TiO2-La2O3 as Photocatalysts in the Degradation of Naproxen

The indiscriminate use of naproxen as an anti-inflammatory has been the leading cause of pollution in sewage effluents. Conversely, titanium dioxide is one of the most promising photocatalyst for the degradation of pollutants. Ti-La mixed oxides containing 0, 1, 3, 5, and 10 wt.% of lanthanum were synthetized by sol-gel and tested as photocatalysts in the degradation of naproxen (NPX). The materials were further characterized by X-ray diffraction (XRD), nitrogen physisorption (BET), scanning electron microscopy (SEM), UV-Vis and Fourier-transform infrared spectroscopy (FT-IR), transmission electron microscopy (TEM), and X-ray photoelectron spectroscopy (XPS). The XRD patterns resembled that of anatase titania. The Eg values, determined from the UV-Vis spectra, vary from 2.07 to 3.2 eV corresponded to pure titania. The photocatalytic activity of these materials showed a degradation of naproxen from 93.6 to 99.8 wt.% after 4 h under UV irradiation.


Introduction
Drugs top the list of emerging contaminants detected in wastewater effluents [1]. The constant increase and chemical stability of these drugs hinders their removal and/or disposal in conventional wastewater treatment plants [2,3], which causes a deterioration of aquatic life [4,5]. Different water treatments have been reported using advanced oxidation processes (AOPs) [6][7][8] for the elimination of these drugs, such as anti-inflammatories [4,9], antipyretics, and analgesics [10,11]. Among the promising AOPs, photocatalysis stands out, since it manages to degrade and mineralize up to 100% of the contaminants [12,13]. Photocatalytic processes are based on the use of nanomaterials that have the ability to transfer charge between the semiconductor and the aqueous solution, of transference, where the redistribution caused by the absorption of photons (hv) generates the hollow electron pair (e − -h). This, when interacting in the medium, generates hydroxyl radicals (OH) and/or superoxide anions that can function as reducing or oxidizing agents to degrade organic molecules [14,15]. Among the most used photocatalysts are oxides such as TiO 2 [16], ZnO [17,18], Nb 2 O 5 [19], CdS [20], and SnO 2 [21], among others [22]; some researchers have even used mixed oxides as nanocomposites to have greater efficiency in the degradation of this type of contaminant, including TiO-ZnO [2,23], TiO 2 -Fe 2 O 3 [24], TiO-Al 2 O 3 [25], Ti-Si-Fe [13], and Ti-Zr [26]. Many studies have focused on using TiO 2 either as a mixed oxide or as support because its use prevents the recombination of the photo-generated hollow electron pair (e − -h) and, thus, reduces the energy band [27][28][29]. Among the nanomaterials used for band gap reduction, rare earths can be found [30,31], such as Ce [32] and La [33]. The use of TiO 2 -CeO in the degradation of dyes, lignin [34], and metronidazole [35] has been reported. TiO 2 -La 2 O 3 has been used to degrade volatile organic compounds [28], dyes [36][37][38], and drugs [39], and various catalysts have been reported to degrade mostly dyes (Table 1). Table 1. Titania-based nanocomposites with rare earths and their photocatalytic activity.

Catalyst Contaminant Type of Light Reaction Time (min) Efficiency Reference
Ti-La Methylene blue UV irradiation 20 20% [28] Ti-Ce Methylene blue 585 nm 20 <10% [28] Ti-Gd Methylene blue UV irradiation 20 20% [28] Ti-Ce Metronidazole Visible light illumination 120 50% [35] Ti-Ce/N/C Metronidazole Visible light illumination 120 100% [35] Ti-Co/Ce/ La/Eu/Sm Methylene blue 440-550 nm 30 >90% [40] TiO 2 Methylene blue 440-550 nm 30 <50% [40] Ti-Nd Methyl orange Simulated solar radiation (SSR) 240 >90% [41] Ti-Eu Methyl orange SSR 240 <70% [41] Ti-Tb Methyl orange SSR 240 <40% [41] TiO 2 Methyl orange SSR 240 <50% [41] Ti-La Methylene blue 484 nm 120 <70% [42] Ti-La/Tb Methylene blue 484 nm 120 80% [42] Ti-Eu Methyl orange Simulated sunlight radiation 20 >95% [34] P25 TiO 2 Methyl orange Simulated sunlight radiation 20 <50% [34] Ti-Sm Lignin Simulated sunlight radiation 90 >95% [34] P25 TiO 2 Lignin Simulated sunlight radiation 90 <80% [34] Ti-La Azo-dye acid orange 7 >420 nm 300 <80% [43] Ti-La-B Azo-dye acid orange 7 >420 nm 300 >90% [43] Ti-Si-La Methylene blue UV irradiation 120 94% [44] TiO 2 Methylene blue UV irradiation 120 <60% [44] Ti/bc-La Methyl orange UV irradiation 300 >85% [37] Ti/bc-Ce Methyl orange UV irradiation 300 >95% [37] Ti/bc Methyl orange UV irradiation 300 <60% [37] C,N,S-TiO 2 Ibuprofen Visible light 300 >99% [45] The effect of the synthesis parameters on the photocatalytic activity of Ti-La materials in the degradation of naproxen (NPX) is herein reported. Lanthanum is used as a dopant of titania for this purpose. The effect of La concentration on the photocatalytic activity is also explored. The samples are labelled as Ti-LaX where X represents the weight percentage of lanthanum.  Figure 1. Independent of the La content, the morphology of all samples is similar [28]. Agglomerates comprised of particles of about 200 nm are noticed in all samples. The elemental composition determined by Energy-dispersive X-ray spectroscopy (EDS) is also included in Figure 1. The calculated atomic ratios are consistent with the XPS results, in agreement with other research works [38,40]. Figure 2 shows the N 2 adsorption-desorption isotherms and the pore size distribution of the Ti-La samples. The isotherms are of type IV and present type H2 hysteresis associated with capillary condensation, Figure 2a. These isotherms are characteristic of mesopore materials [29]. Similar observations were previously noticed by Perez-Larios et al. for TiO 2 -ZnO, Mn/TiO 2 , Ti-Zr, and Ti-Co catalysts [20,24,30,31]. Figure 2b show unimodal pore distribution with a mean pore size which decreases as the lanthanum content augments. Table 2 presents the specific surface areas determined by the BET equation and the mean pore size of Ti-La samples. The presence of lanthanum into TiO 2 increases the specific surface areas proportionally, whereas the mean pore size increases for Ti-La1 compared to that for TiO 2 and then decreases as the lanthanum content augments.  Figure 1. Independent of the La content, the morphology of all samples is similar [28]. Agglomerates comprised of particles of about 200 nm are noticed in all samples. The elemental composition determined by Energy-dispersive X-ray spectroscopy (EDS) is also included in Figure 1. The calculated atomic ratios are consistent with the XPS results, in agreement with other research works [38,40].  Figure 2 shows the N2 adsorption-desorption isotherms and the pore size distribution of the Ti-La samples. The isotherms are of type IV and present type H2 hysteresis associated with capillary condensation, Figure 2a. These isotherms are characteristic of mesopore materials [29]. Similar observations were previously noticed by Perez-Larios et al. for TiO2-ZnO, Mn/TiO2, Ti-Zr, and Ti-Co catalysts [20,24,30,31]. Figure 2b show unimodal pore distribution with a mean pore size which decreases as the lanthanum content augments. Table 2 presents the specific surface areas determined by the BET equation and the mean pore size of Ti-La samples. The presence of lanthanum into TiO2 increases the specific surface areas proportionally, whereas the mean pore size increases for Ti-La1 compared to that for TiO2 and then decreases as the lanthanum content augments.   The diffraction patterns of the calcined samples before the photocatalytic test are shown in Figure 3. All samples present diffractions which resemble those assigned for anatasa titania. The characteristic diffractions of anatase titania are located at 25,37,48,54,55,62,71, and 75 2θ degrees, which correspond to (101), (004), (200), (105), (211), (204), (116), and (311) crystallographic planes, respectively, according to JCPDS charts No. 00-001-0562 [9]. Additional peaks assigned to La 2 O 3 were unnoticed, indicating, as a consequence, a well lanthanum distribution in the titanium oxide structure and/or the presence of small lanthanum oxides species out of the detection limit of the instrument. The most intensive diffraction at about 25.4 degrees 2θ decreases and became broader as the lanthanum content augments [45,46]. The unit cell parameters presented in Table 2 increase as the lanthanum content augments compared to that of pure titania. Furthermore, the crystallite size, D, decreases as function of lanthanum. Our results, therefore, suggest that La 3+ cations are well dispersed in the titanium oxide structure, probably located in the interstitial positions since the ionic radii of La 3+ is greater to that of Ti 4+ , 0.106 and 0.062 nm, respectively [47], and substitution of titanium by lanthanum ions in the TiO 2 lattice is nonviable. However, assuming the presence of very small lanthanum oxides species (out of the detection limit of the XRD equipment) the replacement of lanthanum by titanium in the structure of lanthanum oxides is possible and Ti 3+ can be formed [48,49]. The average crystallite sizes were determined using the Scherrer equation taken the half width (b) of the most intensive peak located at 25.4 degrees 2θ and are reported in Table 2. The average crystallite size of the Ti-La samples is significant reduced compared to that of pure titania (20.58 nm). Table 2 also shows that increasing the lanthanum content a further reduction in the average crystallite size is observed, as previously noticed for similar Ti-La samples [43,50,51]. Ti-La1

Nitrogen Physisorption
Ti-La3 Ti-La5 Ti-La10     Figure 4 shows the band gap energies of the Ti-La samples determined from UV-Vis spectra. The insert in this figure indicates that all samples exhibit an optical absorption at about 400 nm, which can be attributed to the electron transition of the Ti-O [52]. A slightly shift to the red region (2.92 to 3.01 eV) for the Ti-La samples compared to pure TiO 2 resulted after the incorporation of La 3+ cations into TiO 2 lattice [41]. These observations are in agreement with other research works related to doped titania. A shift in the absorption edge to shorter wavenumber is clearly observed [53].   Figure 5 shows the FTIR spectra of TiO2 and Ti-La samples from 4000 to 400 cm −1 . The bands at about 448 cm −1 are associated to Ti-O vibrations [54] and related to the stretching of the Ti-OH and Ti-O bonds [38]. Although the samples were previously calcined before the FT-IR measurement, the bands at about 1600 and 1500 cm −1 correspond to C=O and C=C bonds originated from the residual organic compounds remained after synthesis [55]. The binding energies for Ti 2p are located at about 459 and 465 eV and correspond to Ti 2p3/2 and Ti 2p1/2, respectively. The separation of these two peaks is about 5.5 eV, which is in good agreement with previous publications related to doped titania (figure 6)  Figure 5 shows the FTIR spectra of TiO 2 and Ti-La samples from 4000 to 400 cm −1 . The bands at about 448 cm −1 are associated to Ti-O vibrations [54] and related to the stretching of the Ti-OH and Ti-O bonds [38]. Although the samples were previously calcined before the FT-IR measurement, the bands at about 1600 and 1500 cm −1 correspond to C=O and C=C bonds originated from the residual organic compounds remained after synthesis [55].   Figure 5 shows the FTIR spectra of TiO2 and Ti-La samples from 4000 to 400 cm −1 . The bands at about 448 cm −1 are associated to Ti-O vibrations [54] and related to the stretching of the Ti-OH and Ti-O bonds [38]. Although the samples were previously calcined before the FT-IR measurement, the bands at about 1600 and 1500 cm −1 correspond to C=O and C=C bonds originated from the residual organic compounds remained after synthesis [55]. The binding energies for Ti 2p are located at about 459 and 465 eV and correspond to Ti 2p3/2 and Ti 2p1/2, respectively. The separation of these two peaks is about 5.5 eV, which is in good agreement with previous publications related to doped titania (figure 6) The binding energies for Ti 2p are located at about 459 and 465 eV and correspond to Ti 2p 3/2 and Ti 2p 1/2 , respectively. The separation of these two peaks is about 5.5 eV, which is in good agreement with previous publications related to doped titania ( Figure 6) [49,56,57]. The La 3d5/2 spectrum presents two peaks, one at about 834 eV and a satellite peak at about 839 eV, with a splitting of about 5 eV (Figure 7). Although the interpretation of the XPS spectrum of pure lanthanum oxide is complicated, since La is very sensitive to water and CO2 from air and form a hydroxide and/or carbonate, it is likely that lanthanum as a dopant in our Ti-La samples is chemically more stable, since it is mainly buried into the titanium oxide structure. For a pure La2O3 sample exposed to air at ambient temperature, on which the formation of a hydroxide and/or carbonate have already occurred, the splitting of the two peaks of the La 3d5/2 spectrum is about 3.6 eV [57][58][59][60], whereas in ours, the separation is about 5 eV. The spectrum of the La 3d3/2 also shows two peaks, one about 851 eV and a satellite at about 856 eV. Their positions randomly change as function of the lanthanum content, similarly as those for La 3d5/2. The La 3d 5/2 spectrum presents two peaks, one at about 834 eV and a satellite peak at about 839 eV, with a splitting of about 5 eV (Figure 7). Although the interpretation of the XPS spectrum of pure lanthanum oxide is complicated, since La is very sensitive to water and CO 2 from air and form a hydroxide and/or carbonate, it is likely that lanthanum as a dopant in our Ti-La samples is chemically more stable, since it is mainly buried into the titanium oxide structure. For a pure La 2 O 3 sample exposed to air at ambient temperature, on which the formation of a hydroxide and/or carbonate have already occurred, the splitting of the two peaks of the La 3d 5/2 spectrum is about 3.6 eV [57][58][59][60], whereas in ours, the separation is about 5 eV. The spectrum of the La 3d 3/2 also shows two peaks, one about 851 eV and a satellite at about 856 eV. Their positions randomly change as function of the lanthanum content, similarly as those for La 3d 5/2 .

FT-IR Spectroscopy
The XPS spectra for O 1s of the Ti-La samples are shown in Figure 8. Although oxygen is bound to titanium and lanthanum, its contribution is overlapped, since only one broad peak is present in all spectra. However, accordingly to the difference in electronegativity of both elements, the signal of the oxygen attached to titanium would appear at lower binding energy, compared to that of the oxygen bound to lanthanum [61]. The O 1s spectrum could be deconvoluted in two signals, an intense peak at about 530.5 eV and the other at about 532 eV, for the oxygen bound to titanium and lanthanum, respectively. The small peak at 532 eV would be in agreement with the low lanthanum content in the Ti-La samples [62].  The XPS spectra for O 1s of the Ti-La samples are shown in Figure 8. Although o gen is bound to titanium and lanthanum, its contribution is overlapped, since only broad peak is present in all spectra. However, accordingly to the difference in electron ativity of both elements, the signal of the oxygen attached to titanium would appea lower binding energy, compared to that of the oxygen bound to lanthanum [61]. The O spectrum could be deconvoluted in two signals, an intense peak at about 530.5 eV and other at about 532 eV, for the oxygen bound to titanium and lanthanum, respectively. small peak at 532 eV would be in agreement with the low lanthanum content in the T samples [62].  The presence of C in the XPS spectra, most likely resulting from the organic compounds used during synthesis, are presented in Figure 9. The C 1s spectra show a single broad peak at about 285 eV. However, their positions randomly change accordingly to the lanthanum content. It was reported that for a pure La2O3 sample exposed to air, the C 1s spectrum shows two signals, one at 284.5 and the other at 289.5 eV, assigned to adventi- The presence of C in the XPS spectra, most likely resulting from the organic compounds used during synthesis, are presented in Figure 9. The C 1s spectra show a single broad peak at about 285 eV. However, their positions randomly change accordingly to the lanthanum content. It was reported that for a pure La 2 O 3 sample exposed to air, the C 1s spectrum shows two signals, one at 284.5 and the other at 289.5 eV, assigned to adventitious carbon C 1s and to carbonate C 1s, respectively [63,64]. Since our C 1s spectra present a single broad peak centered at about 285 eV, the formation of lanthanum carbonate after exposing our Ti-La samples to air was excluded. Carbon species which remain after calcination, such as C-O-C and C=O, could well be present in our Ti-La samples. Additional information obtained from XPS is presented in Table 3. The determined weight percentage is in agreement with the theoretical values. The shifts to lower binding energies for Ti-La5 in the Ti 2p and La 3d spectra, compared to other Ti-La samples, are unclear and more studies are required to elucidate these changes [57,60]. TEM images ( Figure 10) show an interplanar spacing of 0.35 nm, corresponding to the (1 0 1) plane lattice space, typical of TiO2 [40]. It can also be seen in nano-globular form with other nanostructures present for the Ti-La samples containing 1 3, 5, and 10 wt.% of Lanthanum, respectively. The differences in contrast between the doped and pure titania confirm the incorporation of La 3+ into the titania structure. The results are also consistent with those obtained from the XRD analysis and SEM. Figure 10 presents some TEM images of the Ti and Ti-La samples before the photocatalytic test. The interplanar spacing of 0.35 nm, corresponding to the (1 0 1) plane of TiO2 Additional information obtained from XPS is presented in Table 3. The determined weight percentage is in agreement with the theoretical values. The shifts to lower binding energies for Ti-La5 in the Ti 2p and La 3d spectra, compared to other Ti-La samples, are unclear and more studies are required to elucidate these changes [57,60]. TEM images ( Figure 10) show an interplanar spacing of 0.35 nm, corresponding to the (1 0 1) plane lattice space, typical of TiO 2 [40]. It can also be seen in nano-globular form with other nanostructures present for the Ti-La samples containing 1 3, 5, and 10 wt.% of Lanthanum, respectively. The differences in contrast between the doped and pure titania confirm the incorporation of La 3+ into the titania structure. The results are also consistent with those obtained from the XRD analysis and SEM.

Phococatalytic Activity
Although the photocatalytic experiments were analyzed every 30 min, Figure 11 only shows the degradation of NPX from 120 min onwards. This was done to increase the visibility of the results at the end of the experiments. Among the samples, the best catalysts   [40], changes as function of the lanthanum content which is in agreement with our XRD results. Further analysis of the TEM images to determine the nanoparticles size distribution was performed by the Imajej software measuring at least 50 nanoparticles [41][42][43].

Phococatalytic Activity
Although the photocatalytic experiments were analyzed every 30 min, Figure 11 only shows the degradation of NPX from 120 min onwards. This was done to increase the visibility of the results at the end of the experiments. Among the samples, the best catalysts after 4 h under UV irradiation were Ti-La1, Degussa P25, and Ti-La3 which show more than 90% degradation of NPX. Conversely, in absence of a catalyst, the decomposition of NPX by photolysis using UV was about 30% after 4 h of reaction, results not shown.
the TiO2, Ti-La1, and Ti-La3 samples turned out to be slightly more efficient than the Ti-La5 and Ti-La10 photocatalysts at 150 min. It is interesting to note that the Ti-La1 and Ti-La3 doping photocatalysts could lead to the formation of titania lattice distortion, which is prone to form the oxygen vacancy, and the vacancy could act as a trapper to inhibit the recombination of photo-generated electron-hole pairs [43]. It is a possibility that an impurity level has formed under the coactions of Ti 3+ and oxygen vacancies by La doping, which could inhibit the recombination of charge carriers. Then, La2O3 on the TiO2 surface could also transfer the electrons to the surface, which is advantageous for photocatalytic activity [42].
It can be seen from Figure 11, that pure TiO2, (1 × 10 −7 min −1 ) showed a better photocatalytic activity than photolysis (9.7 × 10 −6 min −1 ) under UV light, indicating that nano-TiO2 has an effect on accelerating the decomposition rate of NPX under UV light illumination. The nanocomposites Ti-La1 under the same experimental conditions was higher, which reveals that samples are readily excited under light irradiation, and photogenerated holes accelerate the oxidation of NPX before recombination with electrons. This result is consistent with other research [2,43,65,66]. Obviously, on this occasion, the optimal reaction time is 4 h, and the photocatalytic activity (1.5 × 10 −6 min −1 ) was better than that of TiO2. Ti-La1 Ti-La3 Ti-La5 Ti-La10 P25 Figure 11. Naproxen degradation profiles in presence of TiO2 or Ti-La photocatalysts under UV irradiation. Degussa P25 is used as a reference.
The reusability of TiO2 and Ti-La samples was evaluated by repeating the NPX degradation five times using the same catalysts. After a run, the catalyst was separated by filtration and dried in an oven overnight. Each experiment was performed with a fresh NPX solution, using the spent catalyst, during 4 h and under UV irradiation. As it can be seen in Figure 12, the photocatalytic activity of Ti-La1 is still attractive, since it declined about 5% after the fifth run.
Ti-La10 TiO Figure 11. Naproxen degradation profiles in presence of TiO 2 or Ti-La photocatalysts under UV irradiation. Degussa P25 is used as a reference. Figure 11 shows the NPX photodegradation rate curves of the TiO 2 , Ti-La1, Ti-La3, Ti-La5, and Ti-La10 nanophotocatalysts. The adsorption effect can be observed within the first 30 min, where the Ti-La nanocomposite adsorbs more contaminant than that of the bare TiO 2 ; this could be attributed to particle size and pore diameter, and these results are in agreement with previous reports where the Ti-Zr nanocomposite in the same relation at 1% degraded more than the others photocatalysts [56]. Otherwise, it is observed that the TiO 2 , Ti-La1, and Ti-La3 samples turned out to be slightly more efficient than the Ti-La5 and Ti-La10 photocatalysts at 150 min. It is interesting to note that the Ti-La1 and Ti-La3 doping photocatalysts could lead to the formation of titania lattice distortion, which is prone to form the oxygen vacancy, and the vacancy could act as a trapper to inhibit the recombination of photo-generated electron-hole pairs [43]. It is a possibility that an impurity level has formed under the coactions of Ti 3+ and oxygen vacancies by La doping, which could inhibit the recombination of charge carriers. Then, La 2 O 3 on the TiO 2 surface could also transfer the electrons to the surface, which is advantageous for photocatalytic activity [42].
It can be seen from Figure 11, that pure TiO 2 , (1 × 10 −7 min −1 ) showed a better photocatalytic activity than photolysis (9.7 × 10 −6 min −1 ) under UV light, indicating that nano-TiO 2 has an effect on accelerating the decomposition rate of NPX under UV light illumination. The nanocomposites Ti-La1 under the same experimental conditions was higher, which reveals that samples are readily excited under light irradiation, and photogenerated holes accelerate the oxidation of NPX before recombination with electrons. This result is consistent with other research [2,43,65,66]. Obviously, on this occasion, the optimal reaction time is 4 h, and the photocatalytic activity (1.5 × 10 −6 min −1 ) was better than that of TiO 2 .
The reusability of TiO 2 and Ti-La samples was evaluated by repeating the NPX degradation five times using the same catalysts. After a run, the catalyst was separated by filtration and dried in an oven overnight. Each experiment was performed with a fresh NPX solution, using the spent catalyst, during 4 h and under UV irradiation. As it can be seen in Figure 12, the photocatalytic activity of Ti-La1 is still attractive, since it declined about 5% after the fifth run. Ti-La1 Ti-La3 Ti-La5 Ti-La10 P25 Figure 11. Naproxen degradation profiles in presence of TiO2 or Ti-La photocatalysts under UV irradiation. Degussa P25 is used as a reference.
The reusability of TiO2 and Ti-La samples was evaluated by repeating the NPX degradation five times using the same catalysts. After a run, the catalyst was separated by filtration and dried in an oven overnight. Each experiment was performed with a fresh NPX solution, using the spent catalyst, during 4 h and under UV irradiation. As it can be seen in Figure 12, the photocatalytic activity of Ti-La1 is still attractive, since it declined about 5% after the fifth run. The total amount of carbon (TOC) found in the solutions after the photocatalytic tests were also determined and the percentages of mineralization are presented in Table 4. As can be seen, TiO2 and Ti-La1 are not only effective in degrading but also to mineralize NPX. According to the literature, 2-acetyl-6-methoxynaphthalene, detected by HPLC, is the main intermediate during the decomposition of NPX [66]. The total amount of carbon (TOC) found in the solutions after the photocatalytic tests were also determined and the percentages of mineralization are presented in Table 4. As can be seen, TiO 2 and Ti-La1 are not only effective in degrading but also to mineralize NPX. According to the literature, 2-acetyl-6-methoxynaphthalene, detected by HPLC, is the main intermediate during the decomposition of NPX [66]. The mechanism for the activation of TiO 2 or Ti-La photocatalysts, under UV is similar to that previously reported by Chaker et al. for La/TiO 2 catalysts [27]. Briefly, the electrons and holes (e − and h + ) generated during UV irradiation, reach the surface of the catalyst and react with water and oxygen to generate hydroxyl radicals and oxide radical anions through respective oxidative and reductive reactions. These radicals are strong oxidants and were able to decompose and mineralize NPX present in water. The differences in the photocatalytic activity of our samples are due to the physico-chemical properties of the catalysts. The following table summarizes the results presented in the literature concerning the decomposition of NPX on various catalysts. Although the information included in Table 5 is useful, comparison of the catalytic activities is not straightforward since the experimental conditions used in each research work were different.

Conclusions
In summary, TiO 2 and Ti-La oxides were synthesized by sol-gel and tested as photocatalysts in the degradation of naproxen. Results from characterization indicate that the incorporation of lanthanum into titanium oxide increases the specific surface area and reduces the average crystalline size compared to those of pure TiO 2 . In addition, lanthanum atoms are well distributed into the titanium oxide structure, probably located in the interstitial voids of the titania lattice. XPS results show that it is unlikely that lanthanum forms carbonate and/or a hydroxide after exposing our catalysts to air, since lanthanum atoms are well buried into the titanium oxide structure. XPS also shows that carbon species remain on the catalysts even after calcination. The photocatalytic activities of our samples in the degradation of NPX vary as function of the catalyst composition. Our results indicate that low concentration of lanthanum in titanium oxide is very attractive, since Ti-La1 showed the highest photocatalytic activity per mass of catalyst after 4 h and under UV irradiation. Furthermore, it was able to mineralize NPX and presents good stability after five cycles. The positive photocatalytic effects of Ti-La1 are probably related to the decrement in the recombination of electron-hole pairs.

Catalyst Preparation
The Ti-La samples were prepared by sol-gel and the required quantities of titanium (IV) butoxide and lanthanum nitrate to prepare the oxides containing 1.0, 3.0, 5.0 and 10.0 wt.% of La in titania were used. The metal compounds were dissolved in 44 mL of ethanol and 18 mL of deionized water under stirring and the pH of the solution was adjusted to 3 with an aqueous nitric-acid solution. The water/alkoxide molar ratio was 8. The solution was heated up to 70 • C and kept at this temperature for 24 h. The resulting solids were dried at 100 • C during 24 h. After this treatment, the samples were annealed at 500 • C during 4 h, applying a heating rate of 2 • C/min. Pure titania prepared following a similar procedure, and a commercial catalyst, Degussa P25, were used as references. The samples were then ready to be used as photocatalysts in the photocatalytic degradation of NPX.

Characterization Equipment
The surface morphology of the mixed oxides was analyzed using a scanning electron microscope (Tescan, MIRA 3LMU, Boston, MA, USA)) operated at 20 kV and provided with an Energy Dispersive Spectrometer (Bruker, QUANTAX, MA, USA), which was utilized to determine the superficial elemental composition.
The textural properties of the catalytic samples were determined from nitrogen adsorption-desorption experiments using a micromeritics TriStar II Plus equipment (Norcross, GA, USA). The samples were first degassed under vacuum at 200 • C for 3 h. Nitrogen adsorption isotherms were measured at liquid-nitrogen temperature (77 K), and N 2 pressures ranging from 10-6 to 1.0 P/P 0 . The specific surface areas were calculated following the Brunauer-Emmett-Teller (BET) method and the pore size distribution was obtained according to the Barret-Joyner-Halenda (BJH) method, using the Equation (1): where r p is the pore radius and t c the thickness of the adsorbed multilayer film, which is formed prior to pore condensation [61]. The diffraction patterns of the mixed oxides were recorded using an X-ray diffractometer equipped with Cu Kα radiation (λ = 0.154 nm). The diffraction intensities as a function of 2θ were measured between 10 and 90 • , using a step of 0.02 • and counting time of 0.2 s per step. Furthermore, the average particle crystalline size was determined by the Scherer Equation (2) using the most intense peak for each sample: where K is the shape factor (0.89), λ is the X-ray wavelength (0.154 nm), β is the peak broadening at half maximum, and θ is the Bragg angle. Additionally, interplanar spacing (d) can be calculated from the Bragg law Equation (3): The UV-Vis absorption spectra were obtained using a UV-Vis spectrophotometer (Shimadzu UV-2600, Tokio, Japan) coupled with an integration sphere for diffuse reflectance studies. The equipment was calibrated with barium sulfate BaSO 4 as a reference. The optical absorption was measured in the wavelength range of 190 to 900 nm. From the plot, the band gap energy for each sample was calculated using the following empirical Equation (4): where E g is the band gap energy; h is Plank constant equal to 6.626 × 10 −34 J/s; C is the velocity of light equal to 2.99 × 108 m/s; and λ is the detection wavelength. The FTIR spectra of the materials were recorded with an FTIR (Shimadzu, IRTracer-100, Tokyo, Japan) spectrophotometer using attenuated total reflectance (ATR) with a diamond waveguide (XR model). A detector of fast recovery deuterated triglycine sulfate (DTGS) (standard) was used for the analysis. The spectra were recorded at room temperature, with 24 scans and 4 cm −1 of resolution and from 4000 cm −1 to 400 cm −1 .
The XPS spectra were collected using a SPEC Phoibos 150 provided with a monochromatic Al Ka X-ray source (1487 eV). The position of the O 1s peak at 531.0 eV was monitored for each sample to ensure that no binding energy shift occurred due to charging.

Photocatalytic Activity
The reaction was carried out in a 350 mL Pyrex reactor using an aqueous solution of 30 mg/L NPX. For the photocatalytic experiments, 200 mg of photocatalyst was used in each run. The NPX solution was irradiated with a 1 mW*cm −2 ultraviolet lamp of 254 nm wavelength submerged in a quartz tube at pH 7 ± 0.15. In the first 30 min, the absorption capacity of the material was measured under dark conditions. Irradiation was immediately started and samples were taken every 30 min for 4 h. The analysis of the samples was carried out in a Shimadzu model 2600 UV-Vis spectrophotometer at a wavelength of the maximum absorbance for NPX at 230 nm [2]. To determine the reusability, the spent catalyst was recovered, dried, and calcined, before reusing it.
The NPX concentration was calculated from the corresponding maximum absorption wavelength. The absorbance was proportional to the Beer-Lambert law in the range of the studied NPX concentrations. The degradation percentages of the NPX were calculated by the following equation (Equation (5)): where C 0 and C is the initial and the actual concentration of NPX, and corresponds to a change in the current concentration of the NPX relative to the initial value of C 0 over time. Furthermore, the photocatalytic experiments were carried out three times to verify its reproducibility. Furthermore, the total organic carbon in the samples was measured using a TOC-LCSN equipment (Shimadzu, model), and calculated applying the Equation (6).
where TOC is the amount of total organic carbon (mg L −1 ), TC is the amount of total carbon (mg L −1 ), and IC inorganic carbon (mg L −1 ) in aqueous solution.
The quantitative analysis of the reaction kinetics of the degradation of organic compounds can be obtained by fitting the experimental data using the Langmuir-Hinshelwood model and expressed by the following Equation (7): where k is the rate constant, K the equilibrium constant, and c is the concentration of the organic pollutant. Nonetheless, Equation (1) can be simplified due to the low concentration of pollutant Kc < 1 to adapt to the form of a first-order apparent rate Equation (8): where k App is the apparent pseudo-first-order rate constant (min −1 ), C and C 0 are the final and initial organic compound concentrations (mg L −1 ), and t is the reaction time (min).

Conflicts of Interest:
The authors declare no conflict of interest.