reservoir and analysed by LC-MS/MS. The recovery rates of deuterated BPA are determined against a reference solution, and the results were 95 ± 2% (24 h), 105 ± 4% (48 h), and 94 ± 3% (96 h).

2.6. Analytical Procedure for the Determination of the Cumulated BPA Release

The resulting extracts were analysed using an Agilent 1260 Infinity II HPLC system (Agilent Santa Clara, CA, USA) coupled to an AB Sciex QTRAP^® 6500 triple-quadrupole mass spectrometer (Sciex Framingham, MA, USA) equipped with an electrospray ion source (ESI) in negative ionisation mode. A Kinetex^® 2.6 µm XB-C₁₈ 100 Å LC column (150 × 3 mm) from Phenomenex (Aschaffenburg, Germany) and matching precolumn were used for separation. All samples and standards were injected three times and measured in random order. Blank samples were injected between each measurement to prevent carry-over effects. The injection volume was 50 µL. Mass spectrometric parameter: ESI neg., target scan time 0.700 s, scheduled MRM (multiple reaction monitoring). Further MS parameters can be found in the Supplementary Materials.

Three working ranges (small, middle, and high) are derived from a validated method according to DIN ISO 17025, guaranteeing the maintenance of precision and accuracy [12]. The working ranges for BPA in demineralised water are 30–200, 400–2000, and 2400–6200 ng L⁻¹. The concentration of the internal standard was 80, 1000, or 4000 ng L⁻¹ (small, middle, and high) in the calibration and the individual samples. The calibration curves for all three ranges can be found in the Supplementary Materials (SI).

2.7. Weathering Experiment: Pre-Run and Weathering Campaign

After filling the stainless-steel rain reservoir with demineralised water, the weathering chamber was equipped with the empty sample holder and the weathering programme was started. To pre-condition the chamber, the programme runs until a constant BPA background level (between 30 and 90 ng L⁻¹) is reached for at least five days. When the BPA background reaches a constant level, the sample holder is removed from the chamber. After loading the holder with the polycarbonate samples of choice, it is reinserted in the weathering chamber, and the actual weathering campaign starts with the existing water. Each day five water samples were taken from the raining water reservoir. The reservoir is filled up with BPA-free demineralised water to 35 L again to compensate for a possible loss of water during the weathering caused, e.g., by evaporation. One of the five samples is frozen as a reference sample, and one sample is measured to estimate the working range of the calibration and the level of the internal standard. After the addition of d₁₆-Bisphenol A and ¹³C-labelled Bisphenol A, the BPA content of the three samples was analysed by organic isotope dilution calibration LC-MS/MS. A custom-made quality control standard with 50 ng L⁻¹ BPA, d₁₆-BPA, and ¹³C₁₂-BPA is used for quality assurance and is registered in a control chart.

2.8. Characterisation of the Optical Properties of the Polycarbonate Samples

The CIE Tristimulus values were measured by using a Bruins Omega 20 spectrophotometer (Bruins. Instruments, Puchheim, Germany) according to ASTM E313-20:2020 CIE [13]. The standard illuminant and standard observer were D65 and 10°. The yellowness indices are calculated from these values with the coefficients C_x = 1.3013 and C_z = 1.1498 and were measured and calculated after the individual weathering campaign, as well as after 73 and 96 weeks of outdoor weathering. All results are given as mean values of the 12 samples. The error bars represent the standard deviation.

The haze was measured with a haze-gard plus from BYK Instruments (BYK-Gardner GmbH, Geretsried, Germany) according to ASTM D1003-13 [14].

3. Results and Discussion

The moulded samples made of polycarbonate with a very low content of UV absorber were weathered in duplicate for six weeks in the weathering chamber. The higher cumulated released BPA in the 35 L water reservoir is 3.49 ± 0.05 µg L⁻¹ and 397.2 ± 8.2 µg m⁻² (Figure 7) correlated to the complete weathered surface area. Both curves show an s-shaped cumulated release profile with an increasing slope until day 20 (Figure 7). In the beginning, the weathering with UV-A-light (simulated global radiation), heat, and water induces degradation processes, like photooxidation, hydrolysis, and the migration of residual BPA [8,15,16]. This leads to a subsequent release of Bisphenol A. From day 21, the slope becomes smaller and smaller until after 28 days, a steady-state plateau without further release is reached. It is possible that the weathered and thus aged or oxidised layer protects against or slows down further release. Mercea et al. also mention the adsorption of BPA on the PC sample surface in batch experiments when a certain concentration is reached or a remigration into the PC as possible causes [17]. This can be observed here as a steady state or plateau effect.

The two BPA release curves were fitted using a Weibull CDF fit as a statistical method for modelling lifetime data using the Weibull distribution and predicting the probability of failure over time. The fitting parameters of both curves are given in

Table 1. Parameters obtained from Weibull cumulative distribution function fitting for concentration vs. time data of unprotected monolayer sheets.

Parameter	Symbol	Unit	Sample 1	Sample 2
Offset	y0	µg L−1	37 ± 3	−0.8 ± 0.7
Amplitude	A1	µg L−1	3.15 ± 0.1	3.43 ± 0.1
Scale parameter	a	days	16.8 ± 0.1	18.24 ± 0.02
Shape parameter	b	-	1.89 ± 0.02	2.388 ± 0.007
Adj. R2	-	-	0.9985	0.9999
Reduced χ2	-	-	12.9	13.7

according to:

$$y\left(x\right)=y_0+A_1\ast wblcdf\left(x,a,b\right) for x>0$$

(1)

The released BPA related to the weight of the weathered injection moulded samples made of PC with very low UV absorber content is 58 ± 1 µg kg⁻¹. This is lower than the previously published residual monomer content of 1–190 mg kg⁻¹ [18,19,20]. Here, the release only occurs from the layer that is exposed to weathering since the back sides and the edges are protected. If the release is related to the weight of the four mm thick sheet, although the release only occurs from the surface, no comparable results are obtained. Relating the leaching of BPA from the surfaces is a logical conclusion since the environmental impacts only act here. For example, biocide leaching tests show that there is more leaching from horizontally weathered samples than from vertically exposed ones [21]. These results show that there is a direct correlation between the weathered surface area and the observed release of BPA to the raining water.

Haze indicates the scattering of light by a material. The value is unitless and indicates the percentage of transmitted light that is scattered such that its direction deviates by more than a certain angle from the direction of the incident beam [14]. Environmental degradation mechanisms significantly influence haze development in polycarbonate. Increased haze is usually caused by surface irregularities and sometimes by the formation of a dispersed phase with a refractive index different from that of the bulk material [3]. In polycarbonate, these irregularities develop through photo-oxidation reactions that cause molecular weight reduction and chemical structural modifications, creating surface defects within a depth of several micrometres [22]. The haze of the unweathered reference polycarbonate sheets was 0.6. The haze of the polycarbonate weathered in the weathering chamber for six weeks and the polycarbonate weathered outdoors after 14 weeks was 1.4 (Figure 8). Since only one surface is weathered, these irregularities can be observed mostly on one side.

The yellowness index expresses the level of change in the colour of material from colourless to yellow. Yellowing results from a combination of various reactions like photo-oxidation, thermal degradation, or hydrolysis. These reactions can lead to the formation of an additional conjugated double bond within the aromatic system, which is the cause of the colour change. Therefore, the yellowing is not dependent on the received UV energy alone but on various influencing impacts [23].

Figure 9 shows the average values of the yellowness indices of the unweathered reference samples, the samples weathered outdoors (for 73 and 96 weeks), and the samples weathered in the weathering chamber. The error bars represent the standard deviation. The yellowness index measured in transmission is 0.5 for the unweathered and 4.11 after six weeks in the weathering chamber. The yellowness index after 73 and 96 weeks is 3.0 and 5.8. The yellowness index increases for samples weathered in the weathering chamber and for those weathered outdoors. This is to be expected since they are unprotected.

The injection moulded samples made of polycarbonate with a very low content of UV absorber show the same degree of yellowing after an estimated value of 81 weeks outdoors as after six weeks in the weathering chamber. This is an indication that the yellowing in the weathering chamber is 16.2 times faster than outside. This is also a correlation with the following acceleration factor related to the irradiation dose.

The acceleration factor reflects the correlation between specific material properties and their change due to two different tests or setups. The most common correlation is between outdoor and artificial weathering. It is important to note that acceleration factors are strongly material-dependent, with literature reporting values of 12.75 for PC based on the colour shift or Esters ranging from 5 to 15 [24,25,26]. Additionally, acceleration factors also depend on the observed property measured, such as changes in gloss, colour, or mechanical properties. Furthermore, they are influenced by annual weather variations, chamber-to-chamber variations and even the position of samples within the weathering chamber or outdoors. The linkage of the radiation dose in the chamber with the radiation dose outdoors leads to an acceleration factor. With this factor, the time in the weathering chamber can be “translated” to the natural weathering outdoors. A comparison of the radiation dose of the climatic chamber (Equation (2)) with the radiation dose in Central Europe shows that the weathering is accelerated by a factor of 13.6 (Equation (3)).

The irradiance dose at 340 nm in Central Europe (average value of 10 years (2008–2018)) is 2.0 MJ m⁻² for one year (Deutscher Wetterdienst (DWD 2024)). The irradiation dose of the weathering chamber would be for one year of weathering [27]:

$${\mathrm{E}}_{\mathrm{e}\mathrm{q}\mathrm{u}\mathrm{i}\mathrm{v}}={\mathrm{E}}_{340 \mathrm{n}\mathrm{m}} \left[W\ast m^{-2}\right]\ast {\mathrm{t}}_{\mathrm{e}\mathrm{x}\mathrm{p}\mathrm{o}\mathrm{s}\mathrm{u}\mathrm{r}\mathrm{e}} \left[s\right]=0.864 W\ast m^{-2}\ast \mathrm{31,536,000} s=27.25 MJ\ast m^{-2}$$

(2)

If the radiation doses for Central Europe and the weathering chamber are put in relation to each other, an acceleration factor can be obtained:

$$Acceleration factor={\displaystyle \frac{\mathrm{i}\mathrm{r}\mathrm{r}\mathrm{a}\mathrm{d}\mathrm{i}\mathrm{a}\mathrm{t}\mathrm{i}\mathrm{o}\mathrm{n} \mathrm{d}\mathrm{o}\mathrm{s}\mathrm{e} \mathrm{w}\mathrm{e}\mathrm{a}\mathrm{t}\mathrm{h}\mathrm{e}\mathrm{r}\mathrm{i}\mathrm{n}\mathrm{g} \mathrm{c}\mathrm{h}\mathrm{a}\mathrm{m}\mathrm{b}\mathrm{e}\mathrm{r} \mathrm{f}\mathrm{o}\mathrm{r} 1 \mathrm{y}\mathrm{e}\mathrm{a}\mathrm{r}}{\mathrm{i}\mathrm{r}\mathrm{r}\mathrm{a}\mathrm{d}\mathrm{i}\mathrm{a}\mathrm{t}\mathrm{i}\mathrm{o}\mathrm{n} \mathrm{d}\mathrm{o}\mathrm{s}\mathrm{e} \mathrm{o}\mathrm{u}\mathrm{t}\mathrm{d}\mathrm{o}\mathrm{o}\mathrm{r} \mathrm{w}\mathrm{e}\mathrm{a}\mathrm{t}\mathrm{h}\mathrm{e}\mathrm{r}\mathrm{i}\mathrm{n}\mathrm{g} \mathrm{i}\mathrm{n} \mathrm{c}\mathrm{e}\mathrm{n}\mathrm{t}\mathrm{r}\mathrm{a}\mathrm{l} \mathrm{E}\mathrm{u}\mathrm{r}\mathrm{o}\mathrm{p}\mathrm{e} \mathrm{f}\mathrm{o}\mathrm{r} 1 \mathrm{y}\mathrm{e}\mathrm{a}\mathrm{r}}}=13.6$$

(3)

This means that the samples receive the same irradiation dose after six weeks of artificial weathering in the weathering chamber after about 81 weeks of outdoor weathering in Central Europe. This factor correlates with the factor obtained by comparing the yellowness indices. This indicates that the radiation dose has a major influence on the release of BPA. The determined acceleration factor has an uncertainty that includes the variability of the UV-A fluorescence tube lamps, chamber inhomogeneity concerning the irradiation of individual samples, and the analytical quantification of the released Bisphenol A. All influences on the calculated acceleration factor were either determined experimentally or calculated using the manufacturer’s specifications. For the analytical procedure, the average standard deviation of the measurements was considered. The chamber inhomogeneity was evaluated by measuring the yellowness indices from all twelve sample plates in the sample rack. Here, a standard deviation could be determined within the climate chamber at different positions. The variation in the radiation dose at distinct wavelengths (e.g., at 340 nm) was provided by the producer of the fluorescent tube lamps. Finally, the outdoor spectral uncertainty can be disregarded because it is an average value over 10 years and covers the complete variety. In conclusion, an acceleration factor of 13.6 with a standard deviation of 10% can be calculated.

4. Conclusions

Unprotected injection moulded samples made of polycarbonate with a very low content of UV absorber were weathered with the evaluated and established method and the BPA release could be determined under simulated environmental conditions. The amount of released BPA from unprotected polycarbonate is 397.2 ± 8.2 µg m⁻² or 58 ± 1 µg kg⁻¹ within six weeks of accelerated weathering. The BPA-release is substantially lower than the previously published residual monomer content of 1–190 mg kg⁻¹. Weathering primarily mobilises surface-adjacent and already migrated BPA, while most of the residual monomer remains trapped within the polymer matrix. The Weibull cumulative distribution function provided excellent fits (Adj.R² > 0.998) for both replicate experiments, with characteristic s-shaped release profiles showing initial acceleration until day 20, followed by a plateau phase reaching a steady-state after 28 days without further release. The UV-A radiation triggers the formation of an oxidised thin layer at the surface, largely by photo-oxidation, so that no further BPA is released. The initial BPA releases are due to hydrolysis of PC, as in this phase, the oxidised layer is not yet fully established. The plateau effect observed after day 21 suggests that the weathered and oxidised surface layer may protect against or slow down further release.

Applying our experimental release rate to European polycarbonate installations reveals a theoretical maximum BPA emission of 163 kg across the entire EU. This calculation is based on industry data from the European Polycarbonate Sheet Extruders (EPSE), which indicate that approximately 890,471 tonnes of multiwall polycarbonate sheets and 279,492 tonnes of solid polycarbonate sheets were installed over the past 15 years, corresponding to a total outdoor area of 402.8 km² across the European Union. Using our experimentally determined release rate of β = 397.2 ± 8.2 µg m⁻², the theoretical maximum BPA emission calculates as: m = β × A = 405.4 g km⁻² × 402.8 km² = 163,295 g = 163 kg. However, this represents an extreme worst-case scenario, as all commercially installed outdoor polycarbonate sheets incorporate surface protection through UV stabilisers, coatings, or co-extruded protective layers, while our tested samples were deliberately unprotected to simulate regulatory worst-case conditions.

The degree of yellowing after the weathering in the weathering chamber corresponds to the estimated yellowing after around 81 weeks of outdoor weathering in Central Europe conditions. The yellowness index increased from 0.5 (unweathered) to 4.11 (six weeks chamber weathered) and reached 5.8 after 96 weeks outdoors, indicating that the yellowing in the weathering chamber is accelerated by a factor of 16.2 compared to outdoor conditions. The acceleration factor calculated from radiation dose comparison shows that samples receive equivalent irradiation after six weeks of artificial weathering as after 81 weeks of outdoor weathering in Central Europe, yielding an acceleration factor of 13.6. The correlation between the radiation dose-based acceleration (13.6×) and yellowness index-based (16.2×) indicates that radiation dose has a major influence on ageing and confirms the validity of the accelerated weathering approach for predicting long-term environmental behaviour.

It is also relevant to note that only a few commercially available polycarbonate applications are used in an unprotected form even for indoor use. In all applications, various functional additives are used that are intended to prevent thermo-oxidative damage and photooxidation. Also added are UV stabilisers, UV absorbers, and optical brighteners, which are intended to reduce yellowing and visual perception [28,29]. Moreover, most of the relevant outdoor applications are coated. The weathering conditions (continuous irradiation, temperature cycling from −10 °C to 70 °C and alternating wet-dry phases) are harsher than typical outdoor conditions. The hot and cold temperature cycles in the weathering chamber led to an alternation of expansion and contraction of the material. The rain phases following the expansion due to high temperature led to moisture absorption and desorption. All this occurs under the aspect of continuous irradiation.

This can therefore be considered a worst-case scenario approach, ensuring that the material is unlikely to release more BPA in real-life applications. This validated approach provides a standardised procedure for assessing pollutant release from polymer materials under accelerated weathering conditions, which is essential for regulatory evaluation of substances like BPA under REACH restriction proposals.

An accelerated weathering method could be established and validated. However, some obstacles appeared during the work in this study. There was a multitude of factors to consider ensuring reliable results without over- or underdetermination. The determination of Bisphenol A concentrations in the ultra-trace range in a large-scale weathering chamber was a major challenge—from a technical point of view and in terms of sampling, sample preparation and analysis.

The main problem arises from the ubiquitous BPA blank values since potential blank values can fall in the same range as the release of BPA from the weathered samples. To ensure a reliable determination of the Bisphenol A releases from the samples to be weathered using a raining water simulant, all components used in the climate chamber must be free from Bisphenol A. The use of non-contaminated cleaning cloths and laboratory wipes, as well as non-BPA-contaminated cleaning agents (organic solvents, water), is strongly recommended. Commercially available solvents can be contaminated with BPA to varying degrees, depending on the manufacturer, but also on the lot, so these were tested regularly. A self-designed sample rack (aluminium) for backside and edge protection was used to ensure that only one side of the sample comes into contact with rain and UV radiation. Great care was taken to avoid blank value problems and the materials used to wrap the samples (Teflon^® sealing tape and a thin PTFE sheet) were tested on a regular basis. Since the experiments are performed in a commercially available and industrially manufactured climate chamber special attention must be paid to the materials used for construction. Thus, all plastic parts of the chamber must be replaced by stainless steel, if possible. A strict blank value control before and after the different weathering campaigns is also essential. The standard cleaning procedure includes three water changes followed by a hot raining event (70 °C) with continuous UV radiation for two days. After that, a pre-run without the samples was performed to determine the current blank value of the chamber and the rack together and to ensure, that the blank value is stable. A stable blank value (typically between 30 and 90 ng L⁻¹) was maintained for at least five days, confirming the absence of microbiological contamination before sample loading. The sample rack was loaded with the polycarbonate samples when a constant BPA blank value was reached in the raining water reservoir.

Microbial contamination of the rainwater simulant must be prevented to avoid degradation or metabolisation of Bisphenol A during the weathering experiment, which would falsify the measurement results. To overcome this problem, cleaning of the weathering chamber and rainwater reservoir with antimicrobial agents was carried out before starting a new campaign. In addition, the rainwater simulant must be checked for microbial contamination each working day using a special test procedure. Special care must be taken because the water used to refill the chamber must be autoclaved or boiled to ensure that the chamber is not microbially contaminated.

Also, a loss of rainwater was caused by a high-temperature phase within the weathering scenario resulting in water evaporation. The water was refilled, and the loss was documented. The water loss averaged about 400 mL per day, which was specific to the conditions and the weathering chambers used. An evaporation factor was calculated for each day to correct the measured BPA concentrations.

Another critical point was the irradiation of the individual samples during weathering in the climatic chamber. These lamps have a spectrum between 295 and 400 nm with a maximum of 340 nm and an irradiation dose of about 45 W m⁻². In order to simulate only the spectral range relevant to the sun, an additional filter glass was installed which has a cut-off for all wavelengths smaller than 300 nm. It must be ensured that the irradiation dose was stable during the entire weathering period and that no decrease in the irradiation dose was caused by the ageing of the UV-A fluorescent tube lamps. To confirm this, a calibrated digital measuring head for UV radiation was used. This was inserted into the weathering chamber without opening it, and the radiation intensity was measured and documented. The spectral sensitivity of the measuring head meets the DIN 5032 and ISO/CIE 19476 standards [30,31]. For all experiments, a radiation of 45 W m⁻² at 300–400 nm was used. After a switch-on time of 4000 h, the complete set of fluorescent tube lamps was replaced.

The validation of the LC-MS/MS method for the quantification of bisphenol A in water samples was performed according to the standards DIN 32645 and DIN CEN/TS 16800 [32,33]. The collected rainwater samples and blanks are stored at +4 °C until the end of each week. They are mixed with isotopically labelled Bisphenol A (internal standard) and analysed using the validated LC-MS/MS method. To ensure the traceability and quality of the analytical results, a reference sample with a specific and certified concentration of Bisphenol A in water is measured on each working day. The results are documented in a control chart. A full QA plan can be found in the Supplementary Material (SI).

The weathering campaigns are time-consuming, as only one sample material can be weathered at a time and a campaign runs for at least six weeks. If errors arise during the experiment or are detected subsequently, the campaign must be repeated. Maintenance or repair of wearing parts of the climatic chamber leads to a delay of the planned campaigns. Due to the high cost of a weathering chamber and the high space requirement, the acquisition of several weathering chambers that could operate in parallel is economically difficult.