*2.1. Materials and Methods*

Sodium citrate (Na3C6H5O7, Cit), L-Cysteine (C3H7NO2S, L-cys), silver nitrate (AgNO3) and sodium borohydride (NaBH4) have been used as received (reagent grade, Sigma-Aldrich, St. Louis, MO, USA). Metal ions contamination was accomplished by using the following salts:

NaAsO2, NaHAsO4·7H2O, Ca(ClO4)2, Cd(NO3)2, CoCl2·6H2O, CrCl3·6H2O, Cu(NO3)2, FeCl3·6H2O, Hg(NO3)2·H2O, KClO4, Mg(ClO4)2, NaClO4, NdCl3·6H2O, NiCl2·6H2O, Pb(NO3)2, Zn(NO3)2·6H2O. For all the solutions, we used deionized water (electrical conductivity less than 1 μΩ/cm at room temperature) obtained from Millipore Milli-Q water purification system. All the reagents were purchased (from Sigma-Aldrich, St. Louis, MO, USA) and were used without further purification.

UV-Vis spectra were run in H2O solution by using quartz cells with a Shimadzu 2401 PC UV-vis spectrophotometer and by using single-use UV-PMMA cuvettes with Perkin-Elmer Lambda 19 UV-Vis-NIR for sensing test characterization. ATR-FTIR spectra have been recorded for films deposited by casting from water suspension with an FTIR spectrometer (Nicolet iS50, Thermo Fisher Scientific, Madison, WI, USA) equipped with a mid- and far-IR capable diamond ATR accessory. FT-IR spectra were recorded in the range between 350 and 4000 cm<sup>−</sup><sup>1</sup> with a resolution of 4 cm<sup>−</sup>1, a zero-filling factor of 2 and the co-addition of 64 scans. Data were acquired using OMNIC software (version 9.8.372, Thermo Scientific), subtracting the air background spectrum obtained prior to each sample spectrum acquisition. AgNPs were investigated by TEM (Philips Morgagni 268 D electronics, at 80 KV and equipped with a MegaView II CCD camera, Philips Electronics, Eindhoven, The Netherlands) at 10 mg/<sup>L</sup> in MilliQ water. The obtained images were analysed with *ImageJ* for particles size measurement. A total of 180 particles were measured in two portions of two different images and the average size was calculated. Size distribution of AgNPs (50 mg/L) have been investigated by means of DLS (Zetasizer Nano Series, Malvern instruments, Enigma Business Park, Grovewood Rd, UK), combined with the Zetasizer Nano Series software (version 7.02, Particular Sciences) at T = 25.0 ± 0.2 ◦C in milliQ water, as well as media used for toxicity assessment (freshwater, TG 201, and marine water, F/2). Correlation data have been acquired and fitted in analogy to our previous work [55,56]. Ag+ release from the AgNPs has been assessed in both TG 201 and F/2 aqueous media, at 0 h and after 72 h. Six solutions have been prepared: TG 201, F/2, TG 201 + 500 μg AgNPs/L, F/2 + 500 μg AgNPs/L, TG 201 + 7 μg AgNO3/L, F/2 + 7 μg AgNO3/L. Solutions were kept in the same conditions as toxicity tests (22 ± 2 ◦C and 16/8 light-dark photoperiod) and mixed by shaking once a day. An aliquot of each solution was taken at 0 h and after 72 h and centrifuged (5000 g, 40 min, 22 ◦C) by using a centrifugal filter device with a 3 kDa cut-off (Amicon Ultra-15 mL, Millipore, Sigma-Aldrich, St. Louis, MO, USA). The resulting filtrate was acidified with HNO3 (10%) and analysed by ICP-MS (Perkin Elmer NexION 350 spectrometer) for determining Ag concentration.

SR-XPS measurements were performed at the Elettra synchrotron radiation source (Trieste, Italy), using the Materials Science Beamline (MSB), that is positioned at the left end of the bending magne<sup>t</sup> 6.1. MSB is equipped with a plane grating monochromator providing SR light in the 21–1000 eV energy range. The UHV endstation, whose base pressure is of 2 × 10−<sup>10</sup> mbar, is equipped with a SPECS PHOIBOS 150 hemispherical electron analyser, low-energy electron diffraction optics, a dual-anode Mg/Al X-ray source, an ion gun, a sample manipulator with a K-type thermocouple attached to the rear side of the sample. For the here presented experiments we detected photoelectrons emitted by C1s, O1s, S2p, Ag3d, N1s and Hg4f core levels, using a normal emission geometry. In order to maximize signals intensity, we selected a photon energy value of 630 eV (impinging at 60◦) for all elements

except S; in order to maximize the intensity of S2p signals, that was expected to be very low due to element dilution, the S2p core level was measured with photon energy = 350 eV. Charging correction of binding energies (BEs) was done using as a reference the aliphatic C1s (BE 285.0 eV) [57]. To fit core level spectra, we subtracted a Shirley background and then used Gaussian peak functions as signals components [58,59].

NEXAFS experiments were carried out at the BEAR (Bending magne<sup>t</sup> for Emission Absorption and Reflectivity) beamline, installed at the left exit of the 8.1 bending magne<sup>t</sup> and located the ELETTRA third generation storage ring. The beamline optics can deliver photons having energy comprised between 5 eV and 1600 eV with selectable degree of ellipticity. The carbon K-edge spectra were recorded at normal (90◦) and grazing (20◦) incidence relative to the sample surface of the linearly polarized photon beam; however, no angular e ffects were detected. Calibration of the photon energy and resolution was carried out at the K absorption edges of Ar, N2 and Ne. In order to normalize the spectra, a straight line fitting the part of the spectrum below the edge was subtracted and the value recorded at 330.00 eV was assessed to 1.
