Influence of Mg, Cu, and Ni Dopants on Amorphous TiO2 Thin Films Photocatalytic Activity

The present study investigates Mg (0 ÷ 17.5 wt %), Cu (0 ÷ 21 wt %) and Ni (0 ÷ 20.2 wt %) dopants (M-doped) influence on photocatalytic activity of amorphous TiO2 thin films. Magnetron sputtering was used for the deposition of M-doped TiO2 thin films. According to SEM/EDS surface analysis, the magnetron sputtering technique allows making M-doped TiO2 thin films with high uniformity and high dopant dispersion. Photocatalysis efficiency analysis was set in oxalic acid under UV irradiation. In accordance with the TOC (total organic carbon) measurements followed by the apparent rate constant (kapp) results, the dopants’ concentration peak value was dopant-dependent; for Mg/TiO2, it is 0.9% (kapp—0.01866 cm−1), for Cu/TiO2, it is 0.6% (kapp—0.02221 cm−1), and for Ni/TiO2, it is 0.5% (kapp—0.01317 cm−1). The obtained results clearly state that a concentration of dopants in TiO2 between 0.1% and 0.9% results in optimal photocatalytic activity.


Introduction
Photocatalysis and photocatalytic materials based on semiconductors have been studied for more than a decade [1][2][3]. By reason of higher efficiency and future potential in water treatment [4], air purification [5] and even hydrogen production [6], photocatalysis interest increased significantly. The number of articles related to photocatalysis has risen thousands of times over the past 20 years. Additionally, research based on TiO 2 as a photocatalyst among other semiconductors also increased [7]. TiO 2 manufacturing is affordable, which makes it economically favorable, especially when compared with materials having similar properties, such as SnO 2 , CeO 2 , CdS, and WO 3 . These semiconductors are also used for photocatalysis research because of their biocompatibility, stability in various conditions, and capability to generate excitons [8,9]. TiO 2 achieves better photocatalytic activity than ZnO [10] or CdS [11] and under the same conditions; not only does it have better photochemical stability, but TiO 2 is superior photocatalyst compared to WO 3 [12]. TiO 2 has great potential in energy and environmental research, such as uses in lithium-ion batteries, self-cleaning coatings, water purification, or as a catalyst for photocatalytic reactions [13][14][15][16]. As a result of its unique dielectric and optical properties, TiO 2 is considered as a non-hazardous material and could be modified, depending on the requirements for

Films Preparation
Thin films were deposited using the Kurt J. Lesker PVD 75 vacuum system with four magnetron sputtering stages. Argon and oxygen gases (with 99.999% purity) were used during the plasma activation and deposition processes. Two titanium (Ti) targets (99.995% purity), a copper (Cu) target (99.99% purity), a magnesium (Mg) target (99.95% purity), and a nickel (Ni) target (99.995% purity), all purchased from Sigma Aldrich ® , were used for thin-film formation. Power sources including direct current (DC), pulsed DC, and radio frequency (RF) were used. The DC source was used for the Ti and Ni targets, pulsed DC was used for the Mg target, and RF was used for the Cu target. TiO 2 and Cu, Ni, and Mg-doped TiO 2 films were deposited on stainless steel plates (304 series). The substrate was cleaned in pure acetone (for 10 min) using an ultrasonic bath in order to avoid any organic contamination. Thin films were deposited at 300 • C maintaining the ratio of oxygen and argon at 20/80. M-doped TiO 2 were produced using two titanium cathodes (DC source) and one metal cathode (DC, pulsed DC, or RF source) target. Different output power and shutter open/close ratios were used to achieve different concentrations of M in TiO 2 films. The total thickness of the formed thin films was around 100 nm; the structure based on the deposition procedure is shown in Figure 1. Different Cu, Ni, and Mg concentrations in TiO 2 thin films were achieved (Table 1), i.e., Mg, 0 ÷ 17.5 wt %; Cu, 0 ÷ 21 wt %; and Ni, 0 ÷ 20.2 wt %. The concentration of dopants in TiO 2 was analyzed via X-ray photoelectron spectrometer (XPS) and energy-dispersive X-ray spectroscopy (EDS).

Films preparation
Thin films were deposited using the Kurt J. Lesker PVD 75 vacuum system with four magnetron sputtering stages. Argon and oxygen gases (with 99.999% purity) were used during the plasma activation and deposition processes. Two titanium (Ti) targets (99.995% purity), a copper (Cu) target (99.99% purity), a magnesium (Mg) target (99.95% purity), and a nickel (Ni) target (99.995% purity), all purchased from Sigma Aldrich ® , were used for thin-film formation. Power sources including direct current (DC), pulsed DC, and radio frequency (RF) were used. The DC source was used for the Ti and Ni targets, pulsed DC was used for the Mg target, and RF was used for the Cu target. TiO2 and Cu, Ni, and Mg-doped TiO2 films were deposited on stainless steel plates (304 series). The substrate was cleaned in pure acetone (for 10 minutes) using an ultrasonic bath in order to avoid any organic contamination. Thin films were deposited at 300 °C maintaining the ratio of oxygen and argon at 20/80. M-doped TiO2 were produced using two titanium cathodes (DC source) and one metal cathode (DC, pulsed DC, or RF source) target. Different output power and shutter open/close ratios were used to achieve different concentrations of M in TiO2 films. The total thickness of the formed thin films was around 100 nm; the structure based on the deposition procedure is shown in Figure 1. Different Cu, Ni, and Mg concentrations in TiO2 thin films were achieved (Table 1), i.e., Mg, 0÷17.5 wt %; Cu, 0 ÷ 21 wt %; and Ni, 0 ÷ 20.2 wt %. The concentration of dopants in TiO2 was analyzed via X-ray photoelectron spectrometer (XPS) and energy-dispersive X-ray spectroscopy (EDS).

The morphology and structural analysis
The structure of the TiO2 phase was determined using an X-ray diffractometer (XRD) "Bruker D8 Discover" at 2Θ angle in a 20° to 70° range using Cu Kα (λ=0.154059 nm) radiation, 0.01° step, and a Lynx eye position-sensitive detector (PSD).
The surface topography images were obtained via scanning electron microscope (FE-SEM; JEOL, SM-71010) using 8 kV accelerating voltage.
The distribution of elements (mapping) and a high concentration of dopants were measured using an energy-dispersive X-ray spectroscope "BrukerXFlash Quad 5040" (EDS) with an accelerating voltage of 10 kV.
The low concentration of dopants in TiO2 was determined by the X-ray photoelectron spectrometer (XPS, PHI Versaprobe 5000). Measurements were performed using monochromatic Xray radiation (Al Kα, 1486.6 eV), 25 W power, 100 μm beam size, and the measurement angle was 45° during the experiments. The sample charging effect was compensated using the radiation of low energy electrons and ions. The resolution was 1 eV for Survey Spectrum and 0.1 eV for detailed chemical analysis. The pass energy was 187.580 eV. Thin films were sputtered in order to achieve the depth profile of the samples [48].

The Morphology and Structural Analysis
The structure of the TiO 2 phase was determined using an X-ray diffractometer (XRD) "Bruker D8 Discover" at 2Θ angle in a 20 • to 70 • range using Cu Kα (λ = 0.154059 nm) radiation, 0.01 • step, and a Lynx eye position-sensitive detector (PSD).
The surface topography images were obtained via scanning electron microscope (FE-SEM; JEOL, SM-71010) using 8 kV accelerating voltage. The distribution of elements (mapping) and a high concentration of dopants were measured using an energy-dispersive X-ray spectroscope "BrukerXFlash Quad 5040" (EDS) with an accelerating voltage of 10 kV.
The low concentration of dopants in TiO 2 was determined by the X-ray photoelectron spectrometer (XPS, PHI Versaprobe 5000). Measurements were performed using monochromatic X-ray radiation (Al Kα, 1486.6 eV), 25 W power, 100 µm beam size, and the measurement angle was 45 • during the experiments. The sample charging effect was compensated using the radiation of low energy electrons and ions. The resolution was 1 eV for Survey Spectrum and 0.1 eV for detailed chemical analysis. The pass energy was 187.580 eV. Thin films were sputtered in order to achieve the depth profile of the samples [48].

Photocatalysis Efficiency Evaluation
To maximize the efficiency of the photocatalysis process, a transparent solution must be chosen. There are many types of research where methylene blue (MB) [49,50], rhodamine B (RhB), [51,52] or other organic dyes were selected as a solution for this process. In that case, the light absorption of the solution itself and the degradation process happening without the photocatalyst should be considered as an extraneous effect, which could cause slightly inaccurate results of the semiconductor as a photocatalyst [53][54][55]. For this reason, measurements were set in oxalic acid solution (which is as well a product of RhB during the photodegradation process) made from water and oxalic acid powder (with an initial concentration of 100 mg/l) under UV light (light peak at 254 nm, 18 W power). According to the light source used in the experiment, absorption spectra were measured but not analyzed in this manuscript because of the same absorbance at 254 nm for all the samples. In this range of light, metal dopants do not have an effect on the absorption characteristics of TiO 2 . The samples (surface area-1600 mm 2 ) were immersed in the solution and set for 30 min in the dark before every experiment. After that, the samples were irradiated for 80 min, and the solution was continuously stirred using a magnetic stirrer. The distance between the UV lamp and the sample was set to 50 mm, while the distance between the surface of the solution and the sample was 20 mm. Samples of solution were taken after 20, 40, 60, and 80 min of irradiation, and the degradation of the oxalic acid solution was measured in a Shimadzu TOC-L (Shimadzu Corp. Japan) analyzer according to the EN 1484:2002 procedure. To minimize errors, there were three measurements for each sample taken from the solution, and the average concentration was analyzed. The degradation steps of oxalic acid solution in water are shown in Figure 2. Initially, C 2 H 2 O 4 decomposes into formic acid, and with further degradation into the innocuous compounds (H 2 O and CO 2 ), through intramolecular dehydration.

Photocatalysis efficiency evaluation
To maximize the efficiency of the photocatalysis process, a transparent solution must be chosen. There are many types of research where methylene blue (MB) [49,50], rhodamine B (RhB), [51,52] or other organic dyes were selected as a solution for this process. In that case, the light absorption of the solution itself and the degradation process happening without the photocatalyst should be considered as an extraneous effect, which could cause slightly inaccurate results of the semiconductor as a photocatalyst [53][54][55]. For this reason, measurements were set in oxalic acid solution (which is as well a product of RhB during the photodegradation process) made from water and oxalic acid powder (with an initial concentration of 100 mg/l) under UV light (light peak at 254 nm, 18 W power). According to the light source used in the experiment, absorption spectra were measured but not analyzed in this manuscript because of the same absorbance at 254 nm for all the samples. In this range of light, metal dopants do not have an effect on the absorption characteristics of TiO2. The samples (surface area -1600 mm 2 ) were immersed in the solution and set for 30 min in the dark before every experiment. After that, the samples were irradiated for 80 min, and the solution was continuously stirred using a magnetic stirrer. The distance between the UV lamp and the sample was set to 50 mm, while the distance between the surface of the solution and the sample was 20 mm. Samples of solution were taken after 20, 40, 60, and 80 min of irradiation, and the degradation of the oxalic acid solution was measured in a Shimadzu TOC-L (Shimadzu Corp. Japan) analyzer according to the EN 1484:2002 procedure. To minimize errors, there were three measurements for each sample taken from the solution, and the average concentration was analyzed. The degradation steps of oxalic acid solution in water are shown in Figure 2. Initially, C2H2O4 decomposes into formic acid, and with further degradation into the innocuous compounds (H2O and CO2), through intramolecular dehydration. The results were evaluated based on the thermodynamics and kinetics of the process. The driving force of photocatalysis is an organic solution and oxygen molecules. Oxygen is constant in photocatalysis because the experiment is set in the atmospheric ambient and the solution is surrounded by oxygen molecules. The adsorption of organic compounds on the surface of the photocatalyst is the kinetic driving force of the photocatalysis process. Adsorption and desorption equilibrium follows Langmuir isotherm [56,57]. In this case, under UV irradiation, there is an equilibrium in adsorbed and desorbed molecules.
This evaluation is useful when adsorption and desorption are in equilibrium [58], but because the photocatalysis efficiency is low, it could be applied as a universal equation for different solutions: where kapp -apparent rate constant, CO -initial concentration, Ct -concentration in time, and t -time.
Keeping in mind that kapp depends on the time and concentration differences in time, average meanings were calculated for the different concentrations of dopants in TiO2 (Table 1). The results were evaluated based on the thermodynamics and kinetics of the process. The driving force of photocatalysis is an organic solution and oxygen molecules. Oxygen is constant in photocatalysis because the experiment is set in the atmospheric ambient and the solution is surrounded by oxygen molecules. The adsorption of organic compounds on the surface of the photocatalyst is the kinetic driving force of the photocatalysis process. Adsorption and desorption equilibrium follows Langmuir isotherm [56,57]. In this case, under UV irradiation, there is an equilibrium in adsorbed and desorbed molecules. This evaluation is useful when adsorption and desorption are in equilibrium [58], but because the photocatalysis efficiency is low, it could be applied as a universal equation for different solutions: where k app -apparent rate constant, C O -initial concentration, C t -concentration in time, and t-time.
Keeping in mind that k app depends on the time and concentration differences in time, average meanings were calculated for the different concentrations of dopants in TiO 2 (Table 1).

Results and Discussion
The main parameters for a dopant to be effective as a charge trapper are its concentration and dispersion in TiO 2 lattice, as well as its electron configuration. X-ray Diffraction (XRD) analysis shows that deposited TiO 2 thin films were amorphous ( Figure 3).

Results and Discussion
The main parameters for a dopant to be effective as a charge trapper are its concentration and dispersion in TiO2 lattice, as well as its electron configuration. X-ray Diffraction (XRD) analysis shows that deposited TiO2 thin films were amorphous ( Figure 3). According to elemental analysis ( Figure 4), there is a high dispersity of dopant clusters in TiO2 thin films. Cluster dispersity defines photocatalysis efficiency, since dopant clusters act as electron trappers. Based on the results (Table 1) and theory, cluster size and dispersity, as well as photocatalysis efficiency, depend on dopant concentration. Thus, theoretically, the same efficiency could also have a similar dependence on size and dispersity. Table 1. Dopants concentration in samples and apparent rate constant kapp (where OA is an oxalic acid solution). Dopants concentration was analyzed via X-ray photoelectron spectrometer (XPS) and energy-dispersive X-ray spectroscopy (EDS).

2
According to elemental analysis (Figure 4), there is a high dispersity of dopant clusters in TiO 2 thin films. Cluster dispersity defines photocatalysis efficiency, since dopant clusters act as electron trappers. Based on the results (Table 1) and theory, cluster size and dispersity, as well as photocatalysis efficiency, depend on dopant concentration. Thus, theoretically, the same efficiency could also have a similar dependence on size and dispersity.

Results and Discussion
The main parameters for a dopant to be effective as a charge trapper are its concentration and dispersion in TiO2 lattice, as well as its electron configuration. X-ray Diffraction (XRD) analysis shows that deposited TiO2 thin films were amorphous ( Figure 3). According to elemental analysis (Figure 4), there is a high dispersity of dopant clusters in TiO2 thin films. Cluster dispersity defines photocatalysis efficiency, since dopant clusters act as electron trappers. Based on the results (Table 1) and theory, cluster size and dispersity, as well as photocatalysis efficiency, depend on dopant concentration. Thus, theoretically, the same efficiency could also have a similar dependence on size and dispersity. Table 1. Dopants concentration in samples and apparent rate constant kapp (where OA is an oxalic acid solution). Dopants concentration was analyzed via X-ray photoelectron spectrometer (XPS) and energy-dispersive X-ray spectroscopy (EDS).

Mg/TiO2
Cu   Additionally, XPS analysis was set for samples with a low concentration of dopant in TiO 2 . The calibration of XPS spectra was done according to the standard value of C 1s peak at 284.8 eV BE (binding energy). The reference data for XPS peaks were taken from the "Thermo Fisher" (Thermo Fisher Scientific Inc.) database. After the calibration, core level peaks of TiO 2 ( Figure 5c) and dopants ( Figure 5a) were measured. Ti 2p 3/2 and Ti 2p 1/2 peaks are at 458.5 eV and 464.5 eV BE (compared to TiO 2 Ti 2p 3/2 -458.5 eV BE and Ti 2p 1/2 -464.5 eV BE). The principal Mg KLL Auger peak is at 305 eV BE, accompanied with an Mg 1s peak at 1303 eV BE (ref. Mg 1s-1303 eV BE for Mg metal and 1304.5 eV BE for Mg native oxide). Furthermore, it is hard to distinguish Cu/TiO 2 spectra if a Cu 2 O or Cu structure formed because of low concentration and low peak intensity. Cu 2p 1/2 and Cu 2p 3/2 peaks are at the same energy value of 933 eV BE and 955 eV BE (accordingly). There is also a weak satellite at 945 eV BE, which helps to separate Cu metal and Cu 2 O. The Ni 2p 3/2 peak is at 855 eV BE (ref. Ni metal-852.6 eV BE; NiO at 853.7 eV BE). The depth profile of doped TiO 2 thin films is shown in Figure 5b. In agreement with XPS results, dopants clusters are not oxidized, while the main dopant metal peaks are at the same energy value according to the reference. However, based on the deposition parameters, where an oxygen and argon ratio at 20/80 was used, we consider that at some points, the dopant material could be slightly oxidized. Additionally, XPS analysis was set for samples with a low concentration of dopant in TiO2. The calibration of XPS spectra was done according to the standard value of C 1s peak at 284.8 eV BE (binding energy). The reference data for XPS peaks were taken from the "Thermo Fisher" (Thermo Fisher Scientific Inc.) database. After the calibration, core level peaks of TiO2 ( Figure 5c) and dopants (Figure 5a) were measured. Ti 2p3/2 and Ti 2p1/2 peaks are at 458.5 eV and 464.5 eV BE (compared to TiO2 Ti 2p3/2 -458.5 eV BE and Ti 2p1/2 -464.5 eV BE). The principal Mg KLL Auger peak is at 305 eV BE, accompanied with an Mg 1s peak at 1303 eV BE (ref. Mg 1s -1303 eV BE for Mg metal and 1304.5 eV BE for Mg native oxide). Furthermore, it is hard to distinguish Cu/TiO2 spectra if a Cu2O or Cu structure formed because of low concentration and low peak intensity. Cu 2p1/2 and Cu 2p3/2 peaks are at the same energy value of 933 eV BE and 955 eV BE (accordingly). There is also a weak satellite at 945 eV BE, which helps to separate Cu metal and Cu2O. The Ni 2p3/2 peak is at 855 eV BE (ref. Ni metal -852.6 eV BE; NiO at 853.7 eV BE). The depth profile of doped TiO2 thin films is shown in Figure 5b. In agreement with XPS results, dopants clusters are not oxidized, while the main dopant metal peaks are at the same energy value according to the reference. However, based on the deposition parameters, where an oxygen and argon ratio at 20/80 was used, we consider that at some points, the dopant material could be slightly oxidized. The evaluation results (Table 1) show that increasing the concentration of dopant above 1 wt % in TiO2 thin films results in a decrease in the efficiency of the photocatalysis process. Although, if the concentration of dopant is lower than 1 wt %, the efficiency increases in the first 20 min of irradiation, while the optimal amount of concentration depends on the dopant itself. In this study, Cu-doped TiO2 reaches maximum efficiency at 0.6 wt %, Ni-doped TiO2 reaches maximum efficiency at 0.5 wt %, and Mg-doped TiO2 reaches maximum efficiency at 0.9 wt %. Therefore, the optimal concentration The evaluation results (Table 1) show that increasing the concentration of dopant above 1 wt % in TiO 2 thin films results in a decrease in the efficiency of the photocatalysis process. Although, if the concentration of dopant is lower than 1 wt %, the efficiency increases in the first 20 min of irradiation, while the optimal amount of concentration depends on the dopant itself. In this study, Cu-doped TiO 2 reaches maximum efficiency at 0.6 wt %, Ni-doped TiO 2 reaches maximum efficiency at 0.5 wt %, and Mg-doped TiO 2 reaches maximum efficiency at 0.9 wt %. Therefore, the optimal concentration value suggests that there is an optimal cluster size for maximum efficiency. 42 Efficiency dependence on cluster dispersity and size has not been investigated in this study. The degradation of oxalic acid in time at different Mg wt % concentrations is shown in Figure 6a. The integrated area under the curve (gray area) indicates the photocatalysis process efficiency as well as apparent rate constant k app . The smaller the area and the higher the constant, the higher the process efficiency. The calculated values show that the higher efficiency of Mg/TiO 2 as a photocatalyst is achieved when the dopant concentration in the TiO 2 sample is 0.9 wt % (k app = 0.01866 min −1 ). The photocatalysis process dependence on dopant concentration is shown in Figure 6b. Compared to pure TiO 2 thin films, k app drops significantly when the concentration of Mg is increased by more than 1 wt %. There is a peak in concentration where the photocatalysis efficiency reaches maximum values. It could be stated that there should be a peak with a lower concentration of Mg. As Manzanare's research shows, a small amount of Mg dopant in the TiO 2 lattice increases the electron trap (Ti 3+ ) and hole trap (O S ) concentration, providing a low recombination of electron-hole pairs [59]. On the other hand, Mg clusters on the TiO 2 surface can act as an electron donor for TiO 2 . In this way, oxygen can be directly affected by injected electrons and followed by easily formed oxygen vacancies on the unstable TiO 2 surface [60]. Different concentrations of Mg were used; the most efficient concentration was 0.9 wt %, while 1 wt % was lower but still higher than pure TiO 2 .
Materials 2020, 13, x FOR PEER REVIEW 7 of 15 value suggests that there is an optimal cluster size for maximum efficiency. 42 Efficiency dependence on cluster dispersity and size has not been investigated in this study. The degradation of oxalic acid in time at different Mg wt % concentrations is shown in Figure  6a. The integrated area under the curve (gray area) indicates the photocatalysis process efficiency as well as apparent rate constant kapp. The smaller the area and the higher the constant, the higher the process efficiency. The calculated values show that the higher efficiency of Mg/TiO2 as a photocatalyst is achieved when the dopant concentration in the TiO2 sample is 0.9 wt % (kapp = 0.01866 min −1 ). The photocatalysis process dependence on dopant concentration is shown in Figure 6b. Compared to pure TiO2 thin films, kapp drops significantly when the concentration of Mg is increased by more than 1 wt %. There is a peak in concentration where the photocatalysis efficiency reaches maximum values. It could be stated that there should be a peak with a lower concentration of Mg. As Manzanare's research shows, a small amount of Mg dopant in the TiO2 lattice increases the electron trap (Ti 3+ ) and hole trap (OS) concentration, providing a low recombination of electron-hole pairs [59]. On the other hand, Mg clusters on the TiO2 surface can act as an electron donor for TiO2. In this way, oxygen can be directly affected by injected electrons and followed by easily formed oxygen vacancies on the unstable TiO2 surface [60]. Different concentrations of Mg were used; the most efficient concentration was 0.9 wt %, while 1 wt % was lower but still higher than pure TiO2. Accordingly, the same measurements were set with Cu and Ni dopants as well. Depending on the integrated area under the curve (Figure 7a), Cu (0.6 wt %) is more efficient as dopant compared to Mg (0.9 wt %). Furthermore, comparing the same samples, after the first 20 min, 58% of solution degraded with Cu/TiO2 and only 37% degraded with Mg/TiO2 (Figure 7b and Figure 6b). Cu concentration values between 0.5 and 0.7 are considered optimal concentrations where the degradation of oxalic acid increases drastically. Decreasing or increasing the Cu concentration in the TiO2 lattice negatively affects the efficiency of the process. The sample with 0.6 wt % of Cu reaches the maximum efficiency (kapp=0.02221 min −1 ), while pure TiO2 reaches the maximum efficiency at kapp=0.01160 min −1 , and the sample with 0.3 wt % of Cu reaches the maximum efficiency at kapp=0.01373 (Table 1). Bensouici's studies show that the pure TiO2 reaction rate constant is kapp=0.015 min −1 , and it decreases to kapp=0.001 min −1 for Cu 4 wt % and remains stable for 6-10 wt % Cu-doped TiO2 thin films [61]. In this study, there was a quite huge change in concentration between 0.3 wt % and 2 wt %, while in Minsu Jung's research [62], a concentration of 0.5 wt % Cu in the TiO2 lattice reaches the highest H2 production rate, and decreasing or increasing this value lowers the production of H2. However, other research states that Cu is not an effective dopant for TiO2 modification. With the increased concentration of Cu in TiO2, the yield of a product decreases, as compared to pure TiO2 [63]. However, analyzing the dependency of dopant concentration on photocatalysis efficiency, low concentrations must be tested by slowly increasing them. According to this statement, Tashibi et al. Accordingly, the same measurements were set with Cu and Ni dopants as well. Depending on the integrated area under the curve (Figure 7a), Cu (0.6 wt %) is more efficient as dopant compared to Mg (0.9 wt %). Furthermore, comparing the same samples, after the first 20 min, 58% of solution degraded with Cu/TiO 2 and only 37% degraded with Mg/TiO 2 (Figures 6b and 7b). Cu concentration values between 0.5 and 0.7 are considered optimal concentrations where the degradation of oxalic acid increases drastically. Decreasing or increasing the Cu concentration in the TiO 2 lattice negatively affects the efficiency of the process. The sample with 0.6 wt % of Cu reaches the maximum efficiency (k app = 0.02221 min −1 ), while pure TiO 2 reaches the maximum efficiency at k app = 0.01160 min −1 , and the sample with 0.3 wt % of Cu reaches the maximum efficiency at k app = 0.01373 (Table 1). Bensouici's studies show that the pure TiO 2 reaction rate constant is k app = 0.015 min −1 , and it decreases to k app = 0.001 min −1 for Cu 4 wt % and remains stable for 6-10 wt % Cu-doped TiO 2 thin films [61]. In this study, there was a quite huge change in concentration between 0.3 wt % and 2 wt %, while in Minsu Jung's research [62], a concentration of 0.5 wt % Cu in the TiO 2 lattice reaches the highest H 2 production rate, and decreasing or increasing this value lowers the production of H 2 . However, other research states that Cu is not an effective dopant for TiO 2 modification. With the increased concentration of Cu in TiO 2 , the yield of a product decreases, as compared to pure TiO 2 [63]. However, analyzing the dependency of dopant concentration on photocatalysis efficiency, low concentrations must be tested by slowly increasing them. According to this statement, Tashibi et al. analyzed 0.2 wt % and 0.9 wt % of Cu in TiO 2 , skipping the crucial point of approximately 0.5 wt % where photocatalysis efficiency reaches higher values when compared to pure TiO 2 . Despite that, this research explains the inactivity of Cu-doped TiO 2 by the formed CuO on the surface, which decreases the number of active sites of TiO 2 as well as becomes a recombination center and increases the rate of the charge carrier recombination. However, the optimal amount of clusters results in the dispersion of them enough to reduce recombination rates and act as charge trappers. The photodegradation of oxalic acid reaches its maximum efficiency when the Cu concentration is 0.6 wt % in TiO 2 films, and similar to other studies, the efficiency drops when the concentration is raised or lowered. Different thin film deposition techniques, preparation, and materials (solid, liquid, or gas phase) can cause slightly different results in the experiment. Thin-film parameters such as the crystal size, purity, and surface area apart from the crystal phase rely on deposition techniques, as described earlier in this article. However, a different evaluation method can cause varied results in photocatalysis efficiency. analyzed 0.2 wt % and 0.9 wt % of Cu in TiO2, skipping the crucial point of approximately 0.5 wt % where photocatalysis efficiency reaches higher values when compared to pure TiO2. Despite that, this research explains the inactivity of Cu-doped TiO2 by the formed CuO on the surface, which decreases the number of active sites of TiO2 as well as becomes a recombination center and increases the rate of the charge carrier recombination. However, the optimal amount of clusters results in the dispersion of them enough to reduce recombination rates and act as charge trappers. The photodegradation of oxalic acid reaches its maximum efficiency when the Cu concentration is 0.6 wt % in TiO2 films, and similar to other studies, the efficiency drops when the concentration is raised or lowered. Different thin film deposition techniques, preparation, and materials (solid, liquid, or gas phase) can cause slightly different results in the experiment. Thin-film parameters such as the crystal size, purity, and surface area apart from the crystal phase rely on deposition techniques, as described earlier in this article. However, a different evaluation method can cause varied results in photocatalysis efficiency. Ni/TiO2 thin films as photocatalyst results analysis show that the photodegradation of oxalic acid is efficient with 0.5 wt % of Ni in the TiO2 lattice (Figure 8a), but it is similar to pure TiO2 thin films results. After 20 min, only 34% of oxalic acid solution degrades, which is similar to 0.9 wt % Mg, but taking results after 40 min in comparison, for Mg, it is 67% (Figure 6b), and for Ni, it is 43% (Figure 8b). Chen's [64] studies show that the Ni/TiO2 photocatalyst exhibits the highest photocatalytic activity when the optimal Ni concentration is 0.5 wt %. Experimental results, while using oxalic acid as a solution for the examination of photocatalytic properties, and Ni/TiO2 as the photocatalytic show the same results. Keeping in mind that the degradation process depends on the adsorption of the solution on the surface, increasing the concentration of Ni on TiO2 can result in a decrease in surface irrigation. Other research based on Ni concentration in crystal phase TiO2 shows that there is an optimal concentration of Ni on crystal phase TiO2 as well, with which the photocatalytic efficiency increases drastically. The decrease in efficiency based on the concentration is explained by decreasing the number of active sites of TiO2 on the surface with increased Ni concentration [65]. With the right amount of Ni clusters on the TiO2 surface, Ni acts as a co-catalyst by separating and transferring photogenerated charge carriers, thus decreasing the recombination rates compared to those of pure TiO2 [66].
Based on the results in Figure 6-8a, the degradation rate of the oxalic acid solution, while using optimal dopant concentration, drops exponentially. In the first 20-40 min, the degradation rate is higher compared to degradation after 40 min, when the reaction almost stabilizes. After the experiment with Cu (0.3 and 0.6 wt %), Mg (0.9 wt %) and Ni (0.5 wt %) doped TiO2 films, visible surface area degradation was observed. The thin film degradation was not detected using TiO2 with higher dopant concentration. The reactivation of such catalysts has to be taken into account Ni/TiO 2 thin films as photocatalyst results analysis show that the photodegradation of oxalic acid is efficient with 0.5 wt % of Ni in the TiO 2 lattice (Figure 8a), but it is similar to pure TiO 2 thin films results. After 20 min, only 34% of oxalic acid solution degrades, which is similar to 0.9 wt % Mg, but taking results after 40 min in comparison, for Mg, it is 67% (Figure 6b), and for Ni, it is 43% (Figure 8b). Chen's [64] studies show that the Ni/TiO 2 photocatalyst exhibits the highest photocatalytic activity when the optimal Ni concentration is 0.5 wt %. Experimental results, while using oxalic acid as a solution for the examination of photocatalytic properties, and Ni/TiO 2 as the photocatalytic show the same results. Keeping in mind that the degradation process depends on the adsorption of the solution on the surface, increasing the concentration of Ni on TiO 2 can result in a decrease in surface irrigation. Other research based on Ni concentration in crystal phase TiO 2 shows that there is an optimal concentration of Ni on crystal phase TiO 2 as well, with which the photocatalytic efficiency increases drastically. The decrease in efficiency based on the concentration is explained by decreasing the number of active sites of TiO 2 on the surface with increased Ni concentration [65]. With the right amount of Ni clusters on the TiO 2 surface, Ni acts as a co-catalyst by separating and transferring photogenerated charge carriers, thus decreasing the recombination rates compared to those of pure TiO 2 [66].
Based on the results in Figures 6, 7 and 8a, the degradation rate of the oxalic acid solution, while using optimal dopant concentration, drops exponentially. In the first 20-40 min, the degradation rate is higher compared to degradation after 40 min, when the reaction almost stabilizes. After the experiment with Cu (0.3 and 0.6 wt %), Mg (0.9 wt %) and Ni (0.5 wt %) doped TiO 2 films, visible surface area degradation was observed. The thin film degradation was not detected using TiO 2 with higher dopant concentration. The reactivation of such catalysts has to be taken into account considering the economic implications [67]. The strength of TiO 2 films and adherence are relevant for repetitive applications. Investigations on such parameters were done in previous reports [68][69][70][71].
considering the economic implications [67]. The strength of TiO2 films and adherence are relevant for repetitive applications. Investigations on such parameters were done in previous reports [68][69][70][71].
According to EDS mapping results (Figure 4), it can be stated that the PVD method is suitable for the formation of doped TiO2. Figure 4 shows the dispersion uniformity of Cu in TiO2, which means that active centers are dispersed evenly throughout the surface and in inside layers. The top M/TiO2 layer thickness (depending on the dopant) does not exceed 5 nm, and no annealing is done after deposition in order to keep amorphous TiO2. Since there are more charge trappers, a higher number of active centers lead to higher photocatalytic efficiency. Nevertheless, too many active centers can lead to small gaps between them and high recombination rates. It is known that the doping ratio depends on the particle size and cluster forming due to a high number of particles. On the other hand, having too many empty spaces leads to a lack of active centers, and recombination overcomes the charge-trapping process. Correlation between particle size and doping concentration is discussed by Jonathan et al. [72]. The SEM images show that the surface area of doped TiO2 has no defects, which emphasizes that highquality films are formed via the PVD method. Moreover, according to SEM pictures, it can be noticed that the roughness of doped TiO2 is higher than that of pure TiO2 (Figure 9d). Moreover, the surface roughness of 0.6 wt % Cu/TiO2 (Figure 9c) is higher compared to others (Figure 9a,b), which can also lead to higher efficiency because of the increased surface area. The alignment between the electron work function of metal dopants and Fermi energy levels together with the CBM of TiO2 plays a big role in the photocatalysis mechanism. According to EDS mapping results (Figure 4), it can be stated that the PVD method is suitable for the formation of doped TiO 2 . Figure 4 shows the dispersion uniformity of Cu in TiO 2 , which means that active centers are dispersed evenly throughout the surface and in inside layers. The top M/TiO 2 layer thickness (depending on the dopant) does not exceed 5 nm, and no annealing is done after deposition in order to keep amorphous TiO 2 .
Since there are more charge trappers, a higher number of active centers lead to higher photocatalytic efficiency. Nevertheless, too many active centers can lead to small gaps between them and high recombination rates. It is known that the doping ratio depends on the particle size and cluster forming due to a high number of particles. On the other hand, having too many empty spaces leads to a lack of active centers, and recombination overcomes the charge-trapping process. Correlation between particle size and doping concentration is discussed by Jonathan et al. [72]. The SEM images show that the surface area of doped TiO 2 has no defects, which emphasizes that high-quality films are formed via the PVD method. Moreover, according to SEM pictures, it can be noticed that the roughness of doped TiO 2 is higher than that of pure TiO 2 (Figure 9d). Moreover, the surface roughness of 0.6 wt % Cu/TiO 2 ( Figure 9c) is higher compared to others (Figure 9a,b), which can also lead to higher efficiency because of the increased surface area. The alignment between the electron work function of metal dopants and Fermi energy levels together with the CBM of TiO 2 plays a big role in the photocatalysis mechanism.
The electron work function for dopants are Φ Mg = 3.66 eV; Φ Ni = 5.04 ÷ 5.35 eV; Φ Cu = 4.53 ÷ 5.10 eV; and the TiO 2 electron work function is around 4.4 eV (E F − E C = 0.5 eV) [73]. Figure 10 shows a visual representation of the alignment of energy values. The differences in the electron work function work as a barrier and a certain amount of energy is required for an electron to overcome it. For Mg, it is 0.74 eV; for Ni, it is 0.64-0.82 eV; and for Cu, it is 0.13-0.5 eV. According to these results, the Mg dopant works as an electron donor, while Ni acts as an acceptor, and Cu could be a donor and acceptor at the same time, with Fermi energy almost aligned with the work function of Cu. These results coincide with previously stated results of the apparent rate constant, which showed that with a certain amount of Cu in TiO 2 , the photocatalytic activity was highest with the Cu dopant, followed by Mg and Ni [4]. The electron work function for dopants are ΦMg=3.66 eV; ΦNi=5.04 ÷ 5.35 eV; ΦCu=4.53 ÷ 5.10 eV; and the TiO2 electron work function is around 4.4 eV (EF-EC=0.5 eV) [73]. Figure 10 shows a visual representation of the alignment of energy values. The differences in the electron work function work as a barrier and a certain amount of energy is required for an electron to overcome it. For Mg, it is 0.74 eV; for Ni, it is 0.64-0.82 eV; and for Cu, it is 0.13-0.5 eV. According to these results, the Mg dopant works as an electron donor, while Ni acts as an acceptor, and Cu could be a donor and acceptor at the same time, with Fermi energy almost aligned with the work function of Cu. These results coincide with previously stated results of the apparent rate constant, which showed that with a certain amount of Cu in TiO2, the photocatalytic activity was highest with the Cu dopant, followed by Mg and Ni [4].

Conclusions
The present research suggests that the magnetron sputtering technique is suitable for doped TiO2 film deposition because of its high purity, dopant dispersity in films, and the ability to deposit on a wide range of substrates. Mg-, Cu-, and Ni-doped TiO2 thin films were deposited as photocatalysts on alloy substrate, and the photocatalytic activity was determined by oxalic acid degradation under UV irradiation. Analysis revealed that based on the kapp, with an increased dopant concentration (Mg 0.9 wt % → 4.1 wt %; Cu 0.6 wt % → 3.4 wt %; Ni 0.5 wt % → 3.6 wt %) process efficiency significantly drops accordingly (Mg kapp 0.01866 min −1 → 0.00474 min −1 ; Cu kapp 0.02221 min −1 → 0.00631 min −1 ; Ni kapp 0.01317 min −1 → 0.00214 min −1 ). Equivalent results are with decreased The electron work function for dopants are ΦMg=3.66 eV; ΦNi=5.04 ÷ 5.35 eV; ΦCu=4.53 ÷ 5.10 eV; and the TiO2 electron work function is around 4.4 eV (EF-EC=0.5 eV) [73]. Figure 10 shows a visual representation of the alignment of energy values. The differences in the electron work function work as a barrier and a certain amount of energy is required for an electron to overcome it. For Mg, it is 0.74 eV; for Ni, it is 0.64-0.82 eV; and for Cu, it is 0.13-0.5 eV. According to these results, the Mg dopant works as an electron donor, while Ni acts as an acceptor, and Cu could be a donor and acceptor at the same time, with Fermi energy almost aligned with the work function of Cu. These results coincide with previously stated results of the apparent rate constant, which showed that with a certain amount of Cu in TiO2, the photocatalytic activity was highest with the Cu dopant, followed by Mg and Ni [4].

Conclusions
The present research suggests that the magnetron sputtering technique is suitable for doped TiO2 film deposition because of its high purity, dopant dispersity in films, and the ability to deposit on a wide range of substrates. Mg-, Cu-, and Ni-doped TiO2 thin films were deposited as photocatalysts on alloy substrate, and the photocatalytic activity was determined by oxalic acid degradation under UV irradiation. Analysis revealed that based on the kapp, with an increased dopant concentration (Mg 0.9 wt % → 4.1 wt %; Cu 0.6 wt % → 3.4 wt %; Ni 0.5 wt % → 3.6 wt %) process efficiency significantly drops accordingly (Mg kapp 0.01866 min −1 → 0.00474 min −1 ; Cu kapp 0.02221 min −1 → 0.00631 min −1 ; Ni kapp 0.01317 min −1 → 0.00214 min −1 ). Equivalent results are with decreased

Conclusions
The present research suggests that the magnetron sputtering technique is suitable for doped TiO 2 film deposition because of its high purity, dopant dispersity in films, and the ability to deposit on a wide range of substrates. Mg-, Cu-, and Ni-doped TiO 2 thin films were deposited as photocatalysts on alloy substrate, and the photocatalytic activity was determined by oxalic acid degradation under UV irradiation. Analysis revealed that based on the k app , with an increased dopant concentration (Mg 0.9 wt % → 4.1 wt %; Cu 0.6 wt % → 3.4 wt %; Ni 0.5 wt % → 3.6 wt %) process efficiency significantly drops accordingly (Mg k app 0.01866 min −1 → 0.00474 min −1 ; Cu k app 0.02221 min −1 → 0.00631 min −1 ; Ni k app 0.01317 min −1 → 0.00214 min −1 ). Equivalent results are with decreased dopant concentration: Cu 0.6 wt % → 0.3 wt %, efficiency drops k app 0.02221 min −1 → 0.01373 min −1 . High photocatalysis efficiency is achieved with low exciton recombination rates. Since high concentration leads to small gaps between the active centers and the recombination process occurs, low concentration leads to a few active centers formed in the sample. Nevertheless, according to the TOC results, followed by the apparent rate constant, the Cu dopant achieved better photocatalysis efficiency compared to the Mg and Ni dopants. This can be explained by the formation of impurity energy levels in the TiO 2 band gap.