Comparative Removal of Lead and Nickel Ions onto Nanofibrous Sheet of Activated Polyacrylonitrile in Batch Adsorption and Application of Conventional Kinetic and Isotherm Models

Abstract

We investigated the adsorption of lead (Pb²⁺) and nickel (Ni²⁺) ions by electrospun membranes of polyacrylonitrile (PAN) nanofiber activated with NaHCO₃ (PANmod). Analysis by Fourier-transform infrared spectrometry (FTIR), field emission scanning electron microscopy (FE-SEM), and energy dispersive X-ray spectroscopy (EDX) validated the functionalization of PAN nanofibers with NaHCO₃, and the successful agglomeration of Pb²⁺ and Ni²⁺ onto PANmod. After a rapid uptake of the heavy metal ions (15 min), the equilibrium contact time was attained (60 min) following a linear increase of both adsorption capacity and removal efficiency. PANmod showed a better affinity for Ni²⁺ than Pb²⁺. The adsorption on PANmod was best described by the pseudo-second-order kinetic model for both studied models, supporting chemisorption. By varying the solution pH from 2.0 to 9.0, we found that the adsorption capacity followed an increasing trend, reaching a maximum at the pH of 7.0. Despite increasing adsorption capacities, the removal efficiency of both heavy metal ions exhibited a decreasing trend with increase in initial concentrations. The amount of PANmod directly affects the removal efficiency, with 0.7 and 0.2 g being the optimum dose for maximum uptake of Pb²⁺ and Ni²⁺, respectively. The Langmuir model fitted well the Pb²⁺ adsorption data suggesting monolayer adsorption, and the Freundlich model perfectly fitted the Ni²⁺ adsorption data, indicating heterogeneous adsorption. The estimated values of the mean free energy of adsorption in the D–R isotherm indicated a physical adsorption of both heavy metal ions into the surface of the PANmod.

1. Introduction

Rapid industrial development and urbanization introduced large amount of waste effluents into various environmental compartments [1]. Those effluents contain large amounts of toxic metals/metalloids, drastically polluting the ecosystem. Among toxic metals, nickel (Ni²⁺) and lead (Pb²⁺) abound in industrial wastewaters produced during mining, battery and machinery manufacturing, mineral processing, metal coating, and cable sheathing, as well as by steam electric power plants and stainless steel alloys, plating and tanning industries [2,3]. Elevated levels of the non-biodegradable Ni²⁺ and Pb²⁺ ions in freshwater sources pose a significant threat to animal and human health, because they accumulate in living tissues through the food chain [4]. Exposure to metals/metalloids may induce severe deterioration in human health including memory loss, cardiovascular disease, abdominal pain, weakness, chronic bronchitis and cancer of the lung [5]. Therefore, in pursuance of a pollution-free ecosystem, these toxic metals/metalloids should be eliminated from industrial wastewater before it is discharged into freshwater resources.

Various wastewater treatment techniques such as reverse osmosis, ion exchange, chemical precipitation, electrodialysis, chemical coagulation, solvent extraction and adsorption processes have been employed to remove metals/metalloids from industrial wastewater [6,7,8,9,10]. However, most of these methods are either inefficient or costly [11]. Adsorption technology is emerging as the best technology for wastewater purification due to its simplicity of design, low cost, ease in operation, reduced output of harmful sludge, insensitivity to pollutants, reduced energy consumption, high effectivity and environment compatibility [12]. The adsorption efficiency is directly linked to the materials used in the adsorption process. The most important characteristics of the adsorbent are higher surface area, porosity and the chemical structure [13,14]. In addition, the surface polarity and chemical composition of the adsorbent can affect the attraction forces between adsorbent and adsorbate [15]. Therefore, selection of an appropriate adsorbent is of critical importance. Polymer nanofiber membrane-based adsorbents prepared by the electrospinning technique have been widely used for metals/metalloids removal from wastewater, due to their unique properties of high porosity, specific pattern of fine pores, and larger surface area [16,17]. Polyacrylonitrile (PAN) is a synthetic semi-crystalline organic polymer which does not melt under normal condition, but does at 300 °C [18]. Furthermore, PAN can be electrospun due to its optimum stability, great chemical resistance, high solubility, and great mechanical strength [19,20,21]. Because of these unique properties, PAN is a potential candidate as a nano-fibrous adsorbent for wastewater treatment [22]. The adsorption efficiency of membranes made of PAN can be increased by adding various types of nanomaterials to the PAN/DMF solution before electrospinning, or by applying certain chemical treatments after fabrication to increase its functionality [20,23]. The main objective of treating nano-fibrous membranes with chemical treatments or cross-linking of the nano fibrous membrane is that, under specific conditions, the functional groups on the surface of nanofibers are modified into an appropriate form, depending on the nature of the adsorbate which may be involved in chemical reactions with complexation agents of metal ions [24]. Researchers have been using chemically treated nanofiber membranes such as PAN/EDTA [19], and PAN/NaOH-NaHCO₃ [25], for removal of dyes and metals/metalloids ions since the last decade. It has been assumed that the functionalization of the PAN nanofibers surface is the most important factor for improving the adsorption efficiency of adsorbents used in the elimination of metal ions.

The electrospun sheet of a PAN nanofiber treated with NaHCO₃ (PANmod) has not yet been tested for adsorption of heavy metal ions from aqueous systems. This endeavor is the focus of the current study. Our main objective was to functionalize and characterize the PANmod, and determine its feasibility as an efficient adsorbent for the removal of Pb²⁺ and Ni²⁺ from aqueous solutions. Furthermore, we evaluated adsorption performance employing conventional kinetics and isotherm models.

2. Materials and Methods

2.1. Fabrication and Characterization of the PANmod

We prepared a 10% (w/v) solution of polymer PAN ((C₃H₃N)_n; Sigma-Aldrich, St. Louis, MO, USA) and solvent dimethylformamide (DMF, C₃H₇NO; Sigma-Aldrich, 99.8%), which was homogenized at room temperature using a magnetic stirrer (IKA Werke GmbH, Staufen, Germany) for approximately 12 h. The resulting mixture was injected to an electrospinning device (NanoNC Electrospray system, Seoul, Korea) at a flow rate of 1 mL h⁻¹ using a 5 mL glass syringe. The distance between the needle tip and the grounded metal collector was set at 15 cm and the electrospinning device was operated at high voltage (15 kV) supplied by a high voltage generator (Model: ESN-HV30). The resulting sheet was prepared in a rectangular shape (5 × 7 cm²), dipped in 150 mL of 15% NaOH solution (60 min, 60 °C), and then washed several times with distilled water. Next, to neutralize the excess amount of base, the sheet was immersed in 100 mL of 1.0 M HCl solution. Subsequently, the hydrolyzed sheet was treated with 100 mL of 10% NaHCO₃ solution for 6 h, to prepare the final product. The fabricated PANmod sheet was cut into small pieces, to be used as an adsorbent for heavy metal ions. Its chemical groups and surface morphology were examined by Fourier-transform infrared spectrometry (FTIR) (Vertex 70 Bruker, Karlsruhe, Germany), field emission scanning electron microscopy (FE-SEM) (JSM-7600 JEOL, Tokyo, Japan) and energy dispersive X-ray spectroscopy (EDX) (Inca software installed inside FE-SEM).

2.2. Experimental Setup and Analysis

Stock solutions of Pb²⁺ (1000 mg L⁻¹) and Ni²⁺ (1000 mg L⁻¹) were prepared from analytical reagent grade precursors, lead nitrate (Pb(NO₃)₂; Tianjin Benchmark Chemical Reagent Co., Ltd. Tianjin, China) and nickel nitrate hexahyhydrate (Ni(NO₃)₂·6H₂O; Tianjin Benchmark Chemical Reagent Co., Ltd. Tianjin, China), respectively. 1000 mg of each precursor was dissolved in 1.0 L of deionized water. Each stock solution was further diluted, to achieve the initial concentrations of Pb²⁺ or Ni²⁺ required in different batch tests; the solution pH was maintained at various fixed values (2–9) by using either diluted sodium hydroxide (NaOH) or diluted hydrochloric acid (HCl). 50 mL of metal solution with different initial concentrations (5–50 mg L⁻¹) were agitated for fixed time intervals (5–300 min) in a temperature-controlled shaker (Wise Cube orbital, Wisd. ThermoStable IS-20; Daihan Scientific Co. Ltd., Wonju, South Korea), adding the amount of PANmod to achieve the required dose (0.1–0.9 g) in each batch trial. After shaking them in conical flasks (100 mL) for a predetermined time, the resultant mixture was filtered through 0.45 μm filter paper (Whatman™) and the residual concentration of metal was determined using flame atomic absorption spectrometry (FAAS, Thermo Scientific, ICE 3000 Series, Cambridge, United Kingdom). The amount of metal uptake at time t, that is the adsorption capacity (q_t, mg g⁻¹) and the effective of the adsorption of metal ions into PANmod were determined using the following equations:

$$Adsorption capacity=\left(\frac{\left(C_0-C_t\right)·V}{m}\right)$$

(1)

$$Removal (\%)=\left(\frac{C_0-e}{C_0}\right) \times 100,$$

(2)

where C_o (mg L⁻¹) is the concentration of either Pb²⁺ or Ni²⁺ before the batch adsorption and C_e (mg L⁻¹) is the equilibrium concentration of either Pb²⁺ or Ni²⁺. C_t (mg L⁻¹) is the concentration of either metal ions after sorption for a specified contact time and corresponds to q_t (mg g⁻¹) for evaluating the kinetics of the adsorption process. V (L) is the volume of the metal ions’ solution and m (g) is the mass of the PANmod. For the isotherm models, the equilibrium adsorption capacity, q_e (mg g⁻¹), was evaluated from Equation (1), corresponding to the equilibrium concentration of metal ions measured at equilibrium contact time C_e (mg L⁻¹).

3. Results and Discussion

3.1. PANmod Characteristcs

3.1.1. FTIR Analysis

To investigate modification in the functional groups and surface chemistry of PAN nanofibers, pristine and Ni²⁺ and Pb²⁺ loaded PANmod samples were analyzed by the FTIR technique; the acquired spectra are presented in Figure 1. The sharp peak at 1095 cm⁻¹ is ascribed as stretching vibrations of the C−O group, whereas the peaks at 1225 and 1360 cm⁻¹ are designated as vibrations of C−O−C and C−OH, respectively (Figure 1a). Likewise, sharp peaks appearing at 1450, 3319 and 3498 cm⁻¹ reveal the stretching vibrations of −CH₂−, O−H, and N−H groups [26]. The peaks appearing at 1664, 2243, and 2931 cm⁻¹ correspond to stretching vibrations of the C=O, C−N, and C-H groups, respectively [27]. After the PAN nanofiber was treated with NaHCO₃, the spectral peaks appearing at 1095, 1664 and 2931 cm⁻¹ shifted to the lower wave number positions 1074, 1662, and 2925 cm⁻¹, respectively, while other peaks such as 1225, 1360, and 3319 cm⁻¹ shifted to the higher wave number positions 1240, 1362, and 3321 cm⁻¹, respectively [28].

In addition, new low intensity peaks noted at 1508 and 1550 cm⁻¹ were ascribed to the aromatic N-H secondary amine group [28]; other new peaks appearing around 3621 and 3738 cm⁻¹ were ascribed to stretching vibrations of O−H group [29]. These spectral changes indicate the interactions between PAN nanofibers and NaHCO₃ functionalization by the formation of polar interactions between the functional groups of both components. Polar interactions of the functional groups might help in the adsorption of divalent heavy metal ions.

The vibrational spectra of Ni²⁺ and Pb²⁺ loaded PANmod are shown in Figure 1b. In Ni²⁺ loaded PANmod, the 1070 cm⁻¹ spectral peak is ascribed to C−O bonds. The vibrational spectral peaks at 1451, 1657, 2312, 2926, and 3738 cm⁻¹ are ascribed to the −CH₂−, C=O, C−H, and O−H groups, respectively [27]. In the Pb²⁺ loaded PANmod sample spectra, some peaks were shifted and some new peaks were observed, when compared to PANmod. The shifted spectral features are the 1056, 1404, 1442, 1663, 2313, 2930, and 3739 cm⁻¹ peaks. Moreover, the two new peak positions around 678, and 838 cm⁻¹ were ascribed to metal (M= Ni²⁺, Pb²⁺) and oxygen (M−O or M−O−M) interactions [30]. FTIR spectral analysis revealed a successful adsorption of Ni²⁺ and Pb²⁺ into the PANmod surface.

3.1.2. FE-SEM and EDX Analysis

We investigated by FE-SEM and EDX the changes in elemental composition and morphological features of the pure PAN nanofibers, as well as pristine and Ni²⁺ and Pb²⁺ loaded PANmod (Figure 2). Very smooth, cylindrical, randomly arranged, and beads-free nanofibers of pure PAN are clearly seen in Figure 2a. Their EDX analysis (Figure 2a*) reveal that the C and O content in these nanofibers are 96.56% and 3.44%, respectively. The rough and porous surfaces of the PANmod nanofibers are clearly shown by the FE-SEM image in Figure 2b; the EDX analysis (Figure 2b*) of the same nanofibers reveals their successful functionalization. Additionally, the elemental composition of the PANmod show that the contents of C, N, O, Na, and Cl are 62.90, 31.90, 2.54, 1.47, and 1.19%, respectively.

Furthermore, the agglomeration of Ni²⁺ onto nanofibers can clearly be seen in the FE-SEM image of the post-adsorption Ni²⁺ loaded PANmod samples (Figure 2c). The EDX analysis of Ni²⁺ loaded PANmod (Figure 2c*) shows an increase in C (67.61%) and O contents (17.41%), compared with pristine nanofibers; 14.98% of Ni²⁺ are also observed, suggesting the adsorption of Ni²⁺ on these nanofibers. Similarly, the FE-SEM image of the post-adsorption Pb²⁺ loaded PANmod reveals very small particles attached to the nanofibers, probably Pb²⁺ (Figure 2d). Interestingly, the EDX spectrogram of the PANmod-Pb²⁺ (Figure 2d*) shows a reduction in C (59.39%) and N contents (21.50%), and an increase in O content (8.04%), 11.07% of Pb²⁺ suggests the successful adsorption of Pb²⁺ on PANmod-Pb²⁺. This analysis reveals that the functionalization of the PAN nanofibers with NaHCO₃ is quite beneficial for the adsorption of Ni²⁺ and Pb²⁺ metal ions.

3.2. Influence of the Contact Time and Fitting of Kinetic Models to the Adsorption Data of Both Heavy Metal Ions

We estimated the equilibrium contact time for both heavy metal ions by performing time-series batch experiments; Figure 3 shows the variations in both the adsorption capacity and removal efficiency by PANmod at different initial concentrations of Pb²⁺ and Ni²⁺. The solution pH was maintained at 5.0 and 7.0 for Pb²⁺ and Ni²⁺, respectively, by using a fixed amount (0.5 g) of PANmod. Our observations clearly demonstrate a very rapid uptake of both heavy metal ions at any tested initial concentration, during immediate contact with PANmod. In the first 15 min of contact time, one can see almost 70–80% (Figure 3a) and 90% (Figure 3b) of the maximum removal efficiency of, respectively, Pb²⁺ and Ni²⁺ by PANmod. This could be the result of adsorbed metal ions accumulation on the unused surface and free active sites of the PANmod. After rapid uptake of the adsorbate, a nearly linear increase can be observed in both the adsorption capacity and the removal efficiency, equilibrium being attained at about 60 min since contact started. An additional 4 h of contact yielded insignificant (p = 0.01) effects on the adsorption performance (Figure 3), due to the saturation of the absorption sites of the PANmod. For 20 mg L⁻¹ initial metal concentration, the maximum adsorption capacity at 60 min was 12.7 and 37.4 mg g⁻¹ with corresponding removal efficiency of 32 and 92% for Pb²⁺ and Ni²⁺, respectively. This indicates a dominant effect and better efficiency of the absorption performance of the PANmod for Ni²⁺, compared to Pb²⁺.

Conventional kinetic models were applied to the data collected using the heavy metal ions adsorption on PANmod; and these included the pseudo-first-order (PFO, Equation (3)), pseudo-second-order (PSO, Equation (4)), intraparticle diffusion of Weber and Morris (ID-WM, Equation (5)), and Elovich (Equation (6)) kinetic models.

$$q_t=q_e\left(1-\exp \left(-k_{1}t\right)\right)$$

(3)

$$q_t=\frac{{q_e}^2{k}_2·t}{q_e{k}_2·t+1}$$

(4)

$$q_t=K_{ip}{t}^{1/2}+C$$

(5)

$$q_t=\frac{1}{\beta }\ln (1+\alpha \beta t)$$

(6)

In the above equations, q_t (mg g⁻¹) is the amount of adsorbed Pb²⁺ or Ni²⁺, while q_e (mg g⁻¹) is the equilibrium adsorption capacity of the PANmod for metal ions, k₁ (min⁻¹) is the rate constant for the PFO kinetic model and k₂ (mg g⁻¹ min⁻¹) for the PSO kinetic model. The estimated rate constant in the PSO kinetic model is further used to evaluate the initial adsorption rate h (k₂ q_e²; mg g⁻¹ min⁻¹). For the ID-WM kinetic model, the rate constant is expressed by K_ip (mg g⁻¹ min^1/2) and the boundary-layer thickness by C (mg g⁻¹). In the Elovich kinetic model, α (mg g⁻¹ min⁻¹) is the initial adsorption rate constant and β (g mg⁻¹) expresses the chemisorption activation energy. Figure 4 shows the PSO and Elovich kinetic models fitting of the adsorption data for Pb²⁺ and Ni²⁺ with initial concentrations of 10 and 20 mg L⁻¹; both the nonlinear and linearized approaches were used. The solution pH was maintained at 5 ± 0.2 and 7 ± 0.3 for Pb²⁺ and Ni²⁺, respectively, and the dose of PANmod was 0.5 g.

Table 1. Parameters of the nonlinear pseudo-first-order (PFO), pseudo-second-order (PSO), intraparticle diffusion of Weber and Morris (ID-WM), and Elovich kinetics models for the adsorption of Pb²⁺ and Ni²⁺ onto PANmod (Contact time = 0–300 min, PANmod dose = 0.5 g, pH = 5 ± 0.2 and 7 ± 0.2 for Pb²⁺ and Ni²⁺, respectively).

Kinetic Model	Parameter	Initial Metal Conc. (mg L−1)
		Pb2+				Ni2+
		5	10	15	20	10	15	20
	qe exp (mg g−1)	6.24	10.07	12.51	12.67	21.36	29.43	37.41
PFO	qe cal (mg g−1)	6.11	9.73	12.21	12.24	20.68	28.5	36.51
	k1 (min−1)	0.13	0.13	0.14	0.099	0.31	0.31	0.29
	R2	0.9	0.74	0.92	0.93	0.62	0.65	0.7
PSO	qe cal (mg g−1)	6.49	10.42	12.96	13.05	21.4	29.5	37.89
	k2 (g mg−1 min−1)	0.035	0.02	0.019	0.013	0.034	0.024	0.017
	h (mg g−1 min−1)	1.47	2.17	3.19	2.21	15.57	20.89	24.41
	R2	0.98	0.94	0.99	0.9	0.94	0.95	0.98
ID-WM	Kip (mg g−1 min1/2)	0.15	0.27	0.28	0.33	0.21	0.31	0.45
	C (mg g−1)	4.41	6.78	9.07	8.18	18.53	25.38	32.05
	R2	0.51	0.64	0.49	0.6	0.46	0.5	0.55
Elovich	α (mg g−1 min−1)	70.69	65.3	241.7	41.1	2.31 × 107	1.45 × 107	3.57 × 106
	β (g mg−1)	1.51	0.88	0.8	0.66	1.03	0.72	0.51
	R2	0.79	0.89	0.77	0.68	0.74	0.78	0.81

presents the values of the related parameters in all nonlinear models as estimated using the OriginPro 8.5 software at different initial concentrations of both metal ions.

The PSO kinetic model linearized fitting yielded a perfect fit of the model to the adsorption of both heavy metal ions on PANmod at the initial concentrations tested (10 and 20 mg L⁻¹), with a coefficient of determination (R²) close to unity (1.0). The nonlinear fitting of the PSO kinetic model yielded also a good fitting of the adsorption data with R² in the range 0.90–0.98, with Ni²⁺ presenting a relatively good fit compared to the adsorption data of Pb²⁺, at 20 mg L⁻¹. For the Elovich kinetic model, no significant difference between the R² for nonlinear and linearized fitting was observed, with the linearize fitting yielding somewhat better results than the nonlinear approach. In addition, at 20 mg L⁻¹, the models fitted better the adsorption data for Ni²⁺ than that of the Pb²⁺, as was the case in the PSO kinetic model, but at an initial concentration of 10 mg L⁻¹, adsorption data of Pb²⁺ was better described by the Elovich model that that of Ni²⁺.

The results presented in Table 1 for nonlinear fitting with conventional kinetic models show that, compared to the other models, the PSO model described the best the adsorption of both studied models on PANmod, based on the calculated R² values (0.9–0.99). The only exception was found for 20 mg L⁻¹ in the case of Pb²⁺where the PFO kinetic model proved to be slightly better than the PSO model. This has helped to demonstrate that the chemisorption is the governing mechanism of adsorption involving valence forces due to exchange or sharing of electrons between the metal ions and the PANmod. The estimated adsorption capacities matched well with the experimental values for both PFO and PSO kinetic models with PSO yielding a slightly better agreement than the PFO kinetic model at the equilibrium contact time (60 min), as shown in Table 1. The rate constant in the PFO kinetic model was not affected by changing the initial metal concentration of both heavy metal ions (except at 20 mg L⁻¹ of Pb²⁺) with higher values for Ni²⁺, compared to Pb²⁺(Table 1).

For the PSO kinetic model, the rate constant reduced to almost half for both heavy metal ions as their initial concentration was doubled, i.e., 10 to 20 mg L⁻¹, with Ni²⁺ showing higher k₂ values than Pb²⁺ at equal initial concentrations. The initial adsorption rate was much higher for Ni²⁺ compared with Pb²⁺ (h, Table 1), and increased with initial metal concentrations increase for both types of ions. The ID-WM kinetic model did not fit the adsorption data well for neither of the ion types (R² ranged between 0.46 and 0.64, at initial concentrations of 5 to 20 mg L⁻¹). Both rate constant and boundary-layer thickness increased as the initial metal concentrations increased, exhibiting higher values for Ni²⁺ compared to Pb²⁺ at equal initial concentrations, with the exception of K_ip at 10 mg L⁻¹(Table 1). The Elovich kinetic model fitted reasonably the adsorption data on PANmod for both types of ions, for all initial concentrations, as reflected by the R² values in Table 1.

3.3. Influence of the Solution pH, Initial Ion Concentration, and Dose of Adsorbent on Adsorption Performance and Regeneration of PANmod

The variation in absorption capacity and removal of heavy metal ions obtained by changing the pH (2–9) of the metal solution is shown in Figure 5a. The bath test was conducted at the equilibrium contact time (60 min) using the adsorbent dose of 0.5 g of PANmod and 10 and 20 mg L⁻¹ of Pb²⁺ and Ni²⁺, respectively, as initial ion concentrations. The excess H⁺ created a competition of absorb on the PANmod surface, with divalent Pb²⁺ and Ni²⁺ [31,32] resulting a very low uptake for both ion types at low pH values. The maximum adsorption capacity of nearly 15 and 60 mg g⁻¹ for Pb²⁺ and Ni²⁺, respectively, was found at a solution pH of 7.0, following a trend of increasing adsorption capacity with increasing pH due to the lowered amount of H⁺ and reduced competition for divalent Pb²⁺ and Ni²⁺ to absorb on the PANmod surface [32,33]. Subsequently, a decrease in metal uptake indicated the pH of 7.0 as optimum value, as shown in Figure 5a. The percentage removal, after following an increasing trend for both ion types, decreased from 73% to 21% and from 91% to 73% for Pb²⁺ and Ni²⁺, respectively, as the solution pH increased further from 7.0 to 9.0.

The absorption potential of the PANmod was optimized for the initial concentration of heavy metal ions; the observed changes are depicted in Figure 5b. An equilibrium contact time of 60 min was chosen by fixing the dose of PANmod at 0.5 g while a solution pH of 5.0 and 7.0 was maintained for Pb²⁺ and Ni²⁺, respectively. The PANmod uptake of Pb²⁺ increased linearly as the initial metal concentration increased from 5 to 20 mg L⁻¹; a further increase in the initial metal concentration did not affect in the adsorption capacity, which remained almost constant at the maximum value of approximately 13 mg g⁻¹, as shown in Figure 5b. The Ni²⁺ adsorption, however, continued to increase linearly with its initial concentration increase, attaining the maximum adsorption capacity of 56 mg g⁻¹ at the maximum used value of the initial concentration (50 mg L⁻¹). A concentrated solution with high initial concentration of adsorbate will result in quick adsorption and increased adsorption capacity of a fixed amount of the adsorbent (PANmod in this study) due to strong driving forces between divalent metal ions and active adsorbent sites [34,35]. On the other hand, a decreasing trend in the removal efficiency of both types of ions was seen, due to the limited capacity of the fixed amount of PANmod (0.5 g) and, hence, offering a reduced number of active sites to an increased metal concentration. Finally, a 50% difference different in percentage removal was seen when the initial concentration of Pb²⁺ and Ni²⁺ increased from 5 and 10 to 50 mg L⁻¹, respectively (Figure 5b).

Figure 5c shows variation in the adsorption capacity due to changes in the amount of PANmod (0.1–0.9 g). The equilibrium contact time was 60 min for the batch test, and the solution pH was maintained at 5.0 and 7.0, with a fixed initial concentration of 10 and 20 mg L⁻¹, for Pb²⁺ and Ni²⁺, respectively. The optimum dose of PANmod, corresponding to a maximum uptake of Pb²⁺ and Ni²⁺, was found to be 0.7 and 0.2 g, respectively. A decrease in the adsorption capacity when the adsorbent dose further increased is associated with the high surface area of the PANmod available to absorb a fixed amount of metal ions, that is 10 and 20 mg L⁻¹ for Pb²⁺ and Ni²⁺, respectively. The amount of PANmod has a direct effect on the removal efficiency which increased almost linearly and reached ~90% and about 100% for Pb²⁺ and Ni²⁺, respectively, at the maximum tested value of 0.9 g of PANmod. The adsorption capacities of both metal ions, as estimated in the current study, using the PANmod is compared with other similar adsorbents in previous studies (

Table 2. Comparison of adsorption capacities of PANmod for Pb²⁺ and Ni²⁺ with other similar adsorbents.

Adsorbent	Contaminant	Maximum Adsorption Capacity, mg g−1	Reference
PANmod	Pb2+	15	This study
PANmod	Ni2+	60	This study
Polyaniline grafted chitosan	Pb2+	16	[36]
APAN nanofiber mat	Pb2+	60	[37]
PAN/SiO2 composite nanofiber	Ni2+	138.7	[38]
ACNFs PAN/ZnO	Pb2+	120	[39]
PAN–TiO2–APTES	Ni2+	147	[40]
P-PAN fibers	Pb2+	72.5	[41]

).

Batch experiments were conducted aiming at the regeneration of the PANmod, considering its reusability and adsorption capacities were estimated for both heavy metal ions in different sets of desorption experiments at different initial solution pH and temperature. FE-SEM analysis already revealed a strong bonding and affinity between the PANmod surface and the absorbed metal ions (Pb²⁺ and Ni²⁺), so diluted hydrochloric acid (0.1 M HCl) was used as eluent in addition to a mild heat treatment inside an oven for 24 h. The amount of desorbed metal ions was high at low solution pH (2–5) with average desorption rate reaching almost 100% while the amounts of both metal ions also decreased as temperature increased.

3.4. Application of Isotherm Models to Adsorption Data of Pb²⁺ and Ni²⁺

The equilibrium data was evaluated with the commonly employed isotherm models and the biosorption process was analyzed.

Table 3. Mathematical expressions of the nonlinear and linearized two-parameter isotherm models.

Isotherm	Nonlinear	Linearized
Langmuir	$q_e=\frac{q_m{K}_L{C}_e}{\left(1+K_L{C}_e\right)}$ RL = (1+KLC0) −1	$\frac{1}{q_e}=\frac{1}{q_m}+\left(\frac{1}{q_m{K}_L}\right)\frac{1}{C_e}$
Freundlich	$q_e=K_F{c_e}^{\raisebox{1ex}{$1$}\!\left/ \!\raisebox{-1ex}{$n$}\right.}$	$log{q}_e=log{K}_F+\frac{1}{n}log{C}_e$
Dubinin–Radus-Kevich	$q_e=q_m\exp \left(-K_{DR}{\epsilon }^2\right)$ $\epsilon =RT\ln \left(1+\frac{1}{C_e}\right)$	$\ln q_e=\ln q_{m }-K_{DR}{\epsilon }^2$ E = $\frac{1}{\sqrt{2K_{DR}}}$
Halsey	$q_e=\exp \left(\frac{\ln k_{H }-\ln C_e}{n_H}\right)$	$\ln q_{e=}\frac{1}{n_H}\ln k_{H }-\frac{1}{n_H}\ln C_e$
Temkin	$q_e=\frac{RT}{H_{ads}}\ln (K_T C_e)$	$q_e=\frac{RT}{b_T}\ln A_T+\frac{RT}{b_T}\ln C_e$
Harkins–Jura	$q_e={\left(\frac{A_{HJ}}{B_{HJ}-log{C}_e}\right)}^{\frac{1}{2}}$	$\frac{1}{q_e^2}=\left(\frac{B_{HJ}}{A_{HJ}}\right)-\left(\frac{1}{A_{HJ}}\right) log{C}_e$

shows the original (nonlinear) and linearized mathematical expressions of the two-parameter isotherm models; the Sips and Redlich–Peterson (R–P), among the three-parameter isotherm models are expressed by:

$$q_e=\frac{q_m{K}_S{C_e}^{n_S}}{1+K_S{C_e}^{n_S}}$$

(7)

$$q_e=\frac{q_m{K}_S{C_e}^{n_S}}{1+K_S{C_e}^{n_S}}$$

(8)

Parameters C_e (mg L⁻¹) and q_e (mg g⁻¹), which are used in all expressions of Table 3, represent the residual metal ion concentration in the solution and the corresponding equilibrium adsorption capacity. q_m (mg g⁻¹) denotes the estimated maximum adsorption capacity in the Langmuir, Freundlich, Dubinin–Radushkevich (D–R) models (

Table 4. Estimated parameter values in the nonlinear and linearized isotherm models for Pb²⁺ and Ni²⁺ (contact time = 60 min, PANmod dose= 0.5 g, pH = 5 ± 0.2 and 7 ± 0.2 for Pb²⁺ and Ni²⁺, respectively).

Isotherm	Parameter	Pb2+		Ni2+
		Nonlinear	Linear	Nonlinear	Linear
	qe exp, mg g−1	13.05 (against 30 mg L−1)		56 (against 50 mg L−1)
Langmuir	qm, mg g−1	13.67	14.12	49.38	44.25
	KL, L mg−1	0.56	0.44	2.66	11.89
	RL	0.056	0.07	0.0075	0.002
	R2	0.91	0.96	0.62	0.78
Freundlich	qm, mg g−1	12.90	13.75	63.44	64.22
	KF, ((mg g−1)(L mg−1)1/n)	7.74	7.20	31.37	33.44
	n	6.67	5.26	5.56	6.00
	R2	0.64	0.72	0.97	0.97
D–R	qm, mg g−1	12.74	12.7	48.55	44.54
	KDR, (mol kJ−1)2	66 × 10−8	60 × 10−8	15 × 10−8	2 × 10−8
	E, kJ mol−1	0.87	0.91	1.8	5.0
	R2	0.96	0.98	0.31	0.66
Halsey	qe cal, mg g−1	12.3	11.02	54.15	62.3
	nH	−6.7	−5.26	−5.7	−6.00
	KH	0.000	0.65	0.000	0.177
	R2	0.64	0.72	0.97	0.97
Temkin	KT, L mg−1	45.20	45.09	278.57	278.58
	Hads, kJ mol−1	0.56	1.4	0.17	0.4
	R2	0.71	0.74	0.92	0.97
H–J	AHJ, mg g−1	76.9	89.29	175	1428.6
	BHJ	4.5	1.96	2.9	1.71
	R2	0.63	0.65	1.00	0.95
Sips	qm, mg g−1	12.87		64.81
	KS, L g−1	0.29		0.5
	nS	0.76		0.18
	R2	0.97		0.99
R–P	KRP, L g−1	4.73		7.1
	α, L mg−1	0.2		2.25
	β	0.15		0.82
	R2	0.97		0.99

) and the Sips isotherm model (Equation (7)). The dimensionless separation factor R_L (Table 3) in the Langmuir model is used to describe the shape of the isotherm [42] and is dependent on the Langmuir constant, K_L (L mg⁻¹), which shows the affinity between the heavy metal ions and the PANmod surface. In the Freundlich model, the favorability of the adsorption system is shown by the dimensionless n in the range 2–10 [43], while the model constant is expressed as K_F ((mg g⁻¹)(L mg⁻¹)^1/n). The mean free energy of adsorption E (kJ mol⁻¹, Table 3) in the D–R model is relevant to the model constant K_DR (mol kJ^–1)² and is used to describe whether the adsorption system is physical (E < 8 kJ mol⁻¹) or chemical (E > 8 kJ mol⁻¹). T (Kelvin) and R (8.314 J mol⁻¹·K⁻¹) are used in the D–R and Temkin models and represent the absolute temperature and universal gas constant, respectively. The terms n_H and k_H are the exponent and constant in the Halsey isotherm, respectively, while both A_HJ and B_HJ are constants evaluated in the Harkins–Jura (H–J) model. The heat of adsorption and the binding constant at equilibrium in the Temkin model are expressed by H_ads (kJ mol⁻¹) and K_T (L mg⁻¹), respectively (Table 3).

The constant of the Sips model, K_S (L g⁻¹), is related to the energy of adsorption, while the degree of heterogeneity is expressed by the dimensionless number (n_S) in Equation (7). Both α (L mg⁻¹) and dimensionless β (0–1) are parameters used in R–P isotherm with the model’s constant being expressed by K_RP (L g⁻¹) in Equation (8). Table 4 lists the values of the parameters mentioned above for all the isotherm models evaluated, by using the slope and intercept values in the respective linearized plots; OriginPro 8.5 and CurveExpert Professional software were used for the nonlinear approach. The suitability of the nonlinear two-parameter and three-parameter isotherms to the experimental data of both heavy metal ions is shown in Figure 6 for an equilibrium contact time of 60 min, as described earlier. We analyzed the adsorption data for batch tests performed at 30 °C by using 0.5 g of PANmod; the solution pH was maintained at 5 ± 0.2 and 7 ± 0.2 for Pb²⁺ and Ni²⁺, respectively.

As shown in Figure 6a, the Langmuir model poorly fitted the Ni²⁺ adsorption; the same result was obtained by applying the Freundlich model to Pb²⁺ adsorption data (Figure 6b). This is reflected by low R² for both nonlinear and linearized fittings of the Ni²⁺ and Pb²⁺ adsorption in the Langmuir and Freundlich isotherm models, respectively (Table 4). The theoretical adsorption capacities were in good agreement with the experimental values for both the Langmuir and Freundlich isotherm models for Pb²⁺, but a ~10% disagreement between the experimental and calculated values was observed for Ni²⁺ in both models. The values of the separation factor in the Langmuir model, 0 < R_L < 1, and the dimensionless factor n of 2–10 in the Freundlich model indicated that both isotherms fit well the adsorption of Pb²⁺ and Ni²⁺ by both the nonlinear as well as the linearized approach (Table 4). Considerably higher values of the model constants for both Langmuir and Freundlich isotherms were obtained for Ni²⁺, in comparison to Pb²⁺; the K_L values of 2.66 and 11.89 L mg⁻¹ (Table 4) show the better affinity between the PANmod surface and Ni²⁺, compared with Pb²⁺, respectively. On the other hand, high R² values and close agreement of the theoretical adsorption capacities with the experimental values for the Langmuir model suggest a monolayer adsorption of Pb²⁺ on the PANmod surface.

Figure 6c shows that the D–R isotherm fitted the Ni²⁺ adsorption data poorly, which is further reflected by the very low R² value 0.31 (Table 4); however, the model described well the Pb²⁺ adsorption data with both the linear and nonlinear approaches; the calculated adsorption capacities are in close agreement with those of the experimental values. The estimated values of the mean free energy of adsorption, E < 8 kJ mol⁻¹ clearly reflect a physical adsorption of both types of heavy metal ions on the PANmod surface (Table 4). Based on the observed R² values, the Temkin isotherm also represented better the adsorption data of Ni²⁺ compared to Pb²⁺, as shown in Table 4. This reflects the heterogeneous adsorption [44] of Ni²⁺ with uniform distribution of binding energies on the surface of the PANmod. The same is also proposed based on the suitability of the Freundlich isotherm for representing the adsorption data of Ni²⁺.

As in the case of the Temkin isotherm, both the Halsey and H–J models proved to fit better the adsorption data for Ni²⁺ compared to Pb²⁺, as reflected by the high R² values shown in Table 4. Based on the results of the Temkin model, the adsorption of Ni²⁺ on the PANmod occurred with a lower heat of adsorption (0.17 kJ mol⁻¹) but a much higher equilibrium binding constant (288 L mg⁻¹) compared to Pb²⁺ (Table 4). The suitability of both the Halsey and H–J models in describing the adsorption process relates the multilayer adsorption of Ni²⁺ on PANmod with its heteroporous nature, suggesting a heterogeneous pore distribution. Figure 6d illustrates the R–P model excellent fit of the adsorption data of both heavy metal ions; the high values of the R² in Table 4 confirm the same for the other three-parameter isotherms that were used, that is the Sips model, proposing both homogeneous and heterogeneous adsorption of the Ni²⁺ and Pb²⁺ ions on PANmod [45,46,47,48]. In describing the adsorption of Ni²⁺, the Sips model slightly overestimated the theoretical adsorption capacity compared to the experimental value; a lower energy of adsorption and lesser degree of heterogeneity was evaluated for Ni²⁺ in comparison to the adsorption of Pb²⁺.

4. Conclusions

Here, we studied the successful adsorption of the heavy metal ions Pb²⁺ and Ni²⁺ on an electrospun membrane of PAN nanofiber activated with NaHCO₃ (PANmod). The FTIR analysis revealed the polar interactions between PAN nanofibers and NaHCO₃, helping the successful adsorption of Ni²⁺ and Pb²⁺ on the PANmod surface. Furthermore, the FE-SEM images and the elemental analysis using EDX revealed the functionalization of the PAN nanofiber with NaHCO₃, and the agglomeration of Pb² and Ni²⁺ on PANmod.

After rapid uptake of the heavy metal ions by the PANmod during immediate contact (15 min), a linear increase in both the adsorption capacity and removal efficiency was observed; equilibrium was attained 60 min after the contact started. Among the conventional kinetic models, the PSO kinetic model describes best the adsorption of both ion types on the PANmod sample; the calculated R² values (0.9–0.99) closely agreed with the estimated adsorption capacities, and also with the experimental values.

The percentage removal, after increasing for both heavy metal ions from 2 to 7, decreased from 73% to 21% and from 91% to 73% for Pb²⁺ and Ni²⁺, respectively, as the solution pH increased further from 7.0 to 9.0. The uptake of Pb²⁺ and Ni²⁺ by the PANmod increased linearly as the initial concentration of metal ions increased from 5.0 to 20 mg L⁻¹; at the maximum value of the initial ion concentration (50 mg L⁻¹), the maximum adsorption capacities corresponding to the Pb²⁺ and Ni²⁺ ions were 13 and 56 mg g⁻¹, respectively. The optimal dose of PANmod, corresponding to a maximum uptake of Pb²⁺ and Ni²⁺, was found to be 0.7 and 0.2 g, respectively; the adsorption capacity decreased with further increase in the adsorbent dose.

Among the two-parameter isotherm models, the Langmuir model fitted better the Pb²⁺ adsorption data, suggesting a monolayer adsorption. The Freundlich model fitted the Ni²⁺ adsorption on the surface of the PANmod better than that of Pb²⁺, reflecting a heterogeneous adsorption of Ni²⁺ with a uniform distribution of binding energies on the surface of the PANmod. The estimated values of the mean free energy of adsorption for the D–R isotherm, E < 8 kJ mol⁻¹, reflected clearly a physical adsorption of both the heavy metal ions on the surface of the PANmod. As in the case of the Temkin isotherm, both the Halsey and H–J models proved to be better fits of the adsorption data for Ni²⁺ than that of to Pb²⁺, revealing the multilayer adsorption of Ni²⁺ onto PANmod with heteroporous nature and suggesting heterogeneous pore distribution.