E ﬀ ect of Zr Impregnation on Clay-Based Materials for H 2 O 2 -Assisted Photocatalytic Wet Oxidation of Winery Wastewater

: UV-activated Zr-doped composites were successfully produced through the impregnation of Zr on the crystal lattice of di ﬀ erent clay materials by a one-step route. Fixing the amount of Zr available for dopage (4%), the inﬂuence of di ﬀ erent supports, submitted to di ﬀ erent chemical treatments, on the photocatalytic activity of the resulting Zr-doped pillared clay materials (PILC) was assessed. Both chemical characterization and structural characterization suggest that the immobilization of Zr on montmorillonite and PILC structures occurred through isomorphic substitution between Si and Zr in the tetrahedral sheet of the clay material. This structural change was demonstrated by signiﬁcant modiﬁcations on Si-OH stretching vibrations (1016 cm − 1 , 1100 cm − 1 and 1150 cm − 1 ), and resulted in improved textural properties, with an increase in surface area from 8 m 2 / g (natural montmorillonite) to 107 m 2 / g after the pillaring process, and to 118 m 2 / g after the pillaring and Zr-doping processes ((Zr)Al-Cu-PILC). These materials were tested in the UV-photodegradation of agro-industrial wastewater (AIW), characterized by high concentrations of recalcitrant contaminants. After Zr-dopage on AlCu-PILC heterogeneous catalyst, the total organic carbon ( TOC ) removals of 8.9% and 10.4% were obtained through adsorption and 77% and 86% by photocatalytic oxidation, at pH 4 and 7, respectively. These results suggest a synergetic e ﬀ ect deriving from the combination of Zr and Cu on the photocatalytic degradation process. Author Contributions: Conceptualization, V.G. and A.R.T.; methodology, V.G. and A.R.T.; validation, V.G. and A.R.T.; investigation, V.G. and A.R.T.; writing—original draft preparation, V.G. and A.R.T.; writing—review and editing, V.G., A.R.T., M.S.L. and J.A.P.; visualization, V.G., M.S.L. and J.A.P.; supervision, M.S.L. and J.A.P.;


Introduction
Agro-industrial activities are one of the main sources of wastewater pollution and its impact on the environment has received special attention in recent years [1,2]. Winery wastewater (WW) is characterized by high load of recalcitrant organic compounds [1,3], and its unregulated discharge represents a great threat to aquatic ecosystems and human health [4]. In this regard, the development of effective and low cost methods for the treatment of WW is now imperative.
Currently, different techniques have been developed to treat this type of effluent, including adsorption [5], coagulation [6] and biological processes [7]. However, some of the drawbacks include the limited adsorption capacity and the formation of a potential second pollution source, since these processes only transfer contaminants from one phase to another instead of destroying them [8]. Biological degradation is the most common process applied, however, the microbial activity can be inhibited by the recalcitrant character and toxicity of the organic contaminants [9]. To overcome these problems, advanced oxidation processes (AOPs) have been proposed as effective, fast and non-expensive order, it is worth noting that Fermis based model was specifically developed to describe the kinetics of complex systems, involving mixtures of unknown pollutants and several reaction intermediates formed during the photocatalytic process. Therefore, it includes lumped analytical parameters, such as TOC, that can be derived in groups of compounds with different reactivity [42].
In a previous work [42], a lumped kinetic model based on Fermi's equation was developed to describe the TOC histories for the degradation of a dye by catalytic wet peroxide oxidation, as shown in Equation (1) TOC where k TOC corresponds to the apparent reaction rate constant; t * TOC represents the transition time related to the TOC content curve's inflection point, and x TOC corresponds to the fraction of non-oxidazable compounds that are formed during the reaction.
The Lumped kinetic model based on Fermi's equation has successfully described the kinetics of our previous experiments using pillared clays in the H 2 O 2 -assisted photocatalytic wet oxidation of WW and, therefore, it is intended to apply this method in order to describe the kinetics of the WW degradation process using the new proposed materials as heterogeneous catalysts.

Reagents and Winery Wastewater Sampling
ZrOCl 2 .8H 2 O (99%) was supplied by Alfa-Aesar, CuCl 2 .2H 2 O (99%) by Panreac, H 2 O 2 (30% w/v) by Sigma-Aldrich. NaOH and H 2 SO 4 (95%) were both obtained from Analar NORMAPUR. Deionized water was used to prepare the respective solutions. The agro-industrial wastewater (AIW) was collected from a Portuguese winery cellar located in the Douro region (Northeast of Portugal). The main chemical parameters measured are shown in Table 1. Prior to the oxidation process, the wastewater was submitted to a primary treatment, where the suspended solids were removed from the effluent.

Clay Mineral
Natural montmorillonite (MT) was purchased from Fluka, Alfa-Aesar. The chemical composition and main surface properties of natural clay mineral are listed in Tables 2 and 3, respectively. The chemical data was determined by energy dispersive X-ray spectroscopy (EDS/EDAX, FEI QUANTA-400). The total iron expressed as Fe 2 O 3 content in raw-montmorillonite was found to be 4.28%. The cation exchange capacity (CEC) of the mineral fractions was measured following the ammonium acetate method proposed by Chapman [43].

Analytical Techniques
Several physical-chemical parameters were measured in order to characterize the agro-industrial wastewater, namely the chemical oxygen demand (COD), the biological oxygen demand (BOD 5 ), the total organic carbon (TOC) and the total polyphenols (mg gallic acid/L) presented in Table 1. The COD and BOD 5 were determined according to Standard Methods (5220D; 5210D; respectively) [44]. COD analysis was carried out in a COD reactor from HACH Co. and a HACH DR 2400 spectrophotometer was used for colorimetric measurement. Biochemical oxygen demand (BOD 5 ) was determined using a respirometric OxiTop system. pH evolution was followed by means of a pH-meter (HANNA Instruments, Rhode Island, USA). The TOC content (mg C/L) was determined using a Shimadzu TOC-L CSH analyzer (Tokyo, Japan). Total polyphenols were evaluated following the Folin-Ciocalteu method [45].

Catalysts Preparation
The preparation of the pillared clays was carried out following a conventional procedure described in detail by Molina, et al. [46]. AlCu-PILC was prepared through the intercalation between montmorillonite fractions and poly(hydroxy)aluminium (Al 3 (OH) 4 5+ ) and copper Cu 3 (OH) 4 2+ species.
The pillaring solution was prepared by slow addition of a 0.2 M NaOH solution to a mixture of 0.1 M AlCl 3 and 0.1 M CuCl 2 (Cu/(Al+Cu) = 0.1), under constant stirring until the molar ratio OH/Al = 2.5 was reached. The resulting solution was adjusted to pH 6 and was further aged for 8 h at 298 K. The intercalation process was initiated by the addition of a suspension of 0.1 wt.% montmorillonite in deionized water to the pillaring solution, applying the stoichiometry of 10 mmol Al/g clay. The cationic exchange process was carried out at room temperature for 12 h under constant stirring. The resulting suspension was washed by centrifugation with deionized water in order to reach ionic conductivity values lower than 10 µS. After air-drying, the resulted material was calcinated for 2 h at 400 • C. The Al-Cu oligomeric solution was adjusted to pH 6 in order to achieve the higher proportion of oligomeric species: 100% of both Al 3 (OH) 4 5+ and species. The aqueous speciation was calculated by Visual MINTEQ, version 3.0. After the pillaring process, Cu 2+ oligomeric species were converted to the respective metal oxide clusters by dehydration and dehydroxylation along the calcination process. The Al-PILC was prepared following the same procedure adopted to AlCu-PILC, but only using poly(hydroxy)aluminium (Al 3 (OH) 4 5+ ) species.
The Zr-doped catalysts (Zr-MT, (Zr)Al-PILC and (Zr)AlCu-PILC)) were prepared by the incipient wetness impregnation method. The precursor solution was prepared with ZrOCl 2 .8H 2 O in order to Water 2020, 12, 3387 5 of 15 obtain a zirconium load of 4 wt.%. After impregnation process, the doped catalysts were dried at 100 • C overnight and calcinated for 3 h at 400 • C. The results obtained from the chemical characterization (Table 2) confirm that 4 wt.% of zirconium were successfully immobilized on different heterogeneous catalysts.

Catalysts Characterization
The FTIR spectra were obtained by mixing 1 mg natural montmorillonite with 200 mg KBr. The powder mixtures were then inserted into molds and pressed at 10 ton/cm 2 to obtain the transparent pellets. The samples were analyzed with a Bruker Tensor 27 spectrometer and the infrared spectra in transmission mode were recorded in the 4000-400 cm −1 frequency region. The microstructural characterization was carried out by scanning electron microscopy (SEM/ESEM FEI QUANTA 400) and the chemical composition of the different catalysts was estimated (Table 2) using energy dispersive X-ray spectroscopy (EDS/EDAX).
The textural parameters of samples were obtained from N 2 adsorption-desorption isotherms at 77 K using a Micromeritics ASAP 2020 apparatus (Norcross, Georgia, USA). The samples were degassed at 150 • C up to 10 −4 Torr before analysis. The specific surface area (S BET ) was determined by applying the Gurevitsch's rule at a relative pressure p/p 0 = 0.30 and according to the Brunauer, Emmet, Teller (BET) method from the linear part of the nitrogen adsorption isotherms. Different pore volumes were determined by the Barrett, Joyner, Halenda model (BJH model).

Adsorption Tests
Different adsorption tests were carried out in order to predict the amount of organic carbon removed through adsorption. The adsorption batch experiments were carried out at different pH conditions (pH 4.0 and pH 7.0) by adding 3.00 g/L of each heterogeneous catalyst into 500 mL of WW (500 mg C/L). The temperature was kept constant throughout the experiments. After the adsorption runs, the samples were centrifuged and the TOC content of the supernatant solution was measured. The percentage of organic carbon removed through adsorption was calculated according to Equation (2) [47,48]: where TOC 0 is the initial TOC content (mg C/L) and TOC t is TOC value at instant t (mg C/L).

Photocatalytic Experiments and Kinetic Modelling
The photocatalytic experiments were performed in a batch cylindrical photoreactor (600 cm 3 ) equipped with a UV-C low pressure mercury vapour lamp (TNN 15/32)-working power = 15 W (795.8 W/m 2 ) and λ max = 254 nm (Heraeus, Germany). The UV absorption spectrum of the AIW reveals a maximum at ca. 275 nm (with and without the catalysts) and a high absorption at the wavelength where the UV-C lamp emits. In a typical run, 3.0 g/L of catalyst was mixed with 500 mL of the AIW (TOC = 500 mg C/L) for 15 min. After this, a specific amount of H 2 O 2 (98 mM) was added to the suspension and the UV light was turned on at the same time. The initial pH varied from 4.0 to 7.0, and was adjusted by adding 1 M of H 2 SO 4 or 1 M of NaOH. After the reaction has started, 20 mL of solution was withdrawn for TOC measurements at different reaction times, completing a total period of 240 min. The samples were centrifuged and the Zr and Cu concentrations were analyzed by atomic absorption spectroscopy (AAS) using a Thermo Scientific iCE 3000 SERIES. All experiments were performed in triplicate and the observed standard deviation was always less than 5% of the reported values.
A kinetic modelling based on a lumped kinetic model traduced by Fermi's equation was carried out in order to describe the WW degradation process. The experiments were conducted at different pH conditions (pH 4 and pH 7), where temperature, effluent volume, contaminant concentration, H 2 O 2 concentration and catalyst dosage were kept constant. A nonlinear least squares regression, Water 2020, 12, 3387 6 of 15 based on the Levenberg-Marquardt (LM) algorithm, was applied using the OriginPro 8.5 "Sigmoidal Fit Tool". As a result, a unique semi-empirical function is applied to simultaneously describe the initial low TOC conversion (induction period) and subsequent rapid degradation step. Therefore, both the initial transition period and pseudo-first order kinetic period can be expressed with the proposed model [49].

Catalysts Characterization
The X-ray diffractograms corresponding to natural montmorillonite (MT) and to Zr-doped and undoped PILCs, are shown in Figure 1. The hkl reflections associated with MT diffraction pattern are characteristic of a montmorillonite clay mineral with mixed interlayer composition including different proportions of Na + and Ca 2+ ions (13.08 Å). This assumption is in agreement with the chemical characterization data, which shows proportions of 0.95% and 2.54% of CaO and Na 2 O (Table 2), respectively. The MT samples modified with previously synthesized oligomeric species (Cu 3 (OH) 4 2+ and/or Al 3 (OH) 4 5+ show a shift of the basal reflection d001 from 13.08 Å (MT) to 18.02 Å and to 17.01 Å for (Zr)AlCu-PILC and (Zr)Al-PILC, respectively, confirming the insertion of the oligomeric species in the interlayer region of montmorillonite and the successful pillaring process. The higher expansion observed when both Cu-and Al-oligomeric species were intercalated on montmorillonite results from the higher pillars formed, indicating that the number, charge, size and shape of the oligomeric species affect the pillar size. These results were also suggested by the textural properties obtained for these materials (Table 3), since the (Zr)AlCu-PILC has higher surface area (ABET=x) and higher number of total pore volume than (Zr)Al-PILC, suggesting an increase in contact area available for absorption due to the higher pillars formed.
Water 2020, 12, x FOR PEER REVIEW 6 of 15 As a result, a unique semi-empirical function is applied to simultaneously describe the initial low TOC conversion (induction period) and subsequent rapid degradation step. Therefore, both the initial transition period and pseudo-first order kinetic period can be expressed with the proposed model [49].

Catalysts Characterization
The X-ray diffractograms corresponding to natural montmorillonite (MT) and to Zr-doped and undoped PILCs, are shown in Figure 1. The hkl reflections associated with MT diffraction pattern are characteristic of a montmorillonite clay mineral with mixed interlayer composition including different proportions of Na + and Ca 2+ ions (13.08 Å). This assumption is in agreement with the chemical characterization data, which shows proportions of 0.95% and 2.54% of CaO and Na2O (Table  2), respectively. The MT samples modified with previously synthesized oligomeric species (Cu3(OH)4 2+ and/or Al3(OH)4 5+ show a shift of the basal reflection d001 from 13.08 Å (MT) to 18.02 Å and to 17.01 Å for (Zr)AlCu-PILC and (Zr)Al-PILC, respectively, confirming the insertion of the oligomeric species in the interlayer region of montmorillonite and the successful pillaring process. The higher expansion observed when both Cu-and Al-oligomeric species were intercalated on montmorillonite results from the higher pillars formed, indicating that the number, charge, size and shape of the oligomeric species affect the pillar size. These results were also suggested by the textural properties obtained for these materials (Table 3), since the (Zr)AlCu-PILC has higher surface area (ABET=x) and higher number of total pore volume than (Zr)Al-PILC, suggesting an increase in contact area available for absorption due to the higher pillars formed. Comparing both (Zr)AlCu-PILC and AlCu-PILC diffraction patterns, it is possible to observe an identical behaviour, confirming that Zr was probably incorporated into the AlCu-PILC lattice without structural modification. The chemical composition of both samples before and after the Zr-doping process is also in agreement with this previous conclusion, given that the increase in Zr amount in doped-clay minerals is accompanied by a decrease in Si proportion, suggesting the isomorphic substitution between Si and Zr in the tetrahedral sheet of the pillared clay. This mechanism is triggered by the similar ionic radii of both cations, where the new one may have identical or lower ionic charge than the replaced one. In this case, both Si and Zr have similar ionic radii and the same ionic charge (+4) and, therefore, no structural charge was developed and no significant structural changes have occurred. This is particularly important, because once the AlCu-PILC has not been structurally affected by the Comparing both (Zr)AlCu-PILC and AlCu-PILC diffraction patterns, it is possible to observe an identical behaviour, confirming that Zr was probably incorporated into the AlCu-PILC lattice without structural modification. The chemical composition of both samples before and after the Zr-doping process is also in agreement with this previous conclusion, given that the increase in Zr amount in doped-clay minerals is accompanied by a decrease in Si proportion, suggesting the isomorphic substitution between Si and Zr in the tetrahedral sheet of the pillared clay. This mechanism is triggered by the similar ionic radii of both cations, where the new one may have identical or lower ionic charge than the replaced one. In this case, both Si and Zr have similar ionic radii and the same ionic charge (+4) and, therefore, no structural charge was developed and no significant structural changes have occurred. This is particularly important, because once the AlCu-PILC has not been structurally affected by the doping process, the adsorption capacity, which is crucial for its catalytic activity, was also not negatively affected. Moreover, enhanced catalyst stability is expected; once Zr is directly incorporated on the crystal lattice of montmorillonite, the risk of metal leaching is significantly lower.
Comparing both (Zr)-MT and MT spectra, no additional conclusions are achieved, since, after the Zr impregnation process, sample (Zr)-MT was submitted to the calcination process, which resulted in the total interlayer collapse to 9.55 Å by dehydration. Therefore, independently of the position of Zr (tetrahedral sheet or interlayer region) on the MT structure, the structural collapse will occur and avoid additional conclusion by means of X-ray diffraction (XRD). Figure 2 depicts the FTIR spectra obtained before (MT) and after the pillaring process (AlCu-PILC), as well as before (AlCu-PILC) and after the Zr-doping process ((Zr)AlCu-PILC). The results show some structural alterations on montmorillonite after the uptake of metal poly(hydroxy)-complexes and consequent formation of pillars on its internal surface. This is traduced by the decrease in intensity and shift of peaks in the range between 800 and 950 cm −1 , after the pillaring process, which are assigned to Al-OH, Fe-OH and Mg-OH vibration modes, at 916 cm −1 , 877 cm −1 and 849 cm −1 , respectively. These structural changes were only observed for PILC samples, which, according to Zhou et al. [50], can be attributed to the interactions between Al or Al/Cu mixed poly(hydroxy) species and the alumina octahedral layers.
Significant modifications on Si-OH stretching vibrations were observed after the Zr-doping process, due to the shift of the main band from 1016 cm −1 to 1040 cm −1 , and the reduction in intensity of the additional stretching vibrations assigned to the Si-O group, at 1150 cm −1 and 1100 cm −1 , confirming the incorporation of Zr ions directly in the crystal lattice by isomorphic substitution of Si ions in the tetrahedral sheet of montmorillonite. On the other hand, no additional changes in the vibrations associated with the octahedral sheets of montmorillonite (800-950 cm −1 ) were observed after this process.
Water 2020, 12, x FOR PEER REVIEW 7 of 15 doping process, the adsorption capacity, which is crucial for its catalytic activity, was also not negatively affected. Moreover, enhanced catalyst stability is expected; once Zr is directly incorporated on the crystal lattice of montmorillonite, the risk of metal leaching is significantly lower.
Comparing both (Zr)-MT and MT spectra, no additional conclusions are achieved, since, after the Zr impregnation process, sample (Zr)-MT was submitted to the calcination process, which resulted in the total interlayer collapse to 9.55 Å by dehydration. Therefore, independently of the position of Zr (tetrahedral sheet or interlayer region) on the MT structure, the structural collapse will occur and avoid additional conclusion by means of X-ray diffraction (XRD). Figure 2 depicts the FTIR spectra obtained before (MT) and after the pillaring process (AlCu-PILC), as well as before (AlCu-PILC) and after the Zr-doping process ((Zr)AlCu-PILC). The results show some structural alterations on montmorillonite after the uptake of metal poly(hydroxy)-complexes and consequent formation of pillars on its internal surface. This is traduced by the decrease in intensity and shift of peaks in the range between 800 and 950 cm −1 , after the pillaring process, which are assigned to Al-OH, Fe-OH and Mg-OH vibration modes, at 916 cm −1 , 877 cm −1 and 849 cm −1 , respectively. These structural changes were only observed for PILC samples, which, according to Zhou et al. [50], can be attributed to the interactions between Al or Al/Cu mixed poly(hydroxy) species and the alumina octahedral layers.
Significant modifications on Si-OH stretching vibrations were observed after the Zr-doping process, due to the shift of the main band from 1016 cm −1 to 1040 cm −1 , and the reduction in intensity of the additional stretching vibrations assigned to the Si-O group, at 1150 cm −1 and 1100 cm −1 , confirming the incorporation of Zr ions directly in the crystal lattice by isomorphic substitution of Si ions in the tetrahedral sheet of montmorillonite. On the other hand, no additional changes in the vibrations associated with the octahedral sheets of montmorillonite (800-950 cm −1 ) were observed after this process. The specific surface area and total pore volume of the original montmorillonite and doped and undoped materials are shown in Table 3. These results suggest significant alterations on montmorillonite after the pillaring and Zr-doping processes, resulting in significant and progressive increases in surface area and total pore volume. Accordingly, the surface area increased from 8 m 2 /g (MT) to 107 m 2 /g after the pillaring process, and to 118 m 2 /g after the pillaring and Zr-doping processes, whereas the total pore volume increased from 0.05 cm 3 /g (MT) to 0.20 cm 3 /g and to 0.22 The specific surface area and total pore volume of the original montmorillonite and doped and undoped materials are shown in Table 3. These results suggest significant alterations on montmorillonite after the pillaring and Zr-doping processes, resulting in significant and progressive increases in surface Water 2020, 12, 3387 8 of 15 area and total pore volume. Accordingly, the surface area increased from 8 m 2 /g (MT) to 107 m 2 /g after the pillaring process, and to 118 m 2 /g after the pillaring and Zr-doping processes, whereas the total pore volume increased from 0.05 cm 3 /g (MT) to 0.20 cm 3 /g and to 0.22 cm 3 /g for AlCu-PILC and (Zr)AlCu-PILC, respectively. The respective isotherms can be classified as type II, where unrestricted monolayer-multilayer adsorption occurs, and the behaviour of the hysteresis loops can be associated with type H3, which usually corresponds to aggregates of plate-like particles forming slit-like pores, which is in agreement with these material structures.

Adsorption vs. Reaction
Adsorption experiments were carried out using the pillared and Zr-doped catalysts to evaluate the effect of their surface chemistry on the contaminant adsorption and TOC removal. According to previous studies, adsorption plays an important role as the main mechanism involved in the initial induction period, which corresponds to the period necessary for catalyst surface activation [51]. In the present study, part of the mechanism associated with the induction period is probably associated with the adsorption of H 2 O 2 and organic compounds onto the catalyst surface, producing surface complexes which promote the activation of the oxidation process through the generation of HO• radicals. Our previous research assessed, for the first time, the application of natural pillared clays (PILCs: Al-Cu-ST and Al-Fe-ST) as heterogeneous photocatalysts for the H 2 O 2 -assisted treatment of a real AIW [27]. The results indicated that the transition point between the induction period and surface activation and the production of HO• species was directly influenced by the amount of H 2 O 2 initially dosed to the process. Accordingly, a decrease in the transition period (t *) TOC from 136 to 96 min was observed, using Al-Cu-ST as the heterogeneous catalyst (3.0 g/L), when H 2 O 2 concentration increased from 29 to 98 mM, reducing the period necessary for the surface activation and, therefore, the period required to initiate the degradation process. Considering our previous conclusions, the influence of different catalysts, as well as the effect of Zr-dopage on these supports, were evaluated, taking into account the optimal experimental conditions obtained before, namely [H 2 O 2 ] 0 = 98 mM and catalyst dosage = 3.0 g/L.
The evolution of TOC removal through adsorption at different pH conditions and using the different catalysts is shown in Figure 3. As expected, both pH conditions imposed and catalyst textural properties affected the catalyst adsorption capacity. The lowest contaminant adsorption was obtained for Zr-MT and (Zr)Al-PILC, at pH 4 and pH 7, respectively, which correspond to the catalysts with lower surface area (65 m 2 /g and 81 m 2 /g, respectively). On the contrary, the higher adsorption capacity was obtained for (Zr)AlCu-PILC, at both pH conditions, corresponding to the sample with the highest surface area (118 m 2 /g) and total pore volume (0.22 cm 3 /g) obtained. This behaviour has particularly impact on the TOC removal efficiency along the oxidation process (Figure 4), as it is observed that the catalyst with enhanced adsorption capacity at both pH conditions has greater activity in the degradation process. Accordingly, the TOC removals obtained using (Zr)AlCu-PILC/UV after 4 h, corresponds to 77% and 86%, at pH 4 and 7, respectively. The incorporation of Zr on montmorillonite lattice (Zr-MT) has contributed to the significant improvement of TOC removals, when compared with raw-montmorillonite, with an increase from 46% to 60% and from 37% to 61% at pH 4 and 7, respectively. The development of AlCu pillared structures had additional advantages considering the improvement of montmorillonite textural properties, resulting in additional stability and catalyst activity. This is traduced by the increase in TOC removals to 69% and 73% at pH 4 and 7, respectively, with a maximum of 33% removal in the non-catalytic UV-C/H 2 O 2 experiments performed at both pH conditions. As previously observed, the Zr-dopage on AlCu-PILC has also improved its catalytic activity, promoting an increase in TOC removals to 77% and 86%, at pH 4 and pH 7, respectively, suggesting a synergetic effect of both Zr and Cu on the photocatalytic degradation process. In this case, Zr acts as a semiconductor that is excited by photons with an energy greater than its band gap (5.8-7.1 eV), generating electron-hole pairs, which migrates to the photocatalyst surface yielding radical species that can react with organic molecules upon redox reactions. The electron transfer process is enhanced by the successive redox of Cu 2+ , a transition metal which is continuously releasing e-species, induced by the presence of a permanent irradiation source (UV-C).
During these processes, a decrease in catalyst stability, caused by the increase in Cu leaching levels from 0.0 to 1.4 mg/L, was also observed from the acidic to the neutral conditions. However, all the Cu leaching concentrations are lower than or very close to the legal discharge limit imposed by EU legislation (1.0 mg Cu/L), and only 5.0% of the Cu immobilized was released at neutral conditions, confirming that Cu immobilization on pillared clay support was successful.
Water 2020, 12, x FOR PEER REVIEW 9 of 15 improved its catalytic activity, promoting an increase in TOC removals to 77% and 86%, at pH 4 and pH 7, respectively, suggesting a synergetic effect of both Zr and Cu on the photocatalytic degradation process. In this case, Zr acts as a semiconductor that is excited by photons with an energy greater than its band gap (5.8-7.1 eV), generating electron-hole pairs, which migrates to the photocatalyst surface yielding radical species that can react with organic molecules upon redox reactions. The electron transfer process is enhanced by the successive redox of Cu 2+ , a transition metal which is continuously releasing e-species, induced by the presence of a permanent irradiation source (UV-C). During these processes, a decrease in catalyst stability, caused by the increase in Cu leaching levels from 0.0 to 1.4 mg/L, was also observed from the acidic to the neutral conditions. However, all the Cu leaching concentrations are lower than or very close to the legal discharge limit imposed by EU legislation (1.0 mg Cu/L), and only 5.0% of the Cu immobilized was released at neutral conditions, confirming that Cu immobilization on pillared clay support was successful. Comparing both Zr-MT and (Zr)Al-PILC performances at both pH conditions, it is possible to conclude that Zr-MT has shown increased catalytic activity, mainly from 180 min, with TOC removals of 60% at both pH conditions for Zr-MT, and 46% at both pH conditions for (Zr)Al-PILC. This behavior could be explained by the competition between Zr and an excess of Al for the Si tetrahedral sites on (Zr)Al-PILC, which may have hampered the Zr incorporation onto the montmorillonite crystal lattice. The lower catalytic activity observed using (Zr)Al-PILC is even more pronounced in acidic conditions, where the TOC conversion is very similar to MT along the treatment process. This could be explained by the increased adsorption capacity obtained by MT at pH 4.0 (Figure 3), associated with its higher CEC (Table 2), which despite the lower BET surface area of MT, when compared with (Zr)Al-PILC, has contributed to enhanced adsorption capacity and improved catalytic activity. Comparing both Zr-MT and (Zr)Al-PILC performances at both pH conditions, it is possible to conclude that Zr-MT has shown increased catalytic activity, mainly from 180 min, with TOC removals of 60% at both pH conditions for Zr-MT, and 46% at both pH conditions for (Zr)Al-PILC. This behavior could be explained by the competition between Zr and an excess of Al for the Si tetrahedral sites on (Zr)Al-PILC, which may have hampered the Zr incorporation onto the montmorillonite crystal lattice. The lower catalytic activity observed using (Zr)Al-PILC is even more pronounced in acidic conditions, where the TOC conversion is very similar to MT along the treatment process. This could be explained by the increased adsorption capacity obtained by MT at pH 4.0 (Figure 3), associated with its higher CEC (Table 2), which despite the lower BET surface area of MT, when compared with (Zr)Al-PILC, has contributed to enhanced adsorption capacity and improved catalytic activity. improved its catalytic activity, promoting an increase in TOC removals to 77% and 86%, at pH 4 and pH 7, respectively, suggesting a synergetic effect of both Zr and Cu on the photocatalytic degradation process. In this case, Zr acts as a semiconductor that is excited by photons with an energy greater than its band gap (5.8-7.1 eV), generating electron-hole pairs, which migrates to the photocatalyst surface yielding radical species that can react with organic molecules upon redox reactions. The electron transfer process is enhanced by the successive redox of Cu 2+ , a transition metal which is continuously releasing e-species, induced by the presence of a permanent irradiation source (UV-C).
During these processes, a decrease in catalyst stability, caused by the increase in Cu leaching levels from 0.0 to 1.4 mg/L, was also observed from the acidic to the neutral conditions. However, all the Cu leaching concentrations are lower than or very close to the legal discharge limit imposed by EU legislation (1.0 mg Cu/L), and only 5.0% of the Cu immobilized was released at neutral conditions, confirming that Cu immobilization on pillared clay support was successful. Comparing both Zr-MT and (Zr)Al-PILC performances at both pH conditions, it is possible to conclude that Zr-MT has shown increased catalytic activity, mainly from 180 min, with TOC removals of 60% at both pH conditions for Zr-MT, and 46% at both pH conditions for (Zr)Al-PILC. This behavior could be explained by the competition between Zr and an excess of Al for the Si tetrahedral sites on (Zr)Al-PILC, which may have hampered the Zr incorporation onto the montmorillonite crystal lattice. The lower catalytic activity observed using (Zr)Al-PILC is even more pronounced in acidic conditions, where the TOC conversion is very similar to MT along the treatment process. This could be explained by the increased adsorption capacity obtained by MT at pH 4.0 (Figure 3), associated with its higher CEC (Table 2), which despite the lower BET surface area of MT, when compared with (Zr)Al-PILC, has contributed to enhanced adsorption capacity and improved catalytic activity.

Kinetic Study
In order to better understand the effect of the operational conditions on the induction period, the transition time between the induction period and the fast oxidation reaction (t * TOC) was obtained through the fitting of Fermi's equation based on lumped kinetic model to the experimental data. The results obtained for the different parameters are displayed in Table 4, and the fittings obtained for different catalysts are illustrated in Figure 5. The results reveal a good fitting of the kinetic model to the experimental data obtained for different catalysts, with R 2 values higher than 0.983. Considering the different parameters obtained from the modelling, it is assumed that the experimental conditions influenced the kinetic performance of our processes.

Kinetic Study
In order to better understand the effect of the operational conditions on the induction period, the transition time between the induction period and the fast oxidation reaction (t * TOC) was obtained through the fitting of Fermi's equation based on lumped kinetic model to the experimental data. The results obtained for the different parameters are displayed in Table 4, and the fittings obtained for different catalysts are illustrated in Figure 5. The results reveal a good fitting of the kinetic model to the experimental data obtained for different catalysts, with R 2 values higher than 0.983. Considering the different parameters obtained from the modelling, it is assumed that the experimental conditions influenced the kinetic performance of our processes. As observed in Figure 5, the transition point is significantly affected by the pH conditions, as well as the heterogeneous catalyst used. Accordingly, a significant decrease in t * TOC was observed, from 279 to 121 min (pH 7) and from 253 to 137 min (pH 4), using raw-montmorillonite and AlCu-MT, respectively, suggesting a significant reduction in the surface activation period when the heterogeneous catalyst was applied, as well as a quicker production of HO • species. This tendency was even more pronounced after Zr-dopage at neutral conditions, once the transition time has decreased from 121 to 119 min using (Zr)AlCu-PILC.
In both cases, the evolution of the H2O2 concentration along the photocatalytic experiments shows a decrease and subsequent increase in concentration in the first 60 min (Table 5), which suggests a possible adsorption and desorption of H2O2 on pillared montmorillonite. Therefore, the results suggest that the first 60 min of reaction were mainly associated with the formation of surface complexes between the H2O2 and catalyst surface, whereas the additional period, which completes the total induction period, may be associated with the time required for the surface activation, i.e., the time required for catalytic decomposition of the oxidant in the presence of the active phase (production of HO • ).
Concerning the reaction rates of the different photocatalytic processes, higher reaction rates were observed using both (Zr)AlCu-PILC and AlCu-PILC catalysts when compared with the other catalysts applied (Table 4), confirming an improvement of catalytic performance during the oxidation processes. Comparing both catalysts, AlCu-PILC has the higher reaction rates at both pH conditions, 3.54 × 10 −2 min −1 and 2.87 × 10 −2 min −1 , at pH 4 and 7, respectively. However, (Zr)AlCu-PILC (kTOC = 2.58 × 10 −2 min −1 ) has contributed to lower fractions of non-oxidazable compounds formed during the As observed in Figure 5, the transition point is significantly affected by the pH conditions, as well as the heterogeneous catalyst used. Accordingly, a significant decrease in t * TOC was observed, from 279 to 121 min (pH 7) and from 253 to 137 min (pH 4), using raw-montmorillonite and AlCu-MT, respectively, suggesting a significant reduction in the surface activation period when the heterogeneous catalyst was applied, as well as a quicker production of HO • species. This tendency was even more pronounced after Zr-dopage at neutral conditions, once the transition time has decreased from 121 to 119 min using (Zr)AlCu-PILC.
In both cases, the evolution of the H 2 O 2 concentration along the photocatalytic experiments shows a decrease and subsequent increase in concentration in the first 60 min (Table 5), which suggests a possible adsorption and desorption of H 2 O 2 on pillared montmorillonite. Therefore, the results suggest that the first 60 min of reaction were mainly associated with the formation of surface complexes between the H 2 O 2 and catalyst surface, whereas the additional period, which completes the total induction period, may be associated with the time required for the surface activation, i.e., the time required for catalytic decomposition of the oxidant in the presence of the active phase (production of HO • ).
Concerning the reaction rates of the different photocatalytic processes, higher reaction rates were observed using both (Zr)AlCu-PILC and AlCu-PILC catalysts when compared with the other catalysts applied (Table 4), confirming an improvement of catalytic performance during the oxidation processes. Comparing both catalysts, AlCu-PILC has the higher reaction rates at both pH conditions, 3.54 × 10 −2 min −1 and 2.87 × 10 −2 min −1 , at pH 4 and 7, respectively. However, (Zr)AlCu-PILC (k TOC = 2.58 × 10 −2 min −1 ) has contributed to lower fractions of non-oxidazable compounds formed during the reaction (x TOC = 0.11, pH 7), when compared with AlCu-PILC (x TOC = 0.26, pH 7), which is in agreement with the higher TOC removals obtained.

Catalyst Regeneration
Considering the best performance of (Zr)AlCu-PILC, the catalyst reuse capacity was evaluated throughout three consecutive cycles of H 2 O 2 -assisted photocatalytic AIW treatment. The experiments were carried out at pH 7.0, using a catalyst dosage of 3.0 g/L and a H 2 O 2 concentration of 98 mM. The results show that the TOC removal obtained using (Zr)AlCu-PILC corresponds to 86%, 66% and 63% after 240 min, for the first, second and third cycles, respectively. In general, a decrease in efficiency was observed from the first to the second cycle (with a loss of 20% in TOC removal). However, no additional loss of activity was observed from the second to the third cycle. All the leaching concentrations along the different cycles were very close to the legal limits imposed (1.0 mg Cu/L) and tended to decrease as the number of cycles increased, from 1.4 to 0.97 mg/L of Cu, from the first to the third cycle, revealing that the catalyst stability is not affected along the cycles.

Conclusions
Different catalysts submitted to different chemical treatments and/or the Zr-dopage process, were applied in the H 2 O 2 -assisted treatment of recalcitrant winery wastewater in order to evaluate the influence of the surface chemical properties of the doped supports on their adsorption and catalytic properties.
FTIR results show that the incorporation of Zr in the crystal lattice of montmorillonite and PILC through isomorphic substitution between Si and Zr was traduced by significant modifications on Si-OH stretching vibrations, due to the shift of the main band from 1016 cm −1 to 1040 cm −1 , and the decrease in intensity of the additional stretching vibrations assigned to the Si-O group, at, respectively, 1150 cm −1 and 1100 cm −1 .
In general, the results show that Zr-dopage on AlCu-PILC has improved its adsorption and catalytic activity, promoting an increase in TOC removals to 77% and 86%, with 8.85% and 10.35% of TOC removed through adsorption, at pH 4 and pH 7, respectively. It suggests a synergetic effect caused by the combination of Zr and Cu on the photocatalytic degradation process, once the semiconductor electron transfer process is enhanced by the successive redox of Cu(II), induced by the presence of the UV-C irradiation source.
A significant decrease in t * TOC was observed for AlCu-PILC and (Zr)AlCu-PILC, at both pH conditions, suggesting a significant reduction in the surface activation period when the heterogeneous catalyst was applied, as well as a quicker production of HO • species. As a result, higher reaction rates were obtained using both (Zr)AlCu-PILC (2.58 × 10 −2 min −1 ) and AlCu-PILC (k TOC = 3.54 × 10 −2 min −1 ) catalysts, confirming an improvement in catalytic performance along the oxidation processes. Comparing both catalysts, AlCu-PILC has the higher reaction rates at both pH conditions. However, (Zr)AlCu-PILC has contributed to lower fractions of non-oxidazable compounds formed during the reaction (x TOC = 0.11, pH 7.0), making it a more efficient process.