Photocatalytic degradation of methylene blue at nanostructured ZnO thin films

The photocatalytic degradation of the wastewater dye pollutant methylene blue (MB) at ZnO nanostructured porous thin films, deposited by direct current reactive magnetron sputtering on Si substrates, was studied. It was observed that over 4 photocatalytic cycles (0.3 mg · l−1 MB solution, 540 minUV irradiation), the rate constant k of MB degradation decreased by ∼50%, varying in the range (1.54 ÷ 0.78) · 10–9 (mol·l−1·min−1). For a deeper analysis of the photodegradation mechanism, detailed information on the nanostructured ZnO surface morphology and local surface and subsurface chemistry (nonstoichiometry) were obtained by using scanning electron microscopy (SEM) and x-ray photoelectron spectroscopy (XPS) as complementary analytical methods. The SEM studies revealed that at the surface of the nanostructured ZnO thin films a coral reef structure containing polycrystalline coral dendrites is present, and that, after the photocatalytic experiments, the sizes of individual crystallites increased, varying in the range 43 ÷ 76 nm for the longer axis, and in the range 28 ÷ 58 nm for the shorter axis. In turn, the XPS studies showed a slight non-stoichiometry, mainly defined by the relative [O]/[Zn] concentration of ca. 1.4, whereas [C]/[Zn] was ca. 1.2, both before and after the photocatalytic experiments. This phenomenon was directly related to the presence of superficial ZnO lattice oxygen atoms that can participate in the oxidation of the adsorbed MB molecules, as well as to the presence of surface hydroxyl groups acting as hole-acceptors to produce OH· radicals, which can be responsible for the generation of superoxide ions. In addition, after experiments, the XPS measurements revealed the presence of carboxyl and carbonyl functional groups, ascribable to the oxidation by-products formed during the photodegradation of MB.


Introduction
In recent years, due to population growth and worldwide industrial development, the contamination of surface and groundwater has quickly increased. One of the main sources of surface and groundwater contamination is organic dyes: due to their high toxicity together with their nonbiodegradability, they cause not only several environmental hazards but also health-related issues [1].
In this context, one of the main research challenges is the removal of hazardous dyes from wastewater by using treatments that do not present a serious undesired impact on the environment on one hand, and are relatively simple and not too expensive, on the other hand. One of the most common hazardous dyes in wastewater is methylene blue (tetramethylthionine chloride, MB; for chemical structure, refer to figure S1), which is used, among different fields, in the textile industry, as a dye for wool, silk, leather and cotton [2]. However, during the dyeing process, a large portion (up to 15%) [3] is normally transferred into the wastewater. Unfortunately, MB can directly retard the biological activity of various aquatic plants and animals including fishes and other organisms in the surrounding environment [4].
So far, various methodologies have been implemented for wastewater treatment to remove hazardous dyes by using different approaches, which can be divided into two groups, i.e. separation methods, which can include physical and physico-chemical means, and degradation methods, in which biological and chemical processes are involved, as reviewed by Nidheesh et al [5]. Among the separation methods, one can distinguish adsorption processes based on solid adsorbents [6,7], coagulation and flotation [8], and membrane filtration [9]. In turn, within the group of degradation methods, one can distinguish several methods such as, just to cite a few, biodegradation [10], chemical catalytic reactions combined with biological oxidations [11], and advanced oxidation processes (AOPs) [12,13]. Most of the mentioned techniques are associated with various drawbacks, such as the production of toxic by-products, high costs, limited recovery, or high energy consumption [14]. Out of the AOPs, photocatalysis, which makes use of semiconductor oxide photocatalysts to perform pollutant photodegradation in wastewater [14,15], is widely used because it can reduce the concentration of hazardous organic compounds in water. In general, the photocatalytic method consists of the light-activated (UV or visible radiation) production of radical species able to oxidize the pollutants dissolved in water. The entire process can be schematized in the following six steps (see figure 1.): (i) the photocatalyst is activated via the absorption of light, which induces the formation of electron-hole pairs. In turn, these react with water and oxygen dissolved in water to produce reactive oxygen species such as superoxide ions (O 2 -) and hydroxyl radicals (OH·); (ii) the organic pollutant diffuses from the liquid phase to the surface of the photocatalyst, where (iii) it is adsorbed; (iv) the photocatalytic reaction between the adsorbate and the radical species takes place; (v) the products of the photocatalytic reaction are desorbed, and subsequently (vi) they diffuse away from the photocatalyst surface.
Heterogeneous photocatalysis has been proved as a promising method for the degradation of organic pollutants contained in wastewater. The process, as a means of removal of persistent water contaminants such as dyes, pharmaceuticals, and pesticides, has attracted the attention of many researchers in recent years [16][17][18][19].
The main advantage of photocatalysis is related to the fact that ideally, after the pollutant degradation, harmless final products such as CO 2 , H 2 O, and inorganic salts are obtained. This aspect is extremely important with regard to the undesired potential impact of photocatalysis on the environment [20].
Zinc oxide (ZnO) has been used on a large scale as an effective, inexpensive semiconductor photocatalyst for the degradation of a wide range of organic chemicals [31][32][33].
Moreover, it is common knowledge that ZnO is more efficient than TiO 2 for the degradation of water pollutants, despite the fact that both photocatalysts present a similar band gap energy (i.e. ∼3.2 eV) [34]. This is related to the fact that ZnO exhibits a large room temperature exciton (photoinduced electron-hole pairs) binding energy of 60 meV, a high electrical conductivity (∼10 2 Ω −1 ·cm −1 ), and lower electric potentials of the valence and conduction bands (−0.45 to 2.75 eV, respectively) in comparison to TiO 2 (−0.1 to 3.1 eV, respectively). All these properties can explain why, for ZnO, more effective photocatalytic degradation reactions are observed at lower band potentials [35,36].
It should be underlined that the ZnO band gap limits its photocatalytic activity mainly to the UV region (only 3%-4% of solar light). In order to extend its photocatalytic activity also in the visible region (∼44% of solar light) [36], one strategy is the modification of ZnO morphology to obtain various forms of ZnO nanostructures at the micro-and nanoscale, which provide also more active sites for the photocatalytic reactions [37,38].
As a general trend, the size of the crystallites has a primary influence on ZnO photocatalytic activity: nanometrescale crystallites exhibit better performance than submicron ones, which in turn outperform micron-scale ones [39]. This behavior can be explained in terms of specific surface area, which is higher for nanoscale materials, providing a higher number of active sites. On the other hand, the increase of defect concentration induced by the size decrease favors electron-hole recombination processes, resulting in low photocatalytic activities of ultrasmall nanocrystallites [40][41][42][43][44]. An optimum crystallite size that maximises the photocatalytic activity is thus generally observed. As far as morphology is concerned, recent studies indicate that the ZnO (0 0 1) facet accelerates the production of OH· radicals, which are one of the active species responsible for the photocatalytic degradation of dyes [43].
Among the ZnO two-dimensional nanostructures used as photocatalysts, apart from nanoparticles [52][53][54] and nanopowders [55], mainly ZnO thin films are preferred for photocatalytic applications: their use can facilitate the photocatalyst separation process, and the recycling and regeneration steps, which can be cumbersome with non-supported photocatalysts. These thin films are usually prepared by different deposition methods, such as magnetron sputtering [56,57], spray pyrolysis [58,59], sol-gel process [60,61], as well as combined sonochemical and hydrothermal deposition.
Despite the great amount of dedicated research, the mechanism of the photocatalytic degradation of dyes using the various forms of low dimensional ZnO nanostructures remains mostly unknown [71][72][73]. This is probably related to the fact that the techniques commonly used for the characterization of these nanomaterials such as XRD and SEM are not fully surface-sensitive in character.
To achieve efficient photocatalytic degradation, efficient adsorption at the photocatalyst surface is fundamental. This is because adsorption processes take place at the surface and within the subsurface region of semiconductor photocatalysts. These processes directly cause a charge redistribution within the surface space-charge region related to Debye length L D , and induce a surface band bending effect playing a crucial role in the specific surface conduction mechanism. This effect is also observed in the various nanoforms of ZnO, for which the carrier concentration reaches the value of ∼10 18 [cm −3 ], and usually, a surface depletion region with common upward band bending is observed at the depth of 1 nm [74]. Thus, it is evident that, apart from the information concerning the local surface crystallinity and morphology, the detailed information on local surface and subsurface chemistry of the selected low dimensional ZnO nanostructures, including mainly their stoichiometry (combined with the undesired surface contaminants), should be also taken into account for a deeper analysis of the mechanism of photocatalytic degradation of pollutants at their surface.
For this purpose, one of the best methods is x-ray photoemission spectroscopy (XPS), whose information depth is comparable to the Debye length L D . This method has been successfully used for the study, among others, of the surface chemistry of the nanostructured ZnO porous thin films [74,75], exhibiting the highest extension of the specific internal surface of ZnO thin films.
Here, we propose an x-ray photoelectron spectroscopy (XPS)-based analysis for the assessment of photocatalytic degradation products of MB and to study the surface modifications of nanostructured ZnO thin films following the photocatalytic process, on the basis of precise determination of the surface chemical properties of our samples. In addition, the surface morphological properties of our samples were also determined by using scanning electron microscopy (SEM) as a complementary surface analytical method

Chemicals
Si(100) substrate was purchased from Topsil. 99.95%-pure zinc Zn target was purchased from Kurt J Lesker. Isopropanol (VLSI) and acetone (VLSI) were purchased from JT Baker. The HF for the buffered solution was purchased from JT Baker. 5-N argon and 5-N oxygen were purchased from Mocart. Methylene blue (MB, Reag. Ph. Eur) was purchased from Merck. The ultrapure water was produced by a MembraPure Astacus system (MembraPure GmbH, Hennigsdorf, Germany).

Preparation of ZnO porous thin films
The nanostructured ZnO porous thin films used in the photocatalytic studies were prepared at the ŁRN -Institute of Microelectronics and Photonics, Warsaw, Poland. They were deposited at the Si(100) substrate by means of direct current (DC) reactive magnetron sputtering with post-deposition annealing. Prior to the deposition of the nanostructured ZnO thin films, the substrate was first degreased by boiling it in isopropanol and acetone and subsequently bathed in a buffered HF solution in order to remove any native oxide.
The deposition was carried out in a high vacuum reactor (Surrey NanoSystems Gamma 1000 C, UK) using a 99.95%pure zinc Zn target of 75 mm diameter under constant DC power of 80 W in a mixture of 5N-pure argon and oxygen flowing at 10 sccm and 1 sccm, respectively, whereas the total gas pressure during the sputtering process was equal to 0.4 Pa. The respective gas flow was controlled by mass flow controllers (MFC) and the system used a Baratron manometer in a feed-back loop with a gate throttle valve (VAT, Haag, Switzerland) for independent gas total pressure and flow control.
The base pressure in the vacuum deposition chamber prior to deposition was at the order of 10 -5 Pa. It was pumped using a cryogenic pump (Cryo-Torr, CTI-Cryogenics, USA) placed behind the throttle valve. This step yielded a nanostructured porous Zn thin film, as reported earlier [75]. After deposition, the porous Zn thin films were oxidized to zinc oxide in a 5N-pure oxygen flow for 20 min at a temperature of 700°C inside a rapid thermal annealing furnace (SHS 100, Mattson, USA). Other technological details concerning the fabrication of nanostructured ZnO porous thin films were previously described [76].

Characterization of ZnO nanostructured thin films
The surface morphology (grains characteristics, roughness) and surface chemistry (non-stoichiometry and contaminations) of ZnO nanostructured thin films before and after the photocatalytic activity tests were studied by using the SEM and XPS methods, respectively.
In the SEM studies, a high-resolution scanning electron microscope Carl Zeiss Auriga 60 model was used operating at 5 keV, with a lateral resolution of 2 nm. Other experimental details on SEM studies were previously described [75].
In the XPS studies, a commercial XPS spectrometer (SPECS, Berlin, Germany) was equipped, among others, with an x-ray lamp (AlK α , 1486.6 eV, XR-50 model) and a concentric hemispherical analyzer (PHOIBOS-100 Model), was used. Other experimental details on XPS studies were previously described [75,76].

Photocatalytic activity tests
The experiments on photolytic as well as photocatalytic degradation of methylene blue (MB) dye in aqueous solution (pH = 6.8) under UV light were performed for the direct comparison of their efficiency. In both experiments, 60 ml of a 0.3 mg l −1 MB aqueous solution were used.
The photolysis experiments were performed by irradiating the MB solution by means of 5 UV lamps (Philips, TL 8W BLB 1FM/10X25CC, λ em,max = 365 nm) with a measured irradiance ∼ 1.86 mW·cm −2 at the level of the sample. The solution was kept under magnetic stirring during the experiment.
The photocatalytic experiments were performed in the same experimental system and conditions used for the photolysis experiments. The nanostructured ZnO thin films (1 × 1 cm) were placed horizontally in a beaker by means of golden wire support, so as to be completely submerged in the MB solution [77].
Prior to the irradiation, the system was kept in dark conditions for 30 min in order to establish the adsorptiondesorption equilibrium of MB on the surface of ZnO thin films.
A similar procedure was followed for the recycling experiments: after each experiment, the sample was carefully cleaned with ultrapure water, dried with compressed air, and then reused for the subsequent photocatalytic degradation of a fresh MB solution. The degree of degradation of MB was calculated according to the following equation: where c o is the MB concentration at t = 0, and c t is the MB concentration at irradiation time t.

Photocatalytic activity of nanostructured ZnO thin films toward MB
The nanostructured ZnO porous thin films annealed at 700°C were chosen to study the photocatalytic activity of ZnO thin films toward MB, a model dye pollutant, whose chemical structure is shown in figure S1.
Before the UV irradiation step, the ZnO thin film was submerged in a 0.3 mg·l −1 MB solution and kept in dark conditions for 30 min to reach the adsorption-desorption equilibrium. Subsequently, the irradiation with UV light for a total time of 540 min was performed cyclically, after rinsing the ZnO thin film in water and drying it at the end of each cycle.
The adsorption in dark conditions of MB on ZnO thin films resulted in circa 10% over 4 cycles. The UV irradiation of a 0.3 mg l −1 MB solution in the presence of ZnO thin film induced a decrease in the absorption band centered at 663 nm, resulting in a 64% degree of MB degradation after 540 min (figure 2).
The contribution of photolysis of MB was negligible (figure 2(b)) in the total photocatalytic degradation process, suggesting that the ZnO thin film is responsible for the photodegradation of MB.
The ZnO thin film underwent 4 photocatalytic cycles to verify its reusability (the results for the 4th cycle are shown in figure 2). After each cycle, the nanostructured ZnO porous thin film was rinsed with distilled water and subsequently airdried.
The corresponding photocatalytic degradation curves ( figure 2(b)) were analyzed to clarify the kinetics of the photocatalytic degradation of MB in the presence of ZnO porous thin film.
The fitting of the obtained data with a pseudo-first-order kinetic model was poor. This ruled out a 'diffusion-controlled' process, characterized by relatively fast surface reactions and detachment of products, yielding a negligible surface concentration of MB adsorbates, and relatively slow adsorption of MB molecules. Nevertheless, the data were better fitted using a pseudo-zero-order kinetic model, [76,78] which describes a 'surface-reaction limited' process, in which a relatively fast adsorption equilibrium and relatively slow surface reactions occur. The following fitting equation was used: where c t is the MB molar concentration at time t, c o is the MB concentration at t = 0 and k is the zero-order kinetic constant expressed in mol·l −1 ·min −1 . Table 1 shows the calculated values of the photocatalytic reaction rate constant for the 4 cycles. This behavior indicates that in these experimental conditions the surface of the ZnO thin film is saturated with MB molecules and that the ratedetermining step is represented by the chemical reactions occurring on the surface [79].
The comparison of the obtained rate constants with the literature is extremely complex, due to the fact that most of the reports used a pseudo-first-order kinetic model to fit the data. In addition, the operational parameters (e.g. light source, irradiance, and MB initial concentration), the sample deposition techniques, the ZnO thin layer thickness, and the sample area are different, preventing a direct comparison to be performed [80][81][82][83][84][85][86][87][88][89][90]. Nevertheless, the fitting of the data obtained in this work with a pseudo-first-order model yields values of kinetic constants in line with the ones found in the literature (10 -3 min −1 ).
As shown in table 1, the value of the rate constant decreased by 50% after 4 cycles. This behavior can be ascribed to the decrease of free active sites, which can be occupied by unreacted MB and/or by its degradation products. Moreover, morphological modifications of the surface of the ZnO thin films could also occur. In order to clarify the origin of these results, a morphological and compositional analysis of the sample after the photocatalytic experiments were performed.

Surface morphology of nanostructured ZnO porous thin films
The surface morphology (grains characteristics, roughness) of the nanostructured ZnO thin films was determined via SEM measurements. Two specific images are shown in figure 3.
As shown in figure 3, the surface of nanostructured ZnO thin films presents a coral reef structure containing polycrystalline coral dendrites, as recently observed [90]. This morphology is preserved after the photocatalytic studies with no visible mesoscopic qualitative changes ( figure 3(b)). A statistical analysis of the long (c) and short (d) axis lengths of the individual crystallites making up the coral structure shows that their sizes increased after the photocatalytic experiments with the median shifting from 43 to 76 nm for the longer axis

Cycle number
1st cycle 2nd cycle 3rd cycle 4th cycle k (mol·l −1 ·min −1 ) 1.54·10 -9 1.60·10 -9 0.77·10 -9 0.78·10 -9 and from 28 to 58 for the shorter axis. The statistical data were calculated based on the manual determination of crystallite shape and size in the ImageJ processing program. The observed increase in crystallite size could be attributed both to (i) the presence of residual adsorbed MB or of its photodegradation by-products on the thin film surface and to (ii) the effect of UV irradiation [91].

Surface chemistry of nanostructured ZnO thin films
As mentioned above, the surface chemistry of nanostructured ZnO thin films was studied by XPS measurements, in order to assess possible non-stoichiometry and contaminations. The survey spectra combined with the specific O1s and C1s spectral lines were recorded and then analyzed. Figure 4 shows the exemplary XPS survey spectra of the samples in the commonly used binding energy range (1200 eV) before and after the photocatalytic experiments, respectively.
The XPS survey spectra shown in figure 4 mainly contain the contribution from core level lines of O1s, Zn3s, Zn3p, and Zn3d, corresponding to the main elements (O and Zn) of the nanostructured ZnO thin films. Moreover, an evident contribution of C1s XPS line is also observed at 284 eV, which confirms the existence of strong C undesired contamination due to the adsorption of atmospheric CO 2 . Moreover, the additional peaks related to the Auger electron emission lines at ∼570 eV, ∼500 eV, and ∼470 eV, The XPS data analysis was performed using the commonly used analytical procedure based on the relative intensity (height) of the O1s and Zn2p core level lines (peaks), corrected by the transmission function T(E) of the CHA PHOIBOS 100, and finally after taking into account the atomic sensitivity factors (ASF) related to the height of peaks for O1s (O.66), C1s (0.25), and Zn3p (0.4), respectively. A more detailed description of the XPS data analysis was previously published [75,76].
The analysis revealed that before the photocatalytic experiments, the relative concentration of O atoms for the nanostructured ZnO thin films was ∼ 0.25, and the relative concentration of C atoms was ∼ 0.20. Similar values were found after the photocatalytic experiments.
In addition to the procedure described above, the XPS data were analyzed using a more practical and commonly used simplified procedure to assess the surface non-stoichiometry of the nanostructured ZnO thin films by determination of the relative concentration of O and C surface atoms with respect to the Zn atoms. According to these calculations, the relative [O]/[Zn] concentration was found to be ∼ 1.4, whereas the relative [C]/[Zn] concentration was ∼ 1.2, both before and after the photocatalytic experiments.
This suggests that in both cases a significant non-stoichiometry in a surface/subsurface region of the ZnO thin films was present, which is related to the existence of various surface O atoms combined with the surface C atoms. This could be confirmed via the deconvolution procedure of XPS Zn2p 3/2 , O1s and C1s spectral lines. Figure 5 shows the XPS Zn2p double lines of the nanostructured ZnO porous thin films before and after photocatalytic experiments, together with the XPS Zn2p 3/2 spectral line after Gauss deconvolution.   For both cases, the XPS Zn2p lines contain two broadly resolved components at the specific binding energy of about 1045 eV and 1022 eV, respectively, related to the spin-orbit splitting of Zn2p 1/2 and Zn2p 3/2 , respectively. These peaks are separated by ∼ 23.0 eV, which is in good agreement with available reference data [43]. This directly confirms that Zn exists mainly as Zn 2+ corresponding to Zn atoms in the ZnO lattice [74]. However, both XPS Zn2p 3/2 lines are slightly asymmetrical. After their precise deconvolution procedure, two components were found, located at the binding energy of ∼1022.0 eV and ∼1024.0 eV, respectively, which is very clearly visible in figure 5(b). The higher one (93% of peak area), commonly observed for all ZnO samples, can be assigned to the Zn atoms in the ZnO lattice, whereas the smaller one (only 7% of peak area) to the same specific zinc hydroxide species Zn(OH) 2 , recently observed in ZnO nanoparticles by Guo et al [47], in ZnO thin films as well as in the authors' XPS studies of ZnO nanowires [76]. It should be underlined that the presence of zinc hydroxide species Zn(OH) 2 can be expected, keeping in mind that the ZnO samples studied both before and after photocatalytic processes are usually always exposed to natural humidity conditions. Figure 6 shows the XPS spectra of O1s and C1s after Gauss deconvolution of the nanostructured ZnO thin films before photocatalysis. It is possible to observe the presence of two typical oxygen surface bonds related to hydroxide ions OH − and the ZnO lattice oxide ions O 2-, at a binding energy of 531.1 eV and 532.8 eV, respectively, as well as one typical carbon surface bond related to the carbon hydroxyl group C-OH, at the binding energy 285.8 eV.
It has to be highlighted that the information on the specific surface chemical bonds related to the oxygen and carbon atoms at the surface of low dimensional ZnO nanostructures used as photocatalysts in MB degradation is still rather neglected in the available literature.
However, these bonds play an extremely important role in the photocatalytic degradation of MB, since the surface oxide ions O 2can participate in the oxidation of adsorbed surface species [92] and surface oxygen hydroxide ions OH − can capture holes to form OH radicals and can also enhance O 2 adsorption, resulting in an increased production of superoxide ions O 2 via reduction through photogenerated electrons [68].
Interestingly, slightly different sets of surface bonds are observed after the photocatalytic experiments, as can be seen in figure 7.
As for oxygen surface bonds, the contribution of the ZnO lattice oxide ions O 2is still significant after photocatalysis. Nevertheless, the surface hydroxide ions OH − present before the photocatalysis are substituted by oxygen of the carboxyl functional O-C=O groups, at the binding energy 534.1 eV. As for the carbon surface bonds, the carbon hydroxyl group C-OH present before photocatalysis is accompanied by carbonyl functional C=O groups, at the binding energy 287.2 eV, after photocatalysis. These results suggest that carbonyl and carboxyl groups can originate from by-products of the photocatalytic degradation of MB [93]. Moreover, these adsorbed species could play a role as a specific barrier for the intra-particle diffusion of pollutant molecules on the surface of the ZnO porous thin film nanostructures, which could justify the decrease of photocatalytic activity over the irradiation cycles.
Moreover, only slight modifications of the sample morphology were observed after the photocatalytic studies with no detectable mesoscopic qualitative changes in the crystallite shape of individual crystallites of the coral structure in our porous nanostructured ZnO thin films. An increase in the crystallite size was observed after the photocatalytic experiments, with the median shifting from 43 to 76 nm for the longer axis and from 28 to 58 nm for the shorter axis. It is crucial to note that this finding is in evident contrast to the behavior of ZnO nanowires [57,58], ZnO nanorods [59][60][61][62]94], and hierarchical ZnO nanostructures [63][64][65][66]95].

Conclusions
In this paper, the results of the photocatalytic degradation of methylene blue (MB) on ZnO nanostructured porous thin films are described with a special emphasis on the correlation with the morphology and surface chemistry of the employed nanostructured films. These films possess highly extended internal surfaces and were used as the expected most efficient photocatalyst for the MB degradation process, among the other forms of low dimensional ZnO nanomaterials reported in the literature and cited in this article. The performed photocatalytic studies showed that an MB removal of 64% was  achieved after 540 min. UV irradiation. Additionally, the reusability of these nanostructured ZnO thin films was confirmed and compared with other ZnO-based thin films reported in the literature.
The SEM studies showed that the coral dendritic structure of the pristine samples is maintained after 4 cycles of photocatalytic experiments. In addition, the XPS studies showed that the surface of the nanostructured ZnO porous thin films exhibits a slight non-stoichiometry, related to the existence of hydroxide ions OH -. These groups can act as hole-acceptors to produce OH· radicals, which can be responsible for the generation of superoxide ions. Moreover, the superficial ZnO lattice oxide ions can also participate in the oxidation of the adsorbed MB molecules. XPS studies also revealed the presence of residual MB or its photodegradation products on the surface of the nanostructures after the photodegradation experiments, thus explaining the decrease in photocatalytic efficiency over 4 irradiation cycles.
The results of our studies confirmed that properly selected ZnO nanostructures endowed with a high surface roughness can be efficient for the photocatalytic degradation of MB. Moreover, it was confirmed that XPS is a powerful technique to better understand the surface chemistry of the photocatalyst before and after the photodegradation experiments, giving insights into the possible processes occurring during photodegradation.

Data availability statement
The data that support the findings of this study are available upon reasonable request from the authors.