Comparative Study of ZnO Thin Films Doped with Transition Metals (Cu and Co) for Methylene Blue Photodegradation under Visible Irradiation

: We synthesized and characterized both Co-doped ZnO (ZnO:Co) and Cu-doped ZnO (ZnO:Cu) thin ﬁlms. The catalysts’ synthesis was carried out by the sol–gel method while the doctor blade technique was used for thin ﬁlm deposition. The physicochemical characterization of the catalysts was carried out by Raman spectroscopy, scanning electron microscopy (SEM), X-ray di ﬀ raction, and di ﬀ use reﬂectance measurements. The photocatalytic activity was studied under visible irradiation in aqueous solution, and kinetic parameters were determined by pseudo-ﬁrst-order ﬁtting. The Raman spectra results evinced the doping process and suggested the formation of heterojunctions for both dopants. The structural di ﬀ raction patterns indicated that the catalysts were polycrystalline and demonstrated the presence of a ZnO wurtzite crystalline phase. The SEM analysis showed that the morphological properties changed signiﬁcantly, the micro-aggregates disappeared, and agglomeration was reduced after modiﬁcation of ZnO. The ZnO optical bandgap (3.22 eV) reduced after the doping process, these being ZnO:Co (2.39 eV) and ZnO:Co (3.01 eV). Finally, the kinetic results of methylene blue photodegradation reached 62.6% for ZnO:Co thin ﬁlms and 42.5% for ZnO:Cu thin ﬁlms. generation of a heterojunction [64,70,71]. After charge pairs generation, the ROS can be yielded on the catalyst surface and MB degradation starts.


Introduction
Synthetic dyes are commonly used by various industries, especially textile ones. These physically and chemically stable compounds are harmful to the environment. Synthetic dyes are recalcitrant compounds that exhibit high solubility in water and accumulate in both wastewater and industrial effluents [1,2]. Currently, water pollution is one of the major challenges of the modern world, and the recovery of wastewater by conventional methods is not suitable for emergent pollutants [3,4]. Heterogeneous photocatalysis can be satisfactorily applied for the decontamination of natural samples through the photocatalytic degradation of toxic pollutants from complex matrices, such as river water and wastewater [5]. All renewable technologies have become a promising alternative for both energy generation and wastewater treatments. Solar photocatalysis is a suitable option to degrade recalcitrant Figure 1 shows the XRD pattern for the catalysts synthesized in this study. The hexagonal wurtzite phase (JCPDS No. 36−1451) is identified as a crystalline structure for ZnO thin films, with the signals of the diffraction pattern corresponding to those reported by other authors [46]. The doping process did not affect the main signals in the diffraction patterns, as the XRD patterns for ZnO:Cu and ZnO:Co showed signals of a wurtzite ZnO structure. However, the XRD patterns showed a change in the intensity of the signals, suggesting that metal ions could substitute Zn 2+ ions after the doping process. Lima et al. suggested that the change in the intensities could be associated with changes in both (i) grain size due to network defects and (ii) oxygen vacancies [47,48].
We used the Debye-Scherrer equation to calculate the crystalline domain size of the catalysts, using the full width at half maximum (FWHM) for the highest peak (101), with θ being the Scherrer diffraction angle [49]. Although there is no clear tendency between the intensity of the signal and metal doping load (see Figure 1b,c), all samples reduced the grain size of catalysis after the doping process (see Table 1). This could be explained by the incorporation of Co 2+ and Cu 2+ ions as dopants into the ZnO after the doping process [48,49]. Finally, the structural results suggest that ZnO films incorporated metallic ions. This observation was verified by Raman spectroscopy and diffuse reflectance, as described in the next sections.

Raman Study
The Raman spectra of the catalysts are shown in Figure 2. All the peaks correspond with wurtzite-ZnO (C 4 6v ): (i) 97.4 cm −1 (vibrational mode E 2L ), (ii) 340 cm (E 2H -E 2L ), and (iii) 437.0 cm −1 and 581 cm −1 (A 1 vibrational mode) [50,51]. Figure 2a shows the Raman spectra for Cu-doped ZnO thin films. Signals E 2L (~99 cm −1 ) and E 2H (~437 cm −1 ) widen and decrease after the doping process-a behavior that can be explained by the incorporation of Cu 2+ into the ZnO lattice. Additionally, this behavior has been associated with reduction of ZnO crystallinity by the formation of nanocomposites [52]. For greater Cu loads, Figure 2a shows two new signals, the first one located at 298 cm −1 and a second weak one at 614 cm −1 . These two signals can be attributed to modes A 1g and B 2g for CuO, respectively, and this result suggests the formation of a ZnO-CuO heterojunction during the synthesis process. The hydrodynamic stability of the suspension is affected by the concentration of reagents; so for obtaining greater Cu loads, the CuO generation is feasible. This result is in line with previous reports [39]. The Raman spectra of the catalysts are shown in Figure 2. All the peaks correspond with wurtzite-ZnO (C 4 6v): (i) 97.4 cm −1 (vibrational mode E2L), (ii) 340 cm (E2H-E2L), and (iii) 437.0 cm −1 and 581 cm −1 (A1 vibrational mode) [50,51]. Figure 2a shows the Raman spectra for Cu-doped ZnO thin films. Signals E2L (~99 cm −1 ) and E2H (~437 cm −1 ) widen and decrease after the doping process-a behavior that can be explained by the incorporation of Cu 2+ into the ZnO lattice. Additionally, this behavior has been associated with reduction of ZnO crystallinity by the formation of nanocomposites [52]. For greater Cu loads, Figure 2a shows two new signals, the first one located at 298 cm −1 and a second weak one at 614 cm −1 . These two signals can be attributed to modes A1g and B2g for CuO, respectively, and this result suggests the formation of a ZnO-CuO heterojunction during the synthesis process. The hydrodynamic stability of the suspension is affected by the concentration of reagents; so for obtaining greater Cu loads, the CuO generation is feasible. This result is in line with previous reports [39].  Figure 2b shows the Raman spectrum for ZnO:Co; the intensity of A1LO, E2L, and E2H modes decreases for these films. The ZnO:Co thin films (doping load 5%) show four new signals at 490 cm −1 , 526 cm −1 , 626 cm −1 , and 725 cm −1 . These signals could be attributed to the possible presence of Co3O4, and these results confirm both the doping process and the formation of a heterojunction of ZnO-Co3O4 [53,54]. Figure 3 shows SEM images for the catalysts. Figure 3a shows that the ZnO films formed microaggregates (~220 nm) composed of quasi-spherical ZnO nanoparticles (around 40 nm in diameter), and this is a typical result for this material sensitized by the sol-gel method. Figure 3b,c shows that the morphological properties changed significantly after the doping process. Regarding the ZnO:Cu thin films, Figure 3b shows the formation of nanorods. Meanwhile, Figure 3c shows the formation of nanosized elongated particles of various shapes (~100 nm) from the ZnO:Co thin films. Likewise, Figure 3c shows that the agglomeration on the catalyst surface reduced and the microaggregates disappeared. Different nanostructures have been reported for ZnO (e.g., nanorods, nanotubes, nanobelts, nanosprings, nanospirals, nanorings) [55]. It is known that ZnO's morphological properties rely on synthesis conditions, and in our case, it is clear that the metal ions used during synthesis reduced the agglomeration and changed the thin films' morphology [56].     Figure 3 shows SEM images for the catalysts. Figure 3a shows that the ZnO films formed microaggregates (~220 nm) composed of quasi-spherical ZnO nanoparticles (around 40 nm in diameter), and this is a typical result for this material sensitized by the sol-gel method. Figure 3b,c shows that the morphological properties changed significantly after the doping process. Regarding the ZnO:Cu thin films, Figure 3b shows the formation of nanorods. Meanwhile, Figure 3c shows the formation of nanosized elongated particles of various shapes (~100 nm) from the ZnO:Co thin films. Likewise, Figure 3c shows that the agglomeration on the catalyst surface reduced and the microaggregates Catalysts 2020, 10, 528 5 of 13 disappeared. Different nanostructures have been reported for ZnO (e.g., nanorods, nanotubes, nanobelts, nanosprings, nanospirals, nanorings) [55]. It is known that ZnO's morphological properties rely on synthesis conditions, and in our case, it is clear that the metal ions used during synthesis reduced the agglomeration and changed the thin films' morphology [56].

Optical Study
The diffuse reflectance spectra for the catalysts are shown in Figure 4. We used the Kubelka-Munk (KM) remission function for determining the bandgap energy value of the catalysts [57]. The use of the KM remission function makes it possible to obtain an analog to Tauc plots [58,59]:   Figure 5 shows that ZnO had a bandgap value (Eg) of 3.22 eV, a value that corresponds with that reported by Srikant el al. (3.1 eV and 3.2 eV) [60,61]. For doped ZnO catalysts, the Eg value was lower, and this behavior is associated with the reduction of the Fermi level of ZnO by the generation of intragap states. For ZnO:Cu, the modification of the bandgap can be attributed to the induction of 3d states of Cu located inside the bandgap of ZnO [37]. Additionally, the visible light absorption observed for doped ZnO can be attributed to intragap transitions between Cu 3d and Zn 4s states. Furthermore, the ZnO:Co 5% catalyst has a lower bandgap value compared to other catalysts. This reduction is attributed to s-d and p-d exchange interactions between ZnO and Co 2+ ions [62]. The 3d levels of Co 2+ are located within the bandgap of ZnO, which can create new bands at larger wavelengths [63]. Some photoluminescence studies of the transition metal doping ZnO nanoparticles suggest that this important reduction in Eg value is due to oxygen deficiency [64]. Finally, the

Optical Study
The diffuse reflectance spectra for the catalysts are shown in Figure 4. We used the Kubelka-Munk (KM) remission function for determining the bandgap energy value of the catalysts [57]. The use of the KM remission function makes it possible to obtain an analog to Tauc plots [58,59]: Catalysts 2020, 10, x FOR PEER REVIEW 5 of 13

Optical Study
The diffuse reflectance spectra for the catalysts are shown in Figure 4. We used the Kubelka-Munk (KM) remission function for determining the bandgap energy value of the catalysts [57]. The use of the KM remission function makes it possible to obtain an analog to Tauc plots [58,59]:  Figure 5 shows plots for ∝ ) ) versus (hv) and table 2 lists the optical properties of the catalysts. Figure 5 shows that ZnO had a bandgap value (Eg) of 3.22 eV, a value that corresponds with that reported by Srikant el al. (3.1 eV and 3.2 eV) [60,61]. For doped ZnO catalysts, the Eg value was lower, and this behavior is associated with the reduction of the Fermi level of ZnO by the generation of intragap states. For ZnO:Cu, the modification of the bandgap can be attributed to the induction of 3d states of Cu located inside the bandgap of ZnO [37]. Additionally, the visible light absorption observed for doped ZnO can be attributed to intragap transitions between Cu 3d and Zn 4s states. Furthermore, the ZnO:Co 5% catalyst has a lower bandgap value compared to other catalysts. This reduction is attributed to s-d and p-d exchange interactions between ZnO and Co 2+ ions [62]. The 3d levels of Co 2+ are located within the bandgap of ZnO, which can create new bands at larger wavelengths [63]. Some photoluminescence studies of the transition metal doping ZnO nanoparticles    Table 2 lists the optical properties of the catalysts. Figure 5 shows that ZnO had a bandgap value (E g ) of 3.22 eV, a value that corresponds with that reported by Srikant el al. (3.1 eV and 3.2 eV) [60,61]. For doped ZnO catalysts, the E g value was lower, and this behavior is associated with the reduction of the Fermi level of ZnO by the generation of intragap states. For ZnO:Cu, the modification of the bandgap can be attributed to the induction of 3d states of Cu located inside the bandgap of ZnO [37]. Additionally, the visible light absorption observed for doped ZnO can be attributed to intragap transitions between Cu 3d and Zn 4s states. Furthermore, the ZnO:Co 5% catalyst has a lower bandgap value compared to other catalysts. This reduction is attributed to s-d and p-d exchange interactions between ZnO and Co 2+ ions [62]. The 3d levels of Co 2+ are located within the bandgap of ZnO, which can create new bands at larger wavelengths [63]. Some photoluminescence studies of the transition metal doping ZnO nanoparticles suggest that this important reduction in E g value is due to oxygen deficiency [64]. Finally, the formation of nanoheterojunctions in the catalyst surface leads to an enhanced separation of charge carriers, increasing photocatalytic efficiency in addition to the doping process. The generation of these heterostructures has been reported for photocatalytic applications [65].
Catalysts 2020, 10, x FOR PEER REVIEW 6 of 13 carriers, increasing photocatalytic efficiency in addition to the doping process. The generation of these heterostructures has been reported for photocatalytic applications [65].  Figure 6 shows the decrease of MB as a function of time for all tests performed under visible irradiation. The MB concentration did not change after 140 min under visible irradiation, verifying the stability of MB dye. Furthermore, ZnO films did not show photocatalytic activity under visible irradiation (<3%). This result is in accordance with the ZnO bandgap energy value, and this photocatalyst is active only under UV irradiation. The ZnO:Co 5% catalyst reported the highest photocatalytic activity. This result can be explained by the lower bandgap value of the ZnO:Co 5% catalyst compared to other catalysts. The ZnO:Cu catalysts showed less photocatalytic activity than the Co-doped ZnO films. Compared to the Co-doped ZnO films, the bandgap values of this catalyst did not change; however, the best photodegradation result for ZnO:Cu was 42.5%, a value greater than that obtained for the ZnO thin films. The combined effect of the doping process and the heterojunction can explain this behavior. The photodegradation kinetics of MB on catalysts were studied by using the pseudo-first-order model [ Figure 6 shows the decrease of MB as a function of time for all tests performed under visible irradiation. The MB concentration did not change after 140 min under visible irradiation, verifying the stability of MB dye. Furthermore, ZnO films did not show photocatalytic activity under visible irradiation (<3%). This result is in accordance with the ZnO bandgap energy value, and this photocatalyst is active only under UV irradiation. The ZnO:Co 5% catalyst reported the highest photocatalytic activity. This result can be explained by the lower bandgap value of the ZnO:Co 5% catalyst compared to other catalysts. The ZnO:Cu catalysts showed less photocatalytic activity than the Co-doped ZnO films. Compared to the Co-doped ZnO films, the bandgap values of this catalyst did not change; however, the best photodegradation result for ZnO:Cu was 42.5%, a value greater than that obtained for the ZnO thin films. The combined effect of the doping process and the heterojunction can explain this behavior. The photodegradation kinetics of MB on catalysts were studied by using the pseudo-first-order model [66]:

Photocatalytic Study
the Co-doped ZnO films. Compared to the Co-doped ZnO films, the bandgap values of this catalyst did not change; however, the best photodegradation result for ZnO:Cu was 42.5%, a value greater than that obtained for the ZnO thin films. The combined effect of the doping process and the heterojunction can explain this behavior. The photodegradation kinetics of MB on catalysts were studied by using the pseudo-first-order model [66]: (2)  Time (t) is expressed in minutes and k app is the apparent reaction rate constant (min −1 ). Table 2 lists the kinetic parameters for the studied catalysts. Among all the catalysts, the ZnO thin films (k app = 0.2 × 10 −4 min −1 ) showed the lowest k ap value, while the best results were obtained for ZnO:Co 5% (k app = 7.2 × 10 −3 min −1 ) and ZnO:Cu 5% (k app = 4 × 10 −3 min −1 ). In the best case, the kinetic rate constant was 36 times higher than the ZnO thin films. A combined effect could be present: (i) Cu doping in ZnO and (ii) the formation of a nanoheterojunction (ZnO-CuO). This synergic effect could be a reason for the increase in photocatalytic yields. The heterostructure generation for the methyl orange photodegradation under visible light irradiation has been reported before [67]. Table 3 lists other reports for the use of doped ZnO with different metals as catalysts. Our results indicate that the catalysts produced in this study are suitable options for solar photocatalytic applications. The ZnO films did not show photocatalytic activity under visible irradiation. However, two different processes can contribute to photocatalytic degradation under visible irradiation: (i) the intraband transitions as dopants allow the doped ZnO thin films to absorb visible light, generating charge pairs; (ii) CuO and Co 3 O 4 can absorb visible light after the formation of ZnO-CuO and ZnO-Co 3 O 4 heterojunctions, generating charge pairs. In this case, the electron can be transferred to the conduction band of ZnO. After electrons are located at the conduction band of ZnO, different reactive oxygen species (ROS) can be generated (e.g., O − 2 ; OH), starting the MB photodegradation. Scheme 1 shows the general scheme of energetic levels for doped ZnO thin films and the ROS generation.

Synthesis and Characterization
The ZnO synthesis was carried out according to a previous report [69]: Twenty five mL of ammonium hydroxide (NH4OH) (25%-35% w/w) reactive grade (Merck) was placed in a 250 mL glass beaker, then 0.500 M (Zn(CH3COO)2.2H2O) (Merck) was added dropwise at a rate of 1.7 mL.min −1 for 1 hour, at a temperature of 85 °C under constant agitation at 300 rpm. After that, the suspension stood for three days at room temperature then the solid was filtered and dried for 5 h at 100 °C [72,73].
For the ZnO doping process, we used a similar procedure as previously described. While adding Zn 2+ ions, we also added salts of the doping metals (CuSO4.5H2O) (Merck) for copper doping and (Co(CH3COO)2.4H2O) (JT Baker) for the cobalt doping processes. The synthesis of doped ZnO powder was performed at 1.0%, 3.0%, and 5.0%. The thin films were deposited using the doctor blade method, and the suspension was placed in a glass measuring 2 cm high and 2 cm wide. The thin films were heated for 30 min at 90 °C to evaporate the solvent, and finally, the sintering process was performed at 500 °C for 2 h [73,74]. Using this procedure, we obtained thin films (6 μm thickness). The film thickness was measured through a Veeco Dektak 150 profilometer. The physical chemistry properties of the films were studied by X-ray diffraction, diffuse reflectance spectrophotometry, and Raman spectroscopy assay. X-ray diffraction patterns were obtained using a Shimadzu 6000 diffractometer using Cu Kα radiation (λ = 0.15406 nm) as an X-ray source with a diffraction angle in the 2θ range (20-80°). Diffuse reflectance spectra were obtained with a Lambda 4 Perkin-Elmer spectrophotometer equipped with an integration sphere. The compositional properties of the materials were studied by Raman spectroscopy in a DXR device equipped with a 780 nm laser. The morphological properties were studied by scanning electron microscopy, under an excitation energy of 5 and 1 kV. The metallic content of the films was determined by plasma emission spectroscopy using the SM 3120 B technique, EPA 3015A modified for solids (see Supporting Information).

Photocatalytic Test
Methylene blue (MB) was chosen as the pollutant model in this study. The experiments were carried out in a batch reactor using an LED tape as a source of visible radiation (cold white light 17 watts), and the incident photon flow per unit volume Io was 5.8 × 10 −7 Einstein*L −1 s −1 . Before

Synthesis and Characterization
The ZnO synthesis was carried out according to a previous report [69]: Twenty five mL of ammonium hydroxide (NH 4 OH) (25-35% w/w) reactive grade (Merck) was placed in a 250 mL glass beaker, then 0.500 M (Zn(CH 3 COO) 2 ·2H 2 O) (Merck) was added dropwise at a rate of 1.7 mL.min −1 for 1 h, at a temperature of 85 • C under constant agitation at 300 rpm. After that, the suspension stood for three days at room temperature then the solid was filtered and dried for 5 h at 100 • C [72,73].
For the ZnO doping process, we used a similar procedure as previously described. While adding Zn 2+ ions, we also added salts of the doping metals (CuSO 4 ·5H 2 O) (Merck) for copper doping and (Co(CH 3 COO) 2 ·4H 2 O) (JT Baker) for the cobalt doping processes. The synthesis of doped ZnO powder was performed at 1.0%, 3.0%, and 5.0%. The thin films were deposited using the doctor blade method, and the suspension was placed in a glass measuring 2 cm high and 2 cm wide. The thin films were heated for 30 min at 90 • C to evaporate the solvent, and finally, the sintering process was performed at 500 • C for 2 h [73,74]. Using this procedure, we obtained thin films (6 µm thickness). The film thickness was measured through a Veeco Dektak 150 profilometer. The physical chemistry properties of the films were studied by X-ray diffraction, diffuse reflectance spectrophotometry, and Raman spectroscopy assay. X-ray diffraction patterns were obtained using a Shimadzu 6000 diffractometer using Cu Kα radiation (λ = 0.15406 nm) as an X-ray source with a diffraction angle in the 2θ range (20-80 • ). Diffuse reflectance spectra were obtained with a Lambda 4 Perkin-Elmer spectrophotometer equipped with an integration sphere. The compositional properties of the materials were studied by Raman spectroscopy in a DXR device equipped with a 780 nm laser. The morphological properties were studied by scanning electron microscopy, under an excitation energy of 5 and 1 kV. The metallic content of the films was determined by plasma emission spectroscopy using the SM 3120 B technique, EPA 3015A modified for solids (see Supporting Information).

Photocatalytic Test
Methylene blue (MB) was chosen as the pollutant model in this study. The experiments were carried out in a batch reactor using an LED tape as a source of visible radiation (cold white light 17 watts), and the incident photon flow per unit volume I o was 5.8 × 10 −7 Einstein*L −1 s −1 . Before irradiation, the MB solution was kept in the dark for 90 min at 250 rpm to reach adsorption-desorption equilibrium on the catalysts' surface. Photodegradation was carried out using 50 ± 0.025 mL of an MB solution (10 mg·L −1 ) saturated with oxygen at pH 7.0. The concentration of dye was determined through the spectrophotometric method (Thermo Scientific-Genesys 10S) using 665 nm as a fixed wavelength, with a calibration curve (correlation coefficient R = 0.997) for the use of the Lambert-Beer equation.

Conclusions
We synthesized and characterized ZnO thin films doped with Co and Cu. Raman results corroborated the doping process, which suggested the generation of a heterostructure. For the doped ZnO catalysts, the results show a reduction in the E g values (from 3.22 to 2.42 eV for the best catalyst). This behavior is associated with a reduction of the Fermi level of ZnO by the presence of intragap states. The photocatalytic test indicated that doped catalysts had greater photocatalytic activity than unmodified catalysts, which could be attributed to (i) the generation of intraband states for the insertion of Co and Cu into ZnO, and (ii) the generation of ZnO-CuO and ZnO-Co 3 O 4 heterojunctions. Furthermore, the rate of the photodegradation process (ZnO:Co 5%) was 36 times greater than the rate for unmodified ZnO, and the k app values for the catalysts synthesized in this study had a suitable photocatalytic activity compared to other reports.