Enhanced Fe-TiO2 Solar Photocatalysts on Porous Platforms for Water Purification

In this study, polyethylene glycol-modified titanium dioxide (PEG-modified TiO2) nanopowders were prepared using a fast solvothermal method under microwave irradiation, and without any further calcination processes. These nanopowders were further impregnated on porous polymeric platforms by drop-casting. The effect of adding iron with different molar ratios (1, 2, and 5%) of iron precursor was investigated. The characterization of the produced materials was carried out by scanning electron microscopy (SEM), energy-dispersive X-ray spectroscopy (EDS), X-ray diffraction (XRD), X-ray photoelectron spectroscopy (XPS), and Raman spectroscopy. Optical characterization of all the materials was also carried out. SEM showed that pure TiO2 and Fe-TiO2 nanostructures presented similar nanosized and spherical particles, which uniformly covered the substrates. From XRD, pure TiO2 anatase was obtained for all nanopowders produced, which was further confirmed by Raman spectroscopy on the impregnated substrates. XPS and UV–VIS absorption spectroscopy emission spectra revealed that the presence of Fe ions on the Fe-TiO2 nanostructures led to the introduction of new intermediate energy levels, as well as defects that contributed to an enhancement in the photocatalytic performance. The photocatalytic results under solar radiation demonstrated increased photocatalytic activity in the presence of the 5% Fe-TiO2 nanostructures (Rhodamine B degradation of 85% after 3.5 h, compared to 74% with pure TiO2 for the same exposure time). The photodegradation rate of RhB dye with the Fe-TiO2 substrate was 1.5-times faster than pure TiO2. Reusability tests were also performed. The approach developed in this work originated novel functionalized photocatalytic platforms, which were revealed to be promising for the removal of organic dyes from wastewater.


Introduction
As a by-product of the society's rapid development, environmental contaminants, such as dyes, drugs, and pesticides, are widely being released into effluents and represent, not only a risk to human health, but also to the environment, due to their complex structure and high recalcitrance. Spillage of dyes, widely used in various industries, such as watersoluble Rhodamine B (RhB) and Methylene blue (MB), endangers animals, plants, and human beings [1][2][3]. In this regard, a growing interest in the removal of pollution effluents from wastewater through environmentally friendly routes has spurred intensive research over the last few years [4][5][6]. focused on a fast Fe-TiO 2 microwave synthesis without the need to calcinate the samples; the approach developed in this study.
Crystallinity, porosity, surface chemical properties, and surface area of the metal oxides play an important role in the photocatalytic activity and, thus, need to be considered. Larger surface areas generally provide more active sites for reactions to occur, pore channels favor mass diffusion, and high crystallinity can improve metal oxides' stability. Nevertheless, the challenge relies on the balance between crystallinity and the porous structure, since high crystallinity leads to a rearrangement and migrations of atoms, followed by a disarrangement of the porous structure, which, in consequence, causes a decrease in the specific surface area [58].
In order to produce improved materials with high surface area, several studies have focused on the synthesis of controlled porous TiO 2 -based materials with the addition of surfactants for the enhancement of photocatalytic activity [59][60][61][62]. Feng et al., reported enhanced metal-ion-doped TiO 2 photocatalysts under solar light with a large BET (Brunauer-Emmett-Teller) surface area, synthesized by microwave irradiation, through the use of the surfactant P123 [59]. Kubiak et al., also demonstrated superior photocatalytic degradation of phenol and etodolac, with the nanomaterials obtained by microwave irradiation using different surfactants (PEG (Polyethylene glycol), Triton-X and P123) compared to pristine TiO 2 samples [62]. In another research study, Jang et al., also explored the effect of Polyvinylpyrrolidone (PVP), Triton-X, and P123 on the fabrication of mesoporous structured TiO 2 nanoparticles by a microwave-assisted sol-gel method. A higher photocatalytic performance for MB degradation was exhibited compared to P25, attributed to the large surface area and porosity obtained with the addition of surfactants [63].
Different substrates have been used for growing Fe-doped TiO 2 nanostructures, including glass substrates [64,65], polyamide fabric [66], and Si substrates [67], but to the best of our knowledge, this has never been reported for porous water filters. The impregnation of these nanostructures on commercial water filters avoids the limitations and costs associated with the recovery of nanometer-sized particles.
In this study, pure TiO 2 and 1, 2, and 5 mol% Fe-TiO 2 nanostructures were synthesized by a fast surfactant-assisted microwave method (1 h), without any further calcination of the nanostructures, and impregnated using a simple drop-casting technique on polymeric substrates, used as water filters. Enhanced visible light absorption of TiO 2 -based materials on flexible porous water filter substrates was, thus, produced. The surface modification of TiO 2 -based materials with PEG was investigated. The materials produced were systematically characterized by XRD, Raman spectroscopy, SEM coupled with EDS, and XPS. Optical characterization was also carried out for the nanostructures, via UV-VIS absorption and room temperature diffuse reflectance measurements, to determine the optical band gap values. Finally, their efficiency as photocatalysts on water filters was investigated for the degradation of the organic model pollutant, RhB, under solar radiation.

Preparation of Pure and Fe-TiO 2 Nanostructures via Surfactant-Assisted Microwave Irradiation
Titanium (IV) isopropoxide (TTIP), 97% purity, was purchased from Sigma-Aldrich, St. Louis, MO, USA (CAS: 546-  and was used as the titanium dioxide precursor. TTIP was dissolved in absolute anhydrous ethanol from Carlo Erba reagents, Val-de-Reuil, France (CAS: 64-17-5) and oxalic acid anhydrous from Sigma-Aldrich (CAS: 144-62-7). To prepare a solution of 120 mL, 5 mL of a 1 M oxalic acid solution was added to 115 mL of absolute ethanol and stirred for 10 min. Afterwards, 1 mL of polyethylene glycol (PEG 400) from Sigma-Aldrich (CAS: 25322-68-3) was added, followed by dropwise addition of 4 mL TTIP. Then the solution was stirred for 1 h. Microwave synthesis was performed using a CEM microwave digestion system, Matthews, NC, USA (MARS one), which was carried out at 1000 W, 170 ± 10 • C for 1 h. Solution volumes of 40 mL were transferred into Teflon vessels of 75 mL, which were kept sealed. To prepare Fe-TiO 2 nanoparticles, 1, 2, and 5 mol% of iron (III) nitrate nonahydrate purchased from Sigma-Aldrich (CAS: 7782 Nanomaterials 2022, 12, 1005 4 of 23 81-8, 98% purity) were added to the previous solution, before the addition of TTIP. The nominal ratios of Fe precursor, TTIP, were 0.02, 0.03, and 0.07 for 1, 2, and 5 mol% of iron, respectively. The solution was stirred for 1 h, before transferring it to the microwave vessel. The as-synthesized TiO 2 nanopowders were cleaned using a centrifuge at 4000 rpm for 5 min. The powders were washed repeatedly several times with deionized water, followed by isopropyl alcohol (IPA). Finally, the nanomaterials were kept overnight in a desiccator for drying, at 80 • C in vacuum.

Impregnation of Nanoparticles in Polymeric Substrates by Drop-Casting
In this study a simple and low-cost impregnation method was used, which had already been proved to be efficient for the functionalization of the substrates to be tested in photocatalytic experiments [68,69]. Several circular polymeric filters made of MCE (mixed cellulose esters, composed of inert cellulose nitrate and cellulose acetate polymers) with a pore size of 0.22 µm and diameter of 47 mm were tested. The filters were attached to previously cut ordinary glass squares using Kapton tape. These glass squares were used to prevent the polymeric substrates from bending, after complete drying of the droplets. The coffee-ring stain phenomenon of drop-casting can be suppressed by controlling the liquid properties, such as viscosity and pH, and by changing the temperatures of substrate and droplets [70][71][72]. In this study, 40 mg of nanoparticles was added to 9 mL of deionized water. Then, 1 mL of ethylene glycol from Carlo Erba Reagents (CAS: 107-21-1, 99.5% purity) was also added to the previous solution. It is reported that a small amount of ethylene glycol (10-30 vol.%) in colloidal suspensions produces uniform layers [73,74]. Before applying the droplets, and to mitigate the coffee-ring effect, the hydrophilic polymeric substrates were heated in a hot plate at 80 • C. Six layers of 1 mL of the aqueous solution were uniformly applied onto the substrates, leaving the necessary time for each layer to dry. A schematic illustration for the synthesis of the nanoparticles and the impregnation process is presented in Figure 1. 400) from Sigma-Aldrich (CAS: 25322-68-3) was added, followed by dropwise addition of 4 mL TTIP. Then the solution was stirred for 1 h. Microwave synthesis was performed using a CEM microwave digestion system, Matthews, NC, USA (MARS one), which was carried out at 1000 W, 170 ± 10 °C for 1 h. Solution volumes of 40 mL were transferred into Teflon vessels of 75 mL, which were kept sealed. To prepare Fe-TiO2 nanoparticles, 1, 2, and 5 mol% of iron (III) nitrate nonahydrate purchased from Sigma-Aldrich (CAS: 7782 81-8, 98% purity) were added to the previous solution, before the addition of TTIP. The nominal ratios of Fe precursor, TTIP, were 0.02, 0.03, and 0.07 for 1, 2, and 5 mol% of iron, respectively. The solution was stirred for 1 h, before transferring it to the microwave vessel. The as-synthesized TiO2 nanopowders were cleaned using a centrifuge at 4000 rpm for 5 min. The powders were washed repeatedly several times with deionized water, followed by isopropyl alcohol (IPA). Finally, the nanomaterials were kept overnight in a desiccator for drying, at 80 °C in vacuum.

Impregnation of Nanoparticles in Polymeric Substrates by Drop-Casting
In this study a simple and low-cost impregnation method was used, which had already been proved to be efficient for the functionalization of the substrates to be tested in photocatalytic experiments [68,69]. Several circular polymeric filters made of MCE (mixed cellulose esters, composed of inert cellulose nitrate and cellulose acetate polymers) with a pore size of 0.22 μm and diameter of 47 mm were tested. The filters were attached to previously cut ordinary glass squares using Kapton tape. These glass squares were used to prevent the polymeric substrates from bending, after complete drying of the droplets. The coffee-ring stain phenomenon of drop-casting can be suppressed by controlling the liquid properties, such as viscosity and pH, and by changing the temperatures of substrate and droplets [70][71][72]. In this study, 40 mg of nanoparticles was added to 9 mL of deionized water. Then, 1 mL of ethylene glycol from Carlo Erba Reagents (CAS: 107-21-1, 99.5% purity) was also added to the previous solution. It is reported that a small amount of ethylene glycol (10-30 vol.%) in colloidal suspensions produces uniform layers [73,74]. Before applying the droplets, and to mitigate the coffee-ring effect, the hydrophilic polymeric substrates were heated in a hot plate at 80 °C. Six layers of 1 mL of the aqueous solution were uniformly applied onto the substrates, leaving the necessary time for each layer to dry. A schematic illustration for the synthesis of the nanoparticles and the impregnation process is presented in Figure 1.

Photocatalytic Activity
The photocatalytic activities of pure TiO 2 and Fe-TiO 2 nanostructures on water filters were evaluated at RT, considering the degradation of RhB (C 28 H 31 ClN 2 O 3 ) from Sigma-Aldrich under a solar light simulating source. The nanoparticles were impregnated on a substrate to aid in the recovery of nanosized materials. All the experiments were performed according to the international standard, ISO 10678 [49]. For each experiment, the polymeric-based materials were placed on the bottom of the reaction recipient and for each experiment, 50 mL of the RhB solution (5 mg/L) was stirred for 30 min in the dark, to establish absorption-desorption equilibrium. Solar light exposure was carried out by using a LED solar simulator LSH 7320, Irvine, CA, USA (AM 1.5 spectrum) with an intensity of 100 mW/cm 2 . The experiments were conducted under low constant magnetic agitation at 205 rpm. Absorption spectra were recorded using a PerkinElmer lambda 950 UV/VIS/NIR spectrophotometer, Waltham, MA, USA, with different time intervals, up to a total of 5 h. The measurements were performed in the 400-700 nm range every 30 min. The pristine substrate was also exposed with identical conditions.
The reusability experiments consisted in recovering the water filters with further discarding of the total liquid. The recovered filters were dried in air, prior to the next exposure. The recovered materials were then poured into fresh solution and exposed to solar light considering the same exposure times.
Scanning electron microscopy (SEM) images were obtained using a Regulus 8220 Scanning Electron Microscope (Mito, Japan), while for the EDS analyses, a Carl Zeiss AURIGA CrossBeam FIB-SEM workstation (Oberkochen, Germany) was used. The dimensions of the nanoparticles were determined by measuring 60 nanostructures from SEM micrographs with ImageJ software (version 1.51j8).
X-ray photoelectron spectroscopy (XPS) measurements were carried out with a Kratos Axis Supra (Manchester, UK), using monochromated Al Kα irradiation (1486.6 eV). The acquisition of detailed scans was performed with an X-ray power of 225 W and a pass energy of 40 eV. All spectra were shifted to O 1 s at 530 eV [77]. Data analysis was carried out with CasaXPS.
Room temperature (RT) diffuse reflectance measurements, to obtain the optical bandgap, were performed using a Perkin Elmer lambda 950 UV/VIS/NIR spectrophotometer with a diffuse reflectance module (150-mm diameter integrating sphere, internally coated with Spectralon) and a powder sample holder. The calibration of the system was achieved using a standard reflector sample (reflectance, R = 1.00 from Spectralon disk). The reflectance (R) was obtained from 250 to 700 nm.
The absorption measurements were also carried out using the same UV/VIS/NIR spectrophotometer, and the measurements were performed in the 250-700 nm range.
Raman spectroscopy measurements were obtained with an inVia Qontor confocal Raman microscope from Renishaw (Kingswood, UK). A 17-mW He-Ne laser operated at 532 nm with a 10 s exposure time and settings of 5 accumulations. The Raman spectra were recorded in the range of 100-700 cm −1 . The results presented for the impregnated substrates are based on the average of several points taken on their surface. All measurements were performed at room temperature.

Results and Discussion
Pure TiO 2 and 1, 2, and 5 mol% of Fe-TiO 2 nanostructures with PEG were successfully synthesized under microwave irradiation, considering a fast synthesis time (1 h). The characterization of the nanoparticles, in terms of morphology, crystalline structure, and optical properties was carried out. The as-synthesized nanopowders were further impregnated on water filters for photocatalytic tests. The addition of iron on TiO 2 , as well as the efficiency  (Figure 2b-d, respectively). PEG was used as a surfactant for all microwave syntheses. It can be observed that all conditions resulted in nanosized TiO 2 particles with a spherical shape. These nanoparticles appear as agglomerates, since after drying the powder, there is aggregation in larger particles, reaching the micrometer range. The insets in Figure 2 also show the grain size histograms and the Gaussian model with an acceptable fitting for all particle size distributions. The estimated average particle sizes were 18.58 ± 2.73 nm, 18.40 ± 2.26 nm, 17.00 ± 2.57 nm, and 16.54 ± 3.02 nm for pure TiO 2 , and 1, 2, and 5 mol% of Fe-TiO 2 , respectively. A decrease in diameter was observed with the increase of iron content, which has been previously reported [56,78,79]. A more pronounced decrease in the nanoparticle size was achieved when the Fe concentration was above 2 mol%.

Results and Discussion
Pure TiO2 and 1, 2, and 5 mol% of Fe-TiO2 nanostructures with PEG were successfully synthesized under microwave irradiation, considering a fast synthesis time (1 h). The characterization of the nanoparticles, in terms of morphology, crystalline structure, and optical properties was carried out. The as-synthesized nanopowders were further impregnated on water filters for photocatalytic tests. The addition of iron on TiO2, as well as the efficiency of the impregnation method was systematically investigated. Reusability tests within three repeated cycles were also performed. Figure 2 depicts the SEM morphology of the pure TiO2 ( Figure 2a) and 1, 2, and 5 mol% Fe-TiO2 nanoparticles (Figure 2b-d, respectively). PEG was used as a surfactant for all microwave syntheses. It can be observed that all conditions resulted in nanosized TiO2 particles with a spherical shape. These nanoparticles appear as agglomerates, since after drying the powder, there is aggregation in larger particles, reaching the micrometer range. The insets in Figure 2 also show the grain size histograms and the Gaussian model with an acceptable fitting for all particle size distributions. The estimated average particle sizes were 18.58 ± 2.73 nm, 18.40 ± 2.26 nm, 17.00 ± 2.57 nm, and 16.54 ± 3.02 nm for pure TiO2, and 1, 2, and 5 mol% of Fe-TiO2, respectively. A decrease in diameter was observed with the increase of iron content, which has been previously reported [56,78,79]. A more pronounced decrease in the nanoparticle size was achieved when the Fe concentration was above 2 mol%.

XRD Analysis
The as-synthesized nanopowders were also investigated using XRD and the results are depicted in Figure 3. All peaks in the experimental diffractograms indicate the presence of TiO 2 anatase phase, which has Ti −6 (octahedral) and O −3 (trigonal planer) coordination geometry [80,81]. The peaks detected correspond to the planes (101) No peaks associated with rutile or brookite were detected. For the Fe-rich nanostructures, no peaks assignable to a metal oxide phase (Fe 2 O 3 (hematite) or Fe 2 TiO 5 ) were detected. As such, the quantity/percentage of these oxides could be too low to be detected, could be amorphous, or, since Fe 3+ and Ti 4+ have similar ionic radii, Fe 3+ ions Nanomaterials 2022, 12, 1005 7 of 23 could have successfully substituted Ti 4+ ions. No changes in peak intensity, as a function of iron loading, or shifts could be observed [1,49,82]. nation geometry [80,81]. The peaks detected correspond to the planes (101), (004), (200 (105), (211), (204), (116), (220), and (215) at 2θ = 25.3, 37.8, 48.0, 53.8, 54.9, 62.8, 68.9, an 75.0°, respectively. No peaks associated with rutile or brookite were detected. For the F rich nanostructures, no peaks assignable to a metal oxide phase (Fe2O3 (hematite) Fe2TiO5) were detected. As such, the quantity/percentage of these oxides could be too lo to be detected, could be amorphous, or, since Fe 3+ and Ti 4+ have similar ionic radii, Fe ions could have successfully substituted Ti 4+ ions. No changes in peak intensity, as a fun tion of iron loading, or shifts could be observed [1,49,82]. No peaks coming from impurities such as Ti(OH)4 were detected, and the XRD r sults demonstrate that the materials were well crystallized and highly nanostructure [49,52].
The average crystallite sizes of the synthesized nanomaterials were calculated usin the Debye-Scherrer equation. The estimated crystallite sizes for each powder are summ rized in Table 1. Table 1. Estimated crystallite sizes (nm) for TiO2, 1, 2, and 5% mol of Fe-TiO2 nanoparticles wi PEG synthesized by microwave irradiation.

TiO2
1% Fe-TiO2 2% Fe-TiO2 5% Fe-TiO2 Sizes (nm) 6.0 5.7 5.2 4.9 As expected, the obtained values for the particle sizes obtained previously from SE analysis (Section 4.1.1) are superior to the crystallite sizes, since particles are agglomerat of grains, and the grains are composed of several crystallites. Moreover, the increase iron content led to a decrease in the crystallite and particle sizes, respectively. As reporte doping with Fe ions could inhibit the growth of TiO2 anatase crystals, leading to small No peaks coming from impurities such as Ti(OH) 4 were detected, and the XRD results demonstrate that the materials were well crystallized and highly nanostructured [49,52].
The average crystallite sizes of the synthesized nanomaterials were calculated using the Debye-Scherrer equation. The estimated crystallite sizes for each powder are summarized in Table 1. As expected, the obtained values for the particle sizes obtained previously from SEM analysis (Section 4.1.1) are superior to the crystallite sizes, since particles are agglomerates of grains, and the grains are composed of several crystallites. Moreover, the increase of iron content led to a decrease in the crystallite and particle sizes, respectively. As reported, doping with Fe ions could inhibit the growth of TiO 2 anatase crystals, leading to smaller particles [78,79,82]. Hence, it can be suggested that this effect is more pronounced for higher iron contents.

XPS Analysis
The chemical composition of TiO 2 and Fe-TiO 2 nanostructures with PEG was analyzed by XPS. The conditions of 1 and 2 mol% of Fe-TiO 2 showed a weak iron signal; therefore, these data were not included in this study; hence, only the composition of 5 mol% (corresponding to 1.4 at. %) is presented.
Survey spectra of pure TiO 2 and 5% Fe-TiO 2 are shown in Figure 4. XPS results showed that Ti and O were clearly identified for pure TiO 2 , whereas for the Fe-TiO 2 nanostructures, Fe could also be detected. A carbon peak was also visible, related to adventitious carbon at the surface. Figure 5 shows the high resolution XPS spectra of O 1s (deconvoluted) and Ti 2p core levels from pure synthesized TiO 2 nanostructures with PEG and of 5% Fe-TiO 2 .
The chemical composition of TiO2 and Fe-TiO2 nanostructures with PEG was analyzed by XPS. The conditions of 1 and 2 mol% of Fe-TiO2 showed a weak iron signal; therefore, these data were not included in this study; hence, only the composition of 5 mol % (corresponding to 1.4 at. %) is presented.
Survey spectra of pure TiO2 and 5% Fe-TiO2 are shown in Figure 4. XPS results showed that Ti and O were clearly identified for pure TiO2, whereas for the Fe-TiO2 nanostructures, Fe could also be detected. A carbon peak was also visible, related to adventitious carbon at the surface. Figure 5 shows the high resolution XPS spectra of O 1s (deconvoluted) and Ti 2p core levels from pure synthesized TiO2 nanostructures with PEG and of 5% Fe-TiO2.  In Figure 5a,b, four peaks can be observed that correspond to lattice oxygen (A), undercoordinated oxygen either at the surface or close to oxygen vacancies (B), and adsorbed water and organic species (C,D) [83]. An increase of the peaks B-D with respect to peak A is observed in the Fe-TiO2, compared to TiO2. Part of this increase may be ascribed to a higher oxygen vacancy concentration, due to the doping. Due to charge neutrality, the substitution of titanium ions by iron of a lower oxidation state may lead to the formation of compensating defects, such as oxygen vacancies. In both cases, O 1s core level is cen- Figure 5. Deconvolution of XPS O 1s spectra of (a) pure TiO 2 , where features A, B, C and D are visible and correspond to lattice oxygen, undercoordinated oxygen either at the surface or close to oxygen vacancies and adsorbed water and organic species, respectively, and (b) 5% Fe-TiO 2 nanostructures with PEG. Ti 2p peak is also shown in (c) for pure TiO 2 and 5% Fe-TiO 2 nanostructures with PEG.
In Figure 5a,b, four peaks can be observed that correspond to lattice oxygen (A), undercoordinated oxygen either at the surface or close to oxygen vacancies (B), and adsorbed Nanomaterials 2022, 12, 1005 9 of 23 water and organic species (C,D) [83]. An increase of the peaks B-D with respect to peak A is observed in the Fe-TiO 2 , compared to TiO 2 . Part of this increase may be ascribed to a higher oxygen vacancy concentration, due to the doping. Due to charge neutrality, the substitution of titanium ions by iron of a lower oxidation state may lead to the formation of compensating defects, such as oxygen vacancies. In both cases, O 1s core level is centered at around 530 eV, which corresponds to the binding energy of the metal oxide (TiO 2 ). A shoulder is observed at around 533 eV and is related to the OH groups at the surface [84]. Regarding Figure 5c, the two components of the Ti 2p core level that arise from spin orbit-splitting [85] can be observed for pure TiO 2 and Fe-TiO 2 nanostructures, both associated with the Ti 4+ oxidation state. The peak positions and the peak separation of the Ti 2p doublet are in good agreement with previous studies of pure TiO 2 nanostructures [86].
A detailed scan of Fe 2p from 5 mol % of Fe-TiO 2 , is shown in Figure 6a. The pronounced peak around 708.6 eV indicates the presence of Fe 2+ , and a set of fitting parameters accounting for the rich satellite structure of oxidized iron is available in the literature [87]. The deconvolution did not converge using the exact relative constraints for the peak parameter of Fe 2+ (FeO) in ref [87]. However, freeing the area of the peak at the highest binding energy led to convergence and a satisfactory fit quality. The intensity of this satellite peak at 5.9 eV from the center of gravity of the main peaks (another indication for Fe 2+ [88]) is sensitive to the polarizability of the environment of the Fe atoms via the extra-atomic relaxation, one of the relaxation processes related to the photoemission event [89,90]. In contrast to FeO, Fe atoms in the Fe-TiO 2 sample were surrounded by TiO 2 , which has about 40-times higher permittivity than FeO [91,92]. This justifies the different relative intensity of the satellite peak compared to FeO [87] and identifies the iron oxidation state as pure Fe 2+ . Photocatalytic reduction of Fe 3+ to Fe 2+ in a TiO 2 matrix was speculated to occur in vacuum under X-ray irradiation [93], which made the quantification of the iron oxidation state in the original sample unreliable. Since the reduction of iron is the first part of the photocatalytic mechanism explained below, the detection of only Fe 2+ by XPS corroborates the high photocatalytic activity of the sample. which has about 40-times higher permittivity than FeO [91,92]. This justifies the different relative intensity of the satellite peak compared to FeO [87] and identifies the iron oxidation state as pure Fe 2+ . Photocatalytic reduction of Fe 3+ to Fe 2+ in a TiO2 matrix was speculated to occur in vacuum under X-ray irradiation [93], which made the quantification of the iron oxidation state in the original sample unreliable. Since the reduction of iron is the first part of the photocatalytic mechanism explained below, the detection of only Fe 2+ by XPS corroborates the high photocatalytic activity of the sample. In Figure 6b the XPS valence bands of pure TiO2 and Fe-TiO2 nanostructures are represented. No binding energy reference was applied to these spectra (C 1s was measured at 285.8 eV). For Fe-TiO2 nanostructures, four features (A, B, C, and D) are visible at binding energy positions of ~7, 5, 2.5, and 1 eV, whereas pure TiO2 only shows features A and B. These values are consistent with the ones found in the literature [94]. Both A and B features are related to O 2p derived states and correspond respectively to the 'bonding' and 'non-bonding orbital emissions' of TiO2. Feature D is potentially related to the defect state of Ti 3+ 3d. This would support the conclusion above, of a photocatalytic reduction in the XPS chamber [93]. At around 2.5 eV, another feature starts to appear with Fe doping (feature C), related to mixed Fe 3d and Ti 3d derived states very close to the Fermi level In Figure 6b the XPS valence bands of pure TiO 2 and Fe-TiO 2 nanostructures are represented. No binding energy reference was applied to these spectra (C 1 s was measured at 285.8 eV). For Fe-TiO 2 nanostructures, four features (A, B, C, and D) are visible at binding energy positions of~7, 5, 2.5, and 1 eV, whereas pure TiO 2 only shows features A and B. These values are consistent with the ones found in the literature [94]. Both A and B features are related to O 2p derived states and correspond respectively to the 'bonding' and 'non-bonding orbital emissions' of TiO 2 . Feature D is potentially related to the defect state of Ti 3+ 3d. This would support the conclusion above, of a photocatalytic reduction in the XPS chamber [93]. At around 2.5 eV, another feature starts to appear with Fe doping (feature C), related to mixed Fe 3d and Ti 3d derived states very close to the Fermi level [84,94,95]. As seen in Figure 6b, pure TiO 2 has a valence band maximum of~2.5 eV, while for Fe-TiO 2 , a new state was formed within the band gap, with an edge of maximum energy at~−0.07 eV. Additionally, Fe addition led to a shift towards lower binding energies for the maximum edge of the original valence band of TiO 2 (from 2.7 eV for pure TiO 2 to 2.5 eV for 5 mol% of Fe-TiO 2 ) [96]. A similar shift was previously reported [94].

Optical Characterization
To investigate the optical absorption characteristics of pure TiO 2 and Fe-TiO 2 nanostructures with PEG, UV-VIS absorption spectra were recorded (Figure 7a). Pure TiO 2 shows an absorption peak at 354 nm (~3.5 eV) and with an onset around 400 nm (~3.1 eV), which is fairly in line with the anatase band gap excitation [97,98]. Moreover, undoped TiO 2 also did not show any absorption beyond 400 nm. Meanwhile, for Fe-TiO 2 , even though the absorption maximum is placed at the same value obtained for pure TiO 2 (which was also previously reported [99]), the onset of absorption is seen to shift towards higher wavelengths, due to the increasing contribution from a tail of states that extends from the absorption maximum. This observation is likely related to the higher density of defect states present in these materials, contributing to the visible light absorption. Such an increase accompanies the increase in the Fe%. This could, effectively, indicate that the visible light absorption of TiO 2 has been enhanced by the introduction of Fe. Moreover, a small band can be distinguished at around 476 nm, becoming more pronounced with higher concentrations of iron. According to the literature, this visible light absorption can be attributed to two factors: One is the formation of a dopant energy level (Fe 3+ /Fe 4+ ) within the band gap of TiO 2 , which is related to the excitation of 3d electrons of Fe 3+ from the dopant energy level to the TiO 2 conduction band at 415 nm. The second can be ascribed to the d-d transition of Fe(III) or the charge transfer transition between interacting iron ions (broad band at around 500 nm) [44,100]. Considering the XPS result discussed above, the introduction of new intermediate levels at the nanoparticles' surface may be a likely explanation for the emergence of this new absorption band. These levels can act as traps and reduce the rate of recombination [82,101]. Besides that, a further reduction of the nanoparticles' size could accentuate their contribution from surface defects, due to the increase in the surface/volume ratio. The enhancement of visible light absorption can also be seen from the inset of Figure 7a. The powder color changed from white (pure TiO 2 on the left side) to yellow (Fe-TiO 2 on the right side). All of these reasons may, thus, contribute to an improvement in the photocatalytic efficiency of the nanomaterials under visible light.  [82,101]. Besides that, a further reduction of the nanoparticles' size could accentuate their contribution from surface defects, due to the increase in the surface/volume ratio. The enhancement of visible light absorption can also be seen from the inset of Figure 7a. The powder color changed from white (pure TiO2 on the left side) to yellow (Fe-TiO2 on the right side). All of these reasons may, thus, contribute to an improvement in the photocatalytic efficiency of the nanomaterials under visible light.  (1) and (2):  (1) and (2):

The optical band gap energy (E g ) values were estimated based on the UV-VIS diffused reflectance and can be expressed by a Tauc equation and Kubelka-Munk function, represented in Equations
where R is the reflectance of an infinitely thick specimen, K and S are the absorption and scattering coefficients, respectively [102], hv is the photon energy, A is a proportionality constant, and m is a constant exponent, which determines the type of optical transitions (m = 2 is for direct allowed transitions, and m = 1/2 for indirect transitions). The band gap values can be obtained by extrapolating the linear part of the plots relating (F(R)hv) m and hv to (F(R)hv) m = 0 [56,102]. Anatase TiO 2 has an indirect band gap, and so m = 1/2. Figure 7b shows a band gap reduction with an increase of iron concentration, when compared to the pure TiO 2 band gap. The obtained band gap values were 3.1 eV for pure TiO 2 , decreased to 2.9 eV for 1 mol% Fe-TiO 2 , followed by 2.8 and 2.7 eV for the 2 and 5 mol% Fe-TiO 2 nanopowders, respectively. An acceptable agreement was found with earlier studies, in which TiO 2 anatase was reported to have an optical band gap of around 3.2 eV, while for the TiO 2 doped with iron, the value gradually decreased when increasing the dopant concentration [103]. This reduction is probably associated with the formation of new intermediate energy levels [82].

Characterization of the Impregnated Substrates
As previously mentioned, the as-synthesized nanopowders were impregnated on porous water filters, to aid in the recovery of the nanosized photocatalysts. Pure TiO 2 and 5 mol% of Fe-TiO 2 nanopowders were incorporated on the substrates for further photocatalytic experiments. Figure 8 shows the SEM images of the pristine porous substrates, together with the impregnated ones. From Figure 8, it is possible to compare the differences between the substrates, without ( Figure 8a) and with the nanopowders (Figure 8b,c). The substrate presents a micro-sized porosity, which remained after impregnation. Regarding the impregnated substrates, the nanopowders of pure TiO 2 ( Figure 8b) and 5 mol% of Fe-TiO 2 (Figure 8c) formed, by using the drop-casting technique, uniform films of nanoparticles that entirely and uniformly covered the substrates. Some micrometer-sized agglomerates were also observed on the substrate's surfaces. In terms of morphology, and as expected from the SEM analysis of nanopowders, similar nanosized and spherical particles were obtained on the investigated porous substrates in both conditions. Nanomaterials 2022, 12, x 12 of 23 studies, in which TiO2 anatase was reported to have an optical band gap of around 3.2 eV, while for the TiO2 doped with iron, the value gradually decreased when increasing the dopant concentration [103]. This reduction is probably associated with the formation of new intermediate energy levels [82].

Characterization of the Impregnated Substrates
As previously mentioned, the as-synthesized nanopowders were impregnated on porous water filters, to aid in the recovery of the nanosized photocatalysts. Pure TiO2 and 5 mol % of Fe-TiO2 nanopowders were incorporated on the substrates for further photocatalytic experiments. Figure 8 shows the SEM images of the pristine porous substrates, together with the impregnated ones. From Figure 8, it is possible to compare the differences between the substrates, without ( Figure 8a) and with the nanopowders (Figure 8b,c). The substrate presents a micro-sized porosity, which remained after impregnation. Regarding the impregnated substrates, the nanopowders of pure TiO2 (Figure 8b) and 5 mol % of Fe-TiO2 (Figure 8c) formed, by using the drop-casting technique, uniform films of nanoparticles that entirely and uniformly covered the substrates. Some micrometer-sized agglomerates were also observed on the substrate's surfaces. In terms of morphology, and as expected from the SEM analysis of nanopowders, similar nanosized and spherical particles were obtained on the investigated porous substrates in both conditions. EDS analyses were also carried out on these substrates, as shown in Figure S1. Figure  S1a shows the pristine substrates that were only composed of C and O, consistent with a polymeric substrate, Figure S1b,c, respectively. For the impregnated substrates with pure TiO2 nanoparticles ( Figure S1d), a strong presence of Ti ( Figure S1g) is observed, together with C and O ( Figure S1e,f). In the case of impregnated substrates with 5 mol % of Fe-TiO2 nanoparticles (Figure S1h), the presence of Ti ( Figure S1k) along with Fe ( Figure S1l)) is visible, as well as C ( Figure S1i) and O (Figure S1j), which confirms the successful impregnation of the substrates. In EDS analyses, the presence of some micrometer-sized agglomerates on the impregnated substrates was visible, as well as the homogeneous distribution EDS analyses were also carried out on these substrates, as shown in Figure S1. Figure S1a shows the pristine substrates that were only composed of C and O, consistent with a polymeric substrate, Figure S1b,c, respectively. For the impregnated substrates with pure TiO 2 nanoparticles ( Figure S1d), a strong presence of Ti ( Figure S1g) is observed, together with C and O ( Figure S1e,f). In the case of impregnated substrates with 5 mol% of Fe-TiO 2 nanoparticles (Figure S1h), the presence of Ti ( Figure S1k) along with Fe ( Figure S1l)) is visible, as well as C ( Figure S1i) and O ( Figure S1j), which confirms the successful impregnation of the substrates. In EDS analyses, the presence of some micrometer-sized agglomerates on the impregnated substrates was visible, as well as the homogeneous distribution of all elements.
The EDS chemical analysis of the 5 mol % Fe-TiO 2 nanopowders impregnated on the porous substrate is summarized in Table 2. The EDS Fe:Ti atomic ratio obtained was 0.05. This value is in line with the ratio obtained by XPS (0.06).
Raman measurements were also carried out, as this technique allows distinguishing between TiO 2 phases [104] and evaluating the purity of TiO 2 and Fe-TiO 2 nanostructures impregnated on polymeric substrates. The obtained results are presented in Figure 9. The EDS Fe:Ti atomic ratio obtained was 0.05. This value is in line with the ratio obtained by XPS (0.06).
Raman measurements were also carried out, as this technique allows distinguishing between TiO2 phases [104] and evaluating the purity of TiO2 and Fe-TiO2 nanostructures impregnated on polymeric substrates. The obtained results are presented in Figure 9. The Raman spectrum of the pristine substrate is also shown for comparison, demonstrating that there was no contribution associated to the substrate. In the impregnated substrates, five bands corresponding to six active modes of the tetragonal anatase phase can be observed in Figure 9, i.e., 144 cm −1 (Eg), 198 cm −1 (Eg), 393 cm −1 (B1g), 515 cm −1 (B1g + A1g), and 636 cm −1 (Eg) [52,105,106]. The Eg mode is ascribed to symmetric stretching vibration in octahedral TiO6 clusters, while the B1g mode is related to symmetric bending vibration in the same clusters. The A1g mode is assigned to anti-symmetric bending vibra- The Raman spectrum of the pristine substrate is also shown for comparison, demonstrating that there was no contribution associated to the substrate. In the impregnated substrates, five bands corresponding to six active modes of the tetragonal anatase phase can be observed in Figure 9, i.e., 144 cm −1 (E g ), 198 cm −1 (E g ), 393 cm −1 (B 1g ), 515 cm −1 (B 1g + A 1g ), and 636 cm −1 (E g ) [52,105,106]. The Eg mode is ascribed to symmetric stretching vibration in octahedral TiO 6 clusters, while the B 1g mode is related to symmetric bending vibration in the same clusters. The A 1g mode is assigned to anti-symmetric bending vibration and it is reported that the E g mode at 636 cm −1 is attributed to the displacement of symmetric oxygen atoms in Ti-O bonds in the x,y-plane [107,108]. No additional peaks related to iron or iron oxides were detected in the Fe-TiO 2 impregnated substrate, which corroborates the results of XRD.

Photocatalytic Performance
Absorbance spectra were acquired to evaluate the photocatalytic activity of the impregnated substrates in the degradation of RhB under solar radiation at room temperature ( Figure 10). The absorption peak intensity of RhB (occurring at 554 nm [109]) was measured at different irradiation times, to estimate the decrease of RhB content. The photodegradation rate of the RhB can be calculated using the following Equation (3): where A 0 is the light absorbance of the RhB solution before irradiation and after absorptiondesorption equilibrium in the dark, and A is the light absorbance of the RhB solution after irradiation [110,111]. where A 0 is the light absorbance of the RhB solution before irradiation and after absorption-desorption equilibrium in the dark, and A is the light absorbance of the RhB solution after irradiation [110,111]. According to Figure 10a, without a catalyst on the surface of the polymeric substrates, some degradation of the dye was observed, both in the absence and presence of light. As reported in the product description, these polymeric membranes can also be used in aqueous solutions as filters. They possess a high porosity percentage and, thus, act like hydrophilic 'sponges' that can easily capture the RhB dye molecules on the pore sites in a few hours (63% of degradation was achieved in 5 h). This retention of RhB dye molecules is also clearly observed in Figure S2, with the color change of the substrate before (from white, Figure S2a,b) and after photocatalysis (to fuchsia pink, Figure S2c). Nevertheless, since all experiments were conducted with the same substrate, this contribution will be equal for all the materials studied [52]. In the presence of pure TiO2 (Figure 10b), a higher According to Figure 10a, without a catalyst on the surface of the polymeric substrates, some degradation of the dye was observed, both in the absence and presence of light. As reported in the product description, these polymeric membranes can also be used in aqueous solutions as filters. They possess a high porosity percentage and, thus, act like hydrophilic 'sponges' that can easily capture the RhB dye molecules on the pore sites in a few hours (63% of degradation was achieved in 5 h). This retention of RhB dye molecules is also clearly observed in Figure S2, with the color change of the substrate before (from white, Figure S2a,b) and after photocatalysis (to fuchsia pink, Figure S2c). Nevertheless, since all experiments were conducted with the same substrate, this contribution will be equal for all the materials studied [52]. In the presence of pure TiO 2 (Figure 10b), a higher degradation percentage of 87% was achieved for the same time exposure as the substrate without catalyst. On the other hand, the addition of 5 mol% of Fe drastically improved the photocatalytic activity of TiO 2 . In 3.5 h, a RhB degradation of 85% was obtained, whereas for pure TiO 2 , the RhB degradation was 74% for the same exposure time.
As mentioned before, various factors can influence photocatalytic activity, such as the crystalline structure, morphology, porosity, particle size, iron content, and the presence of surface defects [53,112]. Regarding the contribution of the particles' shape and size to the enhanced photocatalytic behavior observed, and as assessed by the SEM images in Figure 2 and estimated from the XRD results, all conditions resulted in spherical nanoparticles in the same size range (around~18 nm from SEM measurements). Since both the size and shape of the nanoparticles are comparable, similar contributions are expected for all materials; thus, this does not explain the observed behavior for the 5% Fe-TiO 2 material.
Moreover, in terms of the TiO 2 phase present, it is well known that anatase TiO 2 shows superior photocatalytic activity to rutile or brookite [113,114]. It has been reported that anatase exhibits a longer lifetime of photogenerated electrons and holes than rutile and brookite [115]. Additionally, the average effective mass of electrons and holes in anatase is smaller than that of rutile and brookite, which favors a faster migration of photogenerated charge carriers from the interior to the surface of anatase; thus, reducing the recombination rate and improving the photocatalytic activity [115]. As revealed by the XRD and Raman spectroscopy results, all synthesized materials have a anatase TiO 2 phase, hence the photocatalytic degradation enhancement of the substrate with Fe-TiO 2 must be owing to the introduction of iron ions into the TiO 2 lattice.
The enhanced photocatalytic behavior observed with the Fe addition may be ascribed to the mechanism of the photocatalytic processes in Fe-TiO 2 material. It is reported that there are two energy levels: the reduction level, which is located below the TiO 2 conduction band (Fe 3+ /Fe 2+ ), and another one, the oxidation level, above the TiO 2 valence band (Fe 3+ /Fe 4+ ). First, the formation of Fe 2+ species occurs due to the migration of photogenerated electrons from TiO 2 to Fe 3+ . These Fe 2+ species are unstable, owing to the loss of the d 5 electronic configuration, and easily oxidize to Fe 3+ by transferring electrons to absorbed O 2 and forming reactive superoxide anions (O 2 − ). Meanwhile, Fe 3+ ions can act as a hole trap, since the Fe 3+ /Fe 4+ energy level is above the TiO 2 valence band and oxidize to Fe 4+ . Fe 4+ then reduces to Fe 3+ by reacting with an OH − group, and hydroxyl radicals (•OH) are formed. These hydroxyl radicals are powerful oxidizing species, which attack the chemical bonds of the surface-adsorbed organic materials. Therefore, Fe 3+ ions can act as trap sites for the photogenerated electrons and holes; thus, suppressing the recombination of those photogenerated charges and ultimately enhancing the photocatalytic activity under visible light [35,56].
However, an important note should be made regarding the amount of iron content that plays an important role in photocatalytic activity [103]. Some studies reported, for instance, a reduction in the photocatalytic activity when the dopant concentration increases. It is stated that multiple trappings of charge carriers take place, which in consequence increases the electron-hole recombination. Thus, fewer charge carriers will be able to reach the surface to degrade the pollutant. Moreover, an excessive concentration of dopant may accumulate on the surface of the catalyst, reducing the penetration depth of light and, consequently, the number of active sites [35]. Nevertheless, this excessive dopant trend does not seem to occur in this study, since improved photocatalytic activity was demonstrated relative to pure TiO 2 .
The generation of intermediate energy levels, as a result of the presence of Fe ions was also confirmed by XPS, which could have served as trapping centers for the photogenerated carriers. As revealed by XPS, the valence state of Fe ions is lower than that of lattice Ti ions. Considering charge balance, oxygen vacancies are likely to have existed in the Fe-TiO 2 sample and are also expected to play important roles in enhanced photocatalysis, since they can act as photoinduced charge traps and adsorption sites; thus, contributing to increasing the lifetime of photogenerated charge carriers and resulting in an improvement of the photocatalytic performance [116][117][118].
Comparing the curves in Figure 11a, it can also be confirmed that the photodegradation was much faster with the Fe-TiO 2 substrate. The reaction kinetics were also investigated using the Langmuir-Hinshelwood kinetic model. This model can be simplified to a firstorder kinetics reaction (4): where k ap is the photodegradation apparent rate constant, t is the time, C 0 is the initial concentration, and C is the concentration at a certain time [111,112]. According to the Lambert-Beer law, the concentration is proportional to the absorbance; thus, it can be Based on (4), the rate constants (min −1 ) can be determined by plottin ln or ln versus time (Figure 11b) [119]. From the slope of the linear regre sions, the rate constants can be obtained. The obtained kinetic parameters (rate constant linear regression coefficients, and half-life times) are summarized in Table 3. From Based on (4), the rate constants k ap (min −1 ) can be determined by plotting − ln C C 0 or − ln A A 0 versus time (Figure 11b) [119]. From the slope of the linear regressions, the rate constants can be obtained. The obtained kinetic parameters (rate constants, linear regression coefficients, and half-life times) are summarized in Table 3. From Figure 11b and Table 3, it can be concluded that the photocatalytic dye degradation follows the first order-kinetics for both TiO 2 and Fe-TiO 2 samples, as the correlation constant (R 2 ) for the fitted lines is above 0.95 [111]. The obtained photodegradation apparent rate constants were found to be 0.007 min −1 and 0.01 min −1 for the substrates with pure TiO 2 and 5% Fe-TiO 2 nanostructures, respectively. The obtained k ap values indicate that the RhB photodegradation with Fe-TiO 2 materials was 1.5-times faster than pure TiO 2 .  98 69 In contrast, the pristine substrate did not reveal a good fit for the pseudo-first-order kinetics equation. In this case, the data showed a satisfactory fitting (R 2 = 0.99), by using a pseudo-second order equation. Equation (5) indicates that the reaction rate is proportional to the concentration of the reactant: Based on (5), the rate constant can be obtained by a linear fit, through the plot of 1 C − 1 C 0 versus time [120]. Figure 11c shows the fitting of the pristine substrate using a pseudosecond order equation. This kinetic reaction trend can be explained by the fast absorption intensity of RhB during the first 120 min, followed by a decrease and stabilization of absorption. Since kinetic rate coefficients are impossible to compare for different pseudoorder processes, half-life times (t 1/2 ) were determined. Half-life times are defined by the time it takes for the concentration of a reactant to reach half of its initial value [121], and can be determined using Equations (6) and (7): For n a = 1: where n a is the apparent or pseudo reaction order, A t is the ratio of the concentration at a given time (t) to initial concentration (C/C 0 ), thus A 0 = 1 [121,122]. As expected, the kinetic studies confirm that Fe-TiO 2 impregnated substrate exhibited the highest photocatalytic degradation of RhB, when compared to pristine and TiO 2impregnated substrates under the same experimental conditions, since it achieved the lowest half-life time.

Reusability Tests
Reusability tests of the photocatalytic materials are of great importance, to check their applicability in real wastewater treatments. To evaluate the possibility of reuse, the best photocatalyst, i.e., 5 mol% of Fe-TiO 2 photocatalyst on the porous substrate, was chosen for the experiments. Five consecutive cycles of RhB photocatalytic degradation were performed under the same experimental conditions and up to 210 min (3.5 h). On Figure S3, the reusability of the pristine substrate is also presented (up to 3 cycles).
In Figure 12a,b, a decrease in the photocatalytic activity is observed. The reaction rate decreased with the number of uses, from 0.01 min −1 in the first cycle, to 0.004 min −1 in the fifth cycle. The reusability results can also be seen in Figure 12c. The impregnated substrate approximately 27% of its degradation efficiency after performing five repeated cy since in the first cycle a dye degradation of 85% was achieved, while in the fifth cyc was 58%. Despite the considerable efficiency loss after three cycles (of around 21% much smaller efficiency loss was observed during the next cycles. This loss could be tributed to the high adsorption phenomenon of RhB, where RhB molecules remain sorbed on the photocatalyst surface, hindering the available pore sites of the subst with the nanostructures for reaction [123,124]. The reusability tests of the pristine s strate confirmed that a significant contribution to this loss of degradation efficiency co from the substrate, with the adsorption of RhB molecules through the cycling tests Figure S3). This phenomenon also occurs on the impregnated substrates decreasing t overall cycling efficiency. The reusability results can also be seen in Figure 12c. The impregnated substrate lost approximately 27% of its degradation efficiency after performing five repeated cycles, since in the first cycle a dye degradation of 85% was achieved, while in the fifth cycle it was 58%. Despite the considerable efficiency loss after three cycles (of around 21%), a much smaller efficiency loss was observed during the next cycles. This loss could be attributed to the high adsorption phenomenon of RhB, where RhB molecules remain adsorbed on the photocatalyst surface, hindering the available pore sites of the substrate with the nanostructures for reaction [123,124]. The reusability tests of the pristine substrate confirmed that a significant contribution to this loss of degradation efficiency comes from the substrate, with the adsorption of RhB molecules through the cycling tests (see Figure S3). This phenomenon also occurs on the impregnated substrates decreasing their overall cycling efficiency.

Conclusions
Pure TiO 2 and Fe-TiO 2 nanostructures were successfully synthesized by a fast surfactantassisted microwave irradiation (1 h), without a calcination step, and impregnated by drop-casting on porous polymeric substrates. The approach used in this study enabled the total covering of the porous substrates and the evaluation of their photocatalytic activity in the degradation of RhB under solar radiation. SEM confirmed the formation of fine particles with a sphere-like appearance and films that uniformly coated the substrates. XRD revealed the presence of pure TiO 2 anatase in all nanostructures, which was further confirmed by Raman spectroscopy on the impregnated substrates. The platforms impregnated with 5 mol% of Fe-TiO 2 nanostructures exhibited an enhanced RhB photodegradation, when compared to pure TiO 2 and the pristine substrates. The highest photocatalytic activity for the Fe-TiO 2 material under solar radiation reached 85% after 3.5 h, compared to 74% with pure TiO 2 . The photodegradation rate of RhB dye with the Fe-TiO 2 substrate was 1.5-times faster than pure TiO 2 . The XPS and UV-VIS results support that the presence of Fe ions led to the introduction of new energy levels, as well as defects, such as oxygen vacancies, that played an important role as traps for the photogenerated carriers, leading to a reduction in the recombination rate, followed by a visible enhancement in the photocatalytic activity.
In summary, this study demonstrated that the synergy between the micro-porosity of the substrates and the surface-modified Fe-TiO 2 nanostructures enhanced the substrates' photocatalytic properties. Flexible and eco-friendly photocatalytic functionalized substrates were produced with potential for wastewater removal.
Supplementary Materials: The following supporting information can be downloaded at: https: //www.mdpi.com/article/10.3390/nano12061005/s1, Figure S1: SEM images of the impregnated polymeric substrates (a) pristine (without nanopowders) (d) with pure TiO 2 and (h) with 5% Fe-TiO 2 nanopowders with the addition of PEG. The corresponding EDS maps of C (b,e,i), O (c,f,j), Ti (g,k) and Fe (l) are also visible; Figure S2: Porous polymeric substrates (water filters) used in this study (a) pristine (b) impregnated substrate before photocatalysis and (c) impregnated substrate after photocatalysis; Figure S3: Reusability of pristine substrate over 3 degradation cycles under solar simulator light.
Author Contributions: M.L.M. was responsible for producing and characterizing the materials, as well as writing the manuscript. A.P. and D.N. were responsible for the overall scientific orientation and revising the manuscript. A.S.R.-M. and J.R. participated in the review and editing. J.D. performed the XPS measurements. E.F. and R.M. were responsible for supervising all the process and getting funding. All authors have read and agreed to the published version of the manuscript.
Funding: This work was financed by national funds from FCT-Fundação para a Ciência e a Tecnologia, I.P., in the scope of the projects LA/P/0037/2020, UIDP/50025/2020, and UIDB/50025/2020 of the Associate Laboratory Institute of Nanostructures, Nanomodelling, and Nanofabrication-i3N. The authors also acknowledge Fundação para a Ciência e a Tecnologia for funding the Project ICARUS under the reference PTDC/EAM-AMB/30989/2017. The work was also partially funded by the Nanomark collaborative project between INCM (Imprensa Nacional-Casa da Moeda) and CEN-IMAT/i3N. Acknowledgments also go to the EC project SYNERGY H2020-WIDESPREAD-2020-5, CSA, proposal No. 952169, and to the European Community's H2020 program under grant agreement No. 787410 (ERC-2018-AdG DIGISMART).

Data Availability Statement:
The authors confirm that the data supporting the findings of this study are available within the article and its Supplementary Materials.

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