TiO₂ nano powder was procured from Fujifilm Wako Pure Chemical Co., Ltd., with the X-ray Fluorescence (XRF) analysis revealing the following chemical composition in percentage values: CO₂: 14.6%, TiO₂: 85.0%, Na₂O: 0.284%, and Fe₂O₃: 0.024%. Prior to its incorporation with PES, the TiO₂ powder underwent a crushing process and was subsequently sieved through a No. 230 sieve. N-Methyl-2-pyrrolidone (NMP) and various other chemicals were acquired from Nacalai Tesque Inc. (Japan) and utilized in their as-received state without additional purification. The polyethersulfone (PES) employed, specifically Ultrason E2010, was sourced from BASF Co. Ltd. (Germany). To prepare the Ammonium-Nitrogen solutions, ammonium bicarbonate (NH₄HCO₃) obtained from Nacalai Tesque Inc., Japan, was dissolved in distilled water, resulting in a working solution with a concentration of 1000 mg/L. Before conducting the adsorption experiments, the ammonium solution was diluted accordingly. In cases where 0.1 M HCl and 0.1 M NaOH solutions were employed, the pH values of the solutions were adjusted as necessary.

In accordance with our previously established procedures [35] the wet-spinning process was employed to fabricate TiO₂-PES composite fibers by utilizing the phase inversion of a PES solution into solid PES [36]. Here, PES was dissolved in NMP at a concentration of 30 wt%. Subsequently, varying amounts of TiO₂ powder (2, 5, 8, and 10 wt%) relative to the weight of PES were introduced. The resultant mixture was extruded through a cylindrical needle with a 0.6 cm-diameter, and an applied pressure of 0.42 MPa was utilized to extrude it into a water coagulation bath maintained at 25˚C. Within this coagulation bath, the composite fibers took form and were subsequently rinsed in excess water at 60˚C to eliminate any residual NMP. Finally, the composite fibers were subjected to drying under vacuum conditions at 60˚C. The fibers produced at 30 wt% PES, with varying TiO₂ content of 0, 2, 5, 8, and 10 wt%, were denoted as PeTi0, PeTi2, PeTi5, PeTi8, and PeTi10, respectively.

To explore the alterations in fiber morphology, scanning electron microscopy (SEM) was conducted using a TM3030Plus HHTC instrument from Japan. The samples were sputter-coated with gold before examination, and imaging was carried out at an accelerating voltage of 15 kV. The surface area of both the TiO₂ and composite fibers was determined using the Brunauer-Emmer-Teller (BET) method, and the measurements were performed with a Tristar II 3020 apparatus (Micromeritics Inc., Tristar II, Shimizu, Japan). Prior to analysis, the samples were subjected to degassing under vacuum conditions overnight at 70˚C. X-ray diffraction (XRD) spectra of limonite and the composite fibers were acquired using a SmartLab X-ray diffractometer (Rigaku Corporation) equipped with Cu Kα radiation (λ = 1.5406 Å). The XRD patterns were recorded over a 2θ range spanning from 10˚ to 70˚, employing parameters of 40 kV and 30 mÅ. For component analysis of TiO₂ and composite fibers, an X-ray fluorescence analyzer (ZSX Primus II, Rigaku Corporation, Japan) was utilized. Pellets with a 10 mm diameter were fabricated from TiO₂, composite fibers, and an aluminum ring using a hydraulic press, applying a pressure of 450 kgf/cm⁻² for the measurements. The tensile strength of the TiO₂ composite fibers was determined with a load measurement device (LTTU500N, Minebea Co. Ltd., Japan) operating at 500 N and 23˚C with a relative humidity of 50%. The gauge length was set at 30 mm, and the crosshead speed was maintained at 2 mm/min. Five specimens were subjected to testing for each sample, and each specimen had a length of 50 mm. Initially, the diameter of each piece was measured using a micrometer (Mitutoyo 103 - 177, Japan) to calculate the cross-sectional surface area. The tensile strength and elongation were subsequently computed using the following equations: tensile strength (MPa) = maximum load/cross sectional area and elongation (%) = 100 × (elongation at rupture/initial gauge length). The density of the composite fibers was determined through the Archimedes method. A dry specimen with a known weight was immersed in excess water until equilibrium was reached with the specimen in the water. The density of the prepared specimens (ρ) was calculated using the formular:

where ρ is the density (g/cm³), W_s denotes the specimen weight in air (g), W_w represents the specimen mass in ethanol (g), ρ₀ is the density of ethanol (g/cm³), and d is the density of air (g/cm³).

Zeta potential measurements were conducted on both the TiO₂ powder and the composite fibers (after they were ground into powder form) within a pH range of 5 - 11. This analysis was carried out using a Zeta-potential analyzer (ELSE1NGK; Otsuka Electronics Co. Ltd., Japan), with distilled water employed as the dispersion medium. The medium’s pH was adjusted through titration, utilizing 0.1 M HCl and 0.1 M NaOH solutions, to determine the zeta potential of the TiO₂ powder and the composite fibers.

The photocatalytic oxidation of ammonia in water, employing either TiO₂ composite fiber or TiO₂ powders, was carried out within a batch water medium. This reaction was conducted under exposure to light, specifically illuminated using a UV-A Hg lamp with a power rating of 25 W, positioned at a fixed distance of 10 cm above the central point of a pyrex glass reactor. The reactor possessed a 5 L capacity, and the average light intensity throughout the experiments approximated 300 µW/cm². To ensure the accuracy of the experiments, the entire system was shielded by an aluminum foil sheet, effectively preventing any interference from external light sources. For the purpose of maintaining well-mixed solutions during the experiments, circulating water was employed. Additionally, an air diffuser was installed at the base of the reactor tube as part of the experimental setup.

To conduct the ammonium adsorption experiments, either 10 g of TiO₂ composite fiber or 1 g of TiO₂ nano powder was individually introduced into a 1 L solution containing NH 4 HCO 3 − . Subsequently, the mixtures were exposed to UV-A lamp irradiation at an intensity of 300 µW/cm² for a duration of 12 hours at a temperature of 25˚C. The adsorption kinetics of NH 4 + were examined using the chosen adsorbents, namely TiO₂, PeTi2, PeTi5, PeTi8, and PeTi10, across varying concentrations ranging from 0.27 to 2.7 mM (equivalent to 5 to 50 ppm). During the experiments, small sample solutions, each comprising less than 10 mL, were withdrawn and stored in opaque vials to minimize potential interference from indoor fluorescent lighting before subsequent analyses. The concentrations of NH 4 + , NO 2 − , were determined using a UV-visible spectrophotometer (Jasco V-750). Specifically, NO 3 − concentrations were ascertained using the NO 3 − 11S LAQUATwin Kit designed for nitrate measurements. The concentrations of NH 4 + , NO 2 − , and NO 3 − were calculated using five-point external standard calibration curves. Standard solutions were freshly prepared on a daily basis, and the analysis of standards was consistently repeated daily. The NH 4 + ions in the samples were detected through a colorimetric method following the 4500-NH₃ F-Phenate method. The colored samples were assessed for absorbance at 640 nm. For the determination of NO 2 − content, a Griess assay was employed, utilizing N-(1-naphthyl)ethylenediamine (NED) as a coupling component in the reaction. The colored samples were analyzed for absorbance at 540 nm. The adsorption equilibrium isotherms were assessed using both the Langmuir and Freundlich models. In the Langmuir model, the equation took the form:

C e q e = C e q m + 1 q m K L (2),

q e is equilibrium adsorption amount (mmol/g), q m is maximum adsorption capacity (mmol/g), and K L is Langmuir constant.

Conversely, for the Freundlich model, the equation adopted was:

ln ( q e ) = ln ( K F ) + 1 n ln ( C e ) (3),

K F , and n are Freundlich constant.

It is well-established that the degradation rate is associated with the formation of hydroxyl radicals (OH•) on the catalyst surface. The quantity of hydroxyl radicals generated is contingent upon several factors, including the catalyst type, catalyst dosage, pH value, and light intensity. In this particular study, the impact of pH, catalyst dosage, and light intensity were also thoroughly investigated. In the context of photo exposure experiments, 10 g of the composite fibers were brought into contact with a 1 L aqueous ammonia solution containing 35 ppm concentration. Subsequently, the UV-A lamp was exposed to the solution for a duration of 96 hours.

Figure 1 provides an insight into the morphology of TiO₂-PES composite fibers

Figure 1. Cross-section of PES fiber and TiO₂ composite fibers at 150 and 500 magnifications.

loaded with varying amounts of TiO₂ within PES fibers. These resultant fibers exhibited an approximate diameter of 500 µm. Examination of their cross-section revealed a highly porous, sponge-like structure with pores smaller than 5 µm at the central region. This structure displayed a finger-like, elongated pore distribution extending from the fiber’s core to its surface. However, it’s noteworthy that the larger finger-like pores transitioned into much smaller structures when the TiO₂ content ranged from 2 to 8 wt%, and they were distributed at approximately 50 µm from the fiber surface. Remarkably, when the TiO₂ content reached 10 wt%, these larger pores nearly vanished, and the fiber cross-section exhibited a generally dense and spongy appearance. These observations suggest that, when the TiO₂ dispersion solution is added to the PES and coagulated in water, the coagulation rate slows down with increasing TiO₂ content, and the solvent exchange rate also becomes less rapid. Furthermore, the XRF elemental analysis presented in Table 1 reveals that the TiO₂ content closely aligns with the quantity of added TiO₂. As the TiO₂ content in the composite fiber increases, the density values exhibit a trend of 0.81, 0.83, 0.87, and 1.07 for 2, 5, 8, and 10 wt% respectively. The BET surface area values, determined through nitrogen adsorption, demonstrate a pattern where the surface area increases for PeTi2, PeTi5, and PeTi8, and then decreases in PeTi10. These findings indicate that the incremental addition of TiO₂ in the composite influences the surface area of the resulting fibers, with the highest surface area noted at 8 wt%. At a TiO₂ content of 10 wt% in the composite fibers, the surface area exhibited a slight decrease, suggesting the initiation of an aggregation process of TiO₂ nano powder. These results are further corroborated by the SEM cross-section images shown in Figure 1.

In Figure 2, N₂ adsorption is presented at various relative pressures for each composite fiber. The results clearly indicate that the N₂ adsorption amount exhibited an increase in proportion to the TiO₂ loading. Notably, the composite fibers displayed a substantial increase in N₂ adsorption, particularly under a relative pressure of 0.1. To provide a quantitative perspective, the BET surface areas of PES and the composite fibers, including PeTi2, PeTi5, PeTi8, and PeTi10, were measured and found to be 18.09, 25.93, 48.19, and 37.76 m²/g, respectively. As shown in Figure 3, XRD patterns for both TiO₂ and the TiO₂ composite fibers are depicted. The main peaks, discerned at 25.3˚, 37.9˚, 48.3˚, 53.9˚, 54.9˚, 62.7˚, 68.84˚, 70.1˚, and 75.03˚, correspond to the facets (101), (004), (200), (105), (211), (204), (113), (220), and (215), respectively. These peak positions align with the characteristic pattern of tetragonal anatase TiO₂ [37]. Importantly, no other diffraction peaks were observed in the XRD pattern, signifying the absence of any chemical reactions between TiO₂ and the PES polymer during the synthesis process.

Table 1. Properties of composites fibers, PES fiber, and TiO₂.

Figure 2. N₂ adsorption of PES and the TiO₂-PES composite fiber.

Figure 3. XRD spectra of the TiO₂-PES composite fiber.

Figure 4. pH effect on the NH 4 + removal by using photocatalyst method with TiO₂ composite fibers. The ammonia photo-oxidation process was conducted across a range of pH values, spanning from 5 to 11. This pH variation was achieved by the addition of either 0.1 N NaOH or 0.1 N HCl solutions, and the results are depicted in the accompanying Figure 4. pH is a crucial parameter in the photocatalytic oxidation of ammonium, as it plays a role in adjusting the surface charge of the photocatalyst. The results clearly illustrate that ammonia removal efficiency was notably lower in acidic pH conditions. This phenomenon can be attributed to the inhibition of hydroxide group production necessary for the formation of hydroxyl radicals under acidic conditions. Moreover, the adsorption of ammonia molecules onto the photocatalyst surface decreased due to the positively charged nature of the photocatalyst surface, leading to repulsive forces. Consequently, the degradation efficiency of ammonia diminishes under acidic pH. Conversely, the removal efficiency of ammonia exhibited a substantial increase, rising from 18.5% to 60%, as the pH was raised from 5 to 11 during a 12-hour period under UV irradiation. This increase can be attributed to the higher concentration of hydroxide ions generated under alkaline pH conditions.

Figure 4. pH effect on the NH 4 + removal by using photocatalyst method with TiO₂ composite fibers.

These hydroxide ions facilitate the reaction with the photo-generated holes, resulting in the formation of hydroxyl radicals. For practical and environmentally friendly considerations, this study chose to operate at pH 8, which demonstrated high removal efficiency, reaching approximately 50%.

Figure 5 illustrates the adsorption isotherms for ammonium on TiO₂ and the TiO₂ composite fibers. Equilibrium concentration data were analyzed using two isothermal models, Langmuir and Freundlich, and the results are presented in Table 2. In the adsorption isotherm, the data fit well to the Langmuir model, with correlation efficiency (R²) values ranging from 0.9401 to 0.9936, while the Freundlich model yielded R² values in the range of 0.8345 to 0.9601. These results suggest that the saturation adsorption on TiO₂ sites within the fibers forms a monolayer. The maximum adsorption capacity (q_m) for ammonium ions on TiO₂ and composite fibers followed this order: PeTi10 > PeTi8 > PeTi5 > PeTi2. This indicates that q_m values are dependent on the porous structure of the adsorbents.

Figure 5. Adsorption isotherm of NH 4 + by using the TiO₂ composite fibers (a), Freundlich adsorption plots (b), Langmuir adsorption isotherm plots (c).

Table 2. Parameters on Langmuir and Freundlich of ammonium adsorption by the composite fibers.

The photocatalytic performance of TiO₂ and the composite fibers was evaluated through the degradation of NH 4 + in an aqueous solution.

In Figure 6, the adsorption of NH 4 + on TiO₂ and PeTi8 was significantly enhanced compared to other samples, indicating that the PeTi8 sample possesses more active sites on its surface. To optimize reaction conditions, the effects of photocatalyst exposure time and illumination intensity on the photocatalytic removal of aqueous NH 4 + by TiO₂ and TiO₂ composite fibers under a UV-A lamp were further investigated. Figure 6 demonstrates the decrease in aqueous NH 4 + concentration when treated with TiO₂ and TiO₂ composite fibers, starting with an initial solution pH of 8 and a light intensity of 300 µW/cm², while the

Figure 6. Conversion concentration between NH 4 + , NO 2 − , and NO 3 − when exposed to UV-A lamp for 96 hours. PeTi0 (a), PeTi2 (b), PeTi5 (c), PeTi7 (d), TiO₂ (f).

photocatalyst dosage was fixed at 10 g for TiO₂ composite fibers and 1 g of TiO₂ nano powder. As the TiO₂ content in the composite fibers changed, the removal of aqueous NH 4 + initially decreased slowly in the absence of light. However, the removal dramatically decreased when the UV-A lamp was activated, observed between 25 to 36 hours. This phenomenon was consistent across all TiO₂ composite fibers. In cases of composite fibers with lower TiO₂ content, the production of reactive oxygen species was limited, affecting their photocatalytic performance. Figure 6(b) confirms this observation, as NH 4 + concentration stabilized after 24 hours of exposure under the UV-A lamp, reaching 0.7 mmol/g. In the case of the highest TiO₂ content in the composite fiber, PeTi10, the NH 4 + concentration was 0.5 mmol/g higher than the tank content of 10 g of PeTi8, which recorded 0.46 mmol/g, due to aggregation within the fiber structures. On the other hand, the results obtained from TiO₂ nano powder exhibited good adsorption ability in the first 6 hours when the UV-A lamp was turned on, but the removal ability gradually decreased and stabilized after 12 hours of exposure under irradiation. Compared to PeTi0, PES fiber showed no adsorption ability for NH 4 + . Additionally, Figure 6 also presents the concentrations of photoconversion products, NO 2 − and NO 3 − , during the photocatalytic process. The high removal ratio of aqueous NH 4 + and the presence of oxidation byproducts ( NO 3 − and NO 2 − ) could not be attributed to the weak volatilization of aqueous NH 4 + under illumination. The results clearly demonstrate that TiO₂ composite fibers can effectively photo-catalytically remove aqueous NH 4 + under the UV-A lamp. The higher NO 3 − yields in the tank content of 1 g of TiO₂ and 10 g of PeTi8 and PeTi10 compared to PeTi2 and PeTi5 are attributed to the appearance of TiO₂ in the photocatalyst process, as documented in previous research. These higher NO 3 − yields provide evidence that heterogeneous reactions dominate at higher TiO₂ concentrations.

Figure 7 depicts the variation in aqueous NH 4 + concentration when treated with TiO₂ and TiO₂ composite fibers under a UV-A lamp with different tank volumes. Altering the tank volume affects the interaction distance between the light source and the pollutant, providing insights into the impact of light intensity at varying distances. With the same initial concentration of NH 4 + in the aqueous solution and a consistent dosage of 10 g, the tank with a 0.25 L volume demonstrated effective NH 4 + removal. Conversely, in the 2 L volume tank, the larger volume limited the utilization of photons in the solution, resulting in a less effective photocatalytic performance. The rate at which photocatalysis activates and forms electron-hole pairs in photochemical reactions is significantly influenced by the intensity of light. The efficiency of pollutant degradation depends on the distribution of light intensity within the reactor.

To explore the impact of light intensity on photocatalytic efficiency at different distances (2 cm, 5 cm, 10 cm, 15 cm, and 20 cm), all experiments were conducted under identical conditions, and the results are presented in Figure 8. Clearly, as the distance between the photocatalyst (PeTi8) and the light source

Figure 7. Conversion of concentration between NH 4 + , NO 2 − , and NO 3 − exposed to UV-A lamp with changing the tank volume. The PeTi8 composite were used, 0.5 L (a), 1 L (b), 1.5 L (c), 2 L (d).

Figure 8. Distance dependent absorption spectra for the photocatalytic degradation of PeTi8 under UV-A light irradiation.

increases, the concentration of hydroxyl radicals (OH•) decreases. Generally, the distance between the light source and the photocatalyst has a significant effect on the reaction kinetics. It was observed that an increased distance led to a reduction in the rate constant, resulting in decreased catalyst activation and diminished photocatalytic degradation of ammonia. Figure 8 presents the OH• concentration at different distances and light intensities after 12 hours of initiating the reaction. Throughout the photocatalyst experiments, a certain quantity of hydroxyl radicals is produced. Changing the distance of the light source also impacts the ability to generate hydroxyl radicals in the aqueous environment. An experiment was conducted under the same conditions to assess the amount of hydroxyl radicals generated when altering the lamp distance from the catalyst. The results, as shown in Figure 8, indicate that as the distance of the lamp increases, the concentration of hydroxyl radicals decreases, with virtually no hydroxyl radicals being produced at a distance of 20 cm.