Two marine bacterial strains, V. harveyi CECT (Spanish Collection of Type Cultures) 5156 and C. lytica CECT 5014 were selected for this study since both species are recognized as early colonizers in the fouling process and are natural inhabitants of the Mediterranean Sea (Karunasagar et al., 1996; Kientz et al., 2016). Bacteria were routinely grown in Marine Broth (MB), or Tripticase Soja Broth (TSB) supplemented with 2.5% of NaCl (TSB-3) at 28 °C for 24 h (V.harveyi), or 48 h (C. lytica). Bacteria were stored at −80°C in MB supplemented with 20% (v/v) glycerol.

The benthic microalgae Nitzschia ovalis BEA (Spanish Bank of Algae) 0423B, selected for this study was routinely grown in flasks containing 100 ml of natural seawater supplemented with Guillard f/2 medium (Guillard, 1975) inoculated with 1 x 10⁵ cells/ml of the microalgae in pre-stationary phase and supplemented with sea water. Cultures were incubated for 7 days in climatic chambers at 21°C under a light intensity of 100 µmol photons m²/s.

High Density Polyethylene (HDPE) was functionalized with Cu₂O, CuPT, CuSCN and/or silica (biocides registered for use in the European Union). For this purpose, mixes were blended and extruded through a co-rotative twin-screw extruder. Once the compounds were obtained, sheets with a thickness of 300 µm were extruded using an L/D 30 single-screw extruder and a 100 mm wide flat head. These sheets were cut into 12 mm diameter discs or 10x22 cm sheets, as required for the assays. Sterilization of 12 mm discs was carried out by UV radiation in a laminar flow cabinet, with 2 hours of radiation total time (1 hour on each side). Rectangular sheets were washed with 96% ethanol before immersion in seawater tanks for biofilm formation assays.

Different sets of functionalized HDPE were prepared, with the compounds/additives separately/individually or combined at different concentrations ( Table 1 ) to carry out the tests on a laboratory or pilot scale.

The morphology of the functionalized HDPE was microscopically characterized by a scanning electron microscopy (SEM) (Hitachi FE-SEM S-4800) with an accelerated voltage of 10 kV. Samples were attached to a metal sample holder using double-sided adhesive tape with carbon and were coated with gold-palladium (Au-Pd) by sputtering to make them electrically conductive.

Biofilm formation assays were performed with selected bacteria at laboratory and pilot scale.

Laboratory scale. The biofilm formed onto the surface of the discs was quantified by a modified crystal violet (CV) staining method (Niu and Gilbert, 2004). Sterilized 12 mm discs were glued to the bottom of 24-wells cell culture plates. Briefly, overnight cultures in modified artificial seawater basal media [50 mM MgSO₄, 10 mM CaCl₂, 300 mM NaCl, 10 mM KCl, 0.0058% (wt/vol) K₂HPO₄, 10 μM FeSO₄, 50 mM Tris-HCl (pH 7.5)] supplemented with 0.3% glycerol as carbon source (ASWgly) (Wang et al., 2010) were used to inoculate fresh ASWgly to a concentration of 1 x 10⁶ cells/ml (V. harveyi) or 1 x 10⁸ cells/ml (C. lytica). Then, cultures were placed into a 24-well cell culture plate (3 wells per strain and treatment) and left for static growth at 30°C 24 or 48 h for V.harveyi or C. lytica, respectively. After incubation, bacterial growth was measured at A595 before carefully removing the media by aspiration and staining the biofilm with CV (1 ml/well) for 30 min. Wells were rinsed with PBS (Phosphate Buffered Saline, pH 7) (1 ml/well) to remove the excess of CV stain until the non-disc control did not show visible traces of the dye and the violet stain was solubilized with 1 ml of 96% ethanol. All the experiments were performed in triplicate. The OD540 of the solubilized biofilm was measured and quantification of biofilm formation was calculated with the formula:

where AB is the OD540 of stained attached bacteria, CW is the OD540 or OD595 of stained control wells with no bacteria and CG is the OD595 of cells growth in suspended culture (Niu and Gilbert, 2004; Naves et al., 2008).

Pilot scale. The experiments were carried out in the aquarium facilities of the SCSIE (Central Support Service for Experimental Research) of the University of Valencia. The environmental conditions (temperature, humidity, etc.) were monitored daily. Assays were performed in 200 l tanks containing 100 l of natural seawater UV-sterilized. Seawater mesocosms were inoculated with an overnight washed culture in TSB-3 to a final concentration of 10⁵ CFU/ml (dilution 1/1000). Sheets were submerged in the inoculated seawater and incubated for 14 days. Bacteria population was monitored periodically by plate counting onto Marine Agar (MA).

Biofilm formation was determined with the CV staining protocol similar to the one described before. Briefly, sheets with antifouling coatings were taken out of the tanks and placed in a vessel containing 0.1% CV solution for 30 min. Then, sheets were rinsed twice with PBS and air dried for 30 min. Afterwards, 6 cm² strips were cut and placed in centrifuge plastic tubes containing 5 ml of absolute ethanol. The amount of biofilm was estimated through the CV resuspended in the ethanol. Three different experiments were performed and two strips for each treatment were recorded.

The algae biofilm assay was performed in 30 ml screw-tubes filled with 9 ml of sterile seawater supplemented with Guillard f/2 medium plus 1 ml of a culture of N. ovalis. Two discs were placed in each tube (one per treatment) and incubated 7 days at 22°C under a light intensity of 100 µmol photons m²/s.

Biomass adhered at the disc was indirectly measured by a colorimetric quantification of total carbohydrates using the phenol-sulfuric acid method, using glucose for the standard curve (Fournier, 2001). Basically, the two discs for treatment were washed with sterile seawater supplemented with Guillard f/2 medium and inserted into a new tube containing 500 µl of 4% phenol and the biomass was resuspended by agitation with the vortex. Later, 2.5 ml of sulfuric acid was added, and the mix was incubated for 5-10 min. The OD490 was measured, and ng of carbohydrates were calculated using the standard curve. Three different experiments were performed for every treatment studied.

Production of the leachate on functionalized HDPE sheets. For this purpose, glass tanks (30 cm height x 15 cm width x 15 cm depth) were filled with 5 l of synthetic marine water generated with Instant Ocean™ sea salts and dechlorinated tap water (final salinity 37 g/l). HDPE sheets (22 x 22 cm) were wrapped forming a cylinder, then a 36 Watt waterproof UV-C germicide lamp (JEBO^®, 20 cm long x 10 cm in diameter) was placed inside ( Figure 1 ). In order to avoid the water stratification and to assure a constant water movement, a 3.8-Watt (Sicce™) circulation pump generating a flow rate of 300 l/h was fixed in the tank bottom. The water temperature was set at 35°C. The total exposure time was 60 d. Samples were taken at 0, 30 and 60 days. Evaporation was checked twice a week and lost water replaced with deionized water.

The ecotoxicological effect of the HDPE formulations was assessed using a battery of three bioassays based on standardized ISO guidelines.

Assays with the bacteria Aliivibrio fischeri. The ISO 11348-3 standard (ISO, 2022a) is based in the determination of the inhibitory effect of water samples on the light emission of A. fischeri (Luminescent bacteria) after exposure time of 15 min using freeze-dried bacteria (Zhang et al., 2020). The procedure was conducted according to the Biotox-STD^® kit user´s manual of EPBI™ implemented in a handheld PD10 luminometer (Kikkoman™).

Acute toxicity test to the marine rotifer Brachionus plicatilis. The ISO 19820 specifies a method for the determination of the lethal effects of toxicants to B. plicatilis after 24 h or 48 h exposure (Jośko and Oleszczuk, 2012). The Rotoxkit M™ bioassay (MicroBioTests Inc.) was used in the study. Neonates of B. plicatilis were hatched from cysts 24 h before the assay. Mortality of organisms was checked after 24 h and 48 h of exposure to the leachates.

Marine algal growth inhibition test with Phaeodactylum tricornutum. The bioassay was conducted following the ISO 10253 guideline (ISO, 2022b), using the commercially available Algaltoxkit Marine™ (Microbiotests Inc.). The algal growth rate (r) was monitored during 24, 48 and 72 h in a spectrophotometric way at OD₆₇₀ by means of 2021 Aurius™ spectrophotometer (CECIL Instruments™).

Bioassay data analysis. For the three bioassays, the 100%, 50%, 25%, 12.5% and 6.25% from the original leachate were assayed. According to ISO guidelines, all tests were conducted in triplicate. A probit regression (Litchfield and Wilcoxon, 1949) was used to determine the 50% Effective Concentration (EC₅₀) in each bioassay: 72 h ECr₅₀ in the algal bioassay, 15´ EC₅₀ for the bacteria and 24-48 h EC₅₀ in the rotifer test.

The 50% effect endpoint values (in % of dilution) were converted into toxic units (TU) according to the formula TU = 100/EC₅₀, 100/ECr₅₀ or 100/LC₅₀ for bacteria, algae, and rotifers, respectively (Sprague and Ramsay, 2011).

Quality assurance. The quality of the results obtained in the bioassays was verified by means of positive controls. A. fischeri, P. tricornutum and B. plicatilis were exposed to a standard toxic in order to assure their sensitivity. For this purpose, potassium dichromate (K₂Cr₂O₇) was used as toxic reference according to ISO guidelines.

The presence of metals in elutriates after the exposure to UV-C irradiation was determined by conducting chemical trace analysis. After each exposure time (30 d and 60 d) water aliquots were taken from each tank and copper (Cu) analyzed in a single quadrupole inductively coupled plasma mass spectrometry instrument (model 7900 ICP-MSAgilent Technologies™). Analyses were performed in the SCSIE.

Bacterial biofilm inhibition, algae colonization and ecotoxicological assays were statistically analyzed by ANOVA tests (p< 0.05). Significant differences between the control and tested groups were detected using Dunett’s test pairwise comparison method, with an α error of 0.05, obtained using RStudio version 2021.9.0.351 (RStudio Team, 2015).

The surface morphology of the different samples was observed with SEM. Micrographs of the HDPE discs are shown in Figure 2 . No differences were observed between samples in the structure and morphology of their membranes.

In a first laboratory approach, the biofilm forming on the surface of the HDPE discs by the selected bacteria was investigated using a crystal violet assay and the BFI was calculated. In both bacterial species, CuPT-functionalized HDPE inhibited biofilm production ( Figure 3 ), reducing it by between 12 and 40% with respect to the control ( Table S1 ). By contrast, HDPE functionalization with CuSCN enhanced biofilm formation, mainly by V. harveyi, under test conditions, whereas no significant changes in biofilm formation were observed on discs treated with silica or Cu₂O.

The second approach was to assess biofilm formation at pilot scale, which also included combined treatments of HDPE. For this purpose, biofilm formation was investigated after 14 days in 100 l seawater mesocosms contaminated with bacterial populations, simulating natural environments (Pujalte et al., 1999). The number of bacteria in all these trials decreased slightly over time. For V. harveyi, the bacterial counts changed over the experimental period from 4.3x10⁵ to 2x10⁵ CFU/ml, and in the case of C. lytica from 7x10⁵ to 9x10⁴ CFU/ml).

We studied treatments with CuPT, (the most efficient additive in our laboratory assays) at a low concentration (2%), combined with Cu₂O or silica, as well as a triple combination. We observed that CuPT was the only treatment that produced a notable reduction (at least 30%) in bacterial biofilm formation ( Figure 4 , Table S2 ), although this effect was not statistically significant. As occurred in the laboratory approach using CuSCN, the double combinations studied in mesocosms enhanced biofilm formation by both bacterial species under test conditions.

Algae biomass formed by the selected diatomic algae was investigated after 7 days of incubation in suitable conditions. Total carbohydrates adhering to each type of functionalized HDPE were quantified and are represented in Figure 5 . The results show that all the additives studied were effective in reducing biofilm formation by algae. The highest inhibition levels were observed for CuPT- or silica-treated HDPE (% reduction from 35 to 95 with respect to untreated HDPE) ( Figure 5 ; Table S2 ).

To evaluate the environmental impact, functionalized HDPE discs ( Table 2 ) were subjected to a leaching test. When Cu₂O or CuPT were added to HDPE, the amounts of copper released into the water registered low values of between 15 and > 130 µg/l of Cu ( Table 2 ).

The leachates were assessed in a ecotoxicological study following a common methodology of evaluation at different levels of the trophic chain in the marine environment. Decomposer (bacteria), primary producer (algae) and consumer (rotifers) were selected. The toxicity of the leachates generated after 30 and 60 days was expressed in toxic units (TU) ( Table 2 ). No toxic effects were found in any of the bioassays at the beginning of the experiment (0 day-leachate), i.e., in the freshly prepared synthetic seawater.

The ANOVA (p< 0.05) showed significant differences in toxicity values between the 30-day and 60-day leaching times for SiO₂-3 with respect to B. plicatilis, as well as in the assays with A. fischeri. However, no statistically significant differences were observed in the P. tricornutum assays when comparing the control with the other functionalized material.

Except for HDPE-SiO₂-3 in the 60-day leachate assayed with rotifers (TU: 43.5 ± 2.7), none of the generated leachates exhibited toxicity. Assays with the toxic standard (K₂Cr₂O₇) showed that the sensitivity of the organisms was in agree with those values stablised by the ISO guidelines. The EC₅₀ for the toxic standard for A. fischeri, P. tricornutum and B. plicatilis were 51.2 mg/l, 17.9 mg/l and 324.3 mg/l, respectively.