Effects of iron ions, doping methods and nanotubular morphology on TiO₂ solar photocatalytic performance

TNT

TiO₂ nanotubes

IMNP-Fe^x+/TiO₂

iron ions loaded on TiO₂ nanoparticles by impregnation

IMNT-Fe³⁺/TNT

Fe³⁺ loaded on TiO₂ nanotubes by impregnation method

PDNP-Fe³⁺/TiO₂

Fe³⁺ loaded on TiO₂ by photodeposition method

Hydr-Fe³⁺/TiO₂

Fe³⁺ loaded on TiO₂ by hydrothermal method

Solv-Fe³⁺/TiO₂

Fe³⁺ loaded on TiO₂ by solvothermal method

Ti_NT

TNT drying time

Ti_NP

TiO₂ drying time

The pharmaceutical industry generates significant pollution in aqueous environments with highly contaminated discharge, and the treatment of this water remains a major challenge (Shooshtari & Ghazi 2017). Pharmaceutical contaminants are not easily biodegradable under aerobic conditions due to the complexity of their chemical structure; therefore, aqueous effluents require specific treatment (Zavareh & Eghbalazar 2017).

Over the past few years, advanced oxidation processes (AOPs) have been widely used to eliminate residues from effluents; they are represented by hydrogen peroxide in the presence of UV radiation or by the photo-Fenton process using iron species; however, water treatment with titanium dioxide as a catalyst represents a large part of the water treatment field (Nezar & Laoufi 2018).

Unfortunately, TiO₂ has two well known drawbacks. (i) Its activation is limited by the ultraviolet light range, which represents 5% of solar energy, and this drawback limits the utilization of this free and renewable energy source. Photon generation for the activation of TiO₂ consumes energy, which makes photocatalytic treatment economically and environmentally costly. (ii) A crucial but overlooked TiO₂ feature is the adsorption capacity, which promotes pollutant-catalyst contact. This drawback is due to the limited TiO₂ adsorption capacity and the difficulty of improving this property by conventional methods. The pollutant adsorption phase is a crucial step in the photocatalytic reaction; this step induces the transfer of pollutant molecules to the catalyst surface, which is characterized by a high density of hydroxyl radicals, the species responsible for pollutant elimination.

To increase the catalyst performance, structural and morphological modifications were made. Structural modifications were made by doping TiO₂ with Fe ions to reduce the band gap and expand photon absorption to the visible light range, while morphological modifications were accomplished by changing the TiO₂ nanoparticle shapes into nanotubes by hydrothermal treatments to increase the specific surface area of the catalyst and therefore the adsorption capacity.

Fe^x+/TiO₂ catalysts in nanoparticle form were prepared using two doping methods, impregnation and photodeposition, while using the impregnation method for TiO₂ nanotube doping (Fe³⁺/TNT). To prepare impregnated Fe^x+/TiO₂ nanoparticles P25/nanotubes, pure TNP/TNT was dispersed in an aqueous solution containing an appropriate amount of Fe²⁺ or Fe³⁺ and stirred in darkness for 6 h. The product was obtained after heating at 80 °C for 12 h, and it was then ground and heat treated at 450 °C for 4 h. The products were named IMNP-Fe^x+/TiO₂ and IMNT-Fe³⁺/TNT. Photodeposited Fe³⁺/TiO₂ nanoparticles were prepared by similar steps; the unique difference in the preparation of photodeposited Fe³⁺/TiO₂ nanoparticles is that the mixtures were stirred under solar radiation instead of in darkness, and the product was named PDNP-Fe³⁺/TiO₂.

For TiO₂ nanotube preparation, 1 g of commercial TiO₂ P25 was dispersed in 60 mL of concentrated sodium hydroxide solution (10 M), followed by mechanical stirring for 2 h. The mixture was transferred into a Teflon-lined stainless-steel autoclave and heated at 150 °C in an oven for 24 h The product was filtered and washed with 0.1 M HCl solution until neutral (pH = 6–7) and was then rinsed with distilled water several times (Su et al. 2018). The product was finally dried in an oven at 80 °C for 36 h, and the synthesis was completed by a calcination process carried out at 450 °C for 4 h.

The hydrothermal and solvothermal methods of doping Fe³⁺-TiO₂ followed the steps mentioned above, except that in the hydrothermal doping process, TiO₂ and the Fe³⁺ precursor were dispersed in 60 mL of distilled water before adding NaOH, while in the solvothermal process, they were dispersed in ethanol (Wang et al. 2017). The resulting products were named Hydr-Fe³⁺/TiO₂ and Solv-Fe³⁺/TiO₂.

The experiments were carried out in a glass parallelepiped solar photoreactor (Figure 1) with dimensions of (L/W/H) = (50/40/5) cm. The reactor was connected to a reservoir with a capacity of 1 L, the flow was ensured by a peristaltic pump, and the tubing and reservoir were covered by aluminum foil to prevent light penetration. To prepare the treated solution, 10 mg of Cefixime was dissolved in 1 L of distilled water, the catalyst was added (0.2 g/L), and agitation was maintained for 10 min before starting the experiments. The variations in Cefixime concentrations during photocatalytic degradation were monitored by HPLC YL9100 using a reverse-phase column (C18).

Cefixime degradation by TiO₂ was evaluated by total organic carbon measurement. The results presented in Figure 2(b) shows that 69% of TOC removal was achieved after 120 min which indicated that Cefixime is efficiently mineralized in presence of 0.2 g/L of TiO₂ photocatalyst. All experiments were performed under the same operating conditions.

• n: number of repetitions

• ⁠: confidence coefficient depends on the number of results n and the desired probability, extracted from student distribution, for (Jackson 2015).

The Raman spectra of TiO₂ P25 and synthesized Fe^x+-TiO₂ are shown in Figure 3. All the undoped and Fe^x+-doped TiO₂ Raman spectra show six scattering peaks located at Δω = 142, 195, 397, 516, 517 and 638 cm⁻¹, which belong to the Raman active vibrational modes of the anatase phase, and a peak located at 446 cm⁻¹, which matches well with the characteristic rutile-type TiO₂ peak. However, no significant difference was observed among all four Raman spectra of Fe^x+-TiO₂ and undoped TiO₂, and no Raman peaks belonging to iron oxide were detected in any of the doped catalyst spectra. This confirms that Fe^x+ cations are well integrated into the TiO₂ substructure. It should be noted that a slight shift towards a higher frequency for the dominant band in Fe^x+-TiO₂ Raman spectra is observed from 142.39 to 142.94 cm⁻¹, and this small displacement is due to the minor change in the crystallinity size of the synthesized catalysts (Ramakrishnan et al. 2018).

FE-SEM and SEM studies were carried out on TiO₂ nanoparticles and hydrothermally synthesized specimens to illustrate the nanometric nature of the products, and the morphologies are shown in Figure 4. The product of the hydrothermal process without acid treatment, represented in Figure 4(b), shows a structure without a clearly defined shape compared to the initial structure of TiO₂ nanoparticles (Figure 4(a)). This structure is composed of nanosheets as a result of a rupture in the O–Ti–O bonds (Eslami et al. 2016), caused by the aggressive environment in which the hydrothermal reaction takes place at high pressure and temperature in an alkaline solution. Figure 4(c1,c2) reveals the formation of the TiO₂ nanotubular morphology after acid treatment. The sample is composed of a three-dimensional TiO₂ nanotube structure with a uniform diameter of 24 nm and a length of a few nanometres; this structure agglomerates into fibrous clusters by forming an entanglement of fine structures. This result indicates that the hydrothermal conditions (150 °C, 10 M NaOH) are adequate and that acid treatment is crucial for the formation of TiO₂ nanotubes.

Figure 4(d) and 4(e) show FE-SEM images of Fe³⁺-doped TiO₂ nanotubes formed by hydrothermal doping methods as well as an image of Fe³⁺-doped nanotubes formed by impregnation. A nanotubular structure appears when using the hydrothermal doping method (Figure 4(e)), though Fe³⁺ cannot be spotted due to the low amount and high dispersion of the dopant. EDS spectrum of Fe³⁺-doped TiO₂ nanotubes (Figure 5(a)) proved the existence of low amount of Fe after the doping process. As shown in Figure 4(d), the Fe³⁺-impregnated nanotube image is similar to the image of TNT, and the sample contains nanotubes with consistent diameters. However, a noticeable decrease in specific surface area was observed (Table 1); this decrease is probably caused by the doping process, which may have led to breakage of the weaker nanotubes because nanotubes were subjected to 6 h of continuous stirring, which may have disrupted the tubular morphology of the most fragile nanotubes. BET results show that the hydrothermal treatment caused the transformation of TiO₂ nanoparticles into nanotubes, which led to an increase in the surface area with a significant decrease in the average pore size of the catalysts, in the same case nanotube doping increased average pore size and decrease the nanotubes surface area.

The XRD patterns of TiO₂ nanoparticles and nanotubes synthesized by the hydrothermal method before and after heat treatment are shown in Figure 6(a). The XRD pattern of the sample dried at 200 °C shows that a new intermediate phase is formed (2θ = 19°, 74°, 90°, 166°), which does not correspond to the two crystalline phases of TiO₂ P25 (anatase and rutile). After the heat treatment of the sample at 450 °C, a reappearance of the anatase structure was observed, and all the sharp peaks seen in the XRD pattern of TiO₂ nanotubes were attributed to the TiO₂ anatase crystalline phase. Therefore, anatase becomes the major crystalline phase.

The XRD patterns of TiO₂ nanoparticles, nanotubes, Fe³⁺-doped nanotubes formed by the impregnation method (IMNT-Fe³⁺/TNT), Fe³⁺-hydrothermal and solvothermal-doped TiO₂ nanotubes (Hydr-Fe³⁺/TiO₂) and (Solv-Fe³⁺/TiO₂) are shown in Figure 6(b). Diffraction peaks correspond to anatase crystalline phases of TiO₂ P25 present in all samples, no impurities of iron oxides are detected, and the structure has not changed from those of the synthesized catalysts, indicating that Fe³⁺ is well integrated in the TiO₂ substructure. It should be noted that the Solv-Fe³⁺/TiO₂ XRD pattern presents peaks identical to those of pure TiO₂ P25 (anatase and rutile) and with the same intensity, which suggests that solvothermal doping does not cause any changes in the TiO₂ structure, unlike the hydrothermal doping that transforms the rutile phase into anatase, leading to rutile peak disappearance.

Measurements of absorbance UV-visible spectroscopy were carried out on the catalysts to assist the optical properties of the doped TiO₂. The results are shown in Figure 7(a). The spectra show that pure TiO₂ has high absorption in the ultraviolet light range (wavelength lower than 400 nm). A typical TiO₂ absorption edge is observed after 315 nm, which represents the threshold at which the catalyst can no longer absorb visible light. After doping TiO₂, the absorption is redshifted towards a higher wavelength to the visible light range. This is a result of the reduced band gap from 3.29 eV for pure TiO₂ to 3.26, 3.24, and 3.22 eV for IMNP-Fe²⁺/TiO₂, IMNP-Fe³⁺/TiO₂, and PDNP-Fe³⁺/TiO₂, respectively, which are caused by Fe²⁺ and Fe³⁺ electrons 2d and 3d, which excite the TiO₂ band gap and shrink it (Asiltürk et al. 2009). Due to the high density of electrons (as a result of ion doping), a new conduction band is formed from these electrons at a position below the conduction band of the catalyst. Therefore, the resulting bandgap is smaller than the original catalyst band gap (Jongprateep et al. 2018). We suggest that the availability of light in the photodeposition method allowed more electrons to be produced by Fe ions, which resulted in a larger density of electrons and thereby greater reduction in the TiO₂ band gap. It can be observed that all the catalysts showed a red shift in absorption towards the visible light region, which is a result of Fe³⁺ doping and of the thermal treatment, as shown in the case of undoped TiO₂ nanotubes. In a comparison of absorption capacity of the catalysts formed by the different doping methods, the solvothermal-doped TiO₂ catalyst has a better visible light absorption capacity than impregnated nanotubes, hydrothermally doped TiO_2, and nanotube catalysts, which means that solvothermal doping has a more positive effect on the optical properties (Table 2) of the catalysts compared to the other doping methods. The hydrothermally doped catalyst exhibits a slightly greater red shift and a higher visible light absorption intensity compared to impregnated nanotubes, while it appears that the impregnation steps adversely affected the optical properties of the final product.

The effects of iron ions Fe²⁺ and Fe³⁺ as TiO₂ cocatalysts in the presence and absence of H₂O₂ on Cefixime degradation are shown in Figure 8, and a quantity of 0.85% (M/M) of both Fe²⁺ and Fe³⁺ was examined (i.e., M_Fex+ = 0.85% M_TiO2). The percentage of Cefixime elimination was approximately 86% in the absence of Fe²⁺ or Fe³⁺, while introducing them in solution brought the abatement to (X = 84.4% in the presence of Fe²⁺) and (X = 93.7% in the presence of Fe³⁺) after 120 min. The combinations of Fe²⁺/H₂O₂ and Fe³⁺/H₂O₂ in the presence of TiO₂ were found to be more effective, and the same degradation rate was achieved for both (X = 98.9% in the presence of Fe²⁺) and (X = 98.9% in the presence of Fe³⁺). Fe²⁺ and Fe³⁺ showed different behaviors in the presence and absence of H₂O₂, [X_TiO2 < X_TiO2/Fe²⁺ < X_TiO2/Fe³⁺ < (X_{TiO2/Fe²⁺/H2O2} = X_{TiO2/Fe³⁺/H2O2})], which is due to the different mechanisms that take place in the presence of each substance.

Introducing Fe³⁺ ions led to a better abatement of Cefixime compared to the introduction of Fe²⁺ due to the ease of radical generation from the process using Fe³⁺, while the production of radicals in the presence of Fe²⁺ ions is limited by the amount of Fe²⁺ remaining after the formation of H₂O₂.

The effect of Fe²⁺ and Fe³⁺ as dopants on the TiO₂ nanoparticle (NP) photocatalytic performance to degrade Cefixime is represented in Figure 9(a). Photocatalytic activity enhancement is observed as a result of TiO₂ doping, and this improvement is less significant in the case of IMNP-Fe²⁺/TiO₂ than in the case of IMNP-Fe³⁺/TiO₂, where the degradation is enhanced from 85.7% to 87.7% and 89.9% for TiO₂, IMNP-Fe²⁺/TiO_2, and IMNP-Fe³⁺/TiO₂, respectively. This increase in the photocatalytic activity of doped TiO₂ can be attributed to the shift in optical absorption to higher wavelengths towards the visible range and a decrease in the band gap energy, as shown in Figure 7. Therefore, doped TiO₂ becomes more sensitive to visible light and can be better activated under solar radiation. IMNT-Fe³⁺/TiO₂ had a higher performance than IMNT-Fe²⁺/TiO₂ because it had a greater shift in optical absorption towards the visible range, as shown in Figure 7, and hence a greater solar radiation absorption capacity. Moreover, Fe³⁺ behaves as an electron and hole trap that induces electron-hole separation as a result of the higher Fe³⁺/Fe²⁺ and Fe³⁺/Fe⁴⁺ redox potentials (Rajagopalan et al. 2017). Fe²⁺ acts as a trap for electrons, and Fe⁴⁺ acts as a trap for holes, giving the electron holes sufficient time to react and generate radicals.

As shown in Figure 9(a), IMNP-Fe³⁺/TiO₂ had a better photocatalytic performance in Cefixime elimination, despite having a lower shift towards the visible light range than did PDNP-Fe³⁺/TiO₂ (Figure 6); it could be that, during photodeposition, Fe³⁺ ions were deposited on the TiO₂ surface as Fe²⁺ and Fe⁴⁺, and after being reduced or oxidized by photogenerated electrons and holes, the absence of light during the impregnation process inhibited electron-hole generation, resulting in the deposition of ions as Fe³⁺, which is a more active iron ion.

The photocatalytic activity of TiO₂ nanotubes was evaluated. Figure 9(b) shows the Cefixime photodegradation curves in which TNP and TNT are used as photocatalysts under solar radiation. In the presence of TiO₂ nanotubes (TNTs), Cefixime abatement reaches 94.9% after 120 min of solar exposure, whereas it reaches 85.7% in the presence of TiO₂ nanoparticles. This improved performance in Cefixime removal is mainly attributed to the morphology of titanium dioxide, which has a crucial role in catalyst performance and an important influence on photocatalytic behavior and is less attributed to the TNT structure.

Regarding the TNT structure, Figure 6 shows the disappearance of the rutile phase and the formation of a well defined anatase crystalline phase of TiO₂, which is more the photoactive form of TiO₂ than the rutile form. This transformation in TiO₂ P25 phases is due to hydrothermal treatment.

Regarding the TNT morphology, TiO₂ nanotubes offer a much larger BET surface area (294 m²/g) than nanoparticles (55 m²/g) (Table 1), which is a very important characteristic because pollutant decomposition occurs at the catalyst interface; a larger surface area means a higher number of active sites and contact surface to adsorb water for radical generation and to adsorb pollutants, which promotes pollutant-radical contact, leading to higher Cefixime oxidation. Therefore, TNT is more efficient in radical formation, which leads to the availability of a large number of hydroxyl radicals on the nanotube surface, as mentioned by Manfroi et al. (2014).

The TiO₂ charge in the form of nanotubes is more exposed to pollutants and light radiation than NPs; indeed, the hollow TNT tube shape allows the transfer of pollutants into the nanotube core, which promotes pollutant diffusion. The longitudinal tube profile improves light absorption (Marien et al. 2016), which makes TNT an efficient light harvester compared to TNP. In addition, Sun et al. (2014) observed that by changing the morphology of TiO₂ nanoparticles into nanotubes, the generated electrons become more efficient as their lifespan is extended and their transfer becomes faster.

Moreover, nanotubes red shift towards the visible region (Figure 7) when nanotubes are not doped, so TNT had improved solar photoactivation characteristics, which was also observed by Alfaro Cruz et al. (2019). This TNT red shift can be attributed to the tube-shaped TiO₂ and the hydrothermal treatment (Camposeco et al. 2016). Considering all of the above, it is obvious that all of these features improve the photocatalytic performance of TiO₂ nanotubes.

The photocatalytic activities of thermally synthesized catalysts under solar irradiation were evaluated for their efficacy in Cefixime removal, and the results are presented in Figure 8(c). All the synthesized catalysts showed higher photocatalytic activity than TiO₂ P25. In 120 min., Cefixime abatement reached 85.7, 89.9, 94.9, 94.9, 97.5, and 100% for TiO₂, IMNP-Fe³⁺/TiO₂, Solv-Fe³⁺/TiO₂, TNT, IMNT-Fe³⁺/TNT, and Hydr-Fe³⁺/TiO₂, respectively. The Hydr-Fe³⁺/TiO₂ catalyst eliminated Cefixime successfully in 120 min.

Nanoparticle catalysts doped by impregnation or by the solvothermal method achieved lower abatement than hydrothermally doped or treated catalysts due to the influence of the doping technique. Hydrothermal doping induced TiO₂ P25 rutile phase conversion into anatase, a reduction in the band gap caused by the hydrothermal process itself as well as by Fe³⁺ doping, and nanotube formation, which made it possible to considerably increase the catalyst specific surface area. Consequently, all these properties of the hydrothermal doping process combined to provide a powerful and efficient catalyst, while these improvements were not achieved using other doping methods, such as solvothermal, impregnation, and photodeposition methods, since these methods have a limited influence on the structure (Figure 6) and physical properties (Table 2) of the catalyst, regardless of the dopant (Fe³⁺) influence.

According to the comparison results, Figure 9(d) shows that the photocatalytic efficiency of iron ion Fe³⁺ changes depending on how it was introduced into the reaction. Comparing four different Fe³⁺/TiO₂-doping methods shows that employing Fe³⁺ as a TiO₂ cocatalyst gives better results than impregnation, photodeposition, and solvothermal-doping processes; indeed, when Fe³⁺ is employed as a dopant, Cefixime is eliminated by the radicals generated through TiO₂ activation, and the role of Fe³⁺ is to reduce the band gap and extend the electron-hole pair separation by catching the electrons. However, when Fe³⁺ is used as a cocatalyst, Cefixime is attacked by radicals generated not only by TiO₂ activation but also by Fe³⁺ hydrolyzation. Thus, the photocatalytic system that used Fe³⁺ as a TiO₂ dopant using photodeposition and impregnation-doping methods performed poorly and was shown to be ineffective compared to that in which Fe³⁺ is used as a cocatalyst. The doping process requires steps that consume time and energy. Therefore, when using Fe³⁺, it is recommended that the impregnation and photodeposition-doping methods be recognized as dispensable. Using the hydrothermal-doping method considerably promotes the photocatalytic performance of TiO₂ in Cefixime elimination; this doping method combines a structural improvement represented by the transformation of the crystals from rutile to anatase form, an optical improvement due to the reduction of the band gap, and a physical improvement attributed to the formation of TiO₂ nanotubes that have a large surface area and are capable of destroying the pollutant.

In this work, we were interested in defining the most effective way to exploit iron ions (Fe²⁺, Fe³⁺) to improve Cefixime removal by employing them either as titanium dioxide cocatalysts (radical generators) or as TiO₂ dopants (band gap reducers).

The results suggested that there was no advantage in using Fe³⁺ ions as dopants by conventional methods such as impregnation, photodeposition or solvothermal methods because these methods consume time and energy; they make the photocatalytic system less efficient, so using Fe³⁺ ions as cocatalysts gives better results than using these doping methods.

Titanium dioxide nanotubes were prepared by a hydrothermal method, and the influence of this method on the synthesized catalyst was evaluated; it showed improved physical, structural and optical properties of TiO₂. Therefore, the hydrothermal doping method was exploited to produce a highly effective catalyst with enhanced properties that allowed complete Cefixime elimination in only 90 min by taking advantage of the dopant Fe³⁺ and TiO₂ nanotube properties.

The authors are thankful for technical help from Amar Manseri a member of the Research Center of Semi-conductor Technology for Energy, CRTSE, Algiers, Algeria and Dr Yasmina Roumila a member of the Laboratory of Electrochemistry-Corrosion, Metallurgy and Inorganic Chemistry, Faculty of Chemistry, USTHB, Algeria.

All relevant data are included in the paper or its Supplementary Information.