**1. Introduction**

Environmental awareness has identified water scarcity as a problem of increasing magnitude in many areas due to its decisive and essential role in life. The increase in world population, changes in consumption patterns, the high demand for water in intensive irrigation agriculture or the frequent events of floods and droughts leads to the depletion of many water resources and the unequal distribution of water in different regions of the planet [1]. In this context, the development of novel analytical methods for the analysis of water and the improvement of existing ones reveal the presence of ECs in both drinking water and wastewater effluents. Some of these compounds may be toxic to terrestrial and aquatic organisms at low concentrations [2]. These compounds represent a scientific-technological challenge, since the existing plants have not been designed for their elimination.

On the other hand, the regulation foreseen for the elimination of this type of contaminants implies that the new facilities must face their efficient removal [3,4]. For example, the release of wastewater with pharmaceutical products and bacteria caused the increased resistance to amoxicillin/clavulanic acid in *Salmonella enterica* strains from 1% to 7–16% within 15 years (2003–2018). Furthermore, other compounds such as endocrine disrupting chemicals (EDCs) can be released into the environment, stored in the organisms due to their recalcitrant properties and stability, and accumulated in certain organs producing long-term effects [5]. In view of the above, it is necessary to ensure treatment systems that are capable of eliminating ECs in the different water matrices due to the adverse effects on human health and ecosystem [6].

Thus, advanced oxidation processes (AOPs) represent an alternative to conventional methods to remove these contaminants. AOPs are physicochemical processes that involve the generation of reactive oxygen species (ROS), which are effective against oxidation of organic matter because they have high oxidation potentials capable of reacting and degrading a wide range of contaminants [7]. In recent years, the use of heterogeneous photocatalysts has been intensively studied for their wide application for environmental protection, with special attention to wastewater treatment [8]. Some authors have studied the degradation of the ECs using different types of catalyst based on metallic oxides such as ZnO or TiO2 for the degradation of pharmaceuticals and personal care products (PPCPs) [9], pesticides [10] or industrial contaminants [7]. However, photocatalysis may be limited by costly energy requirements associated with the use of ultraviolet (UV) lamps, which also have limitations related to low quantum efficiency. Thus, photoactive materials are being developed, whose catalytic activity takes place in the optical window of visible light [11].

When semiconductor materials are used in photocatalysis, the photocatalyst is irradiated with light (hν) of equal or greater energy than its characteristic band gap, so that the electrons (e−) of the valence band (VB) are promoted to the conduction band (CB), thereby generating electron/hole (e−/h+) pairs. Other mechanistic aspects are based on two broad types of simultaneously occurring photochemical reactions on the surface of the catalyst. The first involves reduction, from the photo-induced negative electrons at the CB to the dissolved O2 present in the medium, producing superoxide ions (O2 −•), which can form hydroxyl radicals (OH•) and hydrogen peroxide (H2O2) in acidic medium. H2O2 may also decompose to OH• under irradiation. The second involves oxidation, from the photoinduced positive holes at the VB, which react with H2O or hydroxyl ions (OH−) to produce OH•. The active oxygen species O2 •− and OH•, and h<sup>+</sup> react with organic molecules, triggering their consequent degradation (Figure 1) [12].

**Figure 1.** General mechanism of semiconductor photocatalysis (**left**) and modification of energy levels with the incorporation of silver nanoclusters (**right**).

Among the different photocatalysts, ZnO has received considerable attention due to its exceptional optoelectronic properties, strong oxidation capability, abundance and physicochemical stability [8]. In this context and with a view to their application in wastewater treatment, the photocatalytic efficiency of ZnO nanoparticles (NPs) has been evaluated in the degradation of pharmaceutical and personal care products (PPCPs) [13] or dyes [14].

However, the photocatalytic efficiency of ZnO NPs is often limited by an inefficient absorption of visible light and a rapid recombination rate of e−/h<sup>+</sup> pairs. To avoid these disadvantages, the incorporation of metallic and non-metallic elements inside the crystal lattice of semiconductors reduces the recombination of electron-hole pairs. These compounds act by creating new energy levels between the CB and the VB that act as electron traps. Examples of doping with metallic silver or copper on ZnO or TiO2 were reported to improve photocatalytic efficiencies [7,15–19].

On the other hand, these drawbacks can be circumvented also by depositing noble metals on the surface of the catalyst, which trap electrons of the CB of the semiconductor, thus reducing the possibilities of e−/h<sup>+</sup> recombination and increasing the photocatalytic activity (Figure 1) [7]. In the search for novel materials capable of degrading organic pollutants in water under sunlight, ZnO NPs have been synthesized and functionalized with silver nanoclusters by a simple and green deposition method in water, conducted at ambient conditions.

Nanoclusters of metal elements are particles with low numbers of atoms, from 2 to ≤100 atoms, with sizes below ≈1.5–2 nm, and properties dramatically different from what would be expected from the scaling laws that govern the behavior of bulk and metal nanoparticles [20]. Nanoclusters of metal elements show the presence of discrete energy states and a sizable HOMO-LUMO bandgap, similar to the conduction band–valence band in semiconductors and lose the metallic behavior. This bandgap can be tuned by changing the number of atoms, the type of metal and the supporting material, and they can be used for different catalytic applications (heterogeneous catalysis, photocatalysis and electrocatalysis) [20–23].

In this article, the crystallinity, optical properties and morphology of the nanostructures obtained, ZnO–Ag, have been evaluated. Finally, the photocatalytic activity of ZnO–Ag nanocomposites with different Ag loadings has been studied in the removal of the dye Orange II (OII), used as model compound of organic pollutant, under UVA and white light.

#### **2. Results and Discussion**

### *2.1. Characterization of Catalysts*

A sample containing silver nanoclusters of ≤10 atoms was used for the deposition onto the ZnO NPs. These small nanoclusters show planar geometries, as it can be shown by atomic force microscopy (Figure S1), confirming the presence of nanoclusters of ≤10 atoms [24].

Different Ag loads were applied on the surface of the ZnO NPs, so that four types of ZnO–Ag NCs with an Ag content of 1.3, 2.9, 3.2 and 7.4% (*w*/*w*) were obtained. The samples were structurally characterized by X-ray diffraction patterns. The two crystalline phases present in the samples were metallic silver (Ag, JCPDS PDF-2 card number 04-0783, peaks highlighted with red down-pointing triangles in Figure 2) and zincite (ZnO, JCPDS PDF-2 card number 36-1451, peaks highlighted with black up-pointing triangles in Figure 2) with hexagonal wurtzite structure.

No additional peaks were observed in the patterns, revealing the absence of impurity phases in the catalyst. Furthermore, there was no significant shift of the diffraction peaks, proving that silver atoms did not substitute any Zn sites in the lattice but were deposited onto the surface of ZnO. Figure 3 shows the morphology of the ZnO NPs (left) and ZnO–Ag NCs (right) observed by FE-SEM. ZnO NPs are present in the form of spherical aggregates of different sizes, between 50 and 500 nm. These aggregates are composed of smaller ZnO NPs (10–15 nm). In the case of the ZnO–Ag NCs, the presence of separate Ag nanoparticles along with the spherical aggregates is shown in Figure 3 (right).

**Figure 2.** X-ray diffraction patterns of ZnO NPs and hybrid ZnO–Ag NCs. The main reflections from zincite (ZnO, JCPDS PDF-2 card number 36-1451) and metallic silver (Ag, JCPDS PDF-2 card number 04-0783) are included as drop lines.

**Figure 3.** Scanning electron micrograph of the spherical ZnO aggregates (**left**); Detailed scanning electron micrograph of the ZnO–Ag nanocomposites, showing the presence of small Ag nanoparticles as brighter spots (**right**).

#### *2.2. Influence of Ag Nanoclusters Loading onto Photocatalytic Activity of ZnO*

Experiments were performed using fixed concentrations of a photocatalyst (200 mg L<sup>−</sup>1) and OII (10 mg L−1) in 10 mL of aqueous solutions, under UVA light for 60 min or white light for 180 min (Figure 4). In the absence of photocatalysts, photolysis controls resulted in OII degradation of 9% under UVA and 5% under white light, while adsorption studies of samples kept in dark conditions showed no OII removal (data not shown). When comparing the decolorization results, using ZnO nanoparticles as photocatalyst, no significant improvement in dye removal was observed, with maximum percentages of 16% and 9% under UVA and white light, respectively. As a general rule for all the experiments, OII degradation exhibited accelerated kinetic rates under UVA irradiation, which is attributed to a strong light absorption of these wavelengths by ZnO, while ZnO absorption in the visible region is weaker. It can be noted that the photocatalytic performance of the ZnO NPs improved with the addition of Ag nanoclusters, obtaining 97% and 49% of OII removal in the presence of ZnO–Ag with 1.3% (*w*/*w*) using UVA or white light, respectively. Photocatalytic performance gradually decreased with increasing Ag loads in the ZnO NPs. Therefore, there is evidence of the existence of an optimal silver loading to enhance the photocatalytic activity of the NC. This can be explained by the specific surface of ZnO available to interact with incident light, being lower with increasing concentrations of Ag nanoclusters in the NCs [25]. In fact, the decoration of the ZnO NPs with Ag nanoclusters leads

to a color modification, from white to brownish and grey, due to the formation of Ag nanoparticles, which improves the absorption of the ZnO–Ag NCs in the visible region (Figure 4).

**Figure 4.** Silver loading effect on photocatalytic performance. The values in brackets correspond to the percentage of Ag in each NC (**left**); Aqueous suspensions of ZnO-NPs and ZnO–Ag NCs with different silver loadings (**right**).

However, according to previous findings, above the optimum Ag loading effectively deposited onto the NC, this can lead to an enhancement of the e−/h<sup>+</sup> recombination rate, acting itself as a recombination center, thus contributing to a decrease in photocatalytic efficiency [26].
