Envirronmentally friendly fabrication of Fe₂TiO₅-TiO₂ nanocomposite for enhanced photodegradation of cinnamic acid solution

The XRD pattern of the FTO samples (figure 1(a)) shows diffraction peaks of pseudobrookite Fe₂TiO₅ that appear at 2θ = 18.2°; 25.7°; 27.6°; 32.7°; 37.5°; 40.6°; 46.3°; 49.0°; 52.5°; and 60.1° (JCPDS#41-1432). The peaks at 18.2°; 25.7°; 27.6°; 32.7°; 37.5°; 52.5°; and 60.1° correspond to the (200), (101), (230), (131), (060) and (531) lattice directions of the pseudobrookite Fe₂TiO₅ structure, respectively. Figure 1 shows the appearance of peaks of the TiO₂ anatase phase at 2θ = 25.5°; 36.7°; and 49.3° (JCPDS#21-1272) and the rutile phase at 2θ = 27.3°; 36.1°; 41.5°; 54.5°; 56.4°; 61.9°; 62.7°; 68.4°; and 69.6° (JCPDS#21-1276), but no peaks of Fe₂O₃ were observed. Meanwhile the composition of TiO₂ (w) synthesised in a water medium is pure anatase (2θ = 25.6°, 38.2°, 48.3° and 54.7°) with a crystal size of 34.8 nm, as shown in previous research [10]. For the FTO-Ti samples, XRD analysis shows that the Fe₂TiO₅ peaks appear with the strongest intensity at 2θ = 25.1°, the TiO₂ anatase peak at 25.5°, and the very weak intensity peaks of the rutile phase at 27.2° and 36.3°. There are slightly right peak shifts for the Fe₂TiO₅ pseudobrookite phase in the hybrid sample compared with the pure Fe₂TiO₅, which shows that in general, the crystal structures of TiO₂ and FTO are conserved in the hybrid system and there is only a weak interaction between them.

Zoom In Zoom Out Reset image size

The intensity of the peaks for FTO and TiO₂ in the composite is much weaker than for the parent FTO and TiO₂, which shows that when adding FTO in combination with TiO₂, the crystallisation of both TiO₂ and FTO is poor. However, as can be seen in figure 1, with increasing FTO content, the intensity of the characteristic peaks of FTO increases; this indicates that the FTO crystals exist independently in the sample but are split into small particles by TiO₂ crystals. This result matches a previous report [32] when assuming that Fe₂TiO₅ is maintained when combined with TiO₂. The reduction in crystallisation of the anatase phase in the heterostructured samples is often accompanied by an increase in defects, which is beneficial to obstructing the recombination of photogenerated charges [43].

On the XRD pattern, the most intensive peaks of TiO₂ anatase (25.5°) and FTO (25.7°) almost overlap, and so the existence of TiO₂ anatase phase cannot be confirmed. Additional information regarding the crystalline structure was obtained through Raman spectroscopy (figure 1(b)), which shows the characteristic peaks for the anatase phase at wave numbers 147, 200, 395, 514, and 647 cm^–1 [44]. These vertices include the E_g peaks at 147, 200, and 647 cm^–1, A_g at 514 cm^–1, and B_1g at 395 cm^–1. Asymmetric stretching oscillations are noted as E_g , while asymmetrical and symmetrical flexural vibrations are seen as peaks B_1g and A_g . Meanwhile, on the Raman spectrum of the Fe₂TiO₅ sample, a characteristic peak for TiO₂ in the rutile phase appears weakly at 612 cm^–1 [45]. In addition, several Raman peaks characteristic of Fe₂TiO₅ could be observed at 200, 223, 330, 407 and 657 cm^–1 (ID R140974 spectra from the RRUFF database) [46, 47]. No peak for FeTiO₃ appears in the Raman spectroscopy, showing that Fe in catalysis exists solely as pseudobrookite. It can therefore be concluded that during the synthesis of pseudobrookite, most of the anatase remains, but some is transformed to rutile, confirming the XRD analysis.

The FTIR spectra of the catalysts (figure 1(c)) exhibit fluctuations at 3406 cm^–1 for the surface OH group and at about 1633 cm^{– 1} for the H-OH group. In addition, two absorption bands at 818 cm^–1 and 518 cm^–1 could be attributed to vibrations in the (FeO₆) octahedra [27]. The absorption band of the oscillation of the Ti-O bond can be seen at 675 cm^–1. The absorption peak at around 432 cm⁻¹ could be due to the vertically aligned Ti–O stretching vibrations of the (TiO₆) octahedra [48]. All the results confirm the presence of TiO₂ and pseudobrookite Fe₂TiO₅ in the catalyst [49]. In addition, the signal strength of the OH and H-OH groups as well as the Fe-O and Ti-O bonds of the Fe₂TiO₅-TiO₂ hybrid samples is much stronger compared to pure Fe₂TiO₅. As is known, the surface hydroxyl groups and adsorbed water could be trapped by the photogenerated holes to form hydroxyl radicals and minimise the electron hole recombination that enhances photocatalytic activity. Therefore, it is likely that the hybrid catalyst will exhibit higher activity in CA photodegradation.

The Raman, XRD and FTIR analysis results confirm the existence of TiO₂ anatase and Fe₂TiO₅ phases; however, their crystallinity is poor compared with single-component samples. In the hybrid samples, the Fe₂TiO₅ and TiO₂ crystals are interlaced and divided from each other to form highly dispersed particles. Indeed, the crystal size of TiO₂ anatase decreases from 34.8 nm in the pure TiO₂ sample to 28.5–31.5 nm in the hybrid sample. The crystal size of FTO follows the same trend, decreasing from 31.6 nm in pure FTO to 26.4–27.1 nm in FTO-Ti catalyst, as shown in table 1. The general trend that can be seen in table 1 is that the crystal sizes of both the TiO₂ anatase and Fe₂TiO₅ phases increase as the FTO content in the composite increases. Specifically, the crystal sizes of TiO₂ are 28.5, 29.7 and 31.5 nm, and the FTO crystal sizes are 26.4, 26.9 and 27.1 nm, respectively, for the 10FTO-Ti, 20FTO-Ti, and 30FTO-Ti samples.

All the samples, including the hybrid catalysts, are single-pore mesoporous materials with pore diameters ranging from 10 to 60 nm, as confirmed from pore size distribution analysis (figure S1 (available online at stacks.iop.org/ANSN/12/045015/mmedia)). The average pore diameter of the hybrid sample is approximately 20 nm, slightly smaller than that of TiO₂ (22.4 nm) but slightly larger than that of FTO (18 nm). This property proves that the heterostructure material consists of tightly connected FTO and TiO₂ particles that form a unified system. The similarity in structural properties of Fe₂TiO₅ and TiO₂ facilitates the growth of the lattice-matched Fe₂TiO₅ on TiO₂ [37]. The pore size of the catalysts is larger than the molecular size of cinnamic acid (0.499 nm × 0.945 nm) [50], and this enables CA diffusion deep into the pores.

The SEM images (figure S2) show that the distribution of the catalysts is relatively uniform. The rhombic shape nanoparticles of the Fe₂TiO₅ sample are about 20 nm in size. Agglomerates of different sizes consisting of smaller particles can be observed in SEM (figure S2) and HRTEM images (figure 2). The bright-field image in figure 2(a) indicates that the FTO sample consists of three or four closely accumulated planes. In addition, HRTEM shows some other crystals with larger sizes, possibly of TiO₂ anatase, and rutile.

Zoom In Zoom Out Reset image size

With the Fe₂TiO₅-TiO₂ samples, the particles become more porous and discrete, with sizes ranging from 20 to 35 nm, as can be seen in the SEM images (figure S2). This phenomenon could be due to the coalescence of small particles of Fe₂TiO₅ and TiO₂ into agglomerates [51]. The morphology is confirmed by HRTEM (figure 2(b)), which shows that the Fe₂TiO₅ and TiO₂ particles are closely intermingled in agglomerates. The HRTEM image of the 10FTO-Ti sample indicates the formation of anatase, rutile, and Fe₂TiO₅ phases (figure 2(b)). The inset in figure 2(b) demonstrates that TiO₂ has three (101), (200), and (004) planes of structured anatase with d-spacings of 0.35, 0.31, and 0.33 nm, respectively. The particle size of anatase is estimated to be in the range 10–15 nm. The (111) plane is the majority in the structure of rutile TiO₂ with a d-spacing of 0.29 nm, reaching 15–20 nm in nanoparticle size. In addition, the HRTEM images indicate that the formation of particles of Fe₂TiO₅ was exposed to anatase and rutile of TiO₂. The estimated d-spacings of 0.25 and 0.21 nm are assigned to (111) and (023) crystal planes, respectively, of the Fe₂TiO₅ phase in the 10FTO-Ti sample. The particle size of Fe₂TiO₅ is estimated to be in the range of 8–10 nm. The surface morphology results from the HRTEM method of the catalyst are fairly well matched with the XRD spectrum of the catalysts. The particle size of TiO₂ and Fe₂TiO₅ inferred from HRTEM images was nearly two times smaller than their crystal size, as determined from the XRD pattern. This result confirms the previous observation that there is a random distribution of TiO₂ and FTO particles, leading to the shredding of their crystals. The displacement of Ti⁴⁺ ions by Fe³⁺ ions reduces the overall positive charge in the Fe₂TiO₅ crystal structure. The O²⁻ negative ions decrease accordingly to compensate for the lost positive charge; the oxygen sites will then be absent, resulting in a loss of Fe-O bonds. This decrease will reduce the overall size of the lattice, thereby significantly reducing the particle size. The HRTEM results confirm the close intermixing distribution of TiO₂ and Fe₂TiO₅ particles in the nanocomposite. Furthermore, HRTEM images demonstrate the existence of Fe₂TiO₅, anatase, and rutile TiO₂ phases in this composite material.

The specific surface area and pore volume of pure TiO₂ and FTO samples are low, at 11.5 m² g⁻¹ and 0.007 cm³ g⁻¹ (as shown in table 1), respectively, for FTO, and 13.8 m² g⁻¹ and 0.01 cm³ g⁻¹ for pure TiO₂. An approximate specific surface value of 11.8 m² g⁻¹ for Fe₂TiO₅ was reported by Vasiljevic et al. [52]. By combining FTO with TiO₂, these parameters are enhanced significantly. The parameters for the 10FTO-Ti and 20FTO-Ti samples are not very different. Their specific surface area is 96.0 and 86.3 m² g⁻¹, while the pore volume reaches 0.059 and 0.053 cm³ g⁻¹, respectively (6–8 times higher than the precursors). The sharp increase in specific surface area may be associated with a decrease in crystal size and heterostructure of the hybrid catalyst. The large specific surface area is also beneficial in improving the photocatalytic activity of the semiconductor [36] as there are more active sites on the surface for the reactants.

The elemental composition of FTO and FTO-Ti was determined by EDS analysis. The EDS spectrum of FTO is shown in figure S3(a). The mass percentage of elements Fe, Ti, and O of FTO identified from the EDS spectrum is 43.9%, 28.6%, and 27.6%, respectively, and is consistent with the theoretical mass ratio of Fe₂TiO₅ (46.7%, 20.0%, 33.3%). The Ti mass composition of the FTO sample is a little higher relative to the theoretical value, and this is attributed to the small amounts of TiO₂ anatase and rutile in the sample, as shown by the XRD patterns and Raman spectra. The mass compositions of Fe, Ti, and O in 10FTO-Ti sample of 2.1%, 59.0%, and 38.9%, respectively, are also shown by the EDS spectrum (figure S3(b)). This result is also relatively consistent with the theoretical mass ratio of 4.6%, 56.0% and 39.4%. The results confirm the formation of heterostructured composites.

The UV-Vis absorption spectra of the samples are in figure 3. The band gap energy (E_g ) of semiconductors in general, and the FTO-Ti samples in particular, is calculated using the popular method based on a Tauc plot [53]. The light wavelength threshold can be determined by the following equation:

$$\begin{eqnarray}&&{E}_{g}=hc/\lambda =h\nu .\end{eqnarray} \tag{ 1 }$$

Zoom In Zoom Out Reset image size

The absorption wavelength (λ) of pure FTO was identified from the Tauc plot as 653 nm, which corresponds to band gap energy (E_g ) at 1.9 eV and is close to the results reported in various studies (E_g approx. 2.1) [27, 52, 54].

The Tauc plot of samples of catalysts indicates that the E_G of the 10FTO-Ti and 20FTO-Ti samples is alike at 3.05 eV. Further, the absorption wavelength corresponds to 407 nm. The E_G of the FTO-Ti samples is higher than that of the pure FTO sample but lower than that of pure TiO₂ (3.14 eV), leading to the extension of the region of higher photon absorption. The effect of the Fe dopant can be explained by enhancement of the O⁻²p4 → Fe⁻³d5 electronic transition (t2g) [55]. Unlike the FeOx–TiO₂ bi-functional compound, whose UV-VIS spectrum has two light absorption maxima at 382 nm and 526 nm [56], the UV-Vis spectrum of the hybrid material has only one absorption peak with a maximum at 407 nm. This proves that in this study, a uniform heterostructure was formed. The shift of the band of TiO₂ to longer wavelengths with the addition of FeOx involves the incorporation of Fe(III) into the TiO₂ through Ti–Fe charge transfer bands [56].

From the zeta potential versus the pH of the different solutions indicated in figure S4, the PZC of the 10FTO-Ti composite material was determined to be 6.42, which is lower than that of TiO₂(w) (PZC = 7.24) [10] and equal to the PZC of ATO-Ti (6.4) [41].

Figure S5 shows that the removal efficiency of CA by the catalysts increased over time. It can be inferred from figure 4 that TiO₂ is much more active than FTO. With pure TiO₂, the 60 min degradation of CA reaches 67.1% (and X₉₀ = 74.8%); with pure FTO this value is only 5.0% (X₉₀ = 8.5%). However, when replacing 10% TiO₂ with FTO, a hybrid catalyst 10FTO-Ti is created which possesses superior activity. All the hybrid catalysts exhibit higher activity than FTO, and the two catalysts containing 10% and 20% FTO show higher activity than TiO₂. These results underline the effectiveness of the hybridisation of FTO and TiO₂. The heterostructure catalyst offers outstanding features, such as small crystal size, high specific surface area and pore volume, low band gap energy, and suppression of the recombination of photo-generated electrons and holes [40] to enhance photocatalytic activity. However, the catalytic activity of composite catalysts rises with increasing TiO₂ mass, which is the highly active ingredient for CA degradation by the hybrid catalyst. Specifically, when reducing the TiO₂ content in the FTO-Ti samples from 90% to 80%, 70% and 60%, CA removal efficiency reduces: the 60 min conversion of CA (X₆₀) reduces from 76.4% to 69.9%, 57.7% and 47.1%, and the 90 min removal efficiency (X₉₀) decreases from 89.0% to 61.0%, as shown in figures 4 and S5. Therefore, 10FTO-Ti is the best catalyst to explore further in the next section. The higher activity of the 10FTO-Ti catalyst is most likely due to the smallest crystal size, largest specific surface area and pore volume, highest density of the –OH group and adsorbed water, as well as low band gap energy among the hybrid catalysts.

Zoom In Zoom Out Reset image size

Figure 5(a) shows that the conversion of CA increases significantly when the air flow rate is increased from 0.0 to 0.3 l min⁻¹; and 90 min CA conversion increases from 70.6% to 89.0%. However, when increasing air flow rate up to 0.5 l min⁻¹, the CA conversion is reduced to 69.4%. Dissolved oxygen plays an important role in the photocatalytic reaction by ensuring that there are sufficient electron scavengers present to trap the excited conduction-band electrons from the recombination [57]. However, at high air velocity, foaming occurs, and this impedes the absorption of light by the catalyst particles; also, catalyst particles move to the liquid surface and are dispersed in the bubbles. It was found that less of the catalyst participated in the reaction, and this reduced catalytic efficiency [58, 59]. Hence, an air flow rate value of 0.3 l min⁻¹ is the most suitable for CA photodegradation.

Zoom In Zoom Out Reset image size

As can be seen from figure 5(b), when the catalyst dosage increases from 0.5 g l⁻¹ to 0.75 g l⁻¹, CA conversion at all durations increases significantly, and the CA conversion after 90 min increases from 77.3% to 89.0%. In fact, the efficiency of photodegradation is improved with an increase in catalyst concentration. This improvement can be attributed to increasing active sites by providing more catalyst, which is the semiconductor in the reaction [60]. Consequently, the formation of electron hole pairs on the surface of semiconductors and the high reactive hydroxyl radicals increases, which can deeply oxidise all of the organic molecules adsorbed on the surface of a semiconductor to non-toxic species such as CO₂ and H₂O [61]. However, increasing the catalyst dosage further to 1.0 g l⁻¹, the 90 min conversion of CA is almost unchanged and then decreases to 85.2% as the catalyst content reaches 1.25 g l⁻¹. The lower degradation efficiency at higher catalyst dosage is attributed to the high turbidity of the solution. When the amount of catalyst increases above a saturation level, the excess FTO-Ti particles have a light-screening effect that reduces the catalyst surface area exposed to light illumination and the efficiency of photodegradation [62, 63]. Thus, FTO-Ti content of 0.75 g l⁻¹ was chosen as the most suitable.

The value of X₉₀ gradually decreases from 89.0% to 15.9% as the initial pH of the reaction solution gradually increases from 3.8 to 9.0 (figure 5(c)). Chen et al. [64] reported that an acidic environment is more suitable for photooxidation, and this can be explained based on the results of the point of zero charges of the 10FTO-Ti catalyst (being 6.42). At a solution pH that is less than the PZC, the catalytic surface will be positively charged with H⁺ ions, which should be able to combine with O₂- radicals to generate ${{\rm{HO}}}_{2}^{\cdot }$ radicals [65, 66]. On the other hand, according to the authors [67], an alkaline environment with a high pH significantly reduces the transmission capacity of ions in the reactive solution and reduces the possibility of beneficial free radicals forming. Thus, a pH value of 3.8 supports photodegradation with the 10FTO-Ti catalyst. In summary, the optimum conditions for the 10FTO-Ti catalyst in degrading CA comprised an initial solution pH of 3.8, an air flow rate of 0.3 l min⁻¹, and a catalyst dosage of 0.75 g l⁻¹. In these conditions, 89.0% of the CA is converted after 90 min.

The activity of the 10FTO-Ti catalyst in the photocatalytic degradation of CA gradually decreases in successive reaction batches (figure 6). During the first eight cycles, CA conversion decreased by several percentage points each cycle, and X₉₀ decreased in total by 18.8%, from 89.0% to 70.2% after the first eight cycles. However, in the last two cycles, the value of X₉₀ dropped to 61.1%. The decrease in catalytic activity is due to the adsorption of the agents and the byproducts of the photochemical process onto the catalyst surface [68]. These surfaces become covered, reducing the adsorption of the substances as well as the potential for exposure to the activated photons required for the reaction. The XRD results of the spent catalyst (figure S6) after the reaction show that the crystalline phases of FTO and TiO₂ anatase still exist. However, the crystallinity and intensity of this sample are lower than those of the fresh catalyst.

Zoom In Zoom Out Reset image size

To date, few publications have considered the activity of titanate-TiO₂ composite for the photodegradation of POPs in water. In our previous studies, heterostructure Al₂TiO₅-TiO₂ catalysts were synthesised by a sol-gel-based hydrothermal method in acidic (ATO-Ti(a)) [42] and neutral conditions (ATO-Ti(w)) [41] with different TiO₂ contents. Compared with the corresponding bare titanium oxide (TiO₂) as well as with Al₂TiO₅ (ATO), the hybrid catalysts of optimal composition [33ATO-Ti(a) and 67ATO-Ti(w)] were found to exhibit enhanced photocatalytic activity in CA decomposition under ultraviolet light. In the most favourable conditions, the CA removal efficiency of 33ATO-Ti(a) and 67ATO-Ti(w) is 88.5% and 70.1% after 60 min, respectively, which is much higher than the corresponding Al₂TiO₅ samples (30.4% and 38.0%, respectively) and comparable to the corresponding TiO₂ (90.9% and 67.1%). The hybrid catalyst prepared in a neutral medium [67ATO-Ti(w)] has slightly lower activity than the catalyst fabricated in an acid medium [33ATO-Ti(a)]; however, the use of water as the medium in the synthesis of hybrid materials ATO-Ti has an environmental benefit. Compared with the 67ATO-Ti(w) sample [41], which was also prepared in a neutral aqueous medium, the 10FTO-Ti sample was more active, with a 60 min CA conversion of 76.4% compared with 70.1% in favourable conditions for the 67ATO-Ti(w) and 10FTO-Ti samples. Yunchan Park et al. [69] synthesised Al₂TiO₅-TiO₂ using a nanoparticle deposition system; when combining two catalysts with a composition of 50:50, the conversion rate for methylene blue photodegradation efficiency was found to be about 90% after 7 hours of ultraviolet (UV) irradiation. The highest decolonisation efficiency of rhodamine B (RhB) in visible light in the presence of a catalyst 40 wt% of Fe₂TiO₅ combined with P25 is about 96% after 60 min [36]; however, it should be noted that in this paper, catalytic activity in the mineralisation of RhB was not evaluated.