E ﬀ ect of Synthesis Conditions of Nitrogen and Platinum Co-Doped Titania Films on the Photocatalytic Performance under Simulated Solar Light

: Platinum and nitrogen co-doped titania ﬁlms of di ﬀ erent surface morphologies obtained via a sol-gel process have been tested for tetracycline hydrochloride photocatalytic decomposition under simulated solar light. Titania crystallization to anatase is shown by XRD for all ﬁlms. A shift of the bandgap edge toward the visible region in absorption spectra and, consequently, a narrowing of the bandgap is observed for some ﬁlms doped with nitrogen and / or exposed to UV pretreatment. The surface peculiarities of the samples are presented by an SEM and TEM investigation. The surface saturation by Pt and N with a homogeneous distribution of Pt ions on the surface as well as bulk as established by XPS and EDS data can be achieved with a certain synthesis procedure. The inﬂuence of the platinum content and of the pretreatment procedure on the state and atomic surface concentration of incorporated nitrogen and platinum is studied by XPS analysis: substitutional and interstitial nitrogen, non-metal containing fragments, Pt 0 , Pt 2 + and Pt 4 + ions. The photocatalytic activity of the ﬁlms is ruled by the presence of Pt 2 + ions and N rather than Pt 0 . The formation of the polycrystalline titania structure and Pt 0 nanoparticles (NPs) is conﬁrmed by TEM and electron di ﬀ raction images. The mechanism of primary photocatalytic processes is proposed. 2M1 and 1Pt,N-TiO 2M2 resulting in co-doping with platinum ions of titania synthesized by the sol-gel method. The reduction and oxidation of Pt 2+ to Pt 0 and Pt 4+ can be achieved owing to the following reaction conditions: doping with urea, presence of some organic components, and UV pretreatment. Depending on the synthesis route, the optical properties are varied: a decrease in bandgap energy values is observed for the films by Method and no influence of nitrogen and platinum


Introduction
The growing scientific and technological interest in the construction of new composite materials with unique properties is directed by their applications as sensors, photocatalysts, catalysts and optical issues [1][2][3]. Environmental photocatalysis is one of the recently developing advanced treatment approaches for the removal of anthropogenic chemicals, which appeared as a result of the accelerated development of industrial and manufacturing rates during the last century [4].
The attempts of various TiO 2 modifications in order to manipulate its electronic structure and adjust its absorption to lower energy light have been widely reported [5]. Exposure to visible light, efficient photogenerated charge separation through the narrowing of the material's bandgap, or the formation of additional sublevels within the bandgap were studied and still remain the main goal of titania modification [6]. Numerous synthesis approaches have been reported for the manufacturing of active doped titania using single or double doping [6][7][8][9]. Depending on the nature and the metal ions, especially, of divalent Pt ions accessible for both oxidation and reduction processes and yet acetylacetonate counter-ions could be easily inbuilt into the organic sol framework during the synthesis. The photocatalytic destruction of TC molecules over the films has been carried out to investigate the influence of the surface morphology, chemical composition and nitrogen doping on their photocatalytic efficiency. For this purpose, some different approaches have been used: (i) the different synthesis routes to obtain different surface morphologies and (ii) UV treatment of the intermediate layers of the films to produce Pt 0 nanoparticles (NPs). An addition of Pluronic 123 as a template and acetylacetone as a stabilizing agent provides the formation of a porous structure that is conserved after sintering at 450 • C (Method 1). The smoothed and non-porous surface is observable for films obtained using the other simpler route (Method 2). SEM images of 1Pt,N-TiO 2M1 and 1Pt,N-TiO 2M2 films are presented in Figure 1.

Results
Platinum and nitrogen co-doped titania films of different morphologies ( Figure 1) have been synthesized by the sol-gel method. Platinum (II) acetylacetonate has been used as a precursor of doped metal ions, especially, of divalent Pt ions accessible for both oxidation and reduction processes and yet acetylacetonate counter-ions could be easily inbuilt into the organic sol framework during the synthesis. The photocatalytic destruction of TC molecules over the films has been carried out to investigate the influence of the surface morphology, chemical composition and nitrogen doping on their photocatalytic efficiency. For this purpose, some different approaches have been used: (i) the different synthesis routes to obtain different surface morphologies and (ii) UV treatment of the intermediate layers of the films to produce Pt 0 nanoparticles (NPs). An addition of Pluronic 123 as a template and acetylacetone as a stabilizing agent provides the formation of a porous structure that is conserved after sintering at 450 °С (Method 1). The smoothed and non-porous surface is observable for films obtained using the other simpler route (Method 2). SEM images of 1Pt,N-TiO2M1 and 1Pt,N-TiO2M2 films are presented in Figure 1. The optical behavior of 1 mol.% Pt-doped films obtained by Method 1 and Method 2 ( Figure 2) is critically different. The shift of the absorption band edges to a shorter wavelength than in pure titania film can be caused by the Q-size effect, i.e., the formation of lower particle sizes ( Figure 2a). However, the absorption intensity in the range of 350-450 nm increases for 1Pt-TiO2M1, 1Pt,N-TiO2M1 and 1PtN-TiO2M1UV, which can be a result of the formation of intra bandgap states, due to the presence of dopants' ions in the electronic structure of the semiconductor [27] and can be partially due also to localized surface plasmon resonance in the Pt nanoparticles embedded in the TiO2 [36]. The similar spectral character has been noted for the films containing 0.5 mol.% of Pt ions ( Figure S1a). The films obtained by Method 2 exhibit a high intensity in the visible part of the spectrum (the color of the films were muddy yellow) indicating the different ability to absorb and scatter light that can be connected to the structural features of the materials (Figure 2b). Optical spectra of the films containing 0.5 mol.% of Pt ions showed a much lower intensity in the visible region ( Figure S1b) compared to the films with 1 mol.% Pt. As shown by the XPS data presented below, the atomic content of Pt species on the surface of 1Pt,(N)-TiO2M2 films exceeds the one of 1Pt,(N)-TiO2M1 resulting in not only high intensity of absorption but also light scattering. It is assumed that the formed Pt 0 NPs are accumulated over the film surface leading to a significant contribution in light scattering that is clearly seen in the case of 1Pt,N-TiO2M2UV and 0.5Pt,N-TiO2M2UV (curve 5 in Figures 2b and S1b).
The emission spectrum of the irradiated source ( Figure S1c) shows that the excitation of the semiconductor materials takes place in the spectral region of sunlight. The intensity of the lamp without and with a cut-off filter (0-360 nm; 700-1200 nm) and the wavelength of the daily sunlight were measured by a pyroelectric wattmeter VP-1 in the middle of May (50.474084° N, 30.343643° E). The optical behavior of 1 mol.% Pt-doped films obtained by Method 1 and Method 2 ( Figure 2) is critically different. The shift of the absorption band edges to a shorter wavelength than in pure titania film can be caused by the Q-size effect, i.e., the formation of lower particle sizes ( Figure 2a). However, the absorption intensity in the range of 350-450 nm increases for 1Pt-TiO 2M1 , 1Pt,N-TiO 2M1 and 1PtN-TiO 2M1UV , which can be a result of the formation of intra bandgap states, due to the presence of dopants' ions in the electronic structure of the semiconductor [27] and can be partially due also to localized surface plasmon resonance in the Pt nanoparticles embedded in the TiO 2 [36]. The similar spectral character has been noted for the films containing 0.5 mol.% of Pt ions ( Figure S1a). The films obtained by Method 2 exhibit a high intensity in the visible part of the spectrum (the color of the films were muddy yellow) indicating the different ability to absorb and scatter light that can be connected to the structural features of the materials (Figure 2b). Optical spectra of the films containing 0.5 mol.% of Pt ions showed a much lower intensity in the visible region ( Figure S1b) compared to the films with 1 mol.% Pt. As shown by the XPS data presented below, the atomic content of Pt species on the surface of 1Pt,(N)-TiO 2M2 films exceeds the one of 1Pt,(N)-TiO 2M1 resulting in not only high intensity of absorption but also light scattering. It is assumed that the formed Pt 0 NPs are accumulated over the film surface leading to a significant contribution in light scattering that is clearly seen in the case of 1Pt,N-TiO 2M2UV and 0.5Pt,N-TiO 2M2UV (curve 5 in Figure 2b and and Figure S1b).
The emission spectrum of the irradiated source ( Figure S1c) shows that the excitation of the semiconductor materials takes place in the spectral region of sunlight. The intensity of the lamp without and with a cut-off filter (0-360 nm; 700-1200 nm) and the wavelength of the daily sunlight were measured by a pyroelectric wattmeter VP-1 in the middle of May (50.  The indirect transition bandgap energy values (Ebg) of the films of Method 1 obtained under different conditions are found to be within 3.5-3.7 eV, while the films of Method 2 are characterized by a lowered Ebg, in particular for 1Pt-TiO2UV, 1Pt,N-TiO2 and 1Pt,N-TiO2UV (Table 1). TiO2 and N-TiO2 films are characterized by 3.7 and 3.8 eV, respectively, for Method 1 and 3.6 eV for Method 2 ( Figure  S2). The variation of optical Ebg values is suggested to be connected to the size of TiO2 particles formed at the presented synthesis conditions as a result of a different aggregation degree [11] and/or hydrolysis rate of titanium tetraisopropoxide during sol formation.  Figure 3. The XRD peaks at 31.8° and 45.6° marked by stars in the Figure 3 do not correspond to any crystalline form of titania and platinum oxide and are related to the trace of impurities of unknown nature that require detailed investigation. The N-TiO2 films of Method 1 and Method 2 contain anatase and the mixture of anatase and brookite, respectively ( Figure S3). The indirect transition bandgap energy values (E bg ) of the films of Method 1 obtained under different conditions are found to be within 3.5-3.7 eV, while the films of Method 2 are characterized by a lowered E bg , in particular for 1Pt-TiO 2UV , 1Pt,N-TiO 2 and 1Pt,N-TiO 2UV (Table 1). TiO 2 and N-TiO 2 films are characterized by 3.7 and 3.8 eV, respectively, for Method 1 and 3.6 eV for Method 2 ( Figure S2). The variation of optical E bg values is suggested to be connected to the size of TiO 2 particles formed at the presented synthesis conditions as a result of a different aggregation degree [11] and/or hydrolysis rate of titanium tetraisopropoxide during sol formation.  An XPS analysis was applied to clarify the effect of different synthesis conditions on the composition of the surface layer: (i) different valence states of platinum; (ii) doping efficiency of both Pt and N. Atomic ratios of the elements on the surface obtained by the XPS data showed that the surface of the films of Method 12 contained less Pt and N compared to Method 2 ( Table 2). The surface of the films of Method 2 is more enriched by Pt and N compared to Method 1. Analyzing the elements' ratio on the surface and bulk of 1Pt-TiO2 and 1Pt,N-TiO2, it can be summarized that Pt is more homogeneously distributed in the structures obtained by Method 2. The ratio of Ti and O approached the stoichiometric value for both surface and bulk. The maxima of Ti2p3/2 XPS peaks (Figures 4 and S4) are centered at 458.5-458.7 eV, with a spinorbit split component equal to 5.7 eV, corresponding to the Ti 4+ atom surrounded by oxygen atoms. The slight shift to a lower binding energy (BE) (0.2 eV) of Ti2p3/2 maxima observed for nitrogen-doped films (Method 1) can be caused by nitrogen incorporation in the titania matrix in the form of substitutional and interstitial atoms ( Table 3). The O1s line showed the peaks corresponding to O 2and surface OH groups, at binding energy (BE) values near 530 and 531 eV ( Figures S4 and S5). An XPS analysis was applied to clarify the effect of different synthesis conditions on the composition of the surface layer: (i) different valence states of platinum; (ii) doping efficiency of both Pt and N. Atomic ratios of the elements on the surface obtained by the XPS data showed that the surface of the films of Method 12 contained less Pt and N compared to Method 2 ( Table 2). The surface of the films of Method 2 is more enriched by Pt and N compared to Method 1. Analyzing the elements' ratio on the surface and bulk of 1Pt-TiO 2 and 1Pt,N-TiO 2 , it can be summarized that Pt is more homogeneously distributed in the structures obtained by Method 2. The ratio of Ti and O approached the stoichiometric value for both surface and bulk. The maxima of Ti2p 3/2 XPS peaks ( Figure 4 and Figure S4) are centered at 458.5-458.7 eV, with a spin-orbit split component equal to 5.7 eV, corresponding to the Ti 4+ atom surrounded by oxygen atoms. The slight shift to a lower binding energy (BE) (0.2 eV) of Ti2p 3/2 maxima observed for nitrogen-doped films (Method 1) can be caused by nitrogen incorporation in the titania matrix in the form of substitutional and interstitial atoms ( Table 3). The O1s line showed the peaks corresponding to O 2and surface OH groups, at binding energy (BE) values near 530 and 531 eV ( Figures S4 and S5).   Nitrogen incorporation in the Pt,N-TiO2 structure takes place in its various states as it is evident from N1s XPS lines (Table 3, Figure 5) compared to N-TiO2 ( Figure S4). The most prominent line at BE = 399.8 ± 0.2 еV, observed for all N modified films, overlaps with the value assigned to the nitrogen-containing organic compounds that originated from the pyrolysis products of urea and corresponds to C-N-C fragments [37][38][39]. The maxima at higher BEs (400.5-404.1 еV) belong to the nitrogen surrounding at least one element more electronegative than carbon, corresponding to the О-N-Х bonds, where Х is С or О. The N1s peak at 398.0 ± 0.3 eV is also found in some cases and can be assigned to the formation of Ti-O-N fragments [40]. Lower BE values fixed at 395.8 ± 0.2 and 397.1 еV characterize the formation of common bonds between Ti and N (Ti-N) defined as substitutional (Ns) and interstitial (Ni) atoms, respectively [20,41,42]. The relative intensity of Ns and Ni to other N states is still low and found only for the samples containing 1 mol.% Pt. It has to be pointed out that nitrogen incorporation in the Ns form (BE~396 eV) is only reported for a few instances of nitrogen and metal ions co-modified, titania obtained by the sol-gel method (using urea and RuCl3 [39], ammonia and Sr(NO3)2 [43], dodecylamine and Fe(NO3)3•9H2O [44], tetramethylethylenediamide and Pt 0 /Pt 2+ [45]). Additionally, N1s lines with BE at 397.5 defined by XPS in the samples obtained by the solvothermal method using urea and H2PtCl6 are also reported [46]. Thus, the possibility of N to be incorporated in Ns and Ni forms by means of the sol-gel technique depends on the nature of the co-doped metal ions. Nitrogen incorporation in the Pt,N-TiO 2 structure takes place in its various states as it is evident from N1s XPS lines (Table 3, Figure 5) compared to N-TiO 2 ( Figure S4). The most prominent line at BE = 399.8 ± 0.2 еV, observed for all N modified films, overlaps with the value assigned to the nitrogen-containing organic compounds that originated from the pyrolysis products of urea and corresponds to C-N-C fragments [37][38][39]. The maxima at higher BEs (400.5-404.1 еV) belong to the nitrogen surrounding at least one element more electronegative than carbon, corresponding to the О-N-Хbonds, where Хis C or О. The N1s peak at 398.0 ± 0.3 eV is also found in some cases and can be assigned to the formation of Ti-O-N fragments [40]. Lower BE values fixed at 395.8 ± 0.2 and 397.1 еV characterize the formation of common bonds between Ti and N (Ti-N) defined as substitutional (Ns) and interstitial (Ni) atoms, respectively [20,41,42]. The relative intensity of Ns and Ni to other N states is still low and found only for the samples containing 1 mol.% Pt. It has to be pointed out that nitrogen incorporation in the Ns form (BE~396 eV) is only reported for a few instances of nitrogen and metal ions co-modified, titania obtained by the sol-gel method (using urea and RuCl 3 [39], ammonia and Sr(NO 3 ) 2 [43], dodecylamine and Fe(NO 3 ) 3 •9H 2 O [44], tetramethylethylenediamide and Pt 0 /Pt 2+ [45]). Additionally, N1s lines with BE at 397.5 defined by XPS in the samples obtained by the solvothermal method using urea and H 2 PtCl 6 are also reported [46]. Thus, the possibility of N to be incorporated in Ns and Ni forms by means of the sol-gel technique depends on the nature of the co-doped metal ions. As reported in [47][48][49], the metal ions related to hard Lewis acids (Ti is among them) are bonded to urea via an oxygen atom while the soft Lewis acids (including Pt) form the monodentate ligand via the metal-nitrogen bond and sometimes via the metal-oxygen one (Scheme 1a,b). Taking into account that the titanium-containing sol is acidified to control the hydrolysis rate, the protonation of urea molecules can occur via oxygen atoms preventing Ti 4+ -urea complexation while the interaction via N of the amide group with Pt 2+ can still take place (Scheme 1c). As a result, urea thermolysis can happen not only through a polycondensation reaction with the formation of heterocyclic compounds [50] defined in N1s at 399.8 ± 0.2 eV (Table 3), but also via the urea decomposition resulting in the formation of reactive nitrogen species (ammonia, cyanic acid) [6] due to its complexation to Pt 2+ ions, leading to the formation of Ns and Ni atoms. The XPS spectra of Pt4f lines exhibit complex non-symmetric shaped curves, indicating the presence of various Pt oxidation states ( Figure 6 and Table 4). A spin-orbit coupling is established according to the position of the energy of deconvoluted Pt4f lines, in the range of 3.3-3.4 eV (Table  S1), with the intensity ratio of Pt4f7/2:Pt4f5/2 = 4:3. Since the area of the Pt4f5/2 region exceeded the Pt4f7/2 one, the deconvolution was performed considering the overlapped peak of Al2p (BE values of Al 3+ As reported in [47][48][49], the metal ions related to hard Lewis acids (Ti is among them) are bonded to urea via an oxygen atom while the soft Lewis acids (including Pt) form the monodentate ligand via the metal-nitrogen bond and sometimes via the metal-oxygen one (Scheme 1a,b). Taking into account that the titanium-containing sol is acidified to control the hydrolysis rate, the protonation of urea molecules can occur via oxygen atoms preventing Ti 4+ -urea complexation while the interaction via N of the amide group with Pt 2+ can still take place (Scheme 1c). As a result, urea thermolysis can happen not only through a polycondensation reaction with the formation of heterocyclic compounds [50] defined in N1s at 399.8 ± 0.2 eV (Table 3), but also via the urea decomposition resulting in the formation of reactive nitrogen species (ammonia, cyanic acid) [6] due to its complexation to Pt 2+ ions, leading to the formation of Ns and Ni atoms. As reported in [47][48][49], the metal ions related to hard Lewis acids (Ti is among them) are bonded to urea via an oxygen atom while the soft Lewis acids (including Pt) form the monodentate ligand via the metal-nitrogen bond and sometimes via the metal-oxygen one (Scheme 1a,b). Taking into account that the titanium-containing sol is acidified to control the hydrolysis rate, the protonation of urea molecules can occur via oxygen atoms preventing Ti 4+ -urea complexation while the interaction via N of the amide group with Pt 2+ can still take place (Scheme 1c). As a result, urea thermolysis can happen not only through a polycondensation reaction with the formation of heterocyclic compounds [50] defined in N1s at 399.8 ± 0.2 eV (Table 3), but also via the urea decomposition resulting in the formation of reactive nitrogen species (ammonia, cyanic acid) [6] due to its complexation to Pt 2+ ions, leading to the formation of Ns and Ni atoms. The XPS spectra of Pt4f lines exhibit complex non-symmetric shaped curves, indicating the presence of various Pt oxidation states ( Figure 6 and Table 4). A spin-orbit coupling is established according to the position of the energy of deconvoluted Pt4f lines, in the range of 3.3-3.4 eV (Table  S1), with the intensity ratio of Pt4f7/2:Pt4f5/2 = 4:3. Since the area of the Pt4f5/2 region exceeded the Pt4f7/2 one, the deconvolution was performed considering the overlapped peak of Al2p (BE values of Al 3+ The XPS spectra of Pt4f lines exhibit complex non-symmetric shaped curves, indicating the presence of various Pt oxidation states ( Figure 6 and Table 4). A spin-orbit coupling is established according to the position of the energy of deconvoluted Pt4f lines, in the range of 3.3-3.4 eV (Table S1), with the intensity ratio of Pt4f 7/2 :Pt4f 5/2 = 4:3. Since the area of the Pt4f 5/2 region exceeded the Pt4f 7/2 one, the deconvolution was performed considering the overlapped peak of Al2p (BE values of Al 3+ are Catalysts 2020, 10, 1074 8 of 18 situated at 74.6-75.6 eV) [51] (Figure 7). The presence of Al 3+ ions can be explained by the traces of this element from the reagents used in the synthesis procedure.
Reduction of Pt 2+ to Pt 0 is confirmed by XPS lines with BE at 69.6-71.4 eV (Figures 6 and 7 and Table 4) and EDS mapping images (Figure 8). The large gap in these values can be explained by the different sizes of Pt nanoparticles (NPs) as reported in [52,53], where the shift of the BEs to lower values is assigned to the formation of bigger size NPs.
Catalysts 2019, 9, x FOR PEER REVIEW 8 of 18 are situated at 74.6-75.6 eV) [51] (Figure 7). The presence of Al 3+ ions can be explained by the traces of this element from the reagents used in the synthesis procedure. Reduction of Pt 2+ to Pt 0 is confirmed by XPS lines with BE at 69.6-71.4 eV (Figures 6 and 7 and Table 4) and EDS mapping images (Figure 8). The large gap in these values can be explained by the different sizes of Pt nanoparticles (NPs) as reported in [52,53], where the shift of the BEs to lower values is assigned to the formation of bigger size NPs.    It has to be noted that Pt 0 is fixed for all the films of Method 1 and some of Method 2, as shown by the deconvoluted XPS Pt4f region (examples of spectra are in Figure 7). In the case of Method 1 samples, the reduction process can be caused by the presence of high-molecular weight organic components in the formed sol (Pluronic 123, acetylacetone). No Pt 0 is observed for the Pt-doped films Catalysts 2020, 10, 1074 9 of 18 of Method 2 obtained from the sol containing low-molecular weight organic compounds (iso-propanol). Nevertheless, an effect of urea on Pt 2+ reduction is pronounced for the nitrogen-containing films of Method 2. A likely possibility is that an intermediate of urea decomposition, ammonium cyanate [50], formed during thermal treatment, plays the role of a reducing agent in this process. As evidence, the oxidative forms of nitrogen are detected in N1s spectra in the range of 400.5-404.1 eV.
It has to be noted that Pt 0 is fixed for all the films of Method 1 and some of Method 2, as shown by the deconvoluted XPS Pt4f region (examples of spectra are in Figure 7). In the case of Method 1 samples, the reduction process can be caused by the presence of high-molecular weight organic components in the formed sol (Pluronic 123, acetylacetone). No Pt 0 is observed for the Pt-doped films of Method 2 obtained from the sol containing low-molecular weight organic compounds (isopropanol). Nevertheless, an effect of urea on Pt 2+ reduction is pronounced for the nitrogen-containing films of Method 2. A likely possibility is that an intermediate of urea decomposition, ammonium cyanate [50], formed during thermal treatment, plays the role of a reducing agent in this process. As evidence, the oxidative forms of nitrogen are detected in N1s spectra in the range of 400.5-404.1 eV. It has to be mentioned that the oxidation of Pt 2+ to Pt 4+ ions takes place only in the case of the films of Method 2, independently of nitrogen doping. The oxidation of Pt 2+ ions can be caused by the thermal decomposition of Pt(acac)2 under certain conditions [54,55].
EDS mapping and TEM images (Figure 8) confirm the formation of Pt NPs in the structures of the films of Method 1. In the case of Method 2, no Pt 0 NPs are observed for 1Pt-TiO2n while an additional map scanning at a higher magnification shows rare Pt 0 NP's for 1Pt,N-TiO2M2. It has to be mentioned that the oxidation of Pt 2+ to Pt 4+ ions takes place only in the case of the films of Method 2, independently of nitrogen doping. The oxidation of Pt 2+ ions can be caused by the thermal decomposition of Pt(acac) 2 under certain conditions [54,55].
EDS mapping and TEM images (Figure 8) confirm the formation of Pt NPs in the structures of the films of Method 1. In the case of Method 2, no Pt 0 NPs are observed for 1Pt-TiO 2 n while an additional map scanning at a higher magnification shows rare Pt 0 NP's for 1Pt,N-TiO 2M2 . Catalysts 2019, 9, x FOR PEER REVIEW 10 of 18 The mean sizes of the TiO2 and Pt 0 obtained from TEM images ( Figure 9) and summarized in Table 5 point out a nanoparticle formation in the range of 9-14 nm and 3-4 nm, respectively. No significant size difference is observed for the films synthesized under different conditions, except for 1% Pt-N-TiO2M1UV film. The formation of larger Pt NPs can be explained by the appearance of the initial Pt crystallization centers due to UV treatment. A more even distribution of the nanoparticles was observed for the samples doped only with platinum. No Pt NPs were observed for the 1% Pt-TiO2M2 sample. The electron diffraction patterns ( Figure S6) of the samples prove the formation of polycrystalline TiO2. The observed interplanar distance between adjacent planes (d-spacing) of the single crystals (Table 5) corresponds to the anatase (1 0 1) polymorph of TiO2.
The correlation of the photocatalytic activity of the films in the TC degradation process vs. the Pt oxidation states/atomic contents and atomic nitrogen content is presented in Figure 10. For this purpose, the relative atomic contents (RACs) of Pt species (Table 6) have been obtained by the recalculation of the total Pt atomic contents (the ratios of the elements are presented in Table 2) and the relative intensities were previously presented in Table 4.
Comparing the photocatalytic activity of Method 1 materials (Figure 10a), one can note that the films synthesized using 1 mol.% Pt(acac)2 exhibit a higher conversion of TC molecules. It is also obvious that RACs of Pt species affect the photoperformance of the photocatalysts: the higher the RAC of Pt 0 is, the lower the TC conversion that occurs. In cases of a low RAC of Pt 2+ , the presence of nitrogen seems to be crucial for an increased activity. The low photoactivity of 0.5Pt-TiO2M1UV and 0.5Pt,N-TiO2M1UV is explained by the lowest atomic contents of the doping agents among tested materials. As seen from Figure 10a, the presence of urea and UV pretreatment affects the relative atomic content of Pt 0 NPs [46]. Hence, the activity of Pt-containing films of Method 1 can be caused by some mutual parameters: efficient incorporation of Pt 2+ in the TiO2 lattice, the low content of Pt 0 and doping by nitrogen. The mean sizes of the TiO 2 and Pt 0 obtained from TEM images ( Figure 9) and summarized in Table 5 point out a nanoparticle formation in the range of 9-14 nm and 3-4 nm, respectively. No significant size difference is observed for the films synthesized under different conditions, except for 1% Pt-N-TiO 2M1UV film. The formation of larger Pt NPs can be explained by the appearance of the initial Pt crystallization centers due to UV treatment. A more even distribution of the nanoparticles was observed for the samples doped only with platinum. No Pt NPs were observed for the 1% Pt-TiO 2M2 sample. The electron diffraction patterns ( Figure S6) of the samples prove the formation of polycrystalline TiO 2 . The observed interplanar distance between adjacent planes (d-spacing) of the single crystals (Table 5) corresponds to the anatase (1 0 1) polymorph of TiO 2 .
The correlation of the photocatalytic activity of the films in the TC degradation process vs. the Pt oxidation states/atomic contents and atomic nitrogen content is presented in Figure 10. For this purpose, the relative atomic contents (RACs) of Pt species (Table 6) have been obtained by the recalculation of the total Pt atomic contents (the ratios of the elements are presented in Table 2) and the relative intensities were previously presented in Table 4.
Comparing the photocatalytic activity of Method 1 materials (Figure 10a), one can note that the films synthesized using 1 mol.% Pt(acac) 2 exhibit a higher conversion of TC molecules. It is also obvious that RACs of Pt species affect the photoperformance of the photocatalysts: the higher the RAC of Pt 0 is, the lower the TC conversion that occurs. In cases of a low RAC of Pt 2+ , the presence of nitrogen seems to be crucial for an increased activity. The low photoactivity of 0.5Pt-TiO 2M1UV and 0.5Pt,N-TiO 2M1UV is explained by the lowest atomic contents of the doping agents among tested materials. As seen from Figure 10a, the presence of urea and UV pretreatment affects the relative atomic content of Pt 0 NPs [46]. Hence, the activity of Pt-containing films of Method 1 can be caused by some mutual parameters: efficient incorporation of Pt 2+ in the TiO 2 lattice, the low content of Pt 0 and doping by nitrogen.    Contrary to the films of Method 1, the highest photoactivity of the films of Method 2 (Figure 10b) was observed for nitrogen-doped structures, independently of the molar percentages of Pt(acac) 2 loaded during synthesis-the most active films are 0.5Pt,N-TiO 2M2 and 1Pt,N-TiO 2M2 , containing the highest N atomic concentrations. The lowest activity of the nitrogen-free film containing only Pt 2+ ions, 0.5Pt-TiO 2M2 and 1Pt-TiO 2M2 , could be caused by ineffective light absorption due to the widest bandgaps among non-porous films (Table 1). One more piece of evidence that can affect the efficiency of photoinduced charge separation is the presence of Pt 4+ ions that can be additional traps for photogenerated electrons.  Contrary to the films of Method 1, the highest photoactivity of the films of Method 2 ( Figure  10b) was observed for nitrogen-doped structures, independently of the molar percentages of Pt(acac)2 loaded during synthesis-the most active films are 0.5Pt,N-TiO2M2 and 1Pt,N-TiO2M2, containing the highest N atomic concentrations. The lowest activity of the nitrogen-free film containing only Pt 2+ ions, 0.5Pt-TiO2M2 and 1Pt-TiO2M2, could be caused by ineffective light absorption due to the widest bandgaps among non-porous films (Table 1). One more piece of evidence that can affect the efficiency of photoinduced charge separation is the presence of Pt 4+ ions that can be additional traps for photogenerated electrons.

Discussion
The semiconductor films were obtained by sol-gel methods using different approaches, and their photocatalytic activity was measured by the photodegradation of tetracycline hydrochloride. The enrichment of the film surface of Method 2 by Pt and N as well as the homogeneous distribution of doped metal on the surface and bulk compared to Method 1 was established by XPS and EDS data. The XPS results reveal that nitrogen is mostly presented in C-N-C, O-N-O(C), and Ti-O-N fragments. Nitrogen incorporation in the forms of substitutional and interstitial atoms was realized for 1Pt,N-TiO2M1 and 1Pt,N-TiO2M2 resulting in co-doping with platinum ions of titania synthesized by the solgel method. The reduction and oxidation of Pt 2+ to Pt 0 and Pt 4+ can be achieved owing to the following reaction conditions: doping with urea, presence of some organic components, and UV pretreatment. Depending on the synthesis route, the optical properties are varied: a decrease in bandgap energy values is observed for the films obtained by Method 2 and no influence of nitrogen and platinum

Discussion
The semiconductor films were obtained by sol-gel methods using different approaches, and their photocatalytic activity was measured by the photodegradation of tetracycline hydrochloride. The enrichment of the film surface of Method 2 by Pt and N as well as the homogeneous distribution of doped metal on the surface and bulk compared to Method 1 was established by XPS and EDS data. The XPS results reveal that nitrogen is mostly presented in C-N-C, O-N-O(C), and Ti-O-N fragments. Nitrogen incorporation in the forms of substitutional and interstitial atoms was realized for 1Pt,N-TiO 2M1 and 1Pt,N-TiO 2M2 resulting in co-doping with platinum ions of titania synthesized by the sol-gel method. The reduction and oxidation of Pt 2+ to Pt 0 and Pt 4+ can be achieved owing to the following reaction conditions: doping with urea, presence of some organic components, and UV pretreatment. Depending on the synthesis route, the optical properties are varied: a decrease in bandgap energy values is observed for the films obtained by Method 2 and no influence of nitrogen and platinum doping is noted for Method 1; the light scattering observed in the absorption spectra of the films obtained by Method 2 is attributed to the saturation of the film surface by Pt 0 NPs.
Taking into account that the TC conversion percentage over TiO 2 and N-TiO 2 films, synthesized in the same manner as platinum-doped films, which is found to be 3-4% and 12-13%, respectively, most of the modified films showed much higher photocatalytic activities, suggesting that Pt species are responsible for the increased photocatalytic degradation as well as visible light absorption. The photocatalytic destruction of tetracycline hydrochloride monitored under a simulated solar light is increased over the samples containing a sufficient amount of Pt 2+ ions as well as N on their surface, while the increased content of Pt 0 delays their photoactivity. The most active films are 1Pt-TiO 2M1 (27%) and 1Pt,N-TiO 2M2 (26%), leading to the conclusion that the morphology of the surface has no effect on the photocatalytic degradation of TC molecules over Pt-doped TiO 2 films, as opposed to the oxidation states, atomic contents and the ratio of platinum and nitrogen species in the surface layer. The proposed mechanism of primary photocatalytic pathways is depicted in Scheme 2. The formation of Pt sublevels under the conduction band of titania responsible for the visible light absorption is suggested. A photogenerated electron can be trapped by Pt n+ ions or Pt 0 preventing the recombination process. There is a high probability that these pathways are competitive depending on Pt NPs distribution and content on the surface. It is reported [33] that platinized TiO 2 exhibits a higher photocatalytic activity in the photocatalytic degradation of trichloroethylene, perchloroethylene, dichloroacetate than the sample doped with Pt 2+ /Pt 4+ ions. On the other hand, the increased Pt loading to 1% causes a decrease in the reaction rate of methanol formation over platinized TiO 2 samples [27]. It is suggested that the decrease in photoactivity of the materials takes place due to (i) the narrowing of the space charge layer with higher Pt loading leading to the deeper penetration depth of light resulting in the favorable electron-hole recombination and/or (ii) the light screening by Pt 0 particles causing the less effective electron-hole pair formation. doping is noted for Method 1; the light scattering observed in the absorption spectra of the films obtained by Method 2 is attributed to the saturation of the film surface by Pt 0 NPs. Taking into account that the TC conversion percentage over TiO2 and N-TiO2 films, synthesized in the same manner as platinum-doped films, which is found to be 3-4% and 12-13%, respectively, most of the modified films showed much higher photocatalytic activities, suggesting that Pt species are responsible for the increased photocatalytic degradation as well as visible light absorption. The photocatalytic destruction of tetracycline hydrochloride monitored under a simulated solar light is increased over the samples containing a sufficient amount of Pt 2+ ions as well as N on their surface, while the increased content of Pt 0 delays their photoactivity. The most active films are 1Pt-TiO2M1 (27%) and 1Pt,N-TiO2M2 (26%), leading to the conclusion that the morphology of the surface has no effect on the photocatalytic degradation of TC molecules over Pt-doped TiO2 films, as opposed to the oxidation states, atomic contents and the ratio of platinum and nitrogen species in the surface layer. The proposed mechanism of primary photocatalytic pathways is depicted in Scheme 2. The formation of Pt sublevels under the conduction band of titania responsible for the visible light absorption is suggested. A photogenerated electron can be trapped by Pt n+ ions or Pt 0 preventing the recombination process. There is a high probability that these pathways are competitive depending on Pt NPs distribution and content on the surface. It is reported [33] that platinized TiO2 exhibits a higher photocatalytic activity in the photocatalytic degradation of trichloroethylene, perchloroethylene, dichloroacetate than the sample doped with Pt 2+ /Pt 4+ ions. On the other hand, the increased Pt loading to 1% causes a decrease in the reaction rate of methanol formation over platinized TiO2 samples [27]. It is suggested that the decrease in photoactivity of the materials takes place due to (i) the narrowing of the space charge layer with higher Pt loading leading to the deeper penetration depth of light resulting in the favorable electron-hole recombination and/or (ii) the light screening by Pt 0 particles causing the less effective electron-hole pair formation. After light absorption with energy lower (the lines in the spectrum of the used lamp with a cutoff filter shown in Figure S1 are 364, 404 and 434 nm corresponding to 3.4, 3.1 and 2.9 eV, respectively) than the bandgap energy of titania (step 1), the transition of an electron from the valence band to the sublevel of Pt or from the N sublevel to the conduction band can occur. It appears that Pt ions play the role of trapping centers for photogenerated electrons, with a following reduction in oxygen molecules to superoxide radicals, and Pt oxidation to its previous oxidation state (steps 2 and 4) [25]. As shown by the experimental results, the higher content of Pt 0 has a detrimental effect on photocatalytic processes, suggesting their light screening from the TiO2 surface or preferential charge recombination, as proposed in [27]. The oxidation pathways are strongly dependent on the trapping efficiency of photoformed electrons and a consequence can be the direct interaction of h + with surface OH groups or via the trapping by N species, in the case of N-containing materials (steps 5 and 6). The After light absorption with energy lower (the lines in the spectrum of the used lamp with a cut-off filter shown in Figure S1 are 364, 404 and 434 nm corresponding to 3.4, 3.1 and 2.9 eV, respectively) than the bandgap energy of titania (step 1), the transition of an electron from the valence band to the sublevel of Pt or from the N sublevel to the conduction band can occur. It appears that Pt ions play the role of trapping centers for photogenerated electrons, with a following reduction in oxygen molecules to superoxide radicals, and Pt oxidation to its previous oxidation state (steps 2 and 4) [25]. As shown by the experimental results, the higher content of Pt 0 has a detrimental effect on photocatalytic processes, suggesting their light screening from the TiO 2 surface or preferential charge recombination, as proposed in [27]. The oxidation pathways are strongly dependent on the trapping efficiency of photoformed electrons and a consequence can be the direct interaction of h + with surface OH groups or via the trapping by N species, in the case of N-containing materials (steps 5 and 6). The resulting OH radical oxidizes the TC molecule (step 7). It is most probable that the nitrogen species are able to prevent the charge recombination by trapping a hole, which is confirmed by the high photoreactivity of the nitrogen-doped films of Method 2. Thus, the design of materials with controlled oxidation states of Pt and N in TiO 2 is a prospective strategy to drive the development of solar photocatalysts.

Materials and Methods
The platinum and nitrogen co-doped titania films were synthesized by sol-gel techniques using two different procedures. To obtain porous films, the sols containing the alcoholic (iso-propanol, Reider-de-Haën) solution of titanium tetraisopropoxide (TTIP, Merk, Darmstadt, Germany), three-block copolymer of polyethyleneoxide and polypropyleneoxide (PEO) 20 (PPO) 70 (PEO) 20 (Pluronic 123, Sigma-Aldrich, Schnelldorf, Germany) and acetylacetone (acac, Sigma-Aldrich) were used. Platinum (II) acetylacetonate (Pt(acac) 2 , ACROS, 98%) (0.5 or 1.0 mol.%) dissolved in acetone and urea (5% mol., Sigma-Aldrich) dissolved in ethanol (SC Chimreactiv SRL) were added to the sol as the doping agents. The molar ratio of the main components in the sols was established as Ti(Oi-Pr) 4  . After hydrolysis (1h) of the last third layer, the thermal treatment of the films was performed at 450 • C for 20 min at the heating rate of 7 • C/min in the presence of air. The porous and non-porous pure TiO 2 and N-doped TiO 2 films were obtained in the same manner as Pt-doped ones without adding a solution of Pt(acac) 2 in acetone. No UV pretreatment was performed for these films.
Absorption spectra of the films were recorded with a double beam spectrophotometer (Lambda 35, PerkinElmer) within the wavelength range of 190-1200 nm. The bandgap energy values of doped TiO 2 films were determined according to the modified Tauc method reported in [56]. XRD patterns of the films scratched off from the glass were performed on DRON-4-07 using Cu Кα irradiation (λ = 1.5418 Å). The anatase crystallite size was calculated using the Scherrer equation and was about 14 nm for all Pt-doped films. The surface morphology of the films was monitored by scanning electron microscopy (SEM) using an FEI Inspect S Scanning Electron Microscope at an acceleration voltage of 20kV, in a high vacuum. The samples were coated with a thin Au film. XPS measurements were carried out with an ESCALAB Xi + (Thermo SCIENTIFIC Surface Analysis, Thermo Fisher Scientific, UK) setup equipped with a multichannel hemispherical electron Analyzer (dual X-ray source) working with Al Kα radiation (hν = 1486.2 eV), using C 1s (284.8 eV) as the energy reference. XPS data were recorded on samples that had been outgassed in the prechamber of the instrument at room temperature at a pressure of <2 × 10 −8 Pa to remove chemisorbed water from their surfaces. The surface chemical composition and oxidation state were estimated from XPS spectra by calculating the integral of each peak after a subtraction of the "S-shaped" Shirley-type background. The spectra were fitted with CasaXPS software (Casa Software Ltd, Teignmouth, United Kingdom) using the appropriate experimental sensitivity factors. Structural features of films have been studied with high resolution transmission electron microscopy (TEM) imaging on an analytical JEOL JEM-2800 (JEOL Ltd., Tokyo, Japan) with an accelerating voltage of 200 kV. Samples were allocated on copper grids with lacey carbon film by drop-casting, using ethanol as a transferring agent, after having been scratched from a glass substrate samples using a scalpel. The mean particle size was calculated from HR-TEM images using 30-50 particles. Electron diffraction patterns were acquired by selected area diffraction, and elemental maps were recorded with a large angle high-speed energy dispersive spectrometer (EDS) with an energy resolution of 133.0 eV.
Photocatalytic activity of the films was tested via the degradation of 2 × 10 −5 mol/L TC aqueous solution. The change in TC concentration was monitored with a Lambda 35 UV-vis spectrophotometer (PerkinElmer, Norwalk, USA) every 20 min; the conversion percentage was calculated from the change in the absorption intensity at λ = 357 nm. Prior to photocatalytic reactions, the adsorption equilibrium of the film-liquid system was reached by stirring in the dark. A system (a quartz reactor filled with 40 mL of 2 × 10 −5 mol/L TC and a film under constant stirring and reaction temperature of 25 • C) was irradiated with a 1000 W middle-pressure mercury lamp with the corresponding cut-off filter ( Figure  S1c, spectrum 2) for 90 min. The distance lamp reactor was set at 90 cm. Two blank experiments, the catalytic reaction (under dark conditions in the presence of a film) and the photolysis of TC (without photocatalyst), showed no significant changes in TC absorption spectra. The most active films, 1Pt-TiO 2 p and 1Pt,N-TiO 2 n, have been thricely tested for the repeatability of their photocatalytic performance resulting in ±5% variation of TC conversion.