Polyamide6 (PA6) and tetra-n-butylammonium bromide (TBAB; ≥98%) were obtained from Sigma-Aldrich whereas trifluoroacetic acid (TFA), acetone, potassium dihydrogen phosphate (KH₂PO₄; 99.3%), sulfuric acid, formic acid, and ascorbic acid (99.4%) were purchased from Thermo Fisher Scientific. Lanthanum oxide (La₂O₃, >99.99% 30–50 nm) nanoparticles were purchased from ACS materials. 1,1,1,3,3,3-Hexafluoro-2-propanol (HFIP) was purchased from Oakwood Chemical. All materials were used without further treatment.

The most common solvents to dissolve PA6 include formic acid, mixture of acetone and strong acid such as trifluoracetic acid (TFA), and 1,1,1,3,3,3,3-Hexafluoro-2-propanol (HFIP) (Heikkilä and Harlin, 2008). The solution was then sonicated to disperse the La₂O₃ homogenously. PA6 was then added and stirred until PA6 pellets were fully dissolved. Lastly, TBAB was added to the solution and stirred overnight for homogeneous solution for electrospinning.

The solutions properties (viscosity, surface tension and electrical conductivity) of the solution were measured. The viscosity was measured using a CPA-40 spindle connected to a Brookfield DV2THB viscometer where the solution viscosity was determined to be independent of the shear rate. Thus, the viscosity was measured at 95% torque at each rotational speed ranging from 0.5 rpm to 200 rpm. An automatic surface tensiometer (Shanghai Fangrui Instrument, QBZY-1) with platinum-coated plate was used to measure the surface tension. Solution electrical conductivity was measured using an electrical conductivity probe from Apera Instruments (Al1311, K = 0.1) connected to EZO conductivity circuit (Atlas Scientific), on Tentacle T3 using Raspberry Pi (Whitebox T3, Mkll). All solution property measurements were taken at room temperature prior to electrospinning to correlate them closely to the resulting nanofiber properties.

Electrospinning was carried out by injecting the prepared solution through a 5-mL BD Luer-Lok syringe with a 25-gauge stainless steel needle using a syringe pump (New Era, NE-100). High-voltage power supply was used to apply the voltage between the needle tip and drum collector. The drum collector was wrapped with aluminum foil and rotated around 400 rpm. Electrospinning and environmental conditions including applied voltage, feed rate, temperature, and absolute humidity were fixed at 12 kV, 0.25 mL/h, 23 ± 1°C, and 0.008 ± 0.001 kg H₂O/kg dry air, respectively. To investigate the effect of post thermal treatment, the as-spun nanofiber samples were annealed in a tube furnace (Thermo Scientific Lindberg/Blue M Mini-Mite TF55030) at varying temperatures (i.e., 80, 100, 120, 140, and 200^°C) for 2 h in air.

Morphology of the as-spun nanofiber was observed with a scanning electron microscope (Prisma E SEM, Thermo Fisher Scientific, United States). Prior to analysis, a thin layer of gold was sputtered using Electron Microscopy Sciences 575X over the samples at 20 mA for 30 s to minimize surface charging. Obtained SEM images were imported to ImageJ software to measure the average fiber diameter, which was obtained by measuring the diameter of 50 unique nanofibers. The bead density was also calculated by dividing the total number of beads from a single SEM image by the total area of the image. Fiber fraction was determined by the proportion of nanofibers in the total product, which could include beads and clumps. Transmission electron microscopy (TEM) samples were collected by placing a carbon-coated copper grid directly in front of the drum collector for 1 min during electrospinning. TEM images were captured using 300 (S)TEM Ceta™.

The crystal structures were characterized by using a powder X-Ray diffractometer (Rigaku Miniflex 6 g). Multiple 2 × 2 cm² sheets were coherently layered to ensure X-ray detection of the sample. Data collection was performed at room temperature at λ = 1.5406 Å over a 2θ range of 10°–80° with a step size of 0.025°. The molecular structures were observed with Perkin Elmer Spotlight 200 Fourier-transform infrared (FT-IR) spectrometer. For FT-IR, approximately 0.5 × 0.5 cm² of sample was cut and placed in a pass-through sample holder. FT-IR spectra were then obtained by scanning from 450 to 4,500 cm⁻¹ at a resolution and scan increment of 1 cm⁻¹.

Differential scanning calorimetry (DSC) was performed by using a DSC 2500, TA Instruments, United States. The accurately weighed samples were placed into Al crucible, and which was used for the analysis. The samples were first heated to 250°C, at a rate of 10^°C/min. The sample was then cooled down to 0°C, at a rate of 10 °C/min and reheated to 250°C, at a rate the same rate. This cooling and reheating cycle was repeated twice. Mechanical properties were examined using a Discovery hybrid rheometer (DHR-30, TA Instruments, United States) attached with RH Linear Tension Rectangular Fixture. Composite nanofiber samples were placed between the plates and the samples were pulled apart at a constant linear rate of 1.0 mm per second until 50 mm is reached at room temperature. Mechanical properties were determined using software provided by the manufacturer (TRIOS, TA Instruments).

Batch adsorption experiments were undertaken to determine the adsorption capacity of phosphate. Various initial P concentrations ranging from 2 to 20 mg/L were tested with adsorbent dose of 1 g/L of La₂O_3, relative to the solution volume, at room temperature. The solution was collected using a 3-mL BD Luer-Lok syringe with a syringe attached (0.22 μm, PTFE Teflon filter) at different time intervals (0.5, 1, 2, 5, 10, 15, 240, and 1,440 min). The phosphate concentration of collected samples was measured at 880 nm with a UV-Vis spectrophotometer (Agilent Cary 60) based on the ascorbic acid molybdate blue method (Wang et al., 2021).

The equilibrium adsorption capacity (q_e) was calculated using Equation 1: q e = C o − C e · V m (1) where q_e (mg/g) is the adsorption capacity at time t, C₀ and C_e are the initial and equilibrium phosphate concentrations in mg/L, respectively. V is the volume of the solution in L, and m is the mass of the composite nanofiber mat in g.

The adsorption kinetic was investigated using the pseudo first order and pseudo second order equations (Equations 2, 3, respectively). q t = q e · 1 − exp ⁡ k 1 · t (2) t q t = 1 k 2 q e 2 + t q e (3) where the rate constant of pseudo first order adsorption as k₁ (min⁻¹) and k₂ as rate constant of pseudo second order adsorption (g·mg⁻¹ min⁻¹); q_e is the amount of phosphate adsorption at equilibrium (mg/g); q_t is the amount of phosphate adsorption at time t (min) in mg/g. Equations 4, 5 were used to determine the adsorption isotherms using Langmuir and Freundlich isotherm models: q e = q max K L C e 1 + K L C e (4) q e = K F C e 1 / n (5) where q_max is the maximum adsorption capacity (mg/g); C_e is the equilibrium phosphate concentration (mg/L); n is the parameter of the Freundlich adsorption isotherm; and K_L (L/mg) and K_F ((mg/g) · (L/mg)^1/n) are the equilibrium constants related to the Langmuir and Freundlich adsorption isotherms, respectively.

Initially, La₂O₃ nanoparticles in the Act:TFA mixture showed homogeneity (Figure 1a). However, La₂O₃ nanoparticle settled within a few hours (Figure 1d). Since active adsorbents settled in the electrospinning solution, a limited amount of La₂O₃ nanoparticles may have embedded into electrospun nanofibers. To ensure that the synthesized membrane contains all the ingredients added to the solution, the solvent must be altered. As mentioned previously, there are other solvents that are known to be compatible with PA6: e.g., formic acid and HFIP. Thus, these solvents were also used as electrospinning solutions. Figures 1b,c shows optical images of as-prepared solutions after the solution was stirred overnight which appeared to be homogeneous. Figures 1d—f shows the images after being left to rest state for 5 h. As shown in Figures 1b,d, the color of formic acid based solutions was yellow compared to milky white for Act: TFA and HFIP based solutions. This might be attributed to the interacting between La₂O₃ nanoparticles and formic acid. The solution made with HFIP remained homogeneous after 5 h (Figure 1f).

To understand the solvent effects, all three solutions are used to electrospun. The mass of PA6, lanthanum oxide nanoparticles and TBAB kept constant while the solvents were varied to understand the solvent effect. As expected, solution properties (i.e., viscosity, surface tension and electrical conductivity) varied significantly depending on the solvent used (Supplementary Table S1).

There are various factors to consider when electrospinning nanofibers. First, is the solution electrospinnable with a high fiber fraction instead of electrospraying. Second, all the components that are in the solution should be electrospun onto the nanofiber membrane except the solvents. As shown in Figure 2, all three solvents resulted in electrospinning over electrospraying with a high fiber fraction. At first glance, there are fiber diameter and bead density variations depending on the solvent used. The formic acid-based solution has the highest viscosity (i.e., 1,664 cP) but resulted in the smallest average fiber diameter of 65 nm (Figure 2a). Previous studies have reported that a high viscosity results in a higher average fiber diameter, but the opposite trend was shown for nanofibers synthesized from formic acid. Additionally, higher bead density (0.16 beads/μm²) was observed compared to when other solvents were used. The mixture of acetone and Act:TFA based solution has a viscosity of 90.1 cP and resulted in nanofiber mats with an average fiber diameter of 97 nm and a fiber fraction of 0.93 μm²/μm² (Figure 2b). The HFIP based solution has a viscosity of 69.2 cP and resulted in nanofiber mats with an average fiber diameter of 162 nm and a high fiber fraction of 0.98 μm²/μm² (Figure 2c). In addition to average fiber diameter and fiber fraction, the diameter distribution is shown in Supplementary Figure S1.

To determine the distribution of La₂O₃ nanoparticles within nanofibers, energy dispersive areal mapping was overlaid on top of the secondary electron scanning electron microscope image (Figures 2d–f). The presence of lanthanum was shown in teal and bromide is shown in red against the carbon of PA6 shown in pink. Compared to the nanofiber obtained from the HFIP based solution, the areal mapping of the nanofiber obtained from Act:TFA based solution shows lower intensity of lanthanum (Figure 2e) which was expected due to agglomeration and settling of La₂O₃ nanoparticles. Unlike other solvents, composite nanofibers synthesized from the HFIP solvent show well distributed La₂O₃ nanoparticles embedded in PA6 matrix.

The stable dispersion observed for La₂O₃ in HFIP compared to other solvents might be attributed to the fact that La₂O₃ reacts with formic acid and TFA, but not in HFIP. Specifically, La₂O₃ reacts with formic acid to form lanthanum formate and with TFA to form lanthanum triflate. These reactions prevent La₂O₃ from remaining in its original form prior to the addition of other chemicals in the tri-composite solution. In contrast, the absence of such a reaction in HFIP likely preserves the oxide form, resulting in a more stable dispersion. This is supported by the XRD pattern shown in Figure 3, which compares composite nanofibers synthesized using different solvents. As shown in the figure, the XRD patterns differed significantly based on the solvent. The composite nanofiber synthesized from HFIP showed dominant peaks at 15.7, 27.4, 28.1, 30.0, 39.6, and 48.7° which are attributed to the standard XRD pattern of lanthanum oxide (JCPDS 04—0856). The sample that was synthesized using the Act:TFA mixture solvent showed a few diffraction peaks (28.1° and 48.7°) that matched to La₂O₃, but the diffraction pattern also showed other peaks that could not be correlated to La₂O₃ nanoparticles. The composite nanofibers electrospun from formic acid did not show any diffraction peaks indicating that the La₂O₃ nanoparticles were not presented.

Figure 4a shows the HR-TEM image of representative PA6/La₂O₃/TBAB composite nanofiber (PLT NF) capturing clusters of lanthanum oxide nanoparticles within PA6 nanofibers. In addition to HR-TEM, elemental mappings of nanofibers were conducted to further understand the distribution of La₂O₃ nanoparticles and TBAB within the composite nanofibers. Elemental mapping of lanthanum (Figure 4c) and bromide (Figure 4d) was obtained from the high-angle annular dark field (HAADF) image shown in Figure 4b. La mapping shows that the La₂O₃ are well embedded within PA6 nanofiber (Figure 4c). Bromide mapping from TBAB shows overlapping with La but also distributed across the NF. Previous work showed that TBAB was used to disperse the aggregated nanoparticles as well as to help smaller Fe₂O₃ nanoparticles enrich themselves to the surface of nanofibers. However, aggregated La₂O₃ nanoparticles can be seen within nanofibers in Figure 4a suggest that TBAB did not act as a surfactant or help with surface enrichment of adsorbents. This may be attributed to the fact that the La₂O₃ nanoparticles (30—50 nm) are much larger compared to previously reported work (i.e., Fe₂O₃ nanoparticles with an average particle size of 3 nm).

To benchmark the adsorption capacity and kinetics of composite nanofiber membranes, suspended metal oxide nanoparticles were tested as a control to determine the maximum adsorption capacity which showed the P adsorption capacity of 21.6 mg/g (Supplementary Figure S2). The results from this preliminary testing served as the control for evaluating the PA6/La₂O₃/TBAB composite nanofibers synthesized using various solvents. Figure 5 shows the batch P removal testing of the composite nanofiber. As shown in the figure, the composite nanofiber membrane showed 97.1% phosphate removal compared to suspended nanoparticles which showed 98.3% removal with an initial concentration of 20.3 mg/L. Our previous work investigated the effect of metal oxide content variation—specifically with iron oxide as the active ingredient—on phosphorus removal efficiency (Wang et al., 2021; Choi et al., 2024). That study identified the optimal metal oxide loading for enhanced removal performance. The membrane prepared with formic acid achieved only 23% removal efficiency, while the Act:TFA combination yielded a 62% removal. The membrane synthesized using HFIP exhibited the highest efficiency at 97% in contrast to the pristine polyamide6 nanofiber, which demonstrated only 4.6% removal. All experiments were performed at an initial concentration of 10 mg/L, with adsorption measured at 1,440 min.

The two most popular kinetic models, the pseudo-first-order and pseudo-second-order models (Equations 2, 3, respectively) were utilized to analyze the data. Figure 6 shows the pseudo-second order model fittings of suspended La₂O₃ nanoparticle and La₂O₃ composite NF membranes with various initial phosphate concentrations (i.e., 2, 5, 10, 15 and 20 mg/L). Table 1 shows the summarized kinetic correlation coefficient from the fitted data using Equations 2, 3. Based on the resulting correlation coefficients (i.e., R² values), it can be concluded that La₂O₃ NP and composite NF membrane are best described by a pseudo-second-order kinetic model, which suggests chemisorption via ligand exchange of negatively charged phosphate ions with the lanthanum oxide surface (Kajjumba and Marti, 2022; He et al., 2022; Fang et al., 2018).

The different adsorption isotherm models (i.e., Langmuir and Freundlich) were fitted to understand the adsorption process. The Langmuir isotherm model fits better than the Freundlich isotherm model (Figure 7). The maximum adsorption capacity (q_max) for the composite nanofiber membrane was determined to be 18.8 mg P/mg which is slightly less than the suspended nanoparticles (21.6 mg P/g) (Table 2).

Polyamide exhibits two main crystalline structures, the α and γ phases (Zhang et al., 2011). The mechanical properties such as Young’s modulus are known to be higher for the α phase than for the γ phase (Liu et al., 2007; Dasgupta et al., 1996). While electrospun PA6 nanofibers predominately exist in the γ phase, the γ phase can transform into the α phase through post treatments, such as thermal processes (Ibanes et al., 2006; Park et al., 2005; Lincoln et al., 2004; Lincoln et al., 2001; Murthy et al., 1985; Carrizales et al., 2008; Dersch et al., 2003; Wang et al., 2014; Sanchaniya et al., 2023). During the post thermal process, the polymer chains are further aligned, leading to higher chain orientation, which has shown an increase in tensile strength in previous studies (Sheng et al., 2016; Ramaswamy et al., 2011; Srithep et al., 2013). SEM images of the tri-composite nanofiber membrane as-spun and post-thermally treated at various temperatures (i.e., 80, 100, 120, 140, and 200°C) are shown in Figure 8. As-spun nanofibers (Figure 8a) and samples annealed at various temperatures up to 120°C (Figures 8b–d) appear to remain similar in morphology, including the average fiber diameter (Supplementary Figures S4a–d). The sample annealed at 140°C (Figure 8e) showed nanofibers with a larger fiber diameter that have been measured to have an average fiber diameter of 240 nm (Supplementary Figure S4e) which is almost double that of the as-spun nanofibers (144 nm). Even with the increase in average fiber diameter at 140°C, optical images show no significant changes in color or shape of the nanofibers (Supplementary Figures S3a–e). When the samples were annealed at 200°C, the nanofiber color changed from white to tan (Supplementary Figure S3f) and shrunk in size. Additionally, the sample annealed at 200°C was rigid and brittle to touch compared to samples annealed at other temperatures, which remained flexible. Figure 8f shows the corresponding SEM image, which reveals deformation of nanofibers with an increased fiber diameters at 481 nm (Supplementary Figure S4f).

XRD patterns show the presence of the γ phase (peak at 21.3°) in pristine nanofiber, with an increase in diffraction intensity as the sample was thermally annealed up to 100°C (Figure 9). When the post thermal treatment temperature was increased above 100°C, two additional peaks started to appear around 20.5° and 24.4° which are corresponding to the (200) and (002)/(202) peaks of the α phase (Liu et al., 2007; Zhang et al., 2011). As the annealing temperature was further increased to 120°C and 140°C, the γ phase transformed to the thermally stable α phase which was shown by decreased intensity of the peak present at 21.3°. The sample that was annealed at 200°C showed two distinct α phase peaks and muted γ phase peak. The α and γ phase percentages have been calculated via peak deconvolution using OriginPro and the results are summarized in Table 3. Figure 10a shows that when the sample was annealed at 80°C, the α phase percentage stayed around 12%–15%. Further increase in annealing temperature from 100 to 200°C significantly increased the content up to 85%. D-spacing and crystallite size for α and γ phase peaks were tabulated in Table 3. The d-spacing can vary depending on the annealing temperature and the introduction of additives due to changes in thermal expansion and contraction which affect the length of the hydrogen bond (Ramesh, 1999; Radusch et al., 1994). Calculated d-spacing for α phase peaks were ∼0.43 and ∼0.37 nm for (200) and mixed (002)/(202), respectively. The d-spacing for the γ phase peak was determined to be ∼0.41 nm. In the α phase, the d-spacing of the (200) plane represents the interchain distance within the sheet and the d-spacing of the (002)/(202) plane represents the inter-sheet distance between the sheets. When a peak is shifted to a higher angle, it represents a decrease in the hydrogen bond length which translates to a decrease in d-spacing. Conversely, a peak shift to a lower angle represents an increase in hydrogen bond length due to thermal expansion and an increase in d-spacing (Zhang et al., 2011). Figure 10b shows a trend of increasing d-spacing values for α₁ and γ peaks at 100°C, maintaining this value with a plateau or decrease in d-spacing values at 200°C, respectively. α₂ peak showed an increase in d-spacing value at 100°C but then showed a gradual decrease as the annealing temperature increased to 200°C. While the composite nanofiber membrane were synthesized using different method and has a different composition compared to pristine polyamide 6 films, or nonwoven fibers, the d-spacing values obtained are comparable to previously reported polyamide 6 materials (Liu et al., 2007; Ramesh, 1999; Cho et al., 2012; Ramesh et al., 1994; King, 2004).

In general, the crystallinity of the α phase-dominated PA6 increased during post thermal annealing (Hindeleh and Johnson, 1978; Vasanthan et al., 2009). When there is a decrease in peak width, the crystallite size increases in the direction of the hydrogen bond due to improvement in lattice order (Hindeleh and Johnson, 1978; Vasanthan et al., 2009). Figure 10c shows the changes in PLT NF crystallite size as a function of annealing temperature. The graphs show that there is an initial dip in the crystallite size of the γ-phase up to 80°C, and then a continuous increase until 200°C. The significant decrease in γ-phase crystallite size at 200°C may be attributed to the meta-stable γ- phase transitioning into thermodynamically stable α-phase at higher temperature through melting and recrystallization (Liu et al., 2007; Lincoln et al., 2001).

FTIR spectra for various tri-composite samples annealed at different temperatures are shown in Figure 11a, spanning the range from 450 to 4,500 cm⁻¹. Polyamide 6 typically exist in α- and γ- crystalline forms. The α-phase consists of sheets of fully extended planar chains joined by the hydrogen bonds between antiparallel chains. In contrast, γ- phase comprises sheets of parallel chains linked by hydrogen bond between the adjacent chains (Arimoto et al., 1965). Figure 11b presents an enlarged view focusing on the characteristic polyamide6 peaks between 920 and 990 cm⁻¹. The bands at 930, and 960 cm⁻¹ are attributed to the α- phase, while the band at 973 cm⁻¹ corresponds to the γ-phase (Porubská et al., 2012; Miri et al., 2011; Vasanthan and Salem, 2001). As the annealing temperature increases, the emergence and sharpening of peaks at 930 and 960 cm⁻¹ indicates a transition to a fully planar conformation, representing transformation to the α- phase. Conversely, the gradual disappearance of the peak at 973 cm⁻¹reflects the conversion of the γ-phase into α- phase. Figure 11c shows peak at 1,200 cm⁻¹—assigned to c-c stretching—is present in both phases but appears sharper in the α-phase due to its higher crystallinity and denser hydrogen bonding compared to the γ-phase.

Differential scanning calorimetry (DSC) was used to examine the melting and crystallization behaviors of electrospun pristine PA6 nanofiber mat and PA6/La₂O₃/TBAB composite nanofiber mat. The glass transition temperature (T_g), crystallization temperature (T_c), and melting temperature (T_m) of the two samples are summarized in Table 4. Figure 12 shows first heating, cooling cycle, and reheating cycles of the samples. The glass transition temperature for the composite nanofiber was calculated to be lower than pristine nanofiber, decreasing from 89.8°C to 68.5°C. This could be due to adding lanthanum oxide nanoparticles and TBAB, which adds defects to electrospun polymeric nanofibers (Zhang et al., 2011; Sridhara et al., 2021). The distinct exothermic crystallization peak shown at 186.8 C for electrospun PA6 nanofiber membrane that typically exist in γ-phase can refer to recrystallization to thermostable α-phase. The melting endotherm occurring with a peak temperature T_m ∼ 217.4°C has a shoulder peak around 200°C and these two peaks can correspond to α- and γ-phase which is typical for samples that have multiple phases (Su et al., 2007; Khanna and Kuhn, 1997). The DSC graph of composite nanofiber mat shows a very small and short exothermic crystallization peak around 132.8°C, followed by two broad endothermic melting peaks. When the composite nanofiber was annealed at various temperatures, XRD data showed an increase in alpha phase peak ratios (Figure 10a) as the calculated glass transition temperature (∼68.5°C) exceeded which can be correlated with the exothermic peak that was shown in Figure 12b. In the first heating cycle, a distinct melting peak is shown around 209.2°C followed by smaller melting peak around 250°C The two melting peaks can correspond to melting of γ peak and then thermostable α peaks, respectively. The XRD in Figure 9 shows disappearance of γ peak at 200 °C and it can be assumed that the melting peak shown at 250°C on DSC data (Figure 12b) corresponds to the remaining α-phase. While the first heating cycle shows two melting peaks at 209.2°C and 250°C, the reheating cycle shows two peaks joint, one main and one as a shoulder endothermic melting peak. While the first heating cycle takes the processing and environmental condition into account, the reheating cycle shows the melting behavior that is correlated to controlled and accurate analysis of the sample as the thermal history is erased by the cooling cycle (Mileva et al., 2012). While the electrospun composite nanofiber showed two distinct melting peaks in the first heating cycle, when reheated, it exhibits similar one endothermic melting peak like pristine PA6 NF.

Figure 13 shows a stress vs. strain plot of PLT NF with different post thermal treatment. Various mechanical property parameters such as Young’s modulus, yield strength, ultimate tensile strength, toughness, and strain to fracture have been calculated and the obtained results are listed in Table 5. Young’s modulus initially decreased followed by an increase until 120°C. Further increases in annealing temperature resulted in a decrease in Young’s modulus. While the Young’s modulus for PLT NF annealed at 200°C showed the highest value, this sample lost flexibility and mechanical durability (Su et al., 2007). From this point on, PLT NF annealed at 200°C is excluded from comparing mechanical durability. Yield strength, which is the maximum stress that can be withstood before permanent deformation, showed its maximum value when the sample was annealed at 120°C.

Figure 14 shows a 3D plot of toughness of the material as a function of annealing temperature and calculated α-phase content. Results show that with an increase in annealing temperature and α-phase percentage, an initial increase and then decrease in toughness of the material is observed. The relationship between annealing temperature and calculated α-phase content is represented in red, toughness as a function of annealing temperature in green, and toughness as a function of calculated α-phase content in blue. These three graphs are combined into a 3D representation, illustrating the interdependence of toughness, annealing temperature, and calculated α-phase content. The 3D graph shows that the toughness of the material reached its maximum when samples were annealed at 100°C, with an α-phase content of 25%. Although the mechanical properties of polyamide 6 were more closely correlated to the α phase, PLT NF annealed at 200°C, which had the highest α phase content, showed poor mechanical properties due to the degradation of the polymer, which resulted in the brittleness of the sample. This indicates that although PLT NF annealed at 200°C had the highest α-phase percentage, this sample was not mechanically durable enough to be used as a filtration membrane. In addition to toughness of the material, PLT NF showed maximized Young’s modulus, Yield strength, and ultimate tensile strength when the sample was annealed at 120°C. Based on this observation, PLT NF annealed at 120°C was deemed suitable for optimization in mechanical durability through the post thermal process.

Phosphorus batch testing was also performed to understand the effect of post thermal processing on the adsorption capacity of the membrane. Figure 15 shows a log-log plot of phosphate uptake of various annealed samples with a fixed initial concentration of P at 20.34 mg/L. It shows that the adsorption of PLT NF annealed at 120°C had higher P removal and faster kinetics compared to as-spun PLT NF and PLT NF annealed at 200°C. The decreased phosphorous adsorption observed from the sample heat treated at 200°C could be due to deformed structure of the electrospun nanofiber membrane. As seen in Figure 8f, the SEM image shows deformed nanofibers that are not porous compared to the rest of the samples (Figures 8a–e). The structural deformation can lead to a reduction in specific surface area and porosity, which directly impacts adsorption efficiency (Cui et al., 2020; Zhu et al., 2021; Wang and Hsiao, 2016; HMTShirazi et al., 2022). Figure 16 shows pseudo first-order and pseudo-second-order models fitting of PLT NF annealed at 120°C and PLT NF annealed at 200°C. Table 1 lists the obtained kinetic data. PLT NF annealed at 120°C and 200°C followed pseudo-second-order kinetics like as-spun composite nanofibers. Additionally, PLT NF annealed at 120°C was faster compared to the rest of the samples at 1.49E^-01 g/(mg·min).

Figure 17 shows the fitted curves using the Langmuir and Freundlich isotherm models, and the calculated adsorption capacity is listed in Table 2. The Langmuir isotherm model fits the experimental data better than Freundlich’s isotherm model for all the samples except for PLT NF annealed at 200°C. PLT NF annealed at 120°C showed ∼19.4% and ∼7.5% decreased in q_max, compared to suspended La₂O₃ NP and as-spun PLT nanofibers, respectively. The difference in q_max value may be attributed to the alteration of porosity and surface area during the post thermal treatment process. Despite the lower calculated q_max, value, PLT NF annealed at 120°C was able to remove 97% of P which is same as-spun PLT NF.

The reusability of polyamide nanofiber membrane has been investigated, with various studies reporting a decrease in adsorption efficiency during second to fifth re-adsorption cycles, likely due to the blockage of active sites. However, the adsorption performance remained relatively high, exceeding 80% against bisphenol (Jasni et al., 2017). Zarrini et al., demonstrated that dye removal using polyamide 6 nanofibers exhibited only a 10% reduction in adsorption capacity, confirming their viability as reusable adsorbents (Zarrini et al., 2017). Other studies have also shown that polyamide 6 based nanofibrous membranes maintain good reusability, with no significant changes observed even after five to eight reusability tests (Amaly et al., 2020; Kakoria et al., 2021).