Preparation of Visible-Light Active Oxygen-Rich TiO2 Coatings Using Low Pressure Cold Spraying

Visible-light active photocatalysts in the form of coatings that can be produced using large-scale methods have attracted considerable attention. Here we show a facile approach to deposit coatings using the low pressure cold spray (LPCS) from oxygen-rich amorphous titanium dioxide, which is a structurally-unconventional feedstock powder for LPCS. We synthesized amorphous TiO2, in which we introduced numerous defects, such as oxide groups (peroxy and superoxy) in volume and hydroxyl groups on the surface. Then we deposited as-prepared powder preserving the presence of active groups, which we demonstrated using Raman spectroscopy. To show the activity of the prepared coatings, we perform methylene blue degradation under visible light. Our research shows that it is worth considering the internal atomic structure and surface chemistry of the powders to be preserved after low pressure cold spraying.


Introduction
The formation of TiO 2 photocatalytic coatings has been a constant challenge over the past 30 years [1][2][3]. During this time, titanium dioxide has remained one of the most photocatalytically efficient and highly stable materials, and stayed at a low price. TiO 2 coatings for photocatalytic systems may be produced through several low-temperature preparation techniques: by the transport of the precursor to the substrate using gas or plasma (chemical vapor deposition (CVD) [4], pulsed laser deposition (PLD) [5], atomic layer deposition (ALD) [6], ion sputtering [7]); deposition of liquid phase containing precursors (dip-coating [8], spin coating [9], spray pyrolysis [10]); or deposition of precursors in the solid-state (using screen-printing, doctor blade method [11] or thermal spraying [12]). Due to increasing environmental pollution, greener and more energy-efficient deposition methods that are large-scale and inexpensive are being actively sought. This urge does not change the main requirement of photocatalytic coatings, which is, namely, to retain or create a specific surface area as high as possible [13].
In response to those demands (apart from environmental safety and economic efficiency) the cold-spray method offers preservation of the nanoparticulate structure of the initial feedstock [14][15][16]. Several publications have already been published on various photocatalytic cold-sprayed coatings [12,[15][16][17]. The coatings were cold sprayed most often using the high pressure variant (HPCS), >2 MPa [12], which have an environmental cost that consists of the consumption of gas and electrical power as well as the replacing of parts due to the relatively high rate of nozzle erosive wear [18]. Deposition at lower pressures

Feedstock Powder and Coatings Characterization
The crystal structure of feedstock powder and low pressure cold sprayed coatings was investigated using the X-ray diffractometer Ultima IV (Rigaku, Tokyo, Japan), with CuKα irradiation (λ = 1.54056 Å) for the 2θ ranging from 5 • to 75 • in the 0.05 steps 3 s per each measurement point.
The evaluation of the surface and sections of the samples were conducted using the SEM microscope (Hitachi S-3400 N, Tokyo, Japan). For the cross-sections of coatings, the samples were cut in the middle of their lengths, embedded in the resin and consecutively polished without etching. The topography of the feedstock powder and coatings after spraying was investigated without any additional preparation.
The particle size analysis of the feedstock powder was carried out by laser diffraction using PSA-1190 (Anton Paar GmbH, Graz, Austria). After the initial measurement, the powder was subjected to ultrasound treatment for 30 min to observe the changes in the size of the initial agglomerates.
The surface roughness (Ra, Rz) of the coatings was measured using a profilometer (Form Talysurf 120 L, Taylor-Hobson, Leicester, United Kingdom). The diamond stylus with a radius of 5 µm was transversed at the contact mode with a measuring force of 0.75 mN and a measuring speed of 1 mm/s along tracing length L t = 15 mm and with a cut-off filter λ C = 2.5 mm.
The specific surface of the samples was measured using VHX-6000 digital microscope (Keyence, Osaka, Japan). The magnification was set to 200, which provided the surface area of 1.733 mm × 1.299 mm (2.233 mm 2 ). Three measurements were carried out for all deposited coatings and the mean value was determined for 200 and 600 samples.
The mass of the coatings was established by weighting the initial substrate before and after coating deposition.
Diffuse Reflectance Spectroscopy (DRS) was applied to determine the bandgap value (E g ) of the TiO 2 feedstock powder and coatings using a UV-VIS spectrophotometer equipped with a 75 mm integrating sphere (Specord 210, Analytik Jena, Jena, Germany). The DRS spectra of the sample were measured in the range of 200-800 nm with a Spectralon ® as the reference material. The Kubelka-Munk function (Equation (1)) was used to convert the obtained reflectance (R) into the absorption coefficient (F(R)), and the Tauc's plot ([F(R)hυ] 0.5 vs. hυ) was drawn to determine the bandgap energy (E g ): The Raman spectra of samples were collected using the Raman spectrophotometer LabRam HR800 (Horiba/Jobin-Yvon, Kyoto, Japan) upon Ar+ laser excitation at 514.55 nm with 50 mW laser power within the spectral range from 50 to 4000 cm −1 .
The photocatalytic degradation of methylene blue (MB) was carried out in the TOPT-V reactor equipped with eight quartz vessels with magnetic stirrers, and a low-temperature cooling circulating pump (Toption Instruments Co., Xi'an, China). The TiO 2 covered plates, immersed in MB solution (100 mL, 1·10 −5 M ≈ 3.2 mg/L), were exposed to VIS irradiation emitted from a xenon lamp (300 W) with a UV cut-off filter. The measured light intensity on the surface of TiO 2 was 0.31 mW/cm 2 . To allow the equilibrium adsorption of MB on TiO 2 , the process was initially conducted in the dark for 1 h. During the experiments, aliquots of the solution were collected every hour and the concentration of methylene blue was Coatings 2022, 12, 475 5 of 19 measured using a UV-VIS spectrometer (Specord 210 Plus, Analytik Jena, Jena, Germany) at the wavelength of 668 nm (the detection limit was 0.25 mg/L with RSD ≤ 6%). Due to an insignificant pH decrease, no chemicals were used to maintain pH at a constant level. All studies (MB photocatalytic degradation as well as adsorption in the dark) were conducted in the photoreactor equipped with a circulation cooling system to ensure constant temperatures (21.5 • C) of all solutions in individual experiments during the tests.

Results and Discussion
There are a number of factors that influence the efficiency of the TiO 2 heterogeneous photocatalysis process: i.e., the degree and type of long-range ordering of atoms in the material, morphological properties that mainly determine the active surface of the catalyst, as well as the defects of its structure and surface. The latter factor seems especially important for oxygen-rich TiO 2 , which has been intentionally defected to achieve an additional activity in visible light. Therefore, here we conduct research on the structural characteristics of the manufactured materials (powders and coatings), taking into account the morphology, the degree of crystal order, and the analysis of chemical groups modifying titanium dioxide (e.g., O-O and hydroxyl groups). The high photocatalytic efficiency of yellow powders is well reported in the literature [35][36][37][38][39][40]42], and therefore the main objective of the research is to preserve it after spraying to produce coatings. Here we discuss the feedstock powder characteristic first and then compare the data with the records for coatings sprayed using carrier gas at two different temperatures: 200 and 600 • C.

Crystal Structure Analysed Using X-ray Diffraction (XRD)
At first, we studied the phase composition of oxygen-rich TiO 2 in the form of powder and coatings. In the diffraction pattern of the feedstock powder ( Figure 1 For the 200 coating ( Figure 1, green plot), the diffraction of aluminum is stronger than in the case of the 600 sample. For example, the peak at 38.45 • is approximately 100 times more intense than the background level (inset in Figure 1), which is a result of penetration of the aluminum substrate by X-rays during the diffraction measurement caused by the low thickness of the 200 coating. Apart from the aluminum in the diffractogram of the 200 sample, an intense peak centered at about 25 • can be found, which may seem identical to anatase (detected in the red plot), however, clear differences exist between them: aluminum oxide peak is sharper and more symmetrical than building up anatase and the additional peak in the green plot is shifted toward the right to appear at 2θ equal 25.55 • (ICSD-9770, marked as Al 2 O 3 ). An additional confirmation of the presence of aluminum oxide is that at higher angles all subsequently observed peaks match the peaks characteristic of the alumina phase. Al 2 O 3 can occur in the sample as the residue of substrate grit-blasting or as the passivation product. Here, the residual Al 2 O 3 is more probable, as XRD studies of the substrate after grit-blasting confirm its presence. Interface contamination with blasting medium is a very popular side effect of grit blasting and can be omitted by using crystalline titanium oxide instead of Al 2 O 3 [43]. That said, there are reports on the deposition of TiO 2 coating on Al 2 O 3 substrates which increased their photocatalytic activity [44]. If passivation occurs while depositing a partially amorphous coating, it may provide better adhesion of the coating to the substrate, via overgrowing, through the residual porosity of the coating [24]. However, the detection of both Al and Al 2 O 3 may be evidence of the poor deposition efficiency in 200 sample. due of substrate grit-blasting or as the passivation product. Here, the residual Al2O3 is more probable, as XRD studies of the substrate after grit-blasting confirm its presence. Interface contamination with blasting medium is a very popular side effect of grit blasting and can be omitted by using crystalline titanium oxide instead of Al2O3 [43]. That said, there are reports on the deposition of TiO2 coating on Al2O3 substrates which increased their photocatalytic activity [44]. If passivation occurs while depositing a partially amorphous coating, it may provide better adhesion of the coating to the substrate, via overgrowing, through the residual porosity of the coating [24]. However, the detection of both Al and Al2O3 may be evidence of the poor deposition efficiency in 200 sample. The conducted X-ray diffraction tests showed that spraying the feedstock powder with cold gas has little effect on the long-range arrangement of the material. Even at a higher carrier gas temperature, the feedstock building coating crystallized only partially. At lower temperature, the deposited feedstock remained amorphous. We know from our previous work [23,24], that the amorphous form facilitated the deposition process and initiated the crystallization (observed here for the 600 sample), however, to determine the outcome for yellow TiO2, further structural characterization is needed.

Morphology and Microstructure Analysed Using Scanning Electron Microscopy (SEM) Supported via Particle Size Analysis and Roughness Measurements
The results of XRD diffraction imposes, especially in 200 sample, a significant reflection that originated from the substrate material, encouraging the observation of the The conducted X-ray diffraction tests showed that spraying the feedstock powder with cold gas has little effect on the long-range arrangement of the material. Even at a higher carrier gas temperature, the feedstock building coating crystallized only partially. At lower temperature, the deposited feedstock remained amorphous. We know from our previous work [23,24], that the amorphous form facilitated the deposition process and initiated the crystallization (observed here for the 600 sample), however, to determine the outcome for yellow TiO 2 , further structural characterization is needed.

Morphology and Microstructure Analysed Using Scanning Electron Microscopy (SEM) Supported via Particle Size Analysis and Roughness Measurements
The results of XRD diffraction imposes, especially in 200 sample, a significant reflection that originated from the substrate material, encouraging the observation of the topography and cross-section of the samples. Again, we begin our observation by investigating the morphology of the powder. SEM micrographs ( Figure 2) reveal strongly unsymmetrical agglomerates of TiO 2 . The particle size of as-synthesized powder presented in (Figure 2a,b) is in the wide range of 6.8-212.2 µm (D0.5 = 44.1 µm). Regardless of the size of the agglomerates (Figure 2c), they are porous and covered with smaller, densely packed unsymmetrical submicrometric flocculent particles. The agglomerates lacking flocculent covering seem denser, but still, they are characterized instead by the developed surface ( Figure 2d). A 30-min ultrasound treatment for laser diffraction particle size analysis caused the detachment of the flocculent covering from the agglomerates, thus decreasing the size of the particles to 1.8-36.5 µm (D0.5 = 12.9 µm). This may suggest that the flocculent covering of agglomerates is weakly bonded, and hence it may be rather easily disintegrated, while in deposition (which proceeds in more drastic conditions than ultrasonication) (as in Figure 2d). topography and cross-section of the samples. Again, we begin our observation by investigating the morphology of the powder. SEM micrographs ( Figure 2) reveal strongly unsymmetrical agglomerates of TiO2. The particle size of as-synthesized powder presented in (Figure 2a,b) is in the wide range of 6.8-212.2 µm (D0.5 = 44.1 µm). Regardless of the size of the agglomerates (Figure 2c), they are porous and covered with smaller, densely packed unsymmetrical submicrometric flocculent particles. The agglomerates lacking flocculent covering seem denser, but still, they are characterized instead by the developed surface (Figure 2d). A 30-min ultrasound treatment for laser diffraction particle size analysis caused the detachment of the flocculent covering from the agglomerates, thus decreasing the size of the particles to 1.8-36.5 µm (D0.5 = 12.9 µm). This may suggest that the flocculent covering of agglomerates is weakly bonded, and hence it may be rather easily disintegrated, while in deposition (which proceeds in more drastic conditions than ultrasonication) (as in Figure 2d). It is not only the particle interactions of a substrate that affects the deposition efficiency in cold spray; the surface roughness also has an impact. The irregular geometry introduced via grit-blasting (Ra = 8.84 µm, Rz = 51.72 µm) was developed to fit the size of the powder considered optimal for cold spray deposition. The initial parameters were modified upon the coating deposition process. The morphology of the 200 coating (Figure 3a) shows a rough surface with blunt edges of protruding irregularities. Since the roughness profile becomes more uniform after coating deposition (Ra = 5.91 µm, Rz = 37.12 µm) it would support the tendency of particles to fill the valleys of roughness and flatten only the top of the roughness peaks. However, at the magnification of 100 times, the coating is too thin to be successfully investigated (Figure 3c). On the contrary, the SEM morphology of the 600 sample (Figure 3b), shows an undulating surface. The substantial waviness is characteristic of cold sprayed coatings and results from the plastic It is not only the particle interactions of a substrate that affects the deposition efficiency in cold spray; the surface roughness also has an impact. The irregular geometry introduced via grit-blasting (Ra = 8.84 µm, Rz = 51.72 µm) was developed to fit the size of the powder considered optimal for cold spray deposition. The initial parameters were modified upon the coating deposition process. The morphology of the 200 coating ( Figure 3a) shows a rough surface with blunt edges of protruding irregularities. Since the roughness profile becomes more uniform after coating deposition (Ra = 5.91 µm, Rz = 37.12 µm) it would support the tendency of particles to fill the valleys of roughness and flatten only the top of the roughness peaks. However, at the magnification of 100 times, the coating is too thin to be successfully investigated (Figure 3c). On the contrary, the SEM morphology of the 600 sample (Figure 3b), shows an undulating surface. The substantial waviness is characteristic of cold sprayed coatings and results from the plastic deformation of the surface upon anchoring large self-inflicted agglomerates [45]. The cross-sectional images ( Figure 3d) reveal thick 25-50 µm dense coating, coarser with respect to the grit-blasted substrate (Ra = 12.18 µm, Rz = 67.49 µm). The roughness of the coatings favors the specific surface area. The measured specific surface of the 200 and 600 samples was 2.315 mm 2 and 2.456 mm 2 , respectively. Compared to the surface area measured for the flat sample before grit-blasting (2.253 mm 2 ), the 200 showed an increase in the surface of 3% and the 600 sample-by 9%. The surface area of grit-blasted substrates was 2.366 mm 2 , which means that the deposition of 200 coating flattened the surface, and the 600 coating-roughened it. Yet, at this magnification, good interlocking is observed throughout the entire cross-section. deformation of the surface upon anchoring large self-inflicted agglomerates [45]. The cross-sectional images ( Figure 3d) reveal thick 25-50 µm dense coating, coarser with respect to the grit-blasted substrate (Ra = 12.18 µm, Rz = 67.49 µm). The roughness of the coatings favors the specific surface area. The measured specific surface of the 200 and 600 samples was 2.315 mm 2 and 2.456 mm 2 , respectively. Compared to the surface area measured for the flat sample before grit-blasting (2.253 mm 2 ), the 200 showed an increase in the surface of 3% and the 600 sample-by 9%. The surface area of grit-blasted substrates was 2.366 mm 2 , which means that the deposition of 200 coating flattened the surface, and the 600 coating-roughened it. Yet, at this magnification, good interlocking is observed throughout the entire cross-section. The morphology of the outer layer of both coatings (Figure 4a,b) has a feedstock powder-like structure. The top view of sample 200 resembles a shot-peened structure ( Figure 4a). In the case of ceramic particles, the kinetic energy is transformed into the fragmentation of agglomerated particles instead of plastic deformation [45]. With this regard, probably only larger and denser agglomerates without flocculent cover reached the substrate and break apart upon impact. Generally, smaller particles, such as pieces of weekly bonded covering, accelerate more rapidly in the de Laval nozzle and achieve higher velocities, however, their kinetic energy can be lost in the bow shock region due to its lower mass [46]. The deceleration of the powder below the critical velocity contributes to the bouncing back of particles from the substrate and shock bow [26]. Hence, smaller flocculent particles are not observed in the 200 coating, which may be caused either by its bouncing back or its compaction by other incoming particles. The surface of the 600 sample seems more developed since the powder submicron structure was preserved ( Figure 4b). The increase in the working gas temperature to 600 °C resulted in a greater acceleration of feedstock powder particles (even those flocculent covering-loose or weekly connected to the agglomerates) and therefore improved the impact velocity of particles and deposition efficiency, which is all consistent with the literature [27].
In both cases (Figure 4a,b), isolated discontinuities can be found (yellow arrows), much less frequently in the 600 sample, and hence the cross-sections were prepared to evaluate their severity (Figure 4c,d). In the case of the 200 sample, it was possible to form a very thin ceramic coating of 2-3 µm thickness (Figure 4c). The bond strength of the coating was not high enough to prevent detachment of the deposited particle by elastic spring-back forces during incoming particles bombardment [25]. The incoming particles The morphology of the outer layer of both coatings (Figure 4a,b) has a feedstock powder-like structure. The top view of sample 200 resembles a shot-peened structure ( Figure 4a). In the case of ceramic particles, the kinetic energy is transformed into the fragmentation of agglomerated particles instead of plastic deformation [45]. With this regard, probably only larger and denser agglomerates without flocculent cover reached the substrate and break apart upon impact. Generally, smaller particles, such as pieces of weekly bonded covering, accelerate more rapidly in the de Laval nozzle and achieve higher velocities, however, their kinetic energy can be lost in the bow shock region due to its lower mass [46]. The deceleration of the powder below the critical velocity contributes to the bouncing back of particles from the substrate and shock bow [26]. Hence, smaller flocculent particles are not observed in the 200 coating, which may be caused either by its bouncing back or its compaction by other incoming particles. The surface of the 600 sample seems more developed since the powder submicron structure was preserved ( Figure 4b).
The increase in the working gas temperature to 600 • C resulted in a greater acceleration of feedstock powder particles (even those flocculent covering-loose or weekly connected to the agglomerates) and therefore improved the impact velocity of particles and deposition efficiency, which is all consistent with the literature [27].
In both cases (Figure 4a,b), isolated discontinuities can be found (yellow arrows), much less frequently in the 600 sample, and hence the cross-sections were prepared to evaluate their severity (Figure 4c,d). In the case of the 200 sample, it was possible to form a very thin ceramic coating of 2-3 µm thickness (Figure 4c). The bond strength of the coating was not high enough to prevent detachment of the deposited particle by elastic spring-back forces during incoming particles bombardment [25]. The incoming particles significantly densified the deposited thin ceramic coating, smoothing its surface upon the impact of the new particles. The thickness of the embedded coating was too small to absorb the excess energy, and, as a result, a net of cracks appeared in the coating material, leading to delamination of the coating. The cross-sections of the 600 sample ( Figure 4d) show that it was possible to produce a uniform dense coating of high internal porosity and a rough surface well connected to the substrate. The outer part of the coating consists of loosely connected particles that were probably ejected from the gas stream and stuck on the coating surface. The detachments appear only in the upper part of the coating (yellow arrow, Figure 4d). The coating-substrate interface remains, however, solid.
oatings 2022, 12, x FOR PEER REVIEW 9 of 20 significantly densified the deposited thin ceramic coating, smoothing its surface upon the impact of the new particles. The thickness of the embedded coating was too small to absorb the excess energy, and, as a result, a net of cracks appeared in the coating material, leading to delamination of the coating. The cross-sections of the 600 sample ( Figure 4d) show that it was possible to produce a uniform dense coating of high internal porosity and a rough surface well connected to the substrate. The outer part of the coating consists of loosely connected particles that were probably ejected from the gas stream and stuck on the coating surface. The detachments appear only in the upper part of the coating (yellow arrow, Figure 4d). The coating-substrate interface remains, however, solid. The SEM analysis showed that in lower temperatures of carrier gas, only a very thin coating was possible to be deposited. Too low a level of both thermal and kinetic energy caused insufficient bonding of the coating to the substrate. Despite low adhesion, the coating was not fully detached and maintained contact with the substrate. The 600 coating was relatively thick and porous well connected to the substrate. The open porosity was the result of the agglomerates being broken apart and, as such, they revealed the internal porosity of the feedstock. Such reorganization enabled the surface of agglomerates to be compacted upon the impact on the substrate and the interior to be revealed. On one hand, the porosity may have been the origin of the crack formation, but in the case of the photocatalytic coatings, which were not under mechanical loading, the porosity ensures a high surface area.

Optical Properties Analysed Using Diffuse Reflectance Spectroscopy (DRS)
The optical properties of the feedstock powder and the cold-sprayed coatings were studied by UV-VIS diffuse reflectance spectra (DRS) ( Figure 5). For all samples in the ul- The SEM analysis showed that in lower temperatures of carrier gas, only a very thin coating was possible to be deposited. Too low a level of both thermal and kinetic energy caused insufficient bonding of the coating to the substrate. Despite low adhesion, the coating was not fully detached and maintained contact with the substrate. The 600 coating was relatively thick and porous well connected to the substrate. The open porosity was the result of the agglomerates being broken apart and, as such, they revealed the internal porosity of the feedstock. Such reorganization enabled the surface of agglomerates to be compacted upon the impact on the substrate and the interior to be revealed. On one hand, the porosity may have been the origin of the crack formation, but in the case of the photocatalytic coatings, which were not under mechanical loading, the porosity ensures a high surface area.

Optical Properties Analysed Using Diffuse Reflectance Spectroscopy (DRS)
The optical properties of the feedstock powder and the cold-sprayed coatings were studied by UV-VIS diffuse reflectance spectra (DRS) ( Figure 5). For all samples in the ultraviolet range, almost all incident UV light is absorbed (Figure 5a). The yellow color of the feedstock powder results in higher absorption at 400-500 nm visible as a shoulder to the peak located at~320 nm (UV) [47]. The spectrally uniform absorption through the 400-800 nm range causes the yellow to fade, imparting simultaneously a grey-white finish [47]. The spectrum of the 600 sample shows a spectral response towards the visible region from 500 to 700 nm. The Tauc plots were used to estimate the bandgaps (Figure 5b). The Tauc plot for the feedstock powder reveals the presence of two optical bandgaps: one at 2.54 eV and another one at 2.24 eV. The records indicate that sample 200 has a lower bandgap (2.83 eV) than sample 600 (3.09 eV).
Coatings 2022, 12, x FOR PEER REVIEW 10 of 20 the peak located at ~320 nm (UV) [47]. The spectrally uniform absorption through the 400-800 nm range causes the yellow to fade, imparting simultaneously a grey-white finish [47]. The spectrum of the 600 sample shows a spectral response towards the visible region from 500 to 700 nm. The Tauc plots were used to estimate the bandgaps ( Figure  5b). The Tauc plot for the feedstock powder reveals the presence of two optical bandgaps: one at 2.54 eV and another one at 2.24 eV. The records indicate that sample 200 has a lower bandgap (2.83 eV) than sample 600 (3.09 eV). The edge of adsorption is generally dependent on the crystal structure and the number of active sites (such as defects or dopants) [35]. In the literature, it is shown that the slight decrease in the bandgap could be, for instance, the result of oxygen vacancies and titanium ions present in the sample [48]. In other research, it is stated that the more defected state would lower the bandgap even more [49]. Here, the broadening of the bandgap of coatings with regard to feedstock powder must be directly related to the structural reorganization induced by spraying. The diffraction measurements (Figure 1) showed only a slight tendency to increase the degree of ordering of the titanium dioxide powder as a result of the spraying, while the optical measurements suggest that, in addition, there are also more subtle changes in the chemical structure. Therefore, it is necessary to investigate the structure and chemical composition of materials at the molecular level.

Vibrational Characterization Using Raman Spectroscopy
To gain additional information about the existence of the -O-O-coordination bonds in feedstock powder and coatings, the Raman scattering of all samples were measured ( Figure 6). The summary of Raman analysis is displayed for convenience in Table 1. The edge of adsorption is generally dependent on the crystal structure and the number of active sites (such as defects or dopants) [35]. In the literature, it is shown that the slight decrease in the bandgap could be, for instance, the result of oxygen vacancies and titanium ions present in the sample [48]. In other research, it is stated that the more defected state would lower the bandgap even more [49]. Here, the broadening of the bandgap of coatings with regard to feedstock powder must be directly related to the structural reorganization induced by spraying. The diffraction measurements (Figure 1) showed only a slight tendency to increase the degree of ordering of the titanium dioxide powder as a result of the spraying, while the optical measurements suggest that, in addition, there are also more subtle changes in the chemical structure. Therefore, it is necessary to investigate the structure and chemical composition of materials at the molecular level.

Vibrational Characterization Using Raman Spectroscopy
To gain additional information about the existence of the -O-O-coordination bonds in feedstock powder and coatings, the Raman scattering of all samples were measured ( Figure 6). The summary of Raman analysis is displayed for convenience in Table 1. Due to omitting the drying step at elevated temperatures, we produced feedstock powder in the amorphous form ( Figure 6, black plot). The amorphous structure of the powder has already been proven using XRD measurements (Figure 1), however the Raman spectroscopy is sensitive even to substantial changes in short-range interactions, which allows one to identify individual chemical bonds even in disordered amorphous materials. The low-intensity wide band found at about 400 cm −1 was the only common band for the obtained spectra and anatase phase (B1g), yet due to the fact that no stronger anatase vibrational mode (especially Eg mode at 147 cm −1 [50]) was observed in the spectrum, the formation of anatase in the feedstock powder was excluded. Furthermore, no characteristic bands for anatase could be found in sample 200 ( Figure 6, green plot). The anatase structure (148, 194, 400, 525, 638 cm −1 [50,51]), as identified by XRD ( Figure 1), was detected only in a sample sprayed with carrier gas at 600 °C ( Figure 6, red plot). If there is no evidence to consider crystal forms of TiO2 in feedstock powder and given that in the Raman spectra of samples 200 and 600 not every band is explained so far, one may think that other modes originate from the interaction of Ti atoms with different forms of oxygen formed during synthesis [35].
The most intense band of feedstock powder spectra is located at 531 cm −1 . The mode observed in the range of 524-529 cm −1 can be connected with the stretching vibration of Ti-O2 2− , in which O2 2− is bound to a single Ti 4+ center in a side-on bonding configuration [52], which is called a triangular peroxy titanyl group in some works [53]. The presence of a band at this frequency is considered to be evidence of obtaining an oxygen-rich titanium dioxide powder containing O2 2− species incorporated during synthesis. When analyzing the spectrum of sample 200, a decrease in the intensity of this band is noticeable, which may indicate a partial loss of these active oxygen groups in the coating. In the 600 sample, it is not possible to detect the presence of a Ti-O2 2− band, as even if it exists, it overlaps with a relatively intense A1g anatase band at a frequency of 525 cm −1 .
In search of further evidence of significant modifications of the structure of titanium dioxide, it is worth looking at the band that appears in the case of feedstock powder at Due to omitting the drying step at elevated temperatures, we produced feedstock powder in the amorphous form ( Figure 6, black plot). The amorphous structure of the powder has already been proven using XRD measurements (Figure 1), however the Raman spectroscopy is sensitive even to substantial changes in short-range interactions, which allows one to identify individual chemical bonds even in disordered amorphous materials. The low-intensity wide band found at about 400 cm −1 was the only common band for the obtained spectra and anatase phase (B 1g ), yet due to the fact that no stronger anatase vibrational mode (especially E g mode at 147 cm −1 [50]) was observed in the spectrum, the formation of anatase in the feedstock powder was excluded. Furthermore, no characteristic bands for anatase could be found in sample 200 ( Figure 6, green plot). The anatase structure (148, 194, 400, 525, 638 cm −1 [50,51]), as identified by XRD ( Figure 1), was detected only in a sample sprayed with carrier gas at 600 • C ( Figure 6, red plot). If there is no evidence to consider crystal forms of TiO 2 in feedstock powder and given that in the Raman spectra of samples 200 and 600 not every band is explained so far, one may think that other modes originate from the interaction of Ti atoms with different forms of oxygen formed during synthesis [35].
The most intense band of feedstock powder spectra is located at 531 cm −1 . The mode observed in the range of 524-529 cm −1 can be connected with the stretching vibration of Ti-O 2 2− , in which O 2 2− is bound to a single Ti 4+ center in a side-on bonding configuration [52], which is called a triangular peroxy titanyl group in some works [53]. The presence of a band at this frequency is considered to be evidence of obtaining an oxygen-rich titanium dioxide powder containing O 2 2− species incorporated during synthesis. When analyzing the spectrum of sample 200, a decrease in the intensity of this band is noticeable, which may indicate a partial loss of these active oxygen groups in the coating. In the 600 sample, it is not possible to detect the presence of a Ti-O 2 2− band, as even if it exists, it overlaps with a relatively intense A 1g anatase band at a frequency of 525 cm −1 .
In search of further evidence of significant modifications of the structure of titanium dioxide, it is worth looking at the band that appears in the case of feedstock powder at 686 cm −1 . This band is ascribed to stretching vibrations of Ti-O-O groups [54] or involving two-fold oxygen [55]  . The band at about 700 cm −1 is still visible in the coating sprayed at 600 • C, although it is not as intense as the bands of anatase.
The Raman spectra of the measured samples contain two more bands in the frequency ranges, which in the literature are ascribed to the vibration of the triangular peroxy titanyl group, in which O-O vibrations [39,52,57] or Ti-O vibrations [53,54] occur. The O-O stretching in the feedstock powder is observed at 913 cm −1 , and in coatings, it blueshifts and loses intensity more for coating sprayed at higher temperatures. The intensity of the Ti-O vibrations in triangular titanyl groups (at~1050 cm −1 ) [53] is also reduced due to spraying. Therefore, all three bands assigned in the powder spectrum to the triangular titanyl groups (531, 913 and 1054 cm −1 ) after spraying show a reduced intensity, indicating the thermal instability of the triangular peroxide species, which is consistent with the literature reports [53]. The disintegration of unstable peroxide groups is accompanied by the appearance of superoxide groups in the coatings, which is evidenced by an increase in the intensity of the 700 cm −1 band in the spectra of the 200 coating. The second characteristic superoxide band [58] can be found at 1145 cm −1 of the 200 spectrum. The superoxides are present also in the spectrum of the feedstock powder (~1152 cm −1 ), which shows that H 2 O 2 treatment forms not only peroxide groups. The typical band for superoxide groups (1145, 1152 cm −1 ) is not observed in the 600 coating, however, it is notable that only in the case of the coating sprayed with gas at a temperature of 600 • C can a very characteristic, sharp and relatively intense band at 1552 cm −1 , attributed to molecular oxygen (O 2 ) vibrations, be identified [59]. This observation implies that the active and fairly labile superoxide species present in the powder decomposes with the evolution of oxygen during the spraying of the 600 sample.
Since a correlation was found between the presence and number of hydroxyl groups on the TiO 2 surface and the photocatalytic potential (by helping to form radicals) [41], it is worth examining the measured Raman spectra in this regard. In all samples, broad highfrequency bands can be distinguished and usually assigned to OH groups, however, their intensity is the highest in the case of feedstock powder (green and red plots are y-stretched ×4 and ×15 times, respectively). The irregularity of the peaks may indicate several bands that overlap in that energy shift. The Raman spectra of all samples are characterized by the bands of OH group stretching and bending at~3160-3450 cm −1 (O-H stretching and bending vibrations, [59]) and deformation vibrations at~1620-1640 cm −1 (chemisorbed and/or physisorbed H-O-H [60]), which is evidence of a large amount of water molecules, adsorbed on the surface of titanium dioxide both in feedstock powder and in coatings. Another band associated with hydroxyl groups can be found at 283 cm −1 . It is easy to see that the width and the asymmetry of this band both increase after the spraying. In the literature on oxygen-rich TiO 2 , vibrations in this frequency range are attributed to either Ti-O-H [40,41] or the intrinsic host lattice defects-oxygen vacancies [39,52], which are typical for oxides. Since spraying causes the appearance of two bands in the spectrum of the 600 sample, it can be assumed that the coatings contain both oxygen defects and Ti-OH groups. Additional evidence of oxygen vacancies (in ordered structure) or oxygen deficiency (in disordered structure) can be found at~450 cm −1 [55,61]. in the spectrum of the 600 sample, it can be assumed that the coatings contain both oxygen defects and Ti-OH groups. Additional evidence of oxygen vacancies (in ordered structure) or oxygen deficiency (in disordered structure) can be found at ~450 cm −1 [55,61]. in the spectrum of the 600 sample, it can be assumed that the coatings contain both oxygen defects and Ti-OH groups. Additional evidence of oxygen vacancies (in ordered structure) or oxygen deficiency (in disordered structure) can be found at ~450 cm −1 [55,61]. in the spectrum of the 600 sample, it can be assumed that the coatings contain both oxygen defects and Ti-OH groups. Additional evidence of oxygen vacancies (in ordered structure) or oxygen deficiency (in disordered structure) can be found at ~450 cm −1 [55,61]. in the spectrum of the 600 sample, it can be assumed that the coatings contain both oxygen defects and Ti-OH groups. Additional evidence of oxygen vacancies (in ordered structure) or oxygen deficiency (in disordered structure) can be found at ~450 cm −1 [55,61]. in the spectrum of the 600 sample, it can be assumed that the coatings contain both oxygen defects and Ti-OH groups. Additional evidence of oxygen vacancies (in ordered structure) or oxygen deficiency (in disordered structure) can be found at ~450 cm −1 [55,61]. in the spectrum of the 600 sample, it can be assumed that the coatings contain both oxygen defects and Ti-OH groups. Additional evidence of oxygen vacancies (in ordered structure) or oxygen deficiency (in disordered structure) can be found at ~450 cm −1 [55,61]. in the spectrum of the 600 sample, it can be assumed that the coatings contain both oxygen defects and Ti-OH groups. Additional evidence of oxygen vacancies (in ordered structure) or oxygen deficiency (in disordered structure) can be found at ~450 cm −1 [55,61]. in the spectrum of the 600 sample, it can be assumed that the coatings contain both oxygen defects and Ti-OH groups. Additional evidence of oxygen vacancies (in ordered structure) or oxygen deficiency (in disordered structure) can be found at ~450 cm −1 [55,61]. in the spectrum of the 600 sample, it can be assumed that the coatings contain both oxygen defects and Ti-OH groups. Additional evidence of oxygen vacancies (in ordered structure) or oxygen deficiency (in disordered structure) can be found at ~450 cm −1 [55,61]. in the spectrum of the 600 sample, it can be assumed that the coatings contain both oxygen defects and Ti-OH groups. Additional evidence of oxygen vacancies (in ordered structure) or oxygen deficiency (in disordered structure) can be found at ~450 cm −1 [55,61]. in the spectrum of the 600 sample, it can be assumed that the coatings contain both oxygen defects and Ti-OH groups. Additional evidence of oxygen vacancies (in ordered structure) or oxygen deficiency (in disordered structure) can be found at ~450 cm −1 [55,61]. in the spectrum of the 600 sample, it can be assumed that the coatings contain both oxygen defects and Ti-OH groups. Additional evidence of oxygen vacancies (in ordered structure) or oxygen deficiency (in disordered structure) can be found at ~450 cm −1 [55,61]. in the spectrum of the 600 sample, it can be assumed that the coatings contain both oxygen defects and Ti-OH groups. Additional evidence of oxygen vacancies (in ordered structure) or oxygen deficiency (in disordered structure) can be found at ~450 cm −1 [55,61]. in the spectrum of the 600 sample, it can be assumed that the coatings contain both oxygen defects and Ti-OH groups. Additional evidence of oxygen vacancies (in ordered structure) or oxygen deficiency (in disordered structure) can be found at ~450 cm −1 [55,61]. in the spectrum of the 600 sample, it can be assumed that the coatings contain both oxygen defects and Ti-OH groups. Additional evidence of oxygen vacancies (in ordered structure) or oxygen deficiency (in disordered structure) can be found at ~450 cm −1 [55,61]. in the spectrum of the 600 sample, it can be assumed that the coatings contain both oxygen defects and Ti-OH groups. Additional evidence of oxygen vacancies (in ordered structure) or oxygen deficiency (in disordered structure) can be found at ~450 cm −1 [55,61]. in the spectrum of the 600 sample, it can be assumed that the coatings contain both oxygen defects and Ti-OH groups. Additional evidence of oxygen vacancies (in ordered structure) or oxygen deficiency (in disordered structure) can be found at ~450 cm −1 [55,61]. Considering the potential use of oxygen-rich TiO2 coatings in photocatalytic reactors, the Raman measurements showed that the cold spraying process changes the chemical structure of the deposited powder. There is a visible decrease in the number of triangular peroxy titanyl groups and an increase in superoxide species could be crucial regarding the photocatalytic activity of coatings. The measured optical bandgap energy of the feedstock powder increases after the coating is deposited at 200 °C and experiences an even greater increase when the spraying temperature was 600 °C. In the literature [48,49], the differences in the bandgap width of the oxygen-rich and pure TiO2 without additions are explained as a result of the introduction of oxygen defects into the structure of titanium dioxide, creating additional energy levels above the valence band. The Raman measurements showed that oxygen defects in feedstock powder and coatings have various forms (e.g., of superoxide groups, peroxide groups) and their number and type change depending on the parameters of coating deposition. These changes undoubtedly affect the electronic structure of TiO2. However, the increase in the TiO2 bandgap itself in no way determines the changes in photocatalytic activity. That is due to the loss of the peroxy groups in the coatings being considered to be responsible for the activity of the photocatalyst in visible light, which is compensated by the formation of superoxide groups known as oxidizing agents and radical initiators. Unpaired electrons make superoxides highly reactive, allowing them to oxidize various organic pollutants [65]. Many studies have shown that the hydroxyl groups on the metal oxide are responsible for trapping photogenerated charge carriers, resulting in a reduced rate of the recombination of electron-hole pairs [41]. Thus, the optimistic note is that spraying does not appear to significantly alter the level of hydroxylation of the titanium dioxide surface.

Visible-Light Photocatalytic Activity via Photobleaching of Methylene Blue
The analysis of the Raman spectra revealed that the coatings had a significant number of active oxygen species and hydroxyl groups, which suggests that the coatings should exhibit some photocatalytic activity in visible light. Although the bandgap is not directly connected to photocatalytic performance, it may be useful to select the activation light [41]. The bandgap of 200 sample was 2.8 eV, making the coating a promising candidate. Even if the 600 sample was characterized by a higher bandgap, with its crystal structure (mixed amorphous-anatase phase) it is a good candidate as well. It is important to emphasize that the masses of 200 and 600 coatings were considerably different, which is a result of the various coating thicknesses (Figures 3 and 4). The 600 coatings weighed 35 ± 9 mg and all 200 coatings-were less than 1 mg. We summarize the results in Table 2 Considering the potential use of oxygen-rich TiO2 coatings in photocatalytic reactors, the Raman measurements showed that the cold spraying process changes the chemical structure of the deposited powder. There is a visible decrease in the number of triangular peroxy titanyl groups and an increase in superoxide species could be crucial regarding the photocatalytic activity of coatings. The measured optical bandgap energy of the feedstock powder increases after the coating is deposited at 200 °C and experiences an even greater increase when the spraying temperature was 600 °C. In the literature [48,49], the differences in the bandgap width of the oxygen-rich and pure TiO2 without additions are explained as a result of the introduction of oxygen defects into the structure of titanium dioxide, creating additional energy levels above the valence band. The Raman measurements showed that oxygen defects in feedstock powder and coatings have various forms (e.g., of superoxide groups, peroxide groups) and their number and type change depending on the parameters of coating deposition. These changes undoubtedly affect the electronic structure of TiO2. However, the increase in the TiO2 bandgap itself in no way determines the changes in photocatalytic activity. That is due to the loss of the peroxy groups in the coatings being considered to be responsible for the activity of the photocatalyst in visible light, which is compensated by the formation of superoxide groups known as oxidizing agents and radical initiators. Unpaired electrons make superoxides highly reactive, allowing them to oxidize various organic pollutants [65]. Many studies have shown that the hydroxyl groups on the metal oxide are responsible for trapping photogenerated charge carriers, resulting in a reduced rate of the recombination of electron-hole pairs [41]. Thus, the optimistic note is that spraying does not appear to significantly alter the level of hydroxylation of the titanium dioxide surface.

Visible-Light Photocatalytic Activity via Photobleaching of Methylene Blue
The analysis of the Raman spectra revealed that the coatings had a significant number of active oxygen species and hydroxyl groups, which suggests that the coatings should exhibit some photocatalytic activity in visible light. Although the bandgap is not directly connected to photocatalytic performance, it may be useful to select the activation light [41]. The bandgap of 200 sample was 2.8 eV, making the coating a promising candidate. Even if the 600 sample was characterized by a higher bandgap, with its crystal structure (mixed amorphous-anatase phase) it is a good candidate as well. It is important to emphasize that the masses of 200 and 600 coatings were considerably different, which is a result of the various coating thicknesses (Figures 3 and 4). The 600 coatings weighed 35 ± 9 mg and all 200 coatings-were less than 1 mg. We summarize the results in Table 2 Considering the potential use of oxygen-rich TiO2 coatings in photocatalytic reactors, the Raman measurements showed that the cold spraying process changes the chemical structure of the deposited powder. There is a visible decrease in the number of triangular peroxy titanyl groups and an increase in superoxide species could be crucial regarding the photocatalytic activity of coatings. The measured optical bandgap energy of the feedstock powder increases after the coating is deposited at 200 °C and experiences an even greater increase when the spraying temperature was 600 °C. In the literature [48,49], the differences in the bandgap width of the oxygen-rich and pure TiO2 without additions are explained as a result of the introduction of oxygen defects into the structure of titanium dioxide, creating additional energy levels above the valence band. The Raman measurements showed that oxygen defects in feedstock powder and coatings have various forms (e.g., of superoxide groups, peroxide groups) and their number and type change depending on the parameters of coating deposition. These changes undoubtedly affect the electronic structure of TiO2. However, the increase in the TiO2 bandgap itself in no way determines the changes in photocatalytic activity. That is due to the loss of the peroxy groups in the coatings being considered to be responsible for the activity of the photocatalyst in visible light, which is compensated by the formation of superoxide groups known as oxidizing agents and radical initiators. Unpaired electrons make superoxides highly reactive, allowing them to oxidize various organic pollutants [65]. Many studies have shown that the hydroxyl groups on the metal oxide are responsible for trapping photogenerated charge carriers, resulting in a reduced rate of the recombination of electron-hole pairs [41]. Thus, the optimistic note is that spraying does not appear to significantly alter the level of hydroxylation of the titanium dioxide surface.

Visible-Light Photocatalytic Activity via Photobleaching of Methylene Blue
The analysis of the Raman spectra revealed that the coatings had a significant number of active oxygen species and hydroxyl groups, which suggests that the coatings should exhibit some photocatalytic activity in visible light. Although the bandgap is not directly connected to photocatalytic performance, it may be useful to select the activation light [41]. The bandgap of 200 sample was 2.8 eV, making the coating a promising candidate. Even if the 600 sample was characterized by a higher bandgap, with its crystal structure (mixed amorphous-anatase phase) it is a good candidate as well. It is important to emphasize that the masses of 200 and 600 coatings were considerably different, which is a result of the various coating thicknesses (Figures 3 and 4). The 600 coatings weighed 35 ± 9 mg and all 200 coatings-were less than 1 mg. We summarize the results in Table 2 Considering the potential use of oxygen-rich TiO2 coatings in photocatalytic reactors, the Raman measurements showed that the cold spraying process changes the chemical structure of the deposited powder. There is a visible decrease in the number of triangular peroxy titanyl groups and an increase in superoxide species could be crucial regarding the photocatalytic activity of coatings. The measured optical bandgap energy of the feedstock powder increases after the coating is deposited at 200 °C and experiences an even greater increase when the spraying temperature was 600 °C. In the literature [48,49], the differences in the bandgap width of the oxygen-rich and pure TiO2 without additions are explained as a result of the introduction of oxygen defects into the structure of titanium dioxide, creating additional energy levels above the valence band. The Raman measurements showed that oxygen defects in feedstock powder and coatings have various forms (e.g., of superoxide groups, peroxide groups) and their number and type change depending on the parameters of coating deposition. These changes undoubtedly affect the electronic structure of TiO2. However, the increase in the TiO2 bandgap itself in no way determines the changes in photocatalytic activity. That is due to the loss of the peroxy groups in the coatings being considered to be responsible for the activity of the photocatalyst in visible light, which is compensated by the formation of superoxide groups known as oxidizing agents and radical initiators. Unpaired electrons make superoxides highly reactive, allowing them to oxidize various organic pollutants [65]. Many studies have shown that the hydroxyl groups on the metal oxide are responsible for trapping photogenerated charge carriers, resulting in a reduced rate of the recombination of electron-hole pairs [41]. Thus, the optimistic note is that spraying does not appear to significantly alter the level of hydroxylation of the titanium dioxide surface.

Visible-Light Photocatalytic Activity via Photobleaching of Methylene Blue
The analysis of the Raman spectra revealed that the coatings had a significant number of active oxygen species and hydroxyl groups, which suggests that the coatings should exhibit some photocatalytic activity in visible light. Although the bandgap is not directly connected to photocatalytic performance, it may be useful to select the activation light [41]. The bandgap of 200 sample was 2.8 eV, making the coating a promising candidate. Even if the 600 sample was characterized by a higher bandgap, with its crystal structure (mixed amorphous-anatase phase) it is a good candidate as well. It is important to emphasize that the masses of 200 and 600 coatings were considerably different, which is a result of the various coating thicknesses (Figures 3 and 4). The 600 coatings weighed 35 ± 9 mg and all 200 coatings-were less than 1 mg. We summarize the results in Table 2 Considering the potential use of oxygen-rich TiO2 coatings in photocatalytic reactors, the Raman measurements showed that the cold spraying process changes the chemical structure of the deposited powder. There is a visible decrease in the number of triangular peroxy titanyl groups and an increase in superoxide species could be crucial regarding the photocatalytic activity of coatings. The measured optical bandgap energy of the feedstock powder increases after the coating is deposited at 200 °C and experiences an even greater increase when the spraying temperature was 600 °C. In the literature [48,49], the differences in the bandgap width of the oxygen-rich and pure TiO2 without additions are explained as a result of the introduction of oxygen defects into the structure of titanium dioxide, creating additional energy levels above the valence band. The Raman measurements showed that oxygen defects in feedstock powder and coatings have various forms (e.g., of superoxide groups, peroxide groups) and their number and type change depending on the parameters of coating deposition. These changes undoubtedly affect the electronic structure of TiO2. However, the increase in the TiO2 bandgap itself in no way determines the changes in photocatalytic activity. That is due to the loss of the peroxy groups in the coatings being considered to be responsible for the activity of the photocatalyst in visible light, which is compensated by the formation of superoxide groups known as oxidizing agents and radical initiators. Unpaired electrons make superoxides highly reactive, allowing them to oxidize various organic pollutants [65]. Many studies have shown that the hydroxyl groups on the metal oxide are responsible for trapping photogenerated charge carriers, resulting in a reduced rate of the recombination of electron-hole pairs [41]. Thus, the optimistic note is that spraying does not appear to significantly alter the level of hydroxylation of the titanium dioxide surface.

Visible-Light Photocatalytic Activity via Photobleaching of Methylene Blue
The analysis of the Raman spectra revealed that the coatings had a significant number of active oxygen species and hydroxyl groups, which suggests that the coatings should exhibit some photocatalytic activity in visible light. Although the bandgap is not directly connected to photocatalytic performance, it may be useful to select the activation light [41]. The bandgap of 200 sample was 2.8 eV, making the coating a promising candidate. Even if the 600 sample was characterized by a higher bandgap, with its crystal structure (mixed amorphous-anatase phase) it is a good candidate as well. It is important to emphasize that the masses of 200 and 600 coatings were considerably different, which is a result of the various coating thicknesses (Figures 3 and 4). The 600 coatings weighed 35 ± 9 mg and all 200 coatings-were less than 1 mg. We summarize the results in Table 2 Considering the potential use of oxygen-rich TiO 2 coatings in photocatalytic reactors, the Raman measurements showed that the cold spraying process changes the chemical structure of the deposited powder. There is a visible decrease in the number of triangular peroxy titanyl groups and an increase in superoxide species could be crucial regarding the photocatalytic activity of coatings. The measured optical bandgap energy of the feedstock powder increases after the coating is deposited at 200 • C and experiences an even greater increase when the spraying temperature was 600 • C. In the literature [48,49], the differences in the bandgap width of the oxygen-rich and pure TiO 2 without additions are explained as a result of the introduction of oxygen defects into the structure of titanium dioxide, creating additional energy levels above the valence band. The Raman measurements showed that oxygen defects in feedstock powder and coatings have various forms (e.g., of superoxide groups, peroxide groups) and their number and type change depending on the parameters of coating deposition. These changes undoubtedly affect the electronic structure of TiO 2 . However, the increase in the TiO 2 bandgap itself in no way determines the changes in photocatalytic activity. That is due to the loss of the peroxy groups in the coatings being considered to be responsible for the activity of the photocatalyst in visible light, which is compensated by the formation of superoxide groups known as oxidizing agents and radical initiators. Unpaired electrons make superoxides highly reactive, allowing them to oxidize various organic pollutants [65]. Many studies have shown that the hydroxyl groups on the metal oxide are responsible for trapping photogenerated charge carriers, resulting in a reduced rate of the recombination of electron-hole pairs [41]. Thus, the optimistic note is that spraying does not appear to significantly alter the level of hydroxylation of the titanium dioxide surface.

Visible-Light Photocatalytic Activity via Photobleaching of Methylene Blue
The analysis of the Raman spectra revealed that the coatings had a significant number of active oxygen species and hydroxyl groups, which suggests that the coatings should exhibit some photocatalytic activity in visible light. Although the bandgap is not directly connected to photocatalytic performance, it may be useful to select the activation light [41]. The bandgap of 200 sample was 2.8 eV, making the coating a promising candidate. Even if the 600 sample was characterized by a higher bandgap, with its crystal structure (mixed amorphous-anatase phase) it is a good candidate as well. It is important to emphasize that the masses of 200 and 600 coatings were considerably different, which is a result of the various coating thicknesses (Figures 3 and 4). The 600 coatings weighed 35 ± 9 mg and all 200 coatings-were less than 1 mg. We summarize the results in Table 2. With this information, we performed the methylene blue (MB) degradation experiments under VIS irradiation to investigate the photocatalytic performance of 200 and 600 samples. ** Calculated as the difference between the mass of the coated sample and sample before spraying. *** The differences were lower than 1 mg.
The UV-VIS spectra of methylene blue before and after visible light irradiation ( The difference in the MB degradation efficiency and adsorption can be assigned to a substantially lower mass of the 200 sample (<1 mg, thickness 2-3 µm) than the mass of sample 600 (~34 mg, thickness 25-50 µm) and a different spatial structure. As the interfacial surface is primarily responsible for the effectiveness of all surface processes, the limited deposition of TiO 2 in the 200 sample resulted in a lower specific surface area, and finally in much lower photocatalytic activity. Due to the high porosity of both samples, the higher thickness of the 600 sample was more beneficial in terms of providing a higher specific surface area. Moreover, referring to the efficiency of the studied photocatalytic process, one can note that it was conducted under very low irradiance (0.31 mW/cm 2 ; λ > 460 nm), and for a small catalyst area (400 mm 2 ) immersed in 100 mL of the MB solution. Generally, with an increase in the intensity of radiation, an increase in the number of photons reaching the surface of the photocatalyst can be observed, which in turn causes the increase in the number of decomposed MB molecules. The enhancement of the photocatalytic performance of oxygen-rich TiO 2 can be the surface composition and size distribution rather than the lower bandgaps [41]. Nevertheless, in the presented results, both coatings, which remained mechanically stable after the test, demonstrated the photodegradation of methylene blue. In the case of the 200 sample, it is important to provide an adequate surface specific area, which is also possible using the cold spray method and together with a lower bandgap could impart higher photocatalytic activity. The obtained results reveal the high photocatalytic activity of the 600 sample obtained by the proposed cold spray method and make further research relevant and promising. The difference in the MB degradation efficiency and adsorption can be assigned to a substantially lower mass of the 200 sample (<1 mg, thickness 2-3 µm) than the mass of sample 600 (~34 mg, thickness 25-50 µm) and a different spatial structure. As the interfacial surface is primarily responsible for the effectiveness of all surface processes, the limited deposition of TiO2 in the 200 sample resulted in a lower specific surface area, and finally in much lower photocatalytic activity. Due to the high porosity of both samples, the higher thickness of the 600 sample was more beneficial in terms of providing a higher specific surface area. Moreover, referring to the efficiency of the studied photocatalytic process, one can note that it was conducted under very low irradiance (0.31 mW/cm 2 ; λ > 460 nm), and for a small catalyst area (400 mm 2 ) immersed in 100 mL of the MB solution. Generally, with an increase in the intensity of radiation, an increase in the number of photons reaching the surface of the photocatalyst can be observed, which in turn causes the increase in the number of decomposed MB molecules. The enhancement of the photocatalytic performance of oxygen-rich TiO2 can be the surface composition and size distribution rather than the lower bandgaps [41]. Nevertheless, in the presented results, both coatings, which remained mechanically stable after the test, demonstrated the photodegradation of methylene blue. In the case of the 200 sample, it is important to provide an adequate surface specific area, which is also possible using the cold spray method and together with a lower bandgap could impart higher photocatalytic activity. The obtained results reveal the high photocatalytic activity of the 600 sample obtained by the proposed cold spray method and make further research relevant and promising.

Conclusions
In this study, we low pressure cold sprayed amorphous oxygen-rich titanium dioxide to produce the coatings exhibiting photocatalytic activity in visible light. We analyzed changes in the feedstock powder before and after spraying to understand how the selected parameters influence the efficiency of the heterogeneous TiO 2 photocatalysis process. Using carrier gas at 600 • C, we deposited a thick (25-50 µm), porous coating with a highly developed surface. Spraying with gas preheated to 200 • C led to the formation of a relatively thinner (2-3 µm) coating with a net of discontinuities. We showed that both coatings were effective in the degradation of methylene blue upon the visible light irradiation, and the morphology of both has not changed after the photocatalytic efficiency test. Hence, the chosen powder immobilization method (LPCS) selected as a result of the efficiency of deposition of amorphous materials turned out to be safe for the subtle, yet substantial, physicochemical structure of the oxygen-rich feedstock powder. Consequently, low pressure cold spraying not only represents low-temperature large-scale technology; it also allows for the preservation of the feedstock powder absorption edge extended to visible light.