Physicochemical Properties of Ti 3+ Self-Doped TiO 2 Loaded on Recycled Fly-Ash Based Zeolites for Degradation of Methyl Orange

: The extensive production of coal ﬂy ash by coal combustion is an issue of concern due to its environmental impact. TiO 2 -zeolite composites were synthesized, at low cost, using recycled coal ﬂy ash from a local thermoelectric power plant to produce the zeolite using the hydrothermal method. TiO 2 was loaded by means of the impregnation method using ethanol and titanium isopropoxide between 8.7 and 49.45 wt% TiO 2 . The samples were characterized by X-ray diffraction, Raman, electron spin resonance, high-resolution transmission electron microscopy, N 2 adsorption-desorption, doppler broadening of annihilation radiation, and diffuse reﬂectance techniques, and the photo-catalytic activity of the composites was evaluated according to the degradation of methyl orange under UV light. The results show that TiO 2 crystallizes in the anatase phase with a Ti 3+ oxidation state, without post-treatment. TiO 2 particles were located within the pores of the substrate and on its surface, increasing the surface area of the composites in comparison with that of the substrates. Samples with TiO 2 at 8.7 and 25 wt% immobilized on hydroxysodalite show the highest degradation of methyl orange among all studied materials, including the commercial TiO 2 Degussa P25 under UV light.


Introduction
Pollution by organic compounds originating from industrial waste waters results in a major threat to human health and aquatic life. To solve this issue, the metal oxide-based semiconductor photocatalysts are considered as a solution due to their capacity to promote oxidation-reduction processes under electromagnetic radiation [1,2]. Among the photocatalysts, the TiO 2 is one of the most extensively studied materials due to its physicochemical properties, low cost, abundance, low toxicity, strong oxidation ability, ease of preparation, and stability [3]. To enhance the photocatalytic properties of TiO 2 via light irradiation, procedures, including doping with metal and nonmetal ions and the insertion of intrinsic defects, such as Ti 3+ , to introduce states within the TiO 2 band gap, have been developed Figure 1a depicts the SEM micrograph of CFA; it shows the morphology in which the major particles are spherical in shape, corresponding to cenospheres. Minor particles form agglomerates that correspond to carbonaceous particles. Figure 1b shows the Xray diffraction pattern of CFA, which was indexed with quartz, mullite, maghemite, and hematite as major crystalline phases, with PDF cards: 01-078-2315, 01-083-1881, 00-039-1346, and 00-024-0072, respectively. The broad XRD signal at 2θ < 35 • indicates the presence of amorphous content. The chemical composition of CFA, determined by X-ray fluorescence, shows that the CFA contains mainly SiO 2   fraction pattern of CFA, which was indexed with quartz, mullite, maghemite, and hematite as major crystalline phases, with PDF cards: 01-078-2315, 01-083-1881, 00-039-1346, and 00-024-0072, respectively. The broad XRD signal at 2θ < 35° indicates the presence of amorphous content. The chemical composition of CFA, determined by X-ray fluorescence, shows that the CFA contains mainly SiO2 (62.57 (0.45)%), Al2O3 (24.62 Figure 2 shows the XRD patterns for TiO2, and the H and C substrates. All diffraction peaks for TiO2 were indexed with a I 41/amd space group of anatase (ICSD 24276). The XRD data of the substrates was analyzed by Rietveld refinement, and the identified crystalline phases are listed in Table 1.   Figure 2 shows the XRD patterns for TiO 2 , and the H and C substrates. All diffraction peaks for TiO 2 were indexed with a I 41/amd space group of anatase (ICSD 24276). The XRD data of the substrates was analyzed by Rietveld refinement, and the identified crystalline phases are listed in Table 1. Hydroxysodalite and zeolite P1 correspond to zeolite materials and are the major phases identified in H substrate. Cancrinite is a zeolite material and is the major phase identified in C substrate. According to the DRX results, it can be concluded that the alkaline activation of CFA by NaOH and posterior heat treatment allows for the obtaining of zeolite materials. Figure 3a,b shows the XRD patterns for the HT and CT composites. In both HT and CT composites, the Bragg reflections of the anatase TiO 2 phase and the Bragg reflections of the major phases of the substrates can be distinguished. Hydroxysodalite and zeolite P1 correspond to zeolite materials and are the major phases identified in H substrate. Cancrinite is a zeolite material and is the major phase identified in C substrate. According to the DRX results, it can be concluded that the alkaline activation of CFA by NaOH and posterior heat treatment allows for the obtaining of zeolite materials. Figure 3a,b shows the XRD patterns for the HT and CT composites. In both HT and CT composites, the Bragg reflections of the anatase TiO2 phase and the Bragg reflections of the major phases of the substrates can be distinguished. For HT composites, the intensity ratio I (101)TiO2 /I (211)H of the major reflection of anatase (101), located at around of 25.4 • , and hydroxysodalite (211), located at around of 24.8 • , was calculated; correspondingly, for CT composites, the intensity ratio I (101)TiO2 /I (101)C of the major reflection of anatase (101), located at around of 25.4 • , and cancrinite (101), located at around of 19 • , was calculated ( Figure 4a). Those ratios tend to increase when the amount of TiO 2 increases in both the CT and HT composites. These results suggest that the incorporation of TiO 2 as anatase was successful on both the H and C substrates. For TiO 2 anatase, in both CT and HT systems, the a and c lattice parameters were calculated from the (200) and (101) reflections, respectively, using the equations:

Properties of TiO2-Zeolite Composites
1/d 2 hkl = (h 2 + k 2 )/a 2 + l 2 /c 2 (2) where λ is the wavelength of the incident X-rays, d hkl is the interplanar distance for a tetragonal system, h, k and l are the Miller indices, and a and c are the lattice parameters.  For HT composites, the intensity ratio I(101)TiO2/I(211)H of the major reflection of anatase (101), located at around of 25.4°, and hydroxysodalite (211), located at around of 24.8°, was calculated; correspondingly, for CT composites, the intensity ratio I(101)TiO2/I(101)C of the major reflection of anatase (101), located at around of 25.4°, and cancrinite (101), located at around of 19°, was calculated ( Figure 4a). Those ratios tend to increase when the amount of TiO2 increases in both the CT and HT composites. These results suggest that the incorporation of TiO2 as anatase was successful on both the H and C substrates. For TiO2 anatase, in both CT and HT systems, the a and c lattice parameters were calculated from the (200) and (101) reflections, respectively, using the equations: 1/d 2 hkl = (h 2 + k 2 )/a 2 + l 2 /c 2 (2) where λ is the wavelength of the incident X-rays, dhkl is the interplanar distance for a tetragonal system, h, k and l are the Miller indices, and a and c are the lattice parameters.
(a) (b)  For HT composites, the intensity ratio I(101)TiO2/I(211)H of the major reflection of anatase (101), located at around of 25.4°, and hydroxysodalite (211), located at around of 24.8°, was calculated; correspondingly, for CT composites, the intensity ratio I(101)TiO2/I(101)C of the major reflection of anatase (101), located at around of 25.4°, and cancrinite (101), located at around of 19°, was calculated ( Figure 4a). Those ratios tend to increase when the amount of TiO2 increases in both the CT and HT composites. These results suggest that the incorporation of TiO2 as anatase was successful on both the H and C substrates. For TiO2 anatase, in both CT and HT systems, the a and c lattice parameters were calculated from the (200) and (101) reflections, respectively, using the equations: 1/d 2 hkl = (h 2 + k 2 )/a 2 + l 2 /c 2 (2) where λ is the wavelength of the incident X-rays, dhkl is the interplanar distance for a tetragonal system, h, k and l are the Miller indices, and a and c are the lattice parameters.     4b-d depicts the lattice parameters and cell volume of TiO 2 for the HT and CT composites. The results show that lattice parameter a remains constant in both the CT and HT composites. The lattice parameter c and cell volume V exhibit a slight variation for 8.7, 25, and 49.45 wt% TiO 2 in comparison with the other samples. The crystallite size for immobilized TiO 2 in all the samples was calculated using the Williamson-Hall method, which considers the contribution of crystallite size (β D ) and microstrain (β S ) to the broadening of a peak, (β hkl ) as [30]: where β hkl is the full width at half maximum, K is a constant, λ is the wavelength of the X-ray, ε is the microstrain, and D is the crystallite size; the latter can be calculated from the analysis of the plot of β hkl cosθ vs. 4sinθ. The crystallite size values are listed in Table 2; these values are larger in the HT composites than in the CT composites, indicating that the substrate H favors the growth of TiO 2 . The structure of the anatase in all samples was verified by Raman measurements. Figure 5a,b shows the Raman spectra for the HT and CT composites, respectively; for comparison, the Raman spectra of synthesized TiO 2 is shown. All samples depict the characteristic five peaks of the D 4h 19 I 41/amd space group of the anatase around 144, 197, 399, 519, and 638 cm −1 , corresponding to Eg(1), Eg(2), B1g(1), an overlap of B1g(2) with the A1g and Eg(3) optical modes, respectively [31]. The majority of the samples exhibit a longrange crystalline order, except for CT-8.7. For both the HT and CT composites, a smooth variation of the Raman shift and the linewidth of the most intense Raman peak around 143 cm −1 , in comparison with the values of synthesized TiO 2 , are depicted (Figure 5c,d). These values are similar in both CT and HT for each of the values of the immobilized TiO 2 , except for 25 wt%. According to the literature, the change in the peak position and linewidth are related to defects in the stoichiometry, the quantum size effects, the presence of secondary phases, and the influence of the substrate [32][33][34]. In this work, the results suggest that the change in the peak position and linewidth are mainly due to TiO 2 quantity.
These values are similar in both CT and HT for each of the values of the immobilized TiO2, except for 25 wt.%. According to the literature, the change in the peak position and linewidth are related to defects in the stoichiometry, the quantum size effects, the presence of secondary phases, and the influence of the substrate [32][33][34]. In this work, the results suggest that the change in the peak position and linewidth are mainly due to TiO2 quantity.   Figure 6a,b show the ESR spectra at room temperature between 500 G and 550 G for the substrates, composites, and synthesized TiO 2 . The spectra for H and C substrates show a characteristic resonance for a g value of 2.124 and 2.135, respectively; these are related to the presence of Cu 2+ in the substrate matrices [35]. For synthesized TiO 2 , a resonance at g = 2.000 is observed; this is related to the Ti 3+ electronic configuration due to oxygen vacancies [36][37][38][39][40]. For both the HT and CT composites, the contributions to the ESR spectra from the substrate and TiO 2 related to the Ti 3+ state are observed (Figure 6c,d). Table 2 lists the g value of the resonance, ∆H that corresponds to the distance between peaks (line width), and the relative number of spins calculated by integrating the ESR signal dI/dH twice and comparing this result with a standard sample of 0.11% of KCl with 366 × 10 13 spin/cm. cancies [36][37][38][39][40]. For both the HT and CT composites, the contributions to the ESR spectra from the substrate and TiO2 related to the Ti 3+ state are observed (Figure 6c,d). Table 2 lists the g value of the resonance, ΔH that corresponds to the distance between peaks (line width), and the relative number of spins calculated by integrating the ESR signal dI/dH twice and comparing this result with a standard sample of 0.11% of KCl with 366 × 10 13 spin/cm. According to the relative number of spins, the Ti 3+ ESR signal is higher in HT composites than in CT composites; thus, the H substrate favor the formation of Ti 3+ states in the TiO2 rather than C substrate. The ESR signal tends to increase when the TiO2 amount grows in both the HT and CT composites. The synthesis of TiO2 with oxygen vacancies is one of the most important research topics in the chemistry of metal oxides; therefore, a wide variety of methods have been developed to produce self-modified TiO2 materials [41]. Thus, it is worth noting that the samples produced in this work under the method According to the relative number of spins, the Ti 3+ ESR signal is higher in HT composites than in CT composites; thus, the H substrate favor the formation of Ti 3+ states in the TiO 2 rather than C substrate. The ESR signal tends to increase when the TiO 2 amount grows in both the HT and CT composites. The synthesis of TiO 2 with oxygen vacancies is one of the most important research topics in the chemistry of metal oxides; therefore, a wide variety of methods have been developed to produce self-modified TiO 2 materials [41]. Thus, it is worth noting that the samples produced in this work under the method described above allow the synthesis of TiO 2 with the formation of oxygen vacancies, without post-treatment. Narrow ∆H values-around 5 G reported for TiO 2 at low temperatures-are mainly attributed to Ti 3+ ions in the bulk, rather than on the surface of the TiO 2 particles [42]. In this work, the HT-8.7, HT-25, and HT-33.25 samples show a smaller ∆H value, which may indicate a larger contribution of the Ti 3+ state in the bulk than in the other samples. The presence of Ti 3+ suggests that it contributes to the observed variation in the Raman shift and the Raman linewidth, as a stoichiometry defect. samples. In CT composites, the TiO 2 particles are found in a wide range of sizes, ranging from 14 nm, as in CT-8.7, Figure 7d, to sizes larger than 33 nm. Lattice spacings of around 0.335 nm, 0.254 nm, and 0.208 nm, assigned to (101), (103), and (200), corresponding to anatase, were found, as shown in Figure 7d-f. From these images, it is observed that the interface between the TiO 2 particles and the zeolite substrates varies gradually. It is apparent that the TiO 2 particles were successfully incorporated onto the substrate, forming the TiO 2 -zeolite composites. The nonhomogeneous particle size of TiO 2 , especially those with a particle size smaller than 20 nm, makes an additional contribution to the observed variation in the Raman shift and Raman linewidth, as a quantum effect. The N2 adsorption-desorption isotherms for synthesized TiO2 and the substrates, HT composites, and CT composites are shown in Figure 8a-c, respectively. The curves for all samples exhibit a similar shape, which is characteristic of type IV, according to the IUPAC classification, associated with mesoporous materials. The hysteresis loops are similar for C, H, and CT/HT composites characteristics of type H3, associated with lamellar pore structures with slit-and wedge-shaped pores. For synthesized TiO2, the hysteresis loop is characteristic of type H2, associated with cylindrical and spherical pores. The surface area, SBET, was calculated using a BET equation from isotherms curves, the average pore diameter was calculated by the BJH method using the desorption branch, and the pore volume was calculated at a single point of saturation. Table 2 summarizes the results. The surface area of all CT composites is higher than that of the C substrate. The surface area in the HT composites is higher than that of H substrate except for HT-25 and HT-33.25. In both composite systems the larger surface area was obtained for a loading of 41.3 wt% TiO2. The average pore diameter decreases in both the CT and HT composites in comparison with that of the C and H substrates, respectively. The total pore volume decreases in HT composites in comparison with that of the H substrate, whereas the total pore volume in- The N 2 adsorption-desorption isotherms for synthesized TiO 2 and the substrates, HT composites, and CT composites are shown in Figure 8a-c, respectively. The curves for all samples exhibit a similar shape, which is characteristic of type IV, according to the IUPAC classification, associated with mesoporous materials. The hysteresis loops are similar for C, H, and CT/HT composites characteristics of type H3, associated with lamellar pore structures with slit-and wedge-shaped pores. For synthesized TiO 2 , the hysteresis loop is characteristic of type H2, associated with cylindrical and spherical pores. The surface area, S BET , was calculated using a BET equation from isotherms curves, the average pore diameter was calculated by the BJH method using the desorption branch, and the pore volume was calculated at a single point of saturation. Table 2 summarizes the results. The surface area of all CT composites is higher than that of the C substrate. The surface area in the HT composites is higher than that of H substrate except for HT-25 and HT-33.25. In both composite systems the larger surface area was obtained for a loading of 41.3 wt% TiO 2 . The average pore diameter decreases in both the CT and HT composites in comparison with that of the C and H substrates, respectively. The total pore volume decreases in HT composites in comparison with that of the H substrate, whereas the total pore volume increases in the CT composites in comparison with that of C substrate, except for in the CT-49.45 composite. There is a reduction in the average pore diameter in both the CT and HT composites as compared to that of the C and T substrates, respectively, probably due to pore blocking caused by TiO 2 .
classification, associated with mesoporous materials. The hysteresis loops are similar for C, H, and CT/HT composites characteristics of type H3, associated with lamellar pore structures with slit-and wedge-shaped pores. For synthesized TiO2, the hysteresis loop is characteristic of type H2, associated with cylindrical and spherical pores. The surface area, SBET, was calculated using a BET equation from isotherms curves, the average pore diameter was calculated by the BJH method using the desorption branch, and the pore volume was calculated at a single point of saturation. Table 2 summarizes the results. The surface area of all CT composites is higher than that of the C substrate. The surface area in the HT composites is higher than that of H substrate except for HT-25 and HT-33.25. In both composite systems the larger surface area was obtained for a loading of 41.3 wt% TiO2. The average pore diameter decreases in both the CT and HT composites in comparison with that of the C and H substrates, respectively. The total pore volume decreases in HT composites in comparison with that of the H substrate, whereas the total pore volume increases in the CT composites in comparison with that of C substrate, except for in the CT-49.45 composite. There is a reduction in the average pore diameter in both the CT and HT composites as compared to that of the C and T substrates, respectively, probably due to pore blocking caused by TiO2.  The average of the S parameter between 1 and 5 keV is list parameter is higher in the substrates than in the composites and synt higher in the H than in the C substrate. In these samples, the S param interaction of positrons with pores. According to synthesized TiO2, th attributed to the oxygen vacancies detected by ESR spectroscopy. For the S parameters in synthesized TiO2 is lower than in other HT compo   Table 3. The S parameter is higher in the substrates than in the composites and synthesized TiO 2 , and is higher in the H than in the C substrate. In these samples, the S parameter is related to the interaction of positrons with pores. According to synthesized TiO 2 , the value of S can be attributed to the oxygen vacancies detected by ESR spectroscopy. For the HT composites, the S parameters in synthesized TiO 2 is lower than in other HT composites; for these samples, the S value lies between that of zeolite and synthesized TiO 2 . This result suggests that any quantity of TiO 2 deposits into the pores of the substrate and does not fill its cavities. Regarding the CT composites, the value of S for the composites is close to that of synthesized TiO 2 . These results suggest that a larger number of pores are filled by TiO 2 in the H substrate in comparison with the C substrate, and a larger amount of TiO 2 is immobilized on the surface rather than in the pores in the C substrate in comparison with that in the H substrate. This explains the behavior observed concerning the total pore volume in the HT and CT composites.  Figures 10 and 11 show the S-W plot for synthesized TiO2, substrates and both composite systems, respectively. The slope of the S-W plots is useful to observe the change in the nature of the positron trapping defect. Theoretically, S depends linearly on W for each kind of defect. The lowest W value indicates the response from the surface; the higher value of W indicates the response of the bulk. For the HT and CT composites, it can be considered that all points are represented for the same slope that will indicate one kind of defect (mainly pores), since there is no variation in the positron annihilation fraction in the trapping sites in each of the samples; as a function of incident positron energy, the dispersion is attributed to a variation of the pore sizes, which is not homogeneous. Slope values are summarized in Table 3. It is noted that all samples exhibit a small change in the slope, and there is a trend to decreases the slope when TiO2 increases in the HT composites, whereas there is not a clear trend in the CT composites; this indicates that a variation exists in the positron annihilation fraction as a function of the loading of TiO2 particles, which depends on the substrate; thus, it shows the changing size of the trapping sites, and thus, a change in the pores of the samples.   Figures 10 and 11 show the S-W plot for synthesized TiO 2 , substrates and both composite systems, respectively. The slope of the S-W plots is useful to observe the change in the nature of the positron trapping defect. Theoretically, S depends linearly on W for each kind of defect. The lowest W value indicates the response from the surface; the higher value of W indicates the response of the bulk. For the HT and CT composites, it can be considered that all points are represented for the same slope that will indicate one kind of defect (mainly pores), since there is no variation in the positron annihilation fraction in the trapping sites in each of the samples; as a function of incident positron energy, the dispersion is attributed to a variation of the pore sizes, which is not homogeneous. Slope values are summarized in Table 3. It is noted that all samples exhibit a small change in the slope, and there is a trend to decreases the slope when TiO 2 increases in the HT composites, whereas there is not a clear trend in the CT composites; this indicates that a variation exists in the positron annihilation fraction as a function of the loading of TiO 2 particles, which depends on the substrate; thus, it shows the changing size of the trapping sites, and thus, a change in the pores of the samples. Figures 10 and 11 show the S-W plot for synthesized TiO2, substrates and both composite systems, respectively. The slope of the S-W plots is useful to observe the change in the nature of the positron trapping defect. Theoretically, S depends linearly on W for each kind of defect. The lowest W value indicates the response from the surface; the higher value of W indicates the response of the bulk. For the HT and CT composites, it can be considered that all points are represented for the same slope that will indicate one kind of defect (mainly pores), since there is no variation in the positron annihilation fraction in the trapping sites in each of the samples; as a function of incident positron energy, the dispersion is attributed to a variation of the pore sizes, which is not homogeneous. Slope values are summarized in Table 3. It is noted that all samples exhibit a small change in the slope, and there is a trend to decreases the slope when TiO2 increases in the HT composites, whereas there is not a clear trend in the CT composites; this indicates that a variation exists in the positron annihilation fraction as a function of the loading of TiO2 particles, which depends on the substrate; thus, it shows the changing size of the trapping sites, and thus, a change in the pores of the samples.   Figure 12 displays the UV-Vis DRS spectra for the HT and CT composites; for comparison, the spectra for synthesized TiO2 and the substrates are shown. For both HT and CT composites, the range of the reflection edge increases in comparison with that of synthesized TiO2. The results suggest that the reflection edge observed in the composites is due to TiO2 and is strongly affected by the interaction between TiO2 and the substrate. The band gap energy (Eg) for the synthesized TiO2 and the composites was calculated by a linear fit of the modified Kubelka-Munk equation as a function of the energy ((F(R)*hν) 1/2 vs. hν) and the allowed indirect transition, and the Eg values are listed in Table 2. As can be seen from the table, the calculated Eg for synthesized TiO2 is 2.99 eV, which is lower than 3.2 eV for bulk TiO2 anatase, which may be caused by the Ti 3+ electronic configuration and the crystallite size.  Figure 12 displays the UV-Vis DRS spectra for the HT and CT composites; for comparison, the spectra for synthesized TiO 2 and the substrates are shown. For both HT and CT composites, the range of the reflection edge increases in comparison with that of synthesized TiO 2 . The results suggest that the reflection edge observed in the composites is due to TiO 2 and is strongly affected by the interaction between TiO 2 and the substrate. The band gap energy (E g ) for the synthesized TiO 2 and the composites was calculated by a linear fit of the modified Kubelka-Munk equation as a function of the energy ((F(R)*hν) 1/2 vs. hν) and the allowed indirect transition, and the E g values are listed in Table 2. As can be seen from the table, the calculated E g for synthesized TiO 2 is 2.99 eV, which is lower than 3.2 eV for bulk TiO 2 anatase, which may be caused by the Ti 3+ electronic configuration and the crystallite size.  Figure 13 shows UV-Vis absorption spectra of the degradation of MO using the samples HT-8.7, HT-25, synthesized TiO2, and commercial Degussa P25 TiO2. The spectra display two absorption bands at 270 nm and 463 nm related to the transition of the aromatic rings and the azo bonds, respectively [43], whose intensity decreases with irradiation time, indicating the degradation of MO.  Figure 13 shows UV-Vis absorption spectra of the degradation of MO using the samples HT-8.7, HT-25, synthesized TiO 2 , and commercial Degussa P25 TiO 2 . The spectra display two absorption bands at 270 nm and 463 nm related to the transition of the aromatic rings and the azo bonds, respectively [43], whose intensity decreases with irradiation time, indicating the degradation of MO. Figure 14a,b shows the variation in the relative concentration (C/C 0 ) of MO as a function of the reaction time for the HT and CT composites, respectively. The composites labeled as HT-8.7 and HT-25 presented better photocatalytic behavior than other composites and the commercial Degussa P25 TiO 2 . Thus, after 210 min of photocatalytic reaction, the HT-8.7 and HT-25 composites favored a dye degradation of 84% and 87%, respectively, compared to the 93% and 67% degradation evidenced using synthesized TiO 2 and Degussa P25 TiO 2 , respectively.
On the other hand, nearly 17% of the dye was degraded using the HT-16.85 and HT-33.25 composites; 19% and 30% using HT-41.3 and HT-49.45, respectively; 3% using CT-8.7 and CT-25; 20%, 6%, 8%, and 0% using CT-16.85, CT-41.3, CT-49.45, and CT-6-425-33.25, respectively. Desorption processes shown in any samples, such as HT-33.25 and CT-33.25, can be attributed to a low generation of˙OH and O ·− 2 radicals and/or a high electron-hole recombination rate [44].  Figure 13 shows UV-Vis absorption spectra of the degradation of MO using the samples HT-8.7, HT-25, synthesized TiO2, and commercial Degussa P25 TiO2. The spectra display two absorption bands at 270 nm and 463 nm related to the transition of the aromatic rings and the azo bonds, respectively [43], whose intensity decreases with irradiation time, indicating the degradation of MO.  Figure 14a,b shows the variation in the relative concentration (C/C0) of MO as a function of the reaction time for the HT and CT composites, respectively. The composites labeled as HT-8.7 and HT-25 presented better photocatalytic behavior than other composites and the commercial Degussa P25 TiO2. Thus, after 210 min of photocatalytic reaction, the HT-8.7 and HT-25 composites favored a dye degradation of 84% and 87%, respectively, compared to the 93% and 67% degradation evidenced using synthesized TiO2 and Degussa P25 TiO2, respectively. The total organic carbon analysis was carried out for the reactions involving the materials that evidenced the best photocatalytic behavior. The contents of TOC after 210 min of reaction were 46, 36, 15, and 64% for HT-8.7, HT-25, Degussa P25 TiO 2 , and synthesized TiO 2 , respectively. These results indicate that the mineralization of MO using these materials occurs in the following order: synthesized TiO 2 > HT-8.7 > HT-25 > commercial TiO 2 . The higher degradation of MO using synthesized TiO 2 , followed by the HT composites, can be explained in terms of the presence of Ti states that promote the charge separation process, acting as a holes trap [45]. However, the larger surface area and larger crystallinity of TiO 2 aids in the separation of the photogenerated electron and holes [4]. The degradation efficiency exhibited by the HT-8.7 and HT-25 composites can be explained in terms of a larger generation of˙OH and O ·− 2 radicals due to a larger irradiated surface area and a low electron-hole recombination rate [44]. Usually, the photodegradation of organic dyes by semiconducting oxides is described by the pseudo first order equation [43,46]: where C 0 is the concentration of the dye at adsorption equilibrium, C is the concentration of the dye at time t, and k is the apparent rate constant. The ln(C 0 /C) vs. t plot is displayed in Figure 14c for the HT-8.7, HT-25, and synthesized TiO 2 samples, in which the values of the rate constant k are 7.60 × 10 −3 , 7.50 × 10 −3 , and 7.93 × 10 −3 , respectively.
tion of the reaction time for the HT and CT composites, respectively. The composites labeled as HT-8.7 and HT-25 presented better photocatalytic behavior than other composites and the commercial Degussa P25 TiO2. Thus, after 210 min of photocatalytic reaction, the HT-8.7 and HT-25 composites favored a dye degradation of 84% and 87%, respectively, compared to the 93% and 67% degradation evidenced using synthesized TiO2 and Degussa P25 TiO2, respectively. On the other hand, nearly 17% of the dye was degraded using t 33.25 composites; 19% and 30% using HT-41.3 and HT-49.45, respec 8.7 and CT-25; 20%, 6%, 8%, and 0% using CT-16.85, CT-41.3, CT-33.25, respectively. Desorption processes shown in any samples, such 33.25, can be attributed to a low generation of ˙OH and • radical tron-hole recombination rate [44].
The total organic carbon analysis was carried out for the reactio terials that evidenced the best photocatalytic behavior. The contents of reaction were 46, 36, 15, and 64% for HT-8.7, HT-25, Degussa P25 T TiO2, respectively. These results indicate that the mineralization of M rials occurs in the following order: synthesized TiO2 > HT-8.7 > HT-2 The higher degradation of MO using synthesized TiO2, followed by can be explained in terms of the presence of Ti states that promote t process, acting as a holes trap [45]. However, the larger surface area ity of TiO2 aids in the separation of the photogenerated electron and According to the results obtained in this work and in the literature [43][44][45][46][47][48][49], the photocatalytic degradation of MO can be described as follows ( Figure 15): when TiO 2 is loaded onto the substrate surface, it is illuminated by UV light, and the photogenerated electronhole pair may then migrate to the surface. The electrons on the surface reduce the adsorbed oxygen molecules to O ·− 2 radicals, and the photogenerated holes migrate toward the surface to oxidize the adsorbed water to form˙OH species. The O ·− 2 radicals, in turn, may generate HO 2˙s pecies. Due to the Ti 3+ states, this avoids the electron-hole pair recombination and increases the number of O ·− 2 species generated in the system. The O ·− 2 ,˙OH, and HO 2ṙ adicals are very reactive and decompose the MO molecules adsorbed on the surface of the composite; this process is faster than the velocity of the desorption of the MO molecules, allow for the degradation of the dye. The larger photocatalytic activity in HT-8.7 and HT-25 can be explained in terms of a larger crystallinity degree evidenced in the Raman results, a synergic effect between the substrate and the amount of TiO 2 that lies on a relatively high surface area, with a partial fill up of substrate pores evidenced by N 2 physisorption and DBAR measurements, in comparison with other composites.
Condens. Matter 2022, 7, x FOR PEER REVIEW 17 of 21 the surface to oxidize the adsorbed water to form ˙OH species. The • radicals, in turn, may generate HO2˙ species. Due to the Ti 3+ states, this avoids the electron-hole pair recombination and increases the number of • species generated in the system. The • , ˙OH, and HO2˙ radicals are very reactive and decompose the MO molecules adsorbed on the surface of the composite; this process is faster than the velocity of the desorption of the MO molecules, allow for the degradation of the dye. The larger photocatalytic activity in HT-8.7 and HT-25 can be explained in terms of a larger crystallinity degree evidenced in the Raman results, a synergic effect between the substrate and the amount of TiO2 that lies on a relatively high surface area, with a partial fill up of substrate pores evidenced by N2 physisorption and DBAR measurements, in comparison with other composites. Figure 15. Scheme of photocatalytic mechanism for MO degradation.

Materials
CFA with a size smaller than 38 μm were selected by sieving, and were collected from the Sochagota TermoPaipa IV thermoelectric power plant (Boyacá, Colombia). NaOH (PanReac AppliChem 98%) and deionized water were employed to perform the alkaline activation of CFA. Titanium isopropoxide (Alfa Aesar 97+%) and ethanol (Alfa Aesar 96%) were used to synthetize TiO2.

Properties of CFA
The chemical composition of CFA was determined by using a MagixPro PW-2440 Philips X-ray fluorescence spectrometer. The morphology of CFA was observed by using a Carl Zeiss EVO-MA10 electron microscope at the accelerating voltage of 10 kV. The crystalline phases were identified by using a PANalytical X'Pert's X-ray diffractometer with λKα-Co = 1.789007 Å.

Alkaline Activation of CFA by the Hydrothermal Method
Two different samples, with larger conversions of CFA into zeolite and under conditions of a low synthesis temperature and concentration of NaOH, as described in previous works, were selected for use as substrates for TiO2 immobilization [22,35]. Briefly, CFA was activated by the conventional hydrothermal method using a Teflon ® vessel placed within a steel reactor and covered by a thermal insulation in order to keep the temperature at a constant value, which was monitored with a K thermocouple with the following synthesis parameters:

Materials
CFA with a size smaller than 38 µm were selected by sieving, and were collected from the Sochagota TermoPaipa IV thermoelectric power plant (Boyacá, Colombia). NaOH (PanReac AppliChem 98%) and deionized water were employed to perform the alkaline activation of CFA. Titanium isopropoxide (Alfa Aesar 97+%) and ethanol (Alfa Aesar 96%) were used to synthetize TiO 2 .

Properties of CFA
The chemical composition of CFA was determined by using a MagixPro PW-2440 Philips X-ray fluorescence spectrometer. The morphology of CFA was observed by using a Carl Zeiss EVO-MA10 electron microscope at the accelerating voltage of 10 kV. The crystalline phases were identified by using a PANalytical X'Pert's X-ray diffractometer with λ Kα-Co = 1.789007 Å.

Alkaline Activation of CFA by the Hydrothermal Method
Two different samples, with larger conversions of CFA into zeolite and under conditions of a low synthesis temperature and concentration of NaOH, as described in previous works, were selected for use as substrates for TiO 2 immobilization [22,35]. Briefly, CFA was activated by the conventional hydrothermal method using a Teflon ® vessel placed within a steel reactor and covered by a thermal insulation in order to keep the temperature at a constant value, which was monitored with a K thermocouple with the following synthesis parameters: After the alkaline activation, the solution was filtered, washed with ethanol, filtered again, and dried. In addition, in this work, the as prepared samples were calcined at 500 • C. From a previous procedure, two samples were obtained with zeolite as the major crystalline phase, corresponding to hydroxysodalite, and cancrinite labeled as H and C, respectively; these samples were used as substrates to immobilize the TiO 2 particles.

Synthesis of TiO 2 -Zeolite Composites
Wetness impregnation of TiO 2 over H and C samples was carried out as follows: H and C substrates were placed in ethanol under magnetic stirring at 200 rpm, titanium isopropoxide was added drop by drop with different nominal TiO 2 loading at 60 • C, and the solution was kept under magnetic stirring for 6 h. The as-prepared composites were dried at 90 • C overnight and calcined at 425 • C for 2 h. A sample of TiO 2 was synthesized following this procedure at 425 • C for 6 h. The obtained composites were labelled according to the scheme of Table 4, in which the HT nomenclature refers to TiO 2 (T) impregnated over the substrate with main phase of hydroxysodalite (H) to obtain the composite HT and, the CT label refers to TiO 2 (T) impregnated over the substrate with the main phase of cancrinite (C) to obtain the composite CT.

Characterization of TiO 2 -Zeolite Composites
The structural characterization was performed by X-ray diffraction (XRD) using a Bruker AXS D8 focus spectrometer with λ Cu-Kα = 0.15405 nm. The Raman spectra were obtained in a Horiba Yobin-Yvon T64000 dispersive Raman spectrometer using a laser source of 532 nm and a power of 0.001 W over the sample. Electron spin resonance (ESR) spectra were obtained at room temperature in a Bruker EMX 10-2.7 Plus spectrometer. The morphological properties were observed by high resolution transmission electron microscopy (HRTEM) in a Tecnai F20 Super Twin TMP microscope of FEI; low-temperature N 2 physisorption measurements were performed using Micromeritics ASAP 2020 equipment, with Doppler broadening annihilation radiation of the positrons, and energy measurements (DBAR) were performed in the slow positron beam facility at Washington State University, USA; the 511 keV annihilation peak was recorded using an ORTEC HPGe detector. The optical properties were studied by ultraviolet-visible diffuse reflectance spectroscopy (DRS) using an UV-VIS Cary 50 VARIAN spectrophotometer equipped with an integrating sphere; BaSO 4 was used as the reference material.
The photocatalytic behavior of the materials was studied in the degradation of MO dye (C 14 H 14 N 3 NaO 3 S). For this purpose, 200 mL of a MO aqueous solution at 25 ppm and 50 mg of photocatalytic material were added in a Pyrex reactor. The suspension was maintained under magnetic stirring at 700 rpm, and it was bubbled with constant air flow. The system was irradiated by employing a low-pressure mercury lamp (254 nm, 4.4 mW) protected with a quartz tube, which was immersed in the suspension. To favor the adsorption-desorption equilibrium, the reaction system was maintained in the dark for 30 min. Then, the irradiation was initiated, and aliquots, which were filtered by a membrane filter, were extracted every 30 min. The samples were analyzed by using an Evolution 300 UV-Vis spectrophotometer. Finally, the residual total organic carbon (TOC) was estimated using a TOC/TN analyzer multi N/C 2100.

Conclusions
Alkaline activation of CFA by the hydrothermal method under hydrothermal conditions allows for the obtaining of aluminosilicates with zeolite materials as majority phases, which were then used as substrates to immobilize TiO 2 , whereby it was possible to produce TiO 2 -zeolite composites at low cost. The synthesis method permits the obtaining of TiO 2 in an anatase phase, with a similar lattice parameter and a long-range of crystalline order for most of the samples. The procedure performed to synthesize the TiO 2 -composites allows for the obtaining of anatase TiO 2 with a Ti 3+ state, indicating the presence of oxygen vacancies without subsequent treatment. The textural properties of the composites were improved in comparison with those of the substrates, since, due to TiO 2 incorporation, the surface of the substrate was changed. The results verify that TiO 2 particles were located on the surface and within the pores of the substrate. On the surface, the particles of TiO 2 crystallize in a wide range of particle sizes due to the agglomeration effect. The HT-8.7, HT-25, and synthesized TiO 2 showed a degradation of MO in an aqueous medium under UV light, which was higher than that of commercial Degussa P25 TiO 2 , due to a synergic effect between the substrate and the amount of TiO 2 .