**3. Materials and Methods**

The investigated TiO2 thin films were deposited by Pulsed Laser Deposition (PLD) starting from a 99.9% pure TiO2 target ablated with Nd:YAG ns laser (Continuum) pulses (λ = 532 nm, pulse duration 5–7 ns, 10 Hz repetition rate, fluence on the target about 3.5 J/cm2) in a background O2 atmosphere (P= 8 Pa) and with a fixed target-to-substrate distance of 50 mm, at room temperature. The deposition time was adjusted to obtain a nominal thickness for all the films of about 1000 nm. Thermal annealing in air at 500 ◦C (4 ◦C/min heating ramp, dwell time 2 h) was employed to induce crystallization to the anatase phase. Crystallinity was verified by Raman spectroscopy (not shown).

Different distribution of Au NPs was achieved, depending on the samples (see Table 1). The TiO2/Au sample (i.e., the configuration substrate/Au NPs/TiO2 film) was obtained depositing a TiO2 film with the above described PLD procedure on a substrate covered with Au NPs obtained by Au thermal evaporation. An Edwards E306 thermal evaporator (Edwards) was employed to deposit a 10 nm Au film, followed by thermal de-wetting (air annealing at 500 ◦C with 4 ◦C/min ramp, 2 h dwell) to induce NP growth. The Au/TiO2 sample (i.e., the configuration substrate/TiO2 film/Au NPs) was obtained depositing a TiO2 film by PLD (details above), followed by deposition of Au NPs by PLD, ablating an Au target in 1000 Pa Ar with a laser fluence of 2 J/cm<sup>2</sup> [22]. The Au/TiO2/Au sample was obtained by combining the Au/TiO2 and the TiO2/Au synthesis procedures. The TiO2–Au

sample, in which Au NPs are embedded in the nanoporous TiO2 film, was deposited by PLD ablating a composite Au–TiO2 target with same ablation parameters as for the pure TiO2 film, followed by annealing in air at 500 ◦C (4 ◦C/min ramp, 2 h dwell) to induce TiO2 crystallization and Au NPs formation, as described in detail in [23].

All samples characterized by electrochemical techniques were deposited on Ti plate substrates, while selected samples for SEM and optical measurements were deposited on Si (100) and soda lime glass substrates, respectively.

Optical transmittance spectra were evaluated with a UV-vis-NIR PerkinElmer Lambda 1050 spectrophotometer (PerkinElmer, Waltham, Massachusetts, USA) with a 150 mm diameter integrating sphere in the range 250–2000 nm, illuminating the sample from the glass substrate side and normalizing the spectra with respect to glass transmittance. The haze was obtained as the ratio between the diffuse and the total (i.e., diffuse + direct) transmittance intensity.

SEM images of the synthesized samples were acquired using a field emission scanning electron microscope (FEG-SEM, Zeiss Supra 40, Carl Zeiss Microscopy GmbH, Jena, Germany); measurements were also performed, at the end of the experimental campaign, to check the stability of the samples.

Photo-electrochemical characterization of all the samples was performed in an undivided three-electrode cell; the synthesized materials were used as working electrodes, while platinum constituted the counter electrode and SCE was used as reference. The electrodes were connected to a potentiostat-galvanostat (Metrohm Autolab 302N, Metrohm, Herisau, Switzerland), controlled by NOVA software.

The photo-activity of the samples was tested under a light flux provided by a 300W Xe lamp (Lot Oriel, Darmstadt, Germay) equipped with an AM 1.5G filter.

Aqueous solution 0.1 M KNO3 was used as supporting electrolyte; depending on the experiments, a fixed amount of BPA (25, 50, or 100 ppm) was added to the electrolyte.

For all the cases, the photocurrent was calculated by subtracting the stable value measured in the dark from that obtained under irradiation. Open circuit voltage (OCV) was also monitored, always after 5 minutes of rest, to allow the sample reaching the equilibrium in the solution under the dark (OCVD) or irradiation (OCVL) condition imposed. Open Circuit Photopotential (OCP) was also evaluated as the difference between OCVL–OCVD in supporting electrolyte or in the presence of organic.

Chronoamperometric tests were performed at different overpotential from the OCP of each sample in both KNO3 and different concentrated BPA solutions.

Cyclic voltammograms (CV) were recorded in the potential window between −1.1 V and 1.8 V, at 100 mV/s.

Electrochemical impedance spectroscopy (EIS) were performed in the dark, at negative potential, to investigate on the ability of the samples to possible charge transfer process. The investigated frequency range was from 105 Hz to 10−<sup>1</sup> Hz.
