All chemicals were of analytical grade and used as received without further purification. Methanol, acetonitrile, and formic acid for UPLC-MS analysis were provided by Biosolve (Valkenswaard, Netherlands) whereas ultrapure water was obtained from a Merck MilliQ Integral 5 system (Merck, Darmstadt, Germany). All standards used are listed in Supplementary Table S1, with purity grade and supplier information.

Surface seawater was collected in the Bay of Marseille (France) in July 2021 using pre-combusted glass bottles, then filtered in the laboratory through a 0.22 μm polyethersulfone (PES) filter pre-rinsed with Milli-Q water and filled into 50 mL quartz tubes containing 1 g/L of the test material; i.e., 50 ± 2 mg per tube. The samples were exposed in laboratory conditions to artificial sunlight using a Suntest CPS + system (ATLAS) equipped with a daylight filter according to (Para et al., 2010). Irradiance intensity was set at 300 W/m² (300–800 nm) and temperature at 22°C ± 2°C, both representing environmental conditions as encountered in Marseille (France) during summer (reference data from July 2021: monthly mean solar radiation = 298 W/m², monthly mean surface seawater temperature = 22°C). The quartz tubes were hereby partially submerged in a basin connected to a Minichiller^® (Huber Kältemaschinenbau AG) inducing a constant water flow to ensure the maintenance of the desired temperature. In parallel, samples were kept under dark conditions by using the same experimental set-up and temperature, but in complete darkness. At each sampling timepoint, three replicates of UV exposed and dark condition samples were sacrificed, each. Hereby, 10 mL were collected for dissolved organic carbon (DOC) analysis, acidified (20 μL of 85% H₃PO₄ acid) and stored in glass vials in the dark at 4°C until analysis within a month by using a TOC-5000 total carbon analyzer (Sohrin and Sempéré, 2005). The natural seawater collected at the surface in the Bay of Marseille exhibited a DOC value of 76.8 (±16.4) µM, which is consistent with DOC concentrations previously measured in the same area (72.5–90 µM in summer-fall; Paluselli et al., 2018). This background DOC concentration from the seawater was subtracted from the DOC values observed in the leachates, to further investigate the DOC which derived from the test materials.

The remaining volume was stored in amber glass vials in the dark at 4°C until analysis of organic compounds using UPLC-TOF-MS (see section 2.5). Samples were collected at three sampling timepoints: after (t₁) 66 h, (t₂) 158 h and (t₃) 336 h, representing 6, 14.4 and 30.6 simulated days, respectively. Simulated days were calculated using the following equation from Gewert et al. (2018): S i m u l a t e d   d a y s = T o t a l   i r r a d i a n c e   e x p o s e d M e a n   E u r o p e a n   i r r a d i a n c e   x   365 where the total irradiance exposed is the product of the intensity (in W/m²) and the number of hours of exposure (in h) and the European mean irradiance equals 1,200 kWh/m² per year

Three types of test material were used for this study. The “CMTT” comprises a mix of tire tread from 20 different, new tires (see Supplementary Table S2), which were homogenized through a ball milling in cryogenic conditions (Mixer Mill MM 400, Retsch, Germany) and represents small particles as they can be found in road runoff. The size distribution and particle shape images are provided in Supplementary Figure S1. The two other test materials are made of crumb rubber granulate as used in artificial sports fields or playgrounds. The original material (hereafter called “virgin crumb rubber”, VCR) was sourced from a sports field in Tromsø (Norway) prior to application on the field and has previously been used for leaching and toxicity testing (Halsband et al., 2020).

Finally, part of the VCR was exposed in situ to natural seawater in Tromsø for 12–18 months (hereafter called “weathered crumb rubber”, WCR) to create a test material of high environmental relevance. For this, the particles were placed in a cylindric stainless-steel container covered with a mesh and submerged in surface waters (3–4 m depth). The outside of the container was mechanically cleaned at regular intervals to remove algal and bivalve growth and permit the sunlight to transpierce. While the choice of stainless steel limited the exposure of the contained particles to natural sunlight, it was considered as preferable to plastic materials to avoid a chemical contamination of the test material. The particles exposed for 12 and for 18 months were mixed together to have sufficient material to perform the experiment. After recovery from the sea, the WCR was not modified (i.e., no removal of the biofilm or other cleaning), but stored in dry conditions prior to use.

All glassware was heated (450°C, 6 h) prior to use and covered with burnt aluminum foil. Plastic and rubber materials were avoided at all times. All sample processing steps were conducted in an ISO class 6 clean room equipped for organic trace analysis. Cotton lab coats and nitrile gloves were worn during sample handling. Instrument and water media blank controls performed during the sample analysis were below LOD/LOQ levels or otherwise subtracted from the sample values.

Aqueous samples were analyzed for 19 tire-related chemicals (Supplementary Table S1) by ultra-performance liquid chromatography time-of-flight mass-spectrometry (LC-TOF-MS). The instrument consisted of an ACQUITY UPLC-System (HSS T3 column; 100 × 2.1 mm, 1.7 μm) coupled to a XEVO XS Q-TOF-MS (Waters GmbH, Eschborn, Deutschland) as described elsewhere (Klöckner et al., 2021; Klöckner et al., 2020). A flow rate of 0.45 μL/min, a column temperature of 45°C and a capillary voltage of 0.7 kV (positive ionization mode) were applied. The source temperature was set at 140°C, while the desolvation temperature was set at 550°C. The sampling cone voltage was 20 V and the source offset was 50 V. Nitrogen was used as a cone gas and argon as a collision gas. The desolvation gas flow was set at 950 L/h and the scan time at 0.15 s. A collision energy of 4 eV (molecular ions) and 15–35 eV (fragment analysis) was applied. The mobile phase consisted of A) water with 0.1% formic acid and B) methanol with 0.1% formic acid. The solvent gradient was as follows: 0 min 2% B, 12.25 min 99% B, 15.00 min 99% B, 15.10 min 2% B, 17.00 min 2% B. A volume of 10 µL of the leachates was directly injected into the LC-HRMS system and the first minute was forwarded to waste, not to the MS. The target substances (Supplementary Table S1) were quantified using an external calibration, which was comparable to seawater (checked separately). Therefore, TargetLynx was used with 0.01 Da mass accuracy and 0.1 min accurate retention time and a calibration range from 0.03 to 30 μg/L.

Identification of transformation products (TPs) was done by evaluating the RPLC-MS data in a retention time window of 1–10 min and a mass range of m/z 50–1,200. MarkerLynx was used to perform the peak picking for the molecular ion trace (function 1) with 0.1 min deviation in retention time and 0.01 Da deviation in exact mass. The marker table with relative intensities (to the total marker intensity) was exported to Excel and all further evaluations were performed there. The peaks with a relative intensity above 1 and present in the leachates compared to controls were selected. Sum formulas were assigned by using a mass tolerance of 5 ppm and an elemental composition of C (0–100), H (0–100), N (0–20), O (0–20), S (0–3) and Na (0–2), and by taking into account the fragment ions. The fragment ions were used, if available, to assign the elemental composition. For DPG related TPs the fragments were screened for those containing m/z 212.118 (C₁₃H₁₄N₃), 195.0922 (C₁₃H₁₁N₂), 131.0599 (C₈H₇N₂) and/or 119.06 (C₇H₇N₂), 106.0648 (C₇H₈N). DCH related TPs were identified by having the fragment m/z 180.1755 (C₁₂H₂₂N). The integration of the areas was done in TargetLynx.

All three test materials released DOC into the seawater in the range of 20–500 μg/g within 14 days (Figure 1). Under dark conditions CMTT released the highest amount of organic carbon (140 μg/g; Figure 1A), followed by WCR and VCR (23–28 μg/g). The stronger DOC release from CMTT may in part be due to a higher content of low molecular weight additives in that material (see section 3.2) and its significantly lower mean particle size (180 µm) compared to VCR and WCR (3 mm). A higher DOC concentration (316 μg/g) was previously reported for a freshly pulverized crumb rubber after 3 days of leaching, after which the concentration decreased (Selbes et al., 2015). The authors suggested that freshly generated particles would release more DOC; this may explain their higher values compared to the crumb rubber in this study.

It is noteworthy that the VCR released only slightly higher amounts of organic carbon (20%–25%) than the same material after being exposed to marine water for about 1 year. This suggests that the pool of organic compounds that can be released from crumb rubber is large and is not much depleted within 1 year. The release rate of organic compounds from the rubber into the surrounding water is determined by the (low) diffusion speed of these compounds from the interior of the particles towards the surface (Huntink, 2003). A part of the organic matter released from WCR may, however, also be natural organic matter adsorbed to it during the exposure period.

For CMTT there is a slight decrease in average DOC concentrations from day 7 to day 14 (Figure 1A), which may reflect the onset of biodegradation of some of the organic chemicals released from the material, as the experiments were not run under sterile conditions.

For all three materials parallel exposure to artificial sunlight drastically increased the release of organic matter over 14 days by a factor of 3–5 compared to the dark conditions (Figure 1.). Several hypotheses might be made to explain this observation: 1) The UV exposure could induce an aging and loosening of the rubber matrix, which would increase its accessibility for water and speed up leaching of organic compounds. 2) A sunlight induced transformation of tire constituents with a relatively lower tendency for leaching into TPs of lower molecular weight and higher polarity, with a stronger leaching tendency might occur. 3) While the temperature within the irradiation chamber was controlled (22°C ± 2°C), the black color of the particles might have led to a higher temperature of the particles themselves exposed to artificial sunlight, supporting the leaching compared to the dark conditions. A combination of these aspects as well as other factors may be responsible for the enhancing effect of sunlight exposure.

While the leaching of chemicals from tire material has been studied (Halsband et al., 2020; Tian et al., 2021; Jeong et al., 2022; Müller et al., 2022), the total amount of organic matter measured as DOC has been reported so far only for crumb rubber (Selbes et al., 2015) but not for a smaller dimensional size range of tire particles.

A total of 19 tire-related chemicals were quantified in the leachates based on LC-HRMS analysis (Supplementary Table S1). Among them were the vulcanization accelerators 2-mercaptobenzothiazole (2-MBT), benzothiazole (BT), benzothiazole-2-sulfonic acid (BTSA), 2-hydroxybenzothiazole (2-OHBT), 1,3-diphenylguanidine (DPG) and its degradation product phenylguanidine (PG), and dicyclohexylamine (DCH), as well as the antioxidant N-(1,3-dimethylbutyl)-N′-phenyl-p-phenylenediamine (6-PPD) and its hydrolysis and oxidation products 4-hydroxydiphenylamine (4-HDPA) and 6-PPD quinone (6-PPDQ), respectively.

For the 19 tire related chemicals quantified in the leachates (Supplementary Table S1), their contribution to the DOC amount was calculated, grouped into vulcanization accelerators (15 compounds, including thiazoles and guanidines) and antioxidants (4 phenylenediamine compounds) (Table 1). For all three tire materials, the explainable DOC is always significantly higher for the leachate generated in the dark than under sunlight exposure. This corresponds to the much higher DOC in the leachate with artificial sunlight. Obviously, sunlight exposure leads to the release of larger amounts of yet unknown compounds. sunlight exposure transforms parent chemicals into TPs, many of which are either unknown or not available as reference compounds and are, therefore, not considered in this balance. A typical example for this are the 23 TPs of DPG, which were detected by LC-HRMS screening in the sunlit CMTT leachate, but could not be quantified (Figure 3, Supplementary Table S4). The oxidative transformation of 6-PPD has been shown to lead to an even larger number of TPs (Seiwert et al., 2022), three of which are available as standards and have been included in this study.

In all three tire materials, vulcanization accelerators played a predominant role in the explainable DOC. On the contrary, antioxidants explain only 1% of DOC in most cases. This is not only because of the lower number of quantified antioxidant compounds, but also due to the fast degradability of antioxidants through hydrolysis and oxidation, which may lead to an extremely high number of possible TPs (Seiwert et al., 2022).

For the CMTT leachates (Table 1), the explainable DOC during the UV exposure was 10% in the first timepoint but then decreased to 7% as the amount of DOC significantly increased. In dark conditions, the explainable DOC corresponds to almost the double compared to the UV exposed samples. While the smaller particle size of the CMTT favors the release of all chemicals compared to the coarser VCR and WCR material, this effect can be expected to be stronger for chemicals with a lower diffusion coefficient, i.e., compounds of higher molecular weight, e.g., oligomers of HMMM and poly (1,2-dihydro-2,2,4-trimethylquinoline) (TMQ). These compounds were not included in the target list, which may explain the lower percentage of explained DOC for CMTT.

The WCR material, which has been weathered in marine water for about 1 year, may additionally release natural organic matter into the leachate, which was taken up during this long-time exposure to seawater. As a consequence of these different trends, the highest percentage of “explainable” DOC was obtained for VCR in the dark (55%) and the lowest percentages for CMTT and WCR under sunlight exposure (7%–8%).