Exploiting Direct Current Plasma Electrolytic Oxidation to Boost Photoelectrocatalysis

In this study, we report an investigation of the photoelectrochemical activity of TiO2 films formed by DC plasma electrolytic oxidation (PEO) at a variable potential in a sulfuric acid electrolyte at 0 and 25 ◦C. The surface morphology was mainly determined by the oxide-forming potential. X-Ray Diffraction and Raman analyses showed that the relative amount of the anatase and rutile phases varied from 100% anatase at low potential (110–130 V) to 100% rutile at high potential (180–200 V), while mixed-phase oxide films formed at intermediate potential. Correspondingly, the band gap of the TiO2 films decreased from about 3.20 eV (pure anatase) to 2.94 eV (pure rutile) and was red-shifted about 0.1 eV by reducing the electrolyte temperature from 25 ◦C to 0 ◦C. Glow-Discharge Optical Emission Spectroscopy (GD-OES) and X-ray Photoelectron Spectroscopy (XPS) analyses evidenced S-containing species located preferentially close to the TiO2/Ti interface. The photoelectrochemical activity was assessed by measuring the incident photon-to-current efficiency (IPCE) under Ultraviolet C (UV-C) irradiation, which showed a non-gaussian normal trend as a function of the PEO cell potential, with maximum values exceeding 80%. Photoelectrocatalytic activity was assessed by decolorization of model solutions containing methylene blue. Photoanodes having higher IPCE values showed faster decolorization kinetics.


Introduction
Since the beginning of the 21st century, the application of photoelectrocatalysis (PEC) to water treatment has attracted increasing attention [1]. Compared to photocatalysis, in PEC the application of an anodic bias to the photoanode prevents electron-hole recombination, thereby greatly enhancing the quantum efficiency for the production of OH (hydroxyl) radicals. Additionally, PEC shows the typical advantages of photocatalytic processes, such as non-toxicity, no use of chemicals, potentially no secondary pollution and complete mineralization. Despite the promising features of PEC processes, they are still limited at the laboratory scale. Most of the studies available in literature use synthetic wastewater prepared by the addition of pollutants (either dyes, chemicals or pharmaceuticals) to ultrapure water [2,3]. Recently, some of the present authors have described the degradation of pharmaceuticals in real waste-water matrices, showing that PEC is more rapid, more efficient and less energy demanding compared to photolysis [4,5]. They also reported a remarkable increase of the biodegradability of real pharmaceutical wastewaters by an H 2 O 2 -assisted PEC process, compared to photolysis, photocatalysis and chemical oxidation with hydrogen peroxide [6].
values of TiO 2 films obtained by PEO. Finally, most of the PEO catalysts described in the literature, obtained in alternating current (AC)/pulsed mode, are traditionally applied to obtain thick and compact protective layers on Al and Mg alloys [10]. On the contrary, photoactive oxide coatings should have high surface areas and thicknesses close to the penetration depth of the incident light [38]; therefore, direct current (DC) mode would probably be more appropriate. Lukiyanchuk et al. recently showed that DC PEO is a very promising operation mode with which to obtain high surface area porous morphologies [23]; however, they did not show evidence of the advantage given by DC mode for the synthesis of undoped TiO 2 photoanodes. DC mode operation was also applied by Stojadinovich [37], Bayati [21] and Wan [39], focusing on doping, but no results were shown on the photoelectrochemical activity of the semiconductors, which in this paper is presented as a tool to foresee the photocatalytic activity of the catalysts in water treatment.
The present paper focuses on DC plasma electrolytic oxidation with the aim of contributing to the validation of PEO as a viable route to produce TiO 2 photoanodes and as a step forward in the technological development of photoelectrocatalysis in water treatment. PEO TiO 2 photoanodes are characterized not only in terms of morphology, crystal structure, composition and photocatalytic activity in the decolorization of a commercial dye, but also considering band-gap and incident photon-to-current efficiency (IPCE), trying to correlate IPCE and photocatalytic activity. We recently described the effect of electrolyte temperature during DC PEO of Ti in 0.5 M H 2 SO 4 , showing that tuning this parameter can be an effective way to optimize the phase proportion of anatase and rutile, and the resulting electrochemical photoactivity [17,40]. As a further step in this direction, the present paper explores the combined effect of applied potential and electrolyte temperature on the physical properties of the coatings with the aim of supporting a deeper understanding not only of the photocatalytic activity but also, and more in general, of the photoelectrochemical activity of PEO TiO 2 photoanodes.

Oxide Thickness and Surface Morphology
PEO anodizing conditions were established over the whole range of cell potential used in this work, irrespective of the electrolyte temperature, as revealed by micro-discharge development at the anode and by the coating morphology, as shown further on. The relationship between the thickness of PEO-TiO 2 films derived from GD-OES measurements and the anodizing potential at electrolyte temperatures of 0 • C and 25 • C is shown in Figure 1a. Apparently, the anodizing potential is the main factor controlling the oxide thickness, whereas the electrolyte temperature has an overall negligible effect. At first, in the range of low potential up to about 140 V, the thickness increased slightly from about 1 to 2 µm; then, the increase was much steeper for potential in excess of the latter value. Notably, the increasing trend of thickness and its values with increasing potential agree with the results reported in [41] for oxide coatings formed in a mixed H 2 SO 4 -H 3 PO 4 electrolyte. Expectedly, the charge passed during film growth (Figure 1b)-derived from the integration of the current vs. time profile at varying potential-shows a similar trend to that of the thickness vs. potential, pointing to two different kinetic regimes of oxide growth, at low and high potentials, which were ostensibly associated with relatively lower and higher micro-discharge rates and lifetimes, respectively. As shown in the SEM and AFM micrographs (Figure 2), and also in Figure S2 and Figure S3, respectively, in the Supplementary Material, the morphology of the TiO2 samples is characterized by the typical features of oxides obtained by PEO. Indeed, the oxide layers are obviously porous, with pore diameter increasing with the oxide formation potential, and that being slightly larger on samples anodized at 0 °C compared to 25 °C. Correspondingly, the average surface porosity increased with potential from about 2-3% to 13%, at 25 °C, and to 24%, at 0 °C. Remarkably, SEM images revealed a distinct change in the film morphology at a potential of 140 V or higher, apparently because of spreading and interconnecting porosity; namely, a transition from the porous layer type morphology observed at lower potentials to the sponge-like morphology obtained at a potential of 140 V or higher. Since porosity is mainly due to the occurrence of micro-discharges, along with the extrusion of entrapped gas bubbles through the softened material during the PEO process [42], this morphological transition is the material evidence of the space-time intensification of micro-discharges with increasing potential.

Crystal Structure
Raman spectra of the TiO2 films demonstrated that the TiO2 films are crystalline and consist of pure anatase and rutile phases or mixtures of the same phases depending on the PEO voltage ( Figure   Figure 1. Thicknesses of TiO 2 films (a) and total charge passed during film growth (b) as a function of PEO potential at 0 and 25 • C.
As shown in the SEM and AFM micrographs (Figure 2), and also in Figures S2 and S3, respectively, in the Supplementary Material, the morphology of the TiO 2 samples is characterized by the typical features of oxides obtained by PEO. Indeed, the oxide layers are obviously porous, with pore diameter increasing with the oxide formation potential, and that being slightly larger on samples anodized at 0 • C compared to 25 • C. Correspondingly, the average surface porosity increased with potential from about 2-3% to 13%, at 25 • C, and to 24%, at 0 • C. Remarkably, SEM images revealed a distinct change in the film morphology at a potential of 140 V or higher, apparently because of spreading and interconnecting porosity; namely, a transition from the porous layer type morphology observed at lower potentials to the sponge-like morphology obtained at a potential of 140 V or higher. Since porosity is mainly due to the occurrence of micro-discharges, along with the extrusion of entrapped gas bubbles through the softened material during the PEO process [42], this morphological transition is the material evidence of the space-time intensification of micro-discharges with increasing potential. As shown in the SEM and AFM micrographs (Figure 2), and also in Figure S2 and Figure S3, respectively, in the Supplementary Material, the morphology of the TiO2 samples is characterized by the typical features of oxides obtained by PEO. Indeed, the oxide layers are obviously porous, with pore diameter increasing with the oxide formation potential, and that being slightly larger on samples anodized at 0 °C compared to 25 °C. Correspondingly, the average surface porosity increased with potential from about 2-3% to 13%, at 25 °C, and to 24%, at 0 °C. Remarkably, SEM images revealed a distinct change in the film morphology at a potential of 140 V or higher, apparently because of spreading and interconnecting porosity; namely, a transition from the porous layer type morphology observed at lower potentials to the sponge-like morphology obtained at a potential of 140 V or higher. Since porosity is mainly due to the occurrence of micro-discharges, along with the extrusion of entrapped gas bubbles through the softened material during the PEO process [42], this morphological transition is the material evidence of the space-time intensification of micro-discharges with increasing potential.

Crystal Structure
Raman spectra of the TiO2 films demonstrated that the TiO2 films are crystalline and consist of pure anatase and rutile phases or mixtures of the same phases depending on the PEO voltage (Figure

Crystal Structure
Raman spectra of the TiO 2 films demonstrated that the TiO 2 films are crystalline and consist of pure anatase and rutile phases or mixtures of the same phases depending on the PEO voltage ( Figure 3a). This was further confirmed by XRD patterns of oxide films formed at 0 • C or 25 • C at varying potentials, as shown in Figure S4 in the Supplementary Material. Apparently, as-prepared TiO 2 films were mostly crystalline in structure with varying relative amounts of the anatase and rutile phases depending on the applied potential. However, the presence of a minor amorphous component in the structure cannot be ruled out completely, based on conventional XRD evidence, possibly also because of its distributed nature, as recently suggested in literature [43]. Based on the  (1)), the mass fractions of the anatase and rutile phases, as functions of the potential for oxide films prepared at 0 • C or 25 • C, were calculated ( Figure 3b). The overall trend emerging from these data is that oxide films grown at a potential below 140 V are uniquely or predominantly of the anatase phase, at low and high electrolyte temperatures, respectively, whereas higher PEO potentials entail the formation of oxide coatings with a predominance of the rutile phase. With potential exceeding 130 V, the rutile mass fraction increases rapidly, such that in the high potential range (170 V or higher), almost pure rutile is formed. The reducing trend of the anatase mass fraction with increasing PEO potential, and then with increasing thickness, is in agreement with the literature [18], and can be explained as a result of stronger local heating of the oxide film induced by more intense discharge phenomena. Correspondingly, the effect of the bath temperature on oxide phase structure-namely, a relatively higher fraction of anatase in oxide films grown at 0 • C compared to 25 • C-is limited to the low range of potential (<150 V), in agreement with our previous work [17]. potentials, as shown in Figure S4 in the Supplementary Material. Apparently, as-prepared TiO2 films were mostly crystalline in structure with varying relative amounts of the anatase and rutile phases depending on the applied potential. However, the presence of a minor amorphous component in the structure cannot be ruled out completely, based on conventional XRD evidence, possibly also because of its distributed nature, as recently suggested in literature [43]. Based on the XRD patterns and the Spurr equation (Equation 1), the mass fractions of the anatase and rutile phases, as functions of the potential for oxide films prepared at 0 °C or 25 °C, were calculated ( Figure 3b). The overall trend emerging from these data is that oxide films grown at a potential below 140 V are uniquely or predominantly of the anatase phase, at low and high electrolyte temperatures, respectively, whereas higher PEO potentials entail the formation of oxide coatings with a predominance of the rutile phase. With potential exceeding 130 V, the rutile mass fraction increases rapidly, such that in the high potential range (170 V or higher), almost pure rutile is formed. The reducing trend of the anatase mass fraction with increasing PEO potential, and then with increasing thickness, is in agreement with the literature [18], and can be explained as a result of stronger local heating of the oxide film induced by more intense discharge phenomena. Correspondingly, the effect of the bath temperature on oxide phase structure-namely, a relatively higher fraction of anatase in oxide films grown at 0 °C compared to 25 °C-is limited to the low range of potential (<150 V), in agreement with our previous work [17]. In Figure S5 in the Supplementary Material, the lattice parameters a and c and the c/a ratios for the anatase and rutile phases in TiO2 films are shown. The arrows indicate the corresponding values given by the relevant powder diffraction files. The lattice parameters of the anatase phase reveal a significant distortion of the structure. In particular, the a parameter of the tetragonal cell is significantly smaller compared to the reference value, whereas the c parameter is either higher than (at low potential) or in agreement with the reference value. Correspondingly, in the low potential region, the c/a ratio of the anatase decreased from about 2.60 to 2.52. Conversely, the lattice parameters of the rutile phase are in overall good agreement with the reference values (mind the narrow scale range in the plot of Figure S5 for the lattice parameters of the rutile phase), the largest percentage difference of the c/a ratio with respect to the reference value being about 0.27% (at 150 V and 25 °C). Therefore, the unit cell of the anatase phase is significantly stretched along the c axis direction; namely, for oxide formed at a low temperature and low potential. In Figure S5 in the Supplementary Material, the lattice parameters a and c and the c/a ratios for the anatase and rutile phases in TiO 2 films are shown. The arrows indicate the corresponding values given by the relevant powder diffraction files. The lattice parameters of the anatase phase reveal a significant distortion of the structure. In particular, the a parameter of the tetragonal cell is significantly smaller compared to the reference value, whereas the c parameter is either higher than (at low potential) or in agreement with the reference value. Correspondingly, in the low potential region, the c/a ratio of the anatase decreased from about 2.60 to 2.52. Conversely, the lattice parameters of the rutile phase are in overall good agreement with the reference values (mind the narrow scale range in the plot of Figure  S5 for the lattice parameters of the rutile phase), the largest percentage difference of the c/a ratio with respect to the reference value being about 0.27% (at 150 V and 25 • C). Therefore, the unit cell of the anatase phase is significantly stretched along the c axis direction; namely, for oxide formed at a low temperature and low potential.

In-Depth Composition Profile
In-depth GD-OES analysis of TiO 2 films for some selected samples at different potentials are shown in Figure 4. In addition to the expected signals for oxygen and titanium atoms, a peak corresponding to sulfur in the proximity of the TiO 2 /Ti interface is evident. The sulfur signal showed a wide tail towards the film surface, which is more pronounced at higher anodizing potentials. It can be argued that at the first stage of the anodizing process, a higher amount of sulfur is included in the film, as SO −2 4 -containing species impregnating the pores, S-containing species included in the TiO 2 or both [44]. At higher anodization potentials, the sulfur-containing species increasingly spread across the whole thickness. This might be attributable to more intense and dense sparking; in fact, in these conditions, Catalysts 2020, 10, 325 6 of 14 the melted material repeatedly remix and sulfur-containing species can more easily migrate towards the surface of the film.
corresponding to sulfur in the proximity of the TiO2/Ti interface is evident. The sulfur signal showed a wide tail towards the film surface, which is more pronounced at higher anodizing potentials. It can be argued that at the first stage of the anodizing process, a higher amount of sulfur is included in the film, as -containing species impregnating the pores, S-containing species included in the TiO2 or both [44]. At higher anodization potentials, the sulfur-containing species increasingly spread across the whole thickness. This might be attributable to more intense and dense sparking; in fact, in these conditions, the melted material repeatedly remix and sulfur-containing species can more easily migrate towards the surface of the film. The nature of the S-containing species was further investigated by XPS analysis. Figure 5 shows a wide scan where the position of the main features observed with XPS (photoemission signals from O, Ti and C) are marked with labels. The position of the S 2p feature is highlighted with a vertical dashed line. Panel (b) shows a close-up of the S 2p binding energy (BE) region, characterized by the presence of a feature at a BE of about 169.2 eV. The energy position of this peak is compatible with the superposition of the photoemission features from S 4+ and S 6+ species (at a BE of 168.8 and 169.5 eV, respectively), in agreement with the literature [45]. Therefore, it can be argued that S could be included both as adsorbed SO4 2− and doping element.  The nature of the S-containing species was further investigated by XPS analysis. Figure 5 shows a wide scan where the position of the main features observed with XPS (photoemission signals from O, Ti and C) are marked with labels. The position of the S 2p feature is highlighted with a vertical dashed line. Panel (b) shows a close-up of the S 2p binding energy (BE) region, characterized by the presence of a feature at a BE of about 169.2 eV. The energy position of this peak is compatible with the superposition of the photoemission features from S 4+ and S 6+ species (at a BE of 168.8 and 169.5 eV, respectively), in agreement with the literature [45]. Therefore, it can be argued that S could be included both as adsorbed SO 4 2− and doping element.

Optical Properties
corresponding to sulfur in the proximity of the TiO2/Ti interface is evident. The sulfur signal showed a wide tail towards the film surface, which is more pronounced at higher anodizing potentials. It can be argued that at the first stage of the anodizing process, a higher amount of sulfur is included in the film, as -containing species impregnating the pores, S-containing species included in the TiO2 or both [44]. At higher anodization potentials, the sulfur-containing species increasingly spread across the whole thickness. This might be attributable to more intense and dense sparking; in fact, in these conditions, the melted material repeatedly remix and sulfur-containing species can more easily migrate towards the surface of the film.  [45]. Therefore, it can be argued that S could be included both as adsorbed SO4 2− and doping element.

Optical Properties
The UV-Vis spectra of some representative samples are shown in Figure 6a. The band gap was calculated based on the reflectance UV-Vis spectra after Kubelka-Munk conversion using the Tauc plot method [46], as shown in Figure S6 in the Supplementary Material for some selected processing potential and electrolyte temperature. However, in some cases the band gap could not be reliably derived from Tauc plots due to interference fringe perturbation. The calculated band gap is plotted in Figure 6b as a function of processing potential. Increasing the applied potential, the band gap decreases gradually from values close to 3.2 eV to values close to 2.94 eV, following the modification of the phase structure from pure anatase to rutile phase as shown in Figure 3b.
Catalysts 2020, 10, x FOR PEER REVIEW 7 of 14 The UV-Vis spectra of some representative samples are shown in Figure 6a. The band gap was calculated based on the reflectance UV-Vis spectra after Kubelka-Munk conversion using the Tauc plot method [46], as shown in Figure S6 in the Supplementary Material for some selected processing potential and electrolyte temperature. However, in some cases the band gap could not be reliably derived from Tauc plots due to interference fringe perturbation. The calculated band gap is plotted in Figure 6b as a function of processing potential. Increasing the applied potential, the band gap decreases gradually from values close to 3.2 eV to values close to 2.94 eV, following the modification of the phase structure from pure anatase to rutile phase as shown in Figure 3b.

Photoelectrochemical Activity
The PEA of TiO2 films was assessed by measuring the photocurrent response of the films under electrical bias, i.e., by linear voltammetry in the dark, and under UV-C irradiation from 0 V to 1.5 V vs. Standard Calomel Electrode (SCE). Figure 7a,b shows the photocurrent-potential characteristics of TiO2 coatings obtained at PEO potentials in the range of 110-200 V and electrolyte temperatures of 0 °C and 25 °C, respectively. Overall, the changes in the shapes of the photocurrent curves can be related to the PEO potential and the phase structure of the oxide. For PEO potentials in the range of 110-140 V and at electrolyte temperature of 0 °C, where anatase phase structure is largely predominant (80% or higher mass fraction), the photocurrent density showed an initially steep increase followed by a continuous but much slower raising trend. Similarly, for oxides grown at potentials in the range 110-140 V at 25 °C, the photocurrent density increases continuously with the potential, though at a receding rate; in particular, for oxides formed at 130 and 140 V. In contrast, oxides grown at 160 V or above, having a predominant rutile phase structure and higher thickness, showed a plateau-like region, and poor photo-electrochemical activity, regardless of the electrolyte temperature. In particular, for oxides grown at 150 V, as the phase composition of the oxide changed from a dominant fraction of anatase to a dominant fraction of rutile phase, a well-developed plateau of photocurrent density appeared. As a final remark, it can be noticed that the oxides formed at 140 V (electrolyte at 0 °C) and at 120 V (electrolyte at 25 °C) showed the maximum of the photocurrent density.
For a straightforward comparison and a summary of the photoelectrochemical activity, Figure 8 shows the IPCEs of the different oxides at the potentials of 0.6 V (a) and 1.5 V (b) vs. SCE. Oxide films formed at potentials in the range of 110-150 V showed IPCE values close to or higher than 60%, reaching maxima of 82% and 88% at 110 V (25 °C) and 140 V (0 °C), respectively. The IPCE then decreased abruptly to values lower than 20% for the oxides having a large fraction of the rutile phase and thickness in excess of about 3 µm (Figure 1a). Therefore, the PEA of the TiO2 films can be tuned by controlling either electrolyte temperature or PEO potential. In agreement with previous findings

Photoelectrochemical Activity
The PEA of TiO 2 films was assessed by measuring the photocurrent response of the films under electrical bias, i.e., by linear voltammetry in the dark, and under UV-C irradiation from 0 V to 1.5 V vs. Standard Calomel Electrode (SCE). Figure 7a,b shows the photocurrent-potential characteristics of TiO 2 coatings obtained at PEO potentials in the range of 110-200 V and electrolyte temperatures of 0 • C and 25 • C, respectively. Overall, the changes in the shapes of the photocurrent curves can be related to the PEO potential and the phase structure of the oxide. For PEO potentials in the range of 110-140 V and at electrolyte temperature of 0 • C, where anatase phase structure is largely predominant (80% or higher mass fraction), the photocurrent density showed an initially steep increase followed by a continuous but much slower raising trend. Similarly, for oxides grown at potentials in the range 110-140 V at 25 • C, the photocurrent density increases continuously with the potential, though at a receding rate; in particular, for oxides formed at 130 and 140 V. In contrast, oxides grown at 160 V or above, having a predominant rutile phase structure and higher thickness, showed a plateau-like region, and poor photo-electrochemical activity, regardless of the electrolyte temperature. In particular, for oxides grown at 150 V, as the phase composition of the oxide changed from a dominant fraction of anatase to a dominant fraction of rutile phase, a well-developed plateau of photocurrent density appeared. As a final remark, it can be noticed that the oxides formed at 140 V (electrolyte at 0 • C) and at 120 V (electrolyte at 25 • C) showed the maximum of the photocurrent density.
For a straightforward comparison and a summary of the photoelectrochemical activity, Figure 8 shows the IPCEs of the different oxides at the potentials of 0.6 V (a) and 1.5 V (b) vs. SCE. Oxide films formed at potentials in the range of 110-150 V showed IPCE values close to or higher than 60%, reaching maxima of 82% and 88% at 110 V (25 • C) and 140 V (0 • C), respectively. The IPCE then decreased abruptly to values lower than 20% for the oxides having a large fraction of the rutile phase and thickness in excess of about 3 µm (Figure 1a). Therefore, the PEA of the TiO 2 films can be tuned by controlling either electrolyte temperature or PEO potential. In agreement with previous findings [17], oxide films formed at low temperature show an overall improved performance in terms of both photocurrent response and corresponding IPCE. Catalysts 2020, 10, x FOR PEER REVIEW 8 of 14 [17], oxide films formed at low temperature show an overall improved performance in terms of both photocurrent response and corresponding IPCE.  Summarizing, it was observed that in addition to the effect of the oxide phase structure, namely, the anatase/rutile fraction, the photocurrent response of the PEO-TiO2 films appears to be affected by the film thickness. However, as reported in literature, for a more exhaustive explanation of the PEA of TiO2 coatings, other properties, such as refractive index, scattering coefficient, surface roughness and absorption coefficient, should also be considered [47].

Photoelectrocatalytic Decolorization of Methylene blue
Degradation of Methylene Blue was carried out to verify the photoelectrocatalytic activity of the PEO TiO2 anodes with respect to the corresponding IPCEs. The initial concentration of Methylene Blue was 5 µM and the measurement was conducted in a 150 mL quartz beaker using a threeelectrode cell configuration where TiO2 (exposed area of 1.5 cm 2 ), a saturated calomel electrode and a platinum foil were used as working, reference and counter electrodes, respectively. The irradiation source was the same UV light used for photocurrent measurement. The reaction kinetic constant was calculated following the Langmuir-Hinshelwood model.
In Figure 9, the relative residual concentration of methylene blue (C/C0) is represented as a function of the irradiation time. Decolorization was carried out by photoelectrocatalysis using some Catalysts 2020, 10, x FOR PEER REVIEW 8 of 14 [17], oxide films formed at low temperature show an overall improved performance in terms of both photocurrent response and corresponding IPCE.  Summarizing, it was observed that in addition to the effect of the oxide phase structure, namely, the anatase/rutile fraction, the photocurrent response of the PEO-TiO2 films appears to be affected by the film thickness. However, as reported in literature, for a more exhaustive explanation of the PEA of TiO2 coatings, other properties, such as refractive index, scattering coefficient, surface roughness and absorption coefficient, should also be considered [47].

Photoelectrocatalytic Decolorization of Methylene blue
Degradation of Methylene Blue was carried out to verify the photoelectrocatalytic activity of the PEO TiO2 anodes with respect to the corresponding IPCEs. The initial concentration of Methylene Blue was 5 µM and the measurement was conducted in a 150 mL quartz beaker using a threeelectrode cell configuration where TiO2 (exposed area of 1.5 cm 2 ), a saturated calomel electrode and a platinum foil were used as working, reference and counter electrodes, respectively. The irradiation source was the same UV light used for photocurrent measurement. The reaction kinetic constant was calculated following the Langmuir-Hinshelwood model.
In Figure 9, the relative residual concentration of methylene blue (C/C0) is represented as a function of the irradiation time. Decolorization was carried out by photoelectrocatalysis using some Summarizing, it was observed that in addition to the effect of the oxide phase structure, namely, the anatase/rutile fraction, the photocurrent response of the PEO-TiO 2 films appears to be affected by the film thickness. However, as reported in literature, for a more exhaustive explanation of the PEA of TiO 2 coatings, other properties, such as refractive index, scattering coefficient, surface roughness and absorption coefficient, should also be considered [47].

Photoelectrocatalytic Decolorization of Methylene blue
Degradation of Methylene Blue was carried out to verify the photoelectrocatalytic activity of the PEO TiO 2 anodes with respect to the corresponding IPCEs. The initial concentration of Methylene Blue was 5 µM and the measurement was conducted in a 150 mL quartz beaker using a three-electrode cell configuration where TiO 2 (exposed area of 1.5 cm 2 ), a saturated calomel electrode and a platinum foil were used as working, reference and counter electrodes, respectively. The irradiation source was the same UV light used for photocurrent measurement. The reaction kinetic constant was calculated following the Langmuir-Hinshelwood model.
In Figure 9, the relative residual concentration of methylene blue (C/C 0 ) is represented as a function of the irradiation time. Decolorization was carried out by photoelectrocatalysis using some selected TiO 2 anodes and by photolysis for comparison. The selected TiO 2 anodes were prepared at 0 • C and 140, 150 and 200 V, respectively. As previously mentioned, the TiO 2 photoanode prepared at 200 V consisted mostly of the rutile phase, and the IPCE under UV-C irradiation was below 1%. Correspondingly, under irradiation and electrical bias, this photoanode showed the lowest decolorization rate (k = 0.015), not exceeding the kinetic rate obtained by photolysis (only UV, no photoanode), so a complete decolorization was obtained after 240 min. By reducing the PEO cell potential to 150 V, the TiO 2 film consisted of a mixture of anatase and rutile, and its IPCE was around 75%. Compared to the anode prepared at 200 V, this anode showed a faster decolorization rate (k = 0.021), and complete decolorization was achieved after 120 min. In agreement with the IPCE values reported in Figure 8, the fastest decolorization kinetics were observed by the TiO 2 photoanode synthesized at 140 V (k = 0.030), with complete decolorization after 90 min. The highest kinetic rate constant measured by photoelectrocatalysis was twice as high as photolysis. This is a reasonable result, considering that Methylene Blue is relatively sensitive to UV irradiation and that the catalysts employed in the decolorization test had a limited exposed area (only 1.5 cm 2 ). Some of the authors recently demonstrated that photoelectrocatalysis can outperform photolysis when recalcitrant compounds are targeted and/or when the nature and the concentration of transformation products are considered [3][4][5][6]. In the present investigation, the decolorization of Methylene Blue was meant to demonstrate a correlation between the photoelectrocatalytic performances and the IPCEs of the PEO TiO 2 films. In fact, it was observed that higher IPCE values corresponded to higher kinetic rate constant values, though the correlation between the two parameters is not linear. This lack of linearity might be tentatively explained considering that, in addition to the degree of photoactivation of the catalysts, mass transport into the porous structure can affect the overall degradation kinetics. However, for the selected photoanodes and the targeted pollutant, substantial agreement was observed between the IPCE values and the photoelectrocatalytic activity, confirming that PEO is a viable technique to produce photoactive TiO 2 electrodes and also that measuring the photoelectrochemical activity of the electrodes by linear-sweep photovoltammetry is a promising approach to evaluate and possibly to monitor the photoelectrocatalytic activity of the electrodes. selected TiO2 anodes and by photolysis for comparison. The selected TiO2 anodes were prepared at 0 °C and 140, 150 and 200 V, respectively. As previously mentioned, the TiO2 photoanode prepared at 200 V consisted mostly of the rutile phase, and the IPCE under UV-C irradiation was below 1%. Correspondingly, under irradiation and electrical bias, this photoanode showed the lowest decolorization rate (k = 0.015), not exceeding the kinetic rate obtained by photolysis (only UV, no photoanode), so a complete decolorization was obtained after 240 min. By reducing the PEO cell potential to 150 V, the TiO2 film consisted of a mixture of anatase and rutile, and its IPCE was around 75%. Compared to the anode prepared at 200 V, this anode showed a faster decolorization rate (k = 0.021), and complete decolorization was achieved after 120 min. In agreement with the IPCE values reported in Figure 8, the fastest decolorization kinetics were observed by the TiO2 photoanode synthesized at 140 V (k = 0.030), with complete decolorization after 90 min. The highest kinetic rate constant measured by photoelectrocatalysis was twice as high as photolysis. This is a reasonable result, considering that Methylene Blue is relatively sensitive to UV irradiation and that the catalysts employed in the decolorization test had a limited exposed area (only 1.5 cm 2 ). Some of the authors recently demonstrated that photoelectrocatalysis can outperform photolysis when recalcitrant compounds are targeted and/or when the nature and the concentration of transformation products are considered [3][4][5][6]. In the present investigation, the decolorization of Methylene Blue was meant to demonstrate a correlation between the photoelectrocatalytic performances and the IPCEs of the PEO TiO2 films. In fact, it was observed that higher IPCE values corresponded to higher kinetic rate constant values, though the correlation between the two parameters is not linear. This lack of linearity might be tentatively explained considering that, in addition to the degree of photoactivation of the catalysts, mass transport into the porous structure can affect the overall degradation kinetics. However, for the selected photoanodes and the targeted pollutant, substantial agreement was observed between the IPCE values and the photoelectrocatalytic activity, confirming that PEO is a viable technique to produce photoactive TiO2 electrodes and also that measuring the photoelectrochemical activity of the electrodes by linear-sweep photovoltammetry is a promising approach to evaluate and possibly to monitor the photoelectrocatalytic activity of the electrodes.

Synthesis of TiO 2 Films
TiO 2 coatings were obtained by plasma anodizing of grade I titanium sheets (4 cm 2 area) in 1. The shift of the temperature from the set-point of 0 • C or 25 • C was at most 5 • C, when operating at high values of the cell potential; namely, when the latter was in excess of 180 or 170 V, respectively. Following PEO, the samples were first rinsed with water, and then dried in a stream of air. All PEO samples were prepared in duplicate.

Morphologies and Crystal Structures of TiO 2 Films
The surface morphology was observed using a scanning electron microscopy (SEM, EVO 50, Carl Zeiss Jena GmbH, Jena, DE) instrument. The phase structure and texture of each TiO 2 film was assessed by X-ray diffraction (XRD) and Raman spectroscopy. XRD was carried out using a PW1830 instrument (Malvern Panalytical Ltd, Malvern, U.K. and Almelo, NL) operating in Bragg-Brentano geometry at a potential of 40 kV with a filament current of 40 mA. Spectra were acquired with Cu K α1 radiation in the 2θ range 20 • -90 • at the scanning rate of 2.5 • per minute. The XRD patterns were indexed according to the powder diffraction files released by the International Center for Diffraction Data (U.S.A.) for titanium (PDF 44-1294), anatase (PDF 21-1272) and rutile phases (PDF 21-1276).
To calculate the lattice parameters a and c, a least mean square method was applied using the interplanar spacing of (101), (103) and (200) reflections for the anatase phase, and the (110), (200) and (210) reflections for the rutile phase.
The weight fraction of anatase was calculated according to Equation (1) [48], where I R is the intensity of the strongest rutile reflection, (110), and I A is the intensity of the strongest anatase reflection, (101), Micro-Raman spectra were acquired using a LABRAM HR800 spectrometer from HORIBA, Ltd. (Kyoto, JPN) equipped with a Peltier cooled CCD detector. The λ = 514 nm excitation line was provided by an argon ion laser (Stabilite 2017, Spectra Physics, Santa Clara, U.S.A.). The laser radiation was filtered by an interference filter and focused on the sample through a 50× objective with a 0.75 numerical aperture. The Rayleigh radiation was rejected using an edge filter for the 514 nm laser line.
Surface porosity was evaluated by analyzing several SEM surface micrographs taken at 20,000x magnification using ImageJ analysis software, an open-source image processing program released by the National Institutes of Health, U.S.. Film thickness and in-depth composition profile were assessed by Glow-Discharge Optical Emission Spectroscopy (GD-OES) using a Spectruma GDA750 analyzer (SPECTRO Analytical instruments Inc., Kleve, DE) operating at 700 V in an argon atmosphere at 230 Pa. GDOES analysis was carried on sample areas of ≈2.5 mm diameter. The monitored light emission wavelengths during the analysis were 130 nm and 362 nm for oxygen and titanium, respectively. XPS data were acquired with a Phoibos 150 hemispherical analyzer (SPECS GmbH, Berlin, DE) by exciting the electrons with a Mg-Kα source (hν = 1253.6 eV). The spectrometer is housed in a home-built vacuum system described elsewhere [49] and kept in ultra-high vacuum conditions (base pressure in the low 10 −8 Pa). We accounted for possible charging effects by setting the peak from adventitious carbon to 285 eV [45].

Photoelectrochemical Properties of TiO 2 Films
The photocurrent was measured by linear sweep photo-voltammetry (LSPhV) in a 4.2 mM KCl aqueous solution with and without irradiation. LSPhV was conducted at a scan rate of 5 mV/s at room temperature using a potentiostat/galvanostat (Solartron Analytical ModuLab ECS, Ametek Inc., Berwyn, U.S.). A saturated calomel electrode (SCE) was used as a reference and the counter electrode was a Pt foil.
The irradiation was achieved by means of a 30 W low-pressure Hg vapor lamp (HGL, Helios Italquartz, Cambiago, IT) emitting at λ = 254 nm. The emission spectrum of the light source is shown in Figure S1 in the Supplementary Material. The irradiance on the TiO 2 film at 254 nm was 1.3 W/m 2 as measured by quantum-photo-radiometer (PM200,Thorlabs Inc., Newton, U.S.). The exposed area during PEO and LSPhV was 4 cm 2 . The photocurrent measurements were repeated three times. The incident photon-to-current efficiency (IPCE), which is defined as the number of electrons generated by light in the external circuit divided by the number of incident photons in the middle ultraviolet (λ = 254 nm), was calculated using the following equation: The band-gaps were measured by UV-Vis reflectance spectra using the integrating sphere accessory of a V570 spectrometer (JASCO Corp., Tokyo, JP). No electrical bias was applied to the TiO 2 samples during spectra acquisition.

Conclusions
The results described in this study confirmed that DC PEO can be a viable and industry-friendly route to produce photoactive TiO 2 for water treatment. TiO 2 films were synthesized at a potential ranging from 110 to 200 V and electrolyte temperature of 0 and 5 • C. The total charge supplied during the experiment increased with the processing potential, resulting in higher TiO 2 film thicknesses of up to 11 µm. SEM images showed a homogeneous morphology with sub-micrometric pores. X-ray diffraction revealed that the relative amounts of rutile and anatase phases were mainly affected by the PEO potential; and pure anatase, pure rutile and a mixture of anatase and rutile were obtained at lower potentials, higher potentials and intermediate potentials, respectively. In agreement with the phase structure modification, the band gap also changed from 3.2 eV to 2.94 eV. GD-OES and XPS analyses revealed that S-containing species were preferentially located at the TiO 2 /Ti interface and might consist of both SO 4 2− adsorbed groups and S 6+ doping species. The PEC activity of the films was assessed by photocurrent measurements under polychromatic UV-Vis irradiation and by decolorization tests of model solutions of methylene blue. IPCE reached a maximum value of 88% for the TiO 2 photoanode synthesized at 140 V. Correspondingly, the fastest decolorization kinetics were observed.
Supplementary Materials: The following are available online at http://www.mdpi.com/2073-4344/10/3/325/s1. Figure S1: Emission spectrum of the light source employed during photoelectrochemical characterization. Figure  S2: SEM micrographs of PEO-TiO 2 coatings grown at varying potential, as indicated, during 5 min anodizing time, at 0 and 25 • C. Figure S3: AFM images of TiO 2 films obtained by PEO in potential range of 100-160 V at 0 • C and 25 • C. Figure S4: XRD patterns of TiO 2 films obtained at 0 • C (a) and 25 • C (b). Figure S5: Lattice parameters a, c and c/a ratio for anatase and rutile as a function of potential. Filled and empty symbols refer to 0 • C and 25 • C, respectively. Arrows indicate the reference values of lattice parameters a and c and c/a ratio for anatase and rutile phases. Figure  Funding: This research received no external funding.