*2.2. ELT Characterization and Preparation of Aqueous Suspensions*

In the 20 different solid samples of ELT-dg, 152 chemicals were preliminarily quantified (Table S1, see references to the methods followed for the determination of each chemical). In addition, the ultrastructure of both ELT-dg and ELT-dp was investigated using scanning electron microscopy (SEM). Selected materials were placed on aluminum stubs with a carbon tape, gold-coated and observed with an FE-SEM SIGMA (Zeiss, Oberkochen, Germany) operating at 5 kV with a working distance of 20 mm [46].

According to the concentrations of chemicals detected in the ELT-dg (Table S1) and considering the theoretical ecotoxicity of these substances/elements as identified by the European Union CLP (classification, labelling and packaging) Regulation (e.g., hazard statements H400, H410, H411, H412 and H413), we selected 20 chemicals (Table S2) with a potential ecotoxicological implication. First, to evaluate the effective detection of these substances/elements in the exposure media from ELT-dg suspension, we calculated their theoretical concentrations in a water fraction of 100.0 mg/L ELT-dg. This concentration used for the ecotoxicity tests was chosen according to the Organization for Economic Co-operation and Development (OECD) guidelines [47–50]. According to the theoretical results (Table S2), the majority of the 20 selected chemicals were not detectable in water (concentrations < 1 μg/L), except for Zn, Cobalt (Co), Copper (Cu), 2-mercaptobenzothiazole, 4-t-octylphenol, aniline and N-cyclohexyl-cyclohexanamine. To confirm this aspect experimentally, we prepared a suspension of 100.0 mg/L ELT-dp from the 20 pooled samples of ELT-dg in ultrapure water, mixing for 48 h and filtering through 0.45 μm filters. The ELT-dp was chosen, assuming that its homogenization was better than that of ELT-dg due to the smaller size of particles. Subsequently, we quantified the selected 20 chemicals in the water fraction (Table S2, see references to the methods followed for the determination of each chemical). The analytical results showed that Zn was the most abundant solubilized component (53.4 μg/L; coefficient of variation or relative standard deviation, RSD = 2.0%; Table S2). For this reason, Zn was chosen as the main solubilized and quantified component among the 20 selected chemicals with a possible ecotoxicological effect during the exposure.

The suspensions of ELT-dg and ELT-dp were prepared by mixing these materials in water and settling to separate the coarse components [51]. In particular, to identify the suitable mixing time, two suspensions of 100.0 mg/L of both ELT-dg and ELT-dp were mixed for 144 h in ultrapure water, and the concentration of Zn was assessed at 24, 48, 72 and 144 h. After 72 h of mixing, we obtained 80% Zn solubilization and decided to use this time to mix the suspensions for the preparation of test media. Regarding the settling time, two other suspensions of 100.0 mg/L of both ELT-dg and ELT-dp were mixed for 24 h and decanted for 144 h, and the total organic carbon (TOC; <3 mg/L, determination by cuvette test LCK 385, Hach), UV-VIS Spectrum (no significant absorption was observed in the UV-vis spectrum) and Zn concentrations were assessed. The quantification of Zn showed a stable concentration of this element after 48 h of settling, and this time was used for the suspension sedimentation. Lastly, considering that the majority of coarse ELT-dg and ELT-dp floated on the water surface and only a few amount of debris decanted, we prepared the tested media by serial dilutions of 100.0 mg/L, collecting the water from the center of suspensions using a peristaltic pump, avoiding the collection of coarse ELT-dg and ELT-dp from the surface or the bottom of used containers.

#### *2.3. Characterization of ELT Suspensions*

Because each model organism used in this study has a different exposure medium, we prepared the ELT suspensions separately for each exposure test. Zn concentration was measured during the entire exposures on one sample obtained on water mix from the different exposure tanks. For this reason, the standard deviation for these data was not calculated, and only the RSD, provided by instrumentation on 3 technical measurements, was reported. In detail, water samples were taken from the exposure tanks and acidified with 0.5% *v/v* nitric acid (HNO3), stored at room temperature in dark conditions and analyzed within 7 days. The analyses of Zn were performed using a plasma optical emission spectrometer (ICP/OES; Agilent model 5110). Zn in a multi-element standard solution at 100.0 mg/L was used as reference material to prepare diluted standard solutions. In addition, to evaluate the performance of the method, two aliquots of 40 mL water medium were spiked with Zn concentrations from 1.5 to 80 μg/L. The limit of quantification of the analytical method depended on the basal level of Zn in aqueous media and was 1.5 μg/L in both fish and *D. magna* acute test media, 2.3 μg/L in the algal medium and 13.0 μg/L in the *D. magna* reproduction medium (see the Supplementary Materials for medium composition).

To investigate the presence and size of ELT particles in the test media, we performed both SEM and dynamic light scattering (DLS) analyses. In detail, SEM analysis was applied to 100.0 mg/L of ELT-dg and ELT-dp suspensions, sampling the aliquots from the water column and avoiding the collection of floating or decanted coarse particles. Samples were placed on aluminum stubs with a carbon tape, gold-coated and observed with an FE-SEM SIGMA (Zeiss, Oberkochen, Germany) operating at 5 kV with a working distance of 20 mm [46]. Once the presence of particles in the 100 mg/L suspensions was ascertained, we also investigated this aspect with DLS measurements, using a Malvern Zetasizer Nano ZS instrument (Malvern Instruments Ltd., Malvern, UK) equipped with a solid-state He-Ne laser with a wavelength of 633 nm, collecting the scattered light at 173◦. The measurements on ELT suspensions were repeated 3 times using a disposable cuvette, leaving 30 s for temperature equilibration. The correlograms were fitted by a *non*-negative least square regression using multiexponential decay with Zetasizer Nano Series Software 7.02.
