TiO 2 Modiﬁed Geopolymers for the Photocatalytic Dye Decomposition

: This article studies the photocatalytic activity of three types of industrially produced TiO 2 powder (P25, CG100 and CG300) incorporated into a parent geopolymer matrix, and their pure counterparts, based on the decomposition of Rhodamine B dye. Rhodamine B dye is applied as a model substance because it is frequently used in the textile industry and thus may be present in the wastewater. The average particle size, speciﬁc surface area and mineralogical composition of TiO 2 powders have been determined. The geopolymer matrix works well as a supporting material for the CG100 and P25 pure types of TiO 2 powder as these input materials have better properties such as a higher average particle size, lower speciﬁc surface area, mineralogicalcomposition, etc., than the CG300 TiO 2 powder. These properties (or their combination) affect the photocatalytic activity of the resulting materials, which may thus become advanced composites with an additional purifying ability, e.g., materials that can be used for wastewater treatment or air puriﬁcation.


Introduction
The industrial and transport activities of our civilization produce adverse effect of air and water pollution. This phenomenon poses serious danger with a significant negative impact on human health. The efforts to sustain current life conditions and to protect the environment have led us to search for new materials. Life protection includes the purification of water and permanent care for clean air [1].
Over the past few decades, many researchers across the world [2][3][4][5] have contributed to a new field in environmental protection, using the photocatalytic reactions produced by TiO 2 nanoparticles. It has been discovered that TiO 2 nanoparticles are able to free electrons when illuminated by ultra-violet (UV) radiation, which decomposes organic substances into radicals [6,7]. Photoactivity is influenced by the mineralogical form of TiO 2 -both rutile and anatase forms are photoactive, but anatase exhibits higher photoactivity than rutile [8]. The photoactivity of TiO 2 nanoparticles further depends on their particle size and surface area [9][10][11]. Simultaneously, TiO 2 is able to form a highly hydrophilic surface [12]. Not only does it facilitate the cleaning of pollutants from the surface of support materials, but it also improves freeze-thaw resistance [13]. For its pollutant purification ability, TiO 2

Sample Preparation
The TiO 2 powder was homogenized with the supplied clay (Mefisto L05) and BFS in a dry state for 10 min. The prepared mixture was activated by an aqueous alkaline solution of potassium silicates, which had been prepared by mixing with water (the molar ratios of the geopolymer-matrix components: SiO 2 /Al 2 O 3 = 3.36; K 2 O/Al 2 O 3 = 0.78; H 2 O/K 2 O = 12.61). The potassium activator was selected because the resulting materials have a lower visible efflorescence and higher compressive strength than those with a sodium activator [45,46]. A homogenized mixture of Mefisto L05, TiO 2 powder and BFS was stirred with the alkaline solution for 10 min. Subsequently, the prepared geopolymer mixture was cast into molds and compacted on the vibrating table VSB 40 (Brio Hranice, Czech Republic) for 1 min. The clay/slag ratio was 1:0.33 and 5 g of the TiO 2 powder (CG100, CG300 and P25) was added to 100 g of the parent clay material Mefisto L05. The schematic illustration of photocatalysts preparation is shown in Figure A1 in Appendix A. A geopolymer matrix without the addition of TiO 2 powder was prepared as a reference sample. The above-mentioned ratios of individual components and the alkaline activator were selected on the basis of previous experiments [47,48].
For clarity, Table 1 summarizes the weights of the individual components used for the preparation of the studied materials. All mixtures were poured into molds and placed into plastic bags. On the second day, solid composites were demolded and stored in plastic bags at laboratory temperature and pressure for 28 days. Simultaneously, the mechanical mixture (MM-GP-P25) was made by mixing a milled GP reference sample (0.97 g) with 0.03 g of pure P25 TiO 2 powder in an agate mortar for 10 min.

Methods
The dried solid-geopolymer samples were pulverized by a Retsch friction mill (the particle size of 0-0.2 mm). Afterward, the samples were stored in a sealable plastic bag to prevent an increase in the moisture content. The materials prepared in this way were used in the subsequent analyses.
A non-destructive X-ray fluorescence (XRF) spectrometer (Spectro IQ, Kleve, Germany), equipped with an HOPG Barkla crystal was used to determine chemical compositions. The tested samples were prepared by the pressed-pellet method (4.0 g of the material and 0.9 g of the binding additive HWC Hoechst wax, Germany). For the evaluation, X-LabPro software was utilized.
For the determination of the loss on ignition (LOI), the samples were pre-dried and pulverized (at 105 • C up to a constant weight, the particle size of 0-0.2 mm) and subsequently heated in a muffle furnace (a temperature rise of 10 • C/min) to a temperature of 1000 • C with a dwell of 10 min.
Microscopic studies of the samples were performed using an SEM (scanning electron microscope) equipped with an Apollo EDS (energy dispersive spectrometer) silicon drift Si (Li) detector with a FET preamplifier for chemical element X-ray microanalysis. Data were acquired on EDAX TSL OIM software with ZAF corrections. Analyses were performed on polished blocks and powder obtained by milling or rubbing individual samples.
Particle-size distribution was measured by a CILAS 920L (CILAS, Orléans, France) analyzer with a range of 0.7-400 µm, using material suspension in water and detecting particle sizes via the laser-beam reflection. The mean hydrodynamic diameter of industrial TiO 2 powder, D H , was measured by a 90Plus/BI-MAS analyzer (Brookhaven Instruments Corp., Holtsville, NY, USA) using dynamic light scattering (DLS). D H values were measured for 30 s at 10 replications, and these data were statistically processed according to the ISO 13321 and 22412. Before measurement, the solids were mixed with water by a laboratory shaker for 25 min. For more details, see [49].
The powder X-ray diffraction (XRD) patterns (Cu Kα, λ = 1.5418 Å) were recorded on powdered samples using a D8 Advance diffractometer (Bruker AXS, Karlsruhe, Germany) with a Bragg-Brentano θ-θ goniometer (radius 217.5 mm) equipped with a Ni-beta filter and a LynxEye detector. The scan was performed at room temperature from 5 to 80 • (2 θ) in 0.01 • steps with a counting time of 1 s per step. The mass concentrations of the phases present were calculated by full-pattern matching using EVA software [50,51]. For each selected pattern, EVA adjusted the scaling factor and width parameters by fitting, and the results were transformed into concentrations using the I/I cor factors from the PDF-4+ database [52].
The Fourier transform infrared (FTIR) transmission spectra of the samples were measured by an FTIR Thermo Nicolet Nexus spectrophotometer using the KBr technique with a total number of 128 scans and the resolution of 4 cm −1 . A 3 mm aperture was used for these measurements. The spectra were normalized to the band with the highest intensity. The parent geopolymer material was used as the background for incorporated photocatalysts.

Photocatalytic Activity
The photocatalytic activity of the industrial TiO 2 photocatalysts incorporated into the geopolymer matrix and their pure counterparts under UV light illumination was evaluated based on dye decomposition. Before photocatalytic testing, 0.15 g of the photocatalyst was mixed with 60 ml of a solution of RhB dye and distilled water (in the concentration of 1 × 10 −5 mol·dm −3 , pH = 7.0 measured by pocked pH meter ad 110 ph) and stirred at 500 rpm for 5 min to obtain a uniform dispersion. After this, the dispersion was illuminated by a 250 nm enhanced Mercury Xenon lamp (200 W), housed in a LC-08 (Hamamatsu). The UV photons were generated in three wavelength regions (photon energy, E ph , in parentheses), of λ = 360 nm (E ph = 3.44 eV), λ = 310 nm (E ph = 4 eV) and λ = 250 nm (E ph = 4.96 eV), which cover almost 80% of the light energy output of the lamp. We also used bandpass filters with transmissions in the range of 280-400 or 320-410 nm. For more details as well as for the spectral dependence of the normalized intensity of the light source, see our previous work [53]. The samples were obtained at specific reaction times; before UV-Vis spectroscopy measurements, they were further centrifuged at 6000 rpm for 2 min in order to eliminate possible light scattering on photocatalyst particles. The photocatalytic activity of the photocatalysts was evaluated based on RhB dye decomposition. The absorbance spectra of the RhB dye were measured using a PerkinElmer Lambda 20 spectrophotometer in the spectral region of 400-750 nm, where the main broad band of the RhB dye was found with the absorption maxima at 555 nm. The photocatalytic activity (X) was calculated according to the following equation: X (%) = (A 0 − A)/A 0 × 100, where X is the dye conversion in %, A 0 is the initial dye absorbance at the absorption maxima and A is the dye absorbance at the absorption maxima at the specific reaction time of photocatalysis [54]. The error in the determination of X was estimated to be ± 1%. Furthermore, in order to exclude the possible absorption of dye by the photocatalysts themselves, identical dispersions were analyzed in the same manner under the same experimental conditions, but without illumination.
It should also be emphasized that we have studied the possible decomposition of RhB dye caused by the illumination alone in the absence of a photocatalyst under the same testing conditions.

Characterization of the Studied Materials
The chemical composition (the average of the three measurements) of the raw materials (Mefisto L05 and BFS) is shown in Table A1. The chemical composition of the clay material Mefisto L05 agrees with the data of the supplier,ČLUZ a.s., (Nové Strašecí). This material has a high content of silica and alumina (52.34 and 41.54 wt.%, respectively). The main components of BFS are calcium oxide (39.78 wt.%), silica (20.79 wt.%), alumina (7.36 wt.%) and sulfur trioxide (10.11 wt.%).
The results of particle-size distribution presented in Table A2 demonstrate that the clay material Mefisto L05 has very fine particles, with sizes below 25 µm (d 50 = 3.94 µm). In comparison, BFS has larger particles, with sizes below 45 µm (d 50 = 5.01 µm).
The results of the chemical analysis (XRF) of TiO 2 powder, presented in Table A3, show the difference between the manufacturing in Precheza a.s. and Evonik Industries AG. Based on the residual content of Cl (0.12 wt.% in P25), it was confirmed that TiO 2 P25 is produced by a TiCl 4 vapor-fed flame pyrolysis process [55]. On the other hand, CG100 and CG300 contain a residual amount of SO 3 (1.36 and 1.22 wt.% in CG100 and CG300, respectively), which proves their manufacture by a sulfate technology [56]. The total content of TiO 2 in CG100, CG300 and P25 is 93.49, 91.19 and 97.62 wt.%, respectively. The lowest LOI was observed for P25, which could be caused by the method of manufacturing.
Powder X-ray diffraction patterns of the starting TiO 2 are shown in Figure 1. Pure CG100 and CG300 contain only anatase, P25 contains approximately 87 wt.% of anatase and 13 wt.% of rutile, which agrees with the data of the producers. Crystallite sizes computed by the Scherrer equation are 18 nm, 6 nm and 20 nm for GC100, GC300 and P25, respectively, which is in good agreement with the average size of the particles provided by the producers (see Table A4). We have also characterized the TiO2 particle-size distribution that could occur in a water colloid during sample preparation by dynamic light scattering. It is clear from Figure 2 that TiO2 forms agglomerates with the average particle size in the ranges of 150-350 nm and 720-1700 nm, depending on the TiO2 powder type. These results are comparable with the literature data where the authors dispersed TiO2 in water-see, e.g., [57,58]. To determine the content and distribution of TiO2 in the incorporated samples, two different routes were followed. The first was an analysis of a cut and polished piece in four different places of the identical cut. The elemental analysis shows a comparable content of titanium in all samples treated with the different types of titanium dioxides and its homogeneous distribution within the sample (see Table 2 and also Figure A2 in Appendix A for the representative EDS spectra and SEM image and Table A6). The other route was an analysis of milled materials as an alternative (more relevant to our experiment) preanalysis preparation. To avoid the problem of the electron penetration depth being comparable to grain size, the ratio of the analyte signal (Ti) to the cumulative signal of (Al+Si) was used for seven different grains. This proved the homogeneity of Ti dispersion over the different grains (see Figure 3) and at the same time the comparable Ti content in the different incorporated samples. Furthermore, Figure 4 shows the representative SEM image for GP-P25, with the average particle size of this sample being ≈ 500-600 nm. Agglomerates of P25 TiO2 powder with the average particle size ≈ 1μm, present in a water colloid (see Figure 2), were not observed here and thus were broken during incorporation into We have also characterized the TiO 2 particle-size distribution that could occur in a water colloid during sample preparation by dynamic light scattering. It is clear from Figure 2 that TiO 2 forms agglomerates with the average particle size in the ranges of 150-350 nm and 720-1700 nm, depending on the TiO 2 powder type. These results are comparable with the literature data where the authors dispersed TiO 2 in water-see, e.g., [57,58]. We have also characterized the TiO2 particle-size distribution that could occur in a water colloid during sample preparation by dynamic light scattering. It is clear from Figure 2 that TiO2 forms agglomerates with the average particle size in the ranges of 150-350 nm and 720-1700 nm, depending on the TiO2 powder type. These results are comparable with the literature data where the authors dispersed TiO2 in water-see, e.g., [57,58]. To determine the content and distribution of TiO2 in the incorporated samples, two different routes were followed. The first was an analysis of a cut and polished piece in four different places of the identical cut. The elemental analysis shows a comparable content of titanium in all samples treated with the different types of titanium dioxides and its homogeneous distribution within the sample (see Table 2 and also Figure A2 in Appendix A for the representative EDS spectra and SEM image and Table A6). The other route was an analysis of milled materials as an alternative (more relevant to our experiment) preanalysis preparation. To avoid the problem of the electron penetration depth being comparable to grain size, the ratio of the analyte signal (Ti) to the cumulative signal of (Al+Si) was used for seven different grains. This proved the homogeneity of Ti dispersion over the different grains (see Figure 3) and at the same time the comparable Ti content in the different incorporated samples. Furthermore, Figure 4 shows the representative SEM image for GP-P25, with the average particle size of this sample being ≈ 500-600 nm. Agglomerates of P25 TiO2 powder with the average particle size ≈ 1μm, present in a water colloid (see Figure 2), were not observed here and thus were broken during incorporation into To determine the content and distribution of TiO 2 in the incorporated samples, two different routes were followed. The first was an analysis of a cut and polished piece in four different places of the identical cut. The elemental analysis shows a comparable content of titanium in all samples treated with the different types of titanium dioxides and its homogeneous distribution within the sample (see Table 2 and also Figure A2 in Appendix A for the representative EDS spectra and SEM image and Table A6). The other route was an analysis of milled materials as an alternative (more relevant to our experiment) pre-analysis preparation. To avoid the problem of the electron penetration depth being comparable to grain size, the ratio of the analyte signal (Ti) to the cumulative signal of (Al+Si) was used for seven different grains. This proved the homogeneity of Ti dispersion over the different grains (see Figure 3) and at the same time the comparable Ti content in the different incorporated samples. Furthermore, Figure 4 shows the representative SEM image for GP-P25, with the average particle size of this sample being ≈ 500-600 nm. Agglomerates of P25 TiO 2 powder with the average particle size ≈ 1µm, present in a water colloid (see Figure 2), were not observed here and thus were broken during incorporation into the GP matrix. Figure 4 also shows the successful incorporation of TiO 2 into the geopolymer structure. In contrast to the work of Maiti et al. [54], where the authors found TiO 2 nanoparticles of different sizes in their respective GP matrices, we observed homogenous material without traces of agglomerated TiO 2 . the GP matrix. Figure 4 also shows the successful incorporation of TiO2 into the geopolymer structure. In contrast to the work of Maiti et al. [54], where the authors found TiO2 nanoparticles of different sizes in their respective GP matrices, we observed homogenous material without traces of agglomerated TiO2.   The geopolymer source materials contain impurities such as titanium dioxide (see Table A1), and the prepared parent geopolymer matrix contains the TiO2 (0.94 wt.%)-see Table 3. The studied materials with the industrially produced TiO2 powder incorporated into the GP matrix were prepared by mixing 5 g of TiO2 powder, 100 g of the clay material Mefisto L05, 33 g of the BFS material and the corresponding amounts of alkalis. The chemical composition of the prepared samples was checked using an XRF analysis, which detected 3.42, 2.93 and 3.31 wt.% of TiO2 in GP-CG100, GP-CG300 and GP-P25 materials, respectively ( Table 3). The elemental maps obtained from the electron microscopy also show the homogenous distribution of titanium dioxide over the entire volume of the studied geopolymer with the incorporated TiO2.  the GP matrix. Figure 4 also shows the successful incorporation of TiO2 into the geopolymer structure. In contrast to the work of Maiti et al. [54], where the authors found TiO2 nanoparticles of different sizes in their respective GP matrices, we observed homogenous material without traces of agglomerated TiO2.   The geopolymer source materials contain impurities such as titanium dioxide (see Table A1), and the prepared parent geopolymer matrix contains the TiO2 (0.94 wt.%)-see Table 3. The studied materials with the industrially produced TiO2 powder incorporated into the GP matrix were prepared by mixing 5 g of TiO2 powder, 100 g of the clay material Mefisto L05, 33 g of the BFS material and the corresponding amounts of alkalis. The chemical composition of the prepared samples was checked using an XRF analysis, which detected 3.42, 2.93 and 3.31 wt.% of TiO2 in GP-CG100, GP-CG300 and GP-P25 materials, respectively ( Table 3). The elemental maps obtained from the electron microscopy also show the homogenous distribution of titanium dioxide over the entire volume of the studied geopolymer with the incorporated TiO2. The geopolymer source materials contain impurities such as titanium dioxide (see Table A1), and the prepared parent geopolymer matrix contains the TiO 2 (0.94 wt.%)-see Table 3. The studied materials with the industrially produced TiO 2 powder incorporated into the GP matrix were prepared by mixing 5 g of TiO 2 powder, 100 g of the clay material Mefisto L05, 33 g of the BFS material and the corresponding amounts of alkalis. The chemical composition of the prepared samples was checked using an XRF analysis, which detected 3.42, 2.93 and 3.31 wt.% of TiO 2 in GP-CG100, GP-CG300 and GP-P25 materials, respectively ( Table 3). The elemental maps obtained from the electron microscopy also show the homogenous distribution of titanium dioxide over the entire volume of the studied geopolymer with the incorporated TiO 2 . Powder X-ray diffraction patterns of the milled geopolymers are shown in Figure 5. The parent geopolymer (GP) contains, aside from a large quantity of amorphous geopolymers, quartz (JCPDS No. 00-046-1045), calcite (JCPDS No. 00-047-1748), gehlenite (JCPDS No. 00-035-0755) and anatase (JCPDS No. 04-014-5762). Diffraction lines of these phases have also been observed in the case of geopolymers with added TiO 2 . The intensity of the (101) line of anatase at about 25.3 • in the diffraction pattern of GP-CG300 seems to be similar to that of the GP, whereas in the case of GP-P25 and GP-CG100, the intensity of this diffraction line increases. To confirm this, the peak areas of the most intensive diffraction lines of anatase, calcite and quartz in the diffractograms were determined; they are shown in Table 4. As evident from Table 4, the anatase/quartz and anatase/calcite peak-area ratios for GP-P25 and GP-CG100 samples are about three times larger than in the case of the parent GP, which is in accordance with the increase in theTiO 2 content. On the other hand, the values of the peak-area ratios for the GP-CG300 sample remained practically the same as in the untreated GP, although the TiO 2 content in GP-CG300 also increased to about 3 wt.%. Furthermore, if we compare the proportion of the amorphous phase in the individual samples, it is clear that in the case of GP-CG100 and GP-P25 it was reduced in comparison with the parent geopolymer matrix by the addition of TiO 2 . On the other hand, the proportion of the amorphous phase in the GP-CG300 sample is comparable to the parent geopolymer. Based on these data, we can assume that the added TiO 2 CG300 (probably due to the smaller particle size and thus larger surface area (see Table A4)) reacts with the geopolymer matrix and therefore the addition of CG300 does not lead to an increase in the anatase content in the material.

Photocatalytic Activity
Before the photocatalytic testing, we studied (i) the possible decomposition of RhB under the used UV illumination without the presence of photocatalysts and (ii) the possi ble absorption of RhB dye by the parent geopolymer material (pure or with incorporated TiO2) due to its porous structure. Both of these phenomena could significantly influenc

Photocatalytic Activity
Before the photocatalytic testing, we studied (i) the possible decomposition of RhB under the used UV illumination without the presence of photocatalysts and (ii) the possible absorption of RhB dye by the parent geopolymer material (pure or with incorporated TiO 2 ) due to its porous structure. Both of these phenomena could significantly influence the as-measured photocatalytic activity. Figure A3A-C in Appendix A shows the change of the RhB dye absorption spectra during the illumination under different UV illumination conditions. It is evident that the RhB dye is decomposed under the illumination by UV photons generated either at λ = 250, 310 and 360 nm (without any filter) and 50% of the total intensity ( Figure A3A) or only at λ = 360 nm (using a filter transmitting in the range of 320-410 nm) and 100% of the total light source intensity ( Figure A3B). In both cases, RhB dye absorbance at the absorption maxima gradually decreases during illumination. After 120 min of illumination, the absorbance of the RhB dye decreases from 0.883 to 0.325 and to 0.281, respectively ( Figure A1A,B, respectively). Under milder illumination conditions, with UV photons generated only at λ = 360 nm (using a filter transmitting in the range of 320-410 nm) and 50% of the total light source intensity, the RhB was stable ( Figure A3C). Under these illumination conditions, the absorbance at the absorption maxima at λ = 555 nm of the RhB dye remained the same for 120 min of illumination and the RhB dye did not decompose. Therefore, only these illumination conditions were used in further photocatalytic tests.
The representative change of RhB dye absorption spectra during the illumination under the above-mentioned conditions in the presence of the GP-P25 photocatalyst is shown in Figure 6A. Based on the decrease in the RhB dye absorbance at the absorption maxima (marked as "absorption" in Figure 6A), the RhB dye is absorbed by the porous geopolymer structure, due to the method of synthesis (polycondensation reactions) [27,28,30], before the photocatalytic testing during the photocatalyst dispersion. The absorbance of RhB dye at the absorption maxima (λ = 555 nm) decreased from 0.883 to 0.793 after 5 min of dispersion ( Figure 6A). After this, the light source was switched on and the photocatalytic activity induced by UV illumination was monitored. With an increase in the illumination time, RhB dye absorbance gradually decreased. The absorbance decreased from 0.793 to 0.702 and 0.611 after 60 and 120 min of illumination, respectively (see Figure 6A and the difference between the lines "0 min" and "120 min" marked as "as-measured photocatalytic activity"). We continued to monitor the possible absorption of the RhB dye by the porous structure of the photocatalyst by the analysis of the identical dispersion, but without the illumination. The absorption of the RhB dye continued to proceed as the absorbance of the RhB dye decreased from 0.793 to 0.782 and 0.778 after 60 and 120 min, respectively (see Figure 6B and also Figure A3D). The photocatalytic activity X (%) was calculated according to the following equation: X (%) = (A 0 − A t )/A 0 × 100, where, based on the above-mentioned experiments A 0 , is the initial dye absorbance at the absorption maxima after 5 min of dispersion and A t is the dye absorbance at the absorption maxima at the specific time of photocatalysis corrected by the RhB dye absorption by the porous geopolymer structure. absorbance of the RhB dye decreased from 0.793 to 0.782 and 0.778 after 60 and 120 min, respectively (see Figure 6B and also FigureA3D). The photocatalytic activity X (%) was calculated according to the following equation: X (%) = (A0 − At)/A0 × 100, where, based on the above-mentioned experiments A0, is the initial dye absorbance at the absorption maxima after 5 min of dispersion and At is the dye absorbance at the absorption maxima at the specific time of photocatalysis corrected by the RhB dye absorption by the porous geopolymer structure. The decomposition rate (calculated in the similar way as X) as-measured and corrected by the absorption of the RhB dye by the geopolymer porous structure difference for the GP-P25 photocatalyst during the illumination is shown in Figure 7. After 120 min of illumination, the as-measured and corrected decomposition rate for the GP-P25 photocatalyst were 22.5 and 18%, respectively. The text below discusses only the corrected decomposition rate of the photocatalysts marked with the photocatalytic activity X, evaluated for pure TiO2 or the counterparts incorporated into the parent GP as in the case of GP-P25.  Figure 8A summarizes the photocatalytic activity X (%) during the illumination for as-supplied pure TiO2 photocatalysts (P25, CG100, CG300), the parent geopolymer matrix (GP) and a mechanical mixture of the parent geopolymer with the addition of 3 wt.% of TiO2 P25 (MM-GP-P25). The photocatalytic activity of the parent geopolymer matrix (Figure 8A) is much lower in comparison with the other materials, of only X ≈ 1.4 % after 120 min of illumination. This is in accordance with the literature as the photocatalytic mechanism in geopolymers is usually extrinsic [59]. On the other hand, the commercially used The decomposition rate (calculated in the similar way as X) as-measured and corrected by the absorption of the RhB dye by the geopolymer porous structure difference for the GP-P25 photocatalyst during the illumination is shown in Figure 7. After 120 min of illumination, the as-measured and corrected decomposition rate for the GP-P25 photocatalyst were 22.5 and 18%, respectively. The text below discusses only the corrected decomposition rate of the photocatalysts marked with the photocatalytic activity X, evaluated for pure TiO 2 or the counterparts incorporated into the parent GP as in the case of GP-P25.
respectively (see Figure 6B and also FigureA3D). The photocatalytic activity X (%) was calculated according to the following equation: X (%) = (A0 − At)/A0 × 100, where, based on the above-mentioned experiments A0, is the initial dye absorbance at the absorption maxima after 5 min of dispersion and At is the dye absorbance at the absorption maxima at the specific time of photocatalysis corrected by the RhB dye absorption by the porous geopolymer structure. The decomposition rate (calculated in the similar way as X) as-measured and corrected by the absorption of the RhB dye by the geopolymer porous structure difference for the GP-P25 photocatalyst during the illumination is shown in Figure 7. After 120 min of illumination, the as-measured and corrected decomposition rate for the GP-P25 photocatalyst were 22.5 and 18%, respectively. The text below discusses only the corrected decomposition rate of the photocatalysts marked with the photocatalytic activity X, evaluated for pure TiO2 or the counterparts incorporated into the parent GP as in the case of GP-P25.  Figure 8A summarizes the photocatalytic activity X (%) during the illumination for as-supplied pure TiO2 photocatalysts (P25, CG100, CG300), the parent geopolymer matrix (GP) and a mechanical mixture of the parent geopolymer with the addition of 3 wt.% of TiO2 P25 (MM-GP-P25). The photocatalytic activity of the parent geopolymer matrix (Figure 8A) is much lower in comparison with the other materials, of only X ≈ 1.4 % after 120 min of illumination. This is in accordance with the literature as the photocatalytic mechanism in geopolymers is usually extrinsic [59]. On the other hand, the commercially used  Figure 8A summarizes the photocatalytic activity X (%) during the illumination for as-supplied pure TiO 2 photocatalysts (P25, CG100, CG300), the parent geopolymer matrix (GP) and a mechanical mixture of the parent geopolymer with the addition of 3 wt.% of TiO 2 P25 (MM-GP-P25). The photocatalytic activity of the parent geopolymer matrix ( Figure 8A) is much lower in comparison with the other materials, of only X ≈ 1.4 % after 120 min of illumination. This is in accordance with the literature as the photocatalytic mechanism in geopolymers is usually extrinsic [59]. On the other hand, the commercially used pure TiO 2 P25 photocatalyst reached X = 98.6% within as early as 15min under the given illumination conditions (see Figure 8A). Furthermore, pure CG100 and CG300 photocatalysts reached the comparable photocatalytic activity (X) of 87.0 and 82.5%, respectively, but only after 120 min of illumination. The reasons for the different photocatalytic activity of the pure photocatalyst may be as follows: (i) even though CG100 and CG300 photocatalysts consist of only pure anatase, P25 also contains 13 wt.% of the rutile phase (see Figure 1 and the corresponding text); and (ii) there are significant differences in the average particle size and the specific surface area between individual photocatalysts (see Table A4). According to the literature data [9][10][11]60,61], all of these properties could have an effect on the photocatalytic activity of the photocatalyst. The different manufacturing method for the pure TiO 2 P25 in comparison with CG100 and CG300 photocatalysts (see Table A1 and the results of LOI) could also influence their photocatalytic activity. Furthermore, the mechanical mixture (MM-GP-P25) of the parent geopolymer with the addition of 3 wt.% of TiO 2 P25 also had an almost comparable X, of 77.0% after 120 min of illumination. This mechanical mixture (MM-GP-P25) was prepared for comparison as the pure TiO 2 P25 powder exhibited the highest photocatalytic activity. but only after 120 min of illumination. The reasons for the different photocatalytic activity of the pure photocatalyst may be as follows: (i) even though CG100 and CG300 photocatalysts consist of only pure anatase, P25 also contains 13 wt.% of the rutile phase (see Figure  1 and the corresponding text); and (ii) there are significant differences in the average particle size and the specific surface area between individual photocatalysts (see Table A4). According to the literature data [9][10][11]60,61], all of these properties could have an effect on the photocatalytic activity of the photocatalyst. The different manufacturing method for the pure TiO2 P25 in comparison with CG100 and CG300 photocatalysts (see Table A1 and the results of LOI) could also influence their photocatalytic activity. Furthermore, the mechanical mixture (MM-GP-P25) of the parent geopolymer with the addition of 3 wt.% of TiO2 P25 also had an almost comparable X, of 77.0% after 120 min of illumination. This mechanical mixture (MM-GP-P25) was prepared for comparison as the pure TiO2 P25 powder exhibited the highest photocatalytic activity.  Figure 8B shows the time dependence of photocatalytic activity (X) for photocatalysts incorporated into the parent geopolymer matrix with a comparable content of TiO2 powder (≈3.2 wt.%, see Table 3), labelled as GP-P25, GP-CG100 and GP-CG300. For the possible schematic mechanism of RhB dye decomposition on TiO2 modified metakaolin-based geopolymer see, e.g., [59,62,63]. All materials were milled to a comparable particle size (d50) before testing (see Table A5). Furthermore, the photocatalysts with the higher (≈6.0 wt.%) and lower (≈1.5 wt.%) content of TiO2 incorporated powder were also tested (not shown in Tables and Figures). However, as it did not have a positive effect, with the respect to the TiO2 cost, on the photocatalytic activity (e.g., the X value is equal to 10.5, 18.0 and 20.0% for GP-P25 photocatalyst with ≈1.7, 3.3 and 6.0 wt.% of TiO2, respectively), the photocatalysts containing ≈3.2 wt.% of different TiO2sources were compared. The time dependence of X for GP-P25 and GP-CG100 photocatalysts is almost comparable ( Figure  8B). After 120 min of illumination, the photocatalytic activity reaches 19.5 and 18.0% for the GP-CG100 and GP-P25 photocatalysts, respectively. Evidently, the difference in X between the pure, industrially produced TiO2 P25 and CG100 photocatalysts ( Figure 8A) were eliminated. Furthermore, the X dropped to 11.0% for the GP-CG300 photocatalyst after 120 min of illumination ( Figure 8B). The different photocatalytic activities of the photocatalysts incorporated into the parent geopolymer matrix may be explained as follows: Firstly, the lowest X for the GP-CG300 photocatalyst could be the consequence of the above-mentioned fact that (see Figure 5 and also Table 4 Figure 8B shows the time dependence of photocatalytic activity (X) for photocatalysts incorporated into the parent geopolymer matrix with a comparable content of TiO 2 powder (≈3.2 wt.%, see Table 3), labelled as GP-P25, GP-CG100 and GP-CG300. For the possible schematic mechanism of RhB dye decomposition on TiO 2 modified metakaolin-based geopolymer see, e.g., [59,62,63]. All materials were milled to a comparable particle size (d 50 ) before testing (see Table A5). Furthermore, the photocatalysts with the higher (≈6.0 wt.%) and lower (≈1.5 wt.%) content of TiO 2 incorporated powder were also tested (not shown in Tables and Figures). However, as it did not have a positive effect, with the respect to the TiO 2 cost, on the photocatalytic activity (e.g., the X value is equal to 10.5, 18.0 and 20.0% for GP-P25 photocatalyst with ≈1.7, 3.3 and 6.0 wt.% of TiO 2 , respectively), the photocatalysts containing ≈3.2 wt.% of different TiO 2 sources were compared. The time dependence of X for GP-P25 and GP-CG100 photocatalysts is almost comparable ( Figure 8B). After 120 min of illumination, the photocatalytic activity reaches 19.5 and 18.0% for the GP-CG100 and GP-P25 photocatalysts, respectively. Evidently, the difference in X between the pure, industrially produced TiO 2 P25 and CG100 photocatalysts ( Figure 8A) were eliminated. Furthermore, the X dropped to 11.0% for the GP-CG300 photocatalyst after 120 min of illumination ( Figure 8B). The different photocatalytic activities of the photocatalysts incorporated into the parent geopolymer matrix may be explained as follows: Firstly, the lowest X for the GP-CG300 photocatalyst could be the consequence of the above-mentioned fact that (see Figure 5 and also Table 4) this was the only incorporated photocatalyst not to exhibit an increase in the anatase diffraction lines in comparison with the parent geopolymer matrix. In the GP-CG300 photocatalyst, titanium is thus present in a form other than anatase. To support this idea, we performed FTIR measurements. Figure 9 shows a comparison of FTIR spectra for pure CG300 and CG100 photocatalysts and those incorporated into the parent geopolymer matrix. We compared the results for all the samples in the spectral region of 1300-800 cm −1 , where the pure TiO 2 samples provided the absorption and the absorption of the parent geopolymer matrix in the incorporated photocatalysts was subtracted. The spectra of the pure CG100 photocatalyst and the incorporated photocatalyst GP-CG100 are practically identical. On the other hand, the spectra of the pure CG300 photocatalyst and the incorporated photocatalyst GP-CG300 are slightly different. It is possible to see the broadening for GP-CG300 in comparison with the spectra of the CG300 photocatalyst, which indicates the loss of ordering in the first case. This corroborates the results obtained by the XRD analysis (see Figure 5 and Table 4). Secondly, the different photocatalytic activities of the incorporated photocatalysts could also be caused by the fact that the average particle size (see Table A4) of the pure CG300 was the lowest of the pure TiO 2 photocatalysts. As stated above, this pure TiO 2 photocatalyst could thus be incorporated into the porous geopolymer structure much more easily than the other pure TiO 2 photocatalysts, and thus the support covers the TiO 2 particles, see, e.g., [16,17]. We assume that covered TiO 2 particles are unavailable for the photocatalytic reaction. On the other hand, DLS demonstrated that the average particle size for the pure CG300 photocatalyst during the preparation of its incorporated counterpart is the highest among the pure photocatalysts (Figure 2), which makes available less-active sites for the reaction. The results of the DLS are inversely proportional to the activities of the photocatalysts; therefore, the higher the DLS-average particle size, the lower the photocatalytic activity. The photocatalytic activity of the incorporated photocatalyst GP-CG300 was negatively affected either by incorporation into the geopolymer structure or by self shielding. Furthermore, the photocatalytic activity of the incorporated photocatalysts could also be influenced by other material characteristics, such as the specific surface area (see Table A4) of the TiO 2 additives. samples in the spectral region of 1300-800 cm −1 , where the pure TiO2 samples provided the absorption and the absorption of the parent geopolymer matrix in the incorporated photocatalysts was subtracted. The spectra of the pure CG100 photocatalyst and the incorporated photocatalyst GP-CG100 are practically identical. On the other hand, the spectra of the pure CG300 photocatalyst and the incorporated photocatalyst GP-CG300 are slightly different. It is possible to see the broadening for GP-CG300 in comparison with the spectra of the CG300 photocatalyst, which indicates the loss of ordering in the first case. This corroborates the results obtained by the XRD analysis (see Figure 5 and Table  4). Secondly, the different photocatalytic activities of the incorporated photocatalysts could also be caused by the fact that the average particle size (see Table A4) of the pure CG300 was the lowest of the pure TiO2 photocatalysts. As stated above, this pure TiO2 photocatalyst could thus be incorporated into the porous geopolymer structure much more easily than the other pure TiO2 photocatalysts, and thus the support covers the TiO2 particles, see, e.g., [16,17]. We assume that covered TiO2 particles are unavailable for the photocatalytic reaction. On the other hand, DLS demonstrated that the average particle size for the pure CG300 photocatalyst during the preparation of its incorporated counterpart is the highest among the pure photocatalysts (Figure 2), which makes available lessactive sites for the reaction. The results of the DLS are inversely proportional to the activities of the photocatalysts; therefore, the higher the DLS-average particle size, the lower the photocatalytic activity. The photocatalytic activity of the incorporated photocatalyst GP-CG300 was negatively affected either by incorporation into the geopolymer structure or by self shielding. Furthermore, the photocatalytic activity of the incorporated photocatalysts could also be influenced by other material characteristics, such as the specific surface area (see Table A4) of the TiO2 additives. Figure 9. The FTIR spectra of pure CG100 and CG300 photocatalysts and those incorporated into the parent geopolymer matrix.
The stability of incorporated photocatalysts, which is one of the most important aspects of heterogeneous photocatalysis in general, was tested in the three consecutive cycles for GP-CG100 material, under the above-mentioned illumination conditions. The catalyst was recovered by filtration between each cycle and at the same time the appropriate volume of fresh RhB dye solution was used. After 120 min of illumination in each cycle, the photocatalytic activity reached 19.5, 18.0 and 18.5% showing good stability, with respect to the experimental error in X determination, for the GP-CG100 photocatalyst. A detailed analysis of photocatalysts after each cycle will be explored in our future work.
In summary, the properties of raw TiO2 powder materials such as the average particle size, mineralogical composition, etc., are suitable for the incorporated photocatalysts GP- The stability of incorporated photocatalysts, which is one of the most important aspects of heterogeneous photocatalysis in general, was tested in the three consecutive cycles for GP-CG100 material, under the above-mentioned illumination conditions. The catalyst was recovered by filtration between each cycle and at the same time the appropriate volume of fresh RhB dye solution was used. After 120 min of illumination in each cycle, the photocatalytic activity reached 19.5, 18.0 and 18.5% showing good stability, with respect to the experimental error in X determination, for the GP-CG100 photocatalyst. A detailed analysis of photocatalysts after each cycle will be explored in our future work.
In summary, the properties of raw TiO 2 powder materials such as the average particle size, mineralogical composition, etc., are suitable for the incorporated photocatalysts GP-CG100 and GP-P25 and the parent geopolymer works well for them as a supporting material, which leads to the higher photocatalytic activity of these materials in comparison with the GP-CG300 material. Furthermore, our future work will be focused on the identification of the key properties and/or their relationship for the preparation of photoactive geopolymer-based photocatalyst with incorporated TiO 2 powder and the photocatalytic activity of block samples, which would be more relevant to their possible practical usage.
For comparison, the kinetic data according to the work of Maiti et al. [54] were evaluated. The pseudo first-order rate constants (see Table 5) for each photocatalyst were estimated from the slopes of linear plots ln (A 0 /A) vs. time. With regard to pure photocatalysts, the highest rate constant (0.27 min −1 ) was observed for TiO 2 P25. For CG100 and CG300 photocatalysts, the reaction was approximately 10 times slower, with rate constants of 0.033 min −1 and 0.029 min −1 , respectively ( Table 5). As mentioned above, this may relate to the properties of the photocatalysts (the mineralogical composition, the average particle size, etc.); the decrease in the rate constants of the CG100 and CG300 photocatalysts is in good agreement with their decrease in photocatalytic activity, (X) in comparison with the P25 photocatalyst. Furthermore, the rate constant for the mechanical mixture of the parent geopolymer with the addition of 3 wt.% of TiO 2 P25 was 0.012 min −1 , which is approximately 4.4% of the rate constant of the pure TiO 2 P25 photocatalyst and thus, in good agreement with its content in the mixture. This may be the result of the shielding effect [64]. The rate constant for the parent geopolymer matrix has not been evaluated as this material exhibits very low photocatalytic activity (X ≈ 1.4%). The rate constants for the photocatalysts incorporated into the parent geopolymer matrix were significantly lower in comparison with their pure counterparts. We assume this to be caused by the covering of TiO 2 particles by supporting materials during preparation (see Figure A1 in Appendix A) and such particles are then inactive in the photocatalytic reaction. The lowest rate constant (0.0010 min −1 ) was observed for the GP-CG300 material. For the GP-CG100 and GP-P25 photocatalysts, the rate constants were almost comparable-0.0017 and 0.0018 min −1 , respectively (Table 5). Evidently, the photocatalytic activities X (see the text above) and the rate constants for GP-CG100 and GP-P25 photocatalysts are comparable, although these properties significantly differ for their pure counterparts. The key parameters identification of pure TiO 2 powders that are suitable for the preparation of incorporated counterparts into the parent geopolymer matrix, with a high photocatalytic activity, will be the subject of our future work. Furthermore, the rate constants of incorporated photocatalysts are in good agreement with their photocatalytic activity.

Conclusions
In this work, the different types of industrial TiO 2 powder were incorporated into the parent geopolymer matrix (GP-P25, GP-CG100 and GP-CG300). The successful incorporation of TiO 2 into the parent geopolymer matrix was demonstrated by the SEM analysis, and the elemental analysis revealed its homogeneous distribution within a sample and its comparable content over different samples. The photo-catalytic activity of materials, corrected by the possible absorption and decomposition of the dye, caused by porous geopolymer structures and UV illumination, respectively, was evaluated based on the Rhodamine B dye decomposition. The GP-CG100 and GP-P25 photocatalysts exhibited a higher photocatalytic activity (≈19%) than the GP-CG300 material (11%). While the addition of CG100 and P25 TiO 2 powder leads to an increase in the anatase content of GP-CG100 and GP-P25 materials, the anatase content in GP-CG300 is comparable with the parent GP (see powder XRD). Furthermore, since CG300 has a significantly smaller particle size and higher surface area than the others, the properties of the TiO 2 powder used could affect its interaction with the geopolymer matrix during preparation, which could also influence the photocatalytic performance of the resultant materials. Additionally, the repeated cycles show the good stability of the GP-CG100 photocatalyst. This metakaolin-based geopolymer, with additional TiO 2 , represents an alternative environmentally friendly material with an enhanced purifying ability.       Figure A1. The schematic illustration of photocatalysts preparation. Figure A1. The schematic illustration of photocatalysts preparation. (a) (b) Figure A2. The representative EDS spectra (a)for the corresponding SEM image (cutand polished GP-P25 sample) (b). Figure A3. The change of RhB dye absorbance spectra during illumination by UV photons generated (A) at λ = 250, 310 and 360 nm (without any filter) and 50% of the total intensity, (B) at λ = 360 nm (using a filter transmitting in the range of 320-410 nm) and 100% of the total intensity, (C) at λ = 360 nm (using a filter transmitting in the range of 320-410 nm) and 50% of the total intensity, (D) and a combination of Figure 6A,B.  Figure A3. The change of RhB dye absorbance spectra during illumination by UV photons generated (A) at λ = 250, 310 and 360 nm (without any filter) and 50% of the total intensity, (B) at λ = 360 nm (using a filter transmitting in the range of 320-410 nm) and 100% of the total intensity, (C) at λ = 360 nm (using a filter transmitting in the range of 320-410 nm) and 50% of the total intensity, (D) and a combination of Figure 6A,B.