Preparation of flexible electrospun AOPAN/PVDF membranes for removing Pb2+ from water

Electrospun AOPAN/PVDF composite nanofiber membranes for metal ions treatment have been prepared by coaxial electrospinning. AOPAN shell layer was modified chemically for adsorbing metal ions by chelation, whereas the chemically stable PVDF inner core was for maintaining mechanical stability. Polymer concentration and applied voltage had obvious influence on the characteristics of the fibers’ structure, morphology and strength. The amidoxime reaction was sensitive to the pH value of solution, and it was found that the alkaline condition hindered the reaction. The characterization by SEM, FTIR and XRD showed that the AOPAN/PVDF membrane retained the core–shell structure integrity after chemical modification. In the static and dynamic adsorption experiments, the mechanical strength of the AOPAN/PVDF membrane did not change obviously within 5 cycles of adsorption and regeneration. In addition, the AOPAN/PVDF membranes showed a certain level of efficiency in removal of Pb2+ in aqueous solution; the adsorption capacities of the membranes in the 5th run were higher than 45% of the adsorption of the corresponding fresh membranes. The work provides a potential approach for preparing membranes having good feasibility for practical application in adsorption separation of metal ions.


Introduction
Heavy metals are among the pollutants that have toxicological effects on humans (Järup 2003). Some heavy metals are associated with destroying cellular structure and thus life-threatening. Therefore, progressively stricter limit for the levels of environmental heavy metals have been enforced (Jobby et al. 2018;Wu et al. 2019;Wang et al. 2017). Water is for everybody indispensably, and water pollution has always been the focus in environmental protection. A series of exogenous and endogenous control have been investigated and/or applied for tackling heavy metals in aquatic system (Fu and Wang 2011). They are mainly relying on chemical precipitation, coagulation and flocculation, membrane filtration, ion exchange, electrochemical methods and adsorption (). All these technologies are featured with flexible design and easy operation. The strength of adsorption lies in the combination of its selectivity toward minor components and high removal efficiency. Therefore, it is of particular interest for treating heavy metals at low concentrations and generally adopted as a polishing step for wastewater processing (Xiaolei and P.J. J. A. 2013).
For practical application, a good adsorbent should have large capacity, fast adsorption rate, facile separation from water and easy regeneration (reactivate adsorption capacity by desorption) (Fu and Wang 2011). In the past decade, nanofibrous membrane (or mat in some cases) by electrospinning emerges as a new type of adsorbent. The membranes could be used either solely as adsorbents or in a membrane separation mode. With the latter, the characteristics of a separation by membrane make the process for metal recovery more continuous. Polymers dominate the materials used for producing nanofibrous membrane by electrospinning. With appropriate adjustment of material and fabrication, large specific surface area and high porosity could be delivered, which satisfies the thermodynamic and kinetic requirements in adsorption separation (Wang 1 3 51 Page 2 of 15 et al. 2018). The PAA/SA nanofibrous membranes exhibited a maximum adsorption capacity up to 591.7 mg/g. Compared with powder or particle adsorbents, the assembly in a membrane structure ensures easier detachment from water. Moreover, high degree of adsorbent regeneration and metal recovery has been reported for some nanofibrous membranes in a relatively easy manner (Hong et al. 2015;Zhang et al. 2019). These points are meaningful in terms of both metal resource sustainability and waste reduction. Many conventional adsorbents are not separable and/or regenerable; they become heavy metal laden solid waste after utilization and their disposal commonly involves high cost (Sethurajan et al. 2018). Obviously, nanofibrous membranes have a big potential for heavy metal treatment.
In spite of the achievement in separation performance, large-scale implementation of nanofibrous membrane is challenged by inferior mechanical durability (Huang et al. 2013). This is firstly attributed to several factors in the process of spinning, including material intrinsic property (Al-Saleh and Sundararaj 2011), fiber arrangement (Hou et al. 2005), inter-fiber fusion and inter-molecular interaction, etc. (Obaid et al. 2016). The potential solutions to these problems are summarized in several researches and reviews (Pereao et al. 2019). Another more critical aspect is the post-treatment by chemical modification to give the membrane customized functionality on fiber surface (Pereao et al. 2019;DAISUKE ISHIMURA, Y.M.A.H. 1998). Kim et al. treated polyacrylonitrile (PAN) fibers by plasma to generate functional groups such as -COOH and -COOR in research about PM 2.5 removal. The fibers collapsed when the duration of modification exceeds 120 s (Kim et al. 1996). Hydroxide sulfonyl groups (-SO 3 H) have the capacity for ion exchange. Kwak's research showed that polyethersulfone (PES) fiber mats were destroyed when sulfonation lasted for more than 150 min (Kwak et al. 2013). Yet another chemical group that has been followed with great interest is amidoxime (AO), one of the most effective structure toward chelating a wide group of metal ions (e.g., U 6+ , Pb 2+ , Cd 2+ , Cu 2+ ) (Aguila et al. 2019;Yin et al. 2018;Zheng et al. 2019;Ren et al. 2018). It is generally formed by reaction between nitrile groups and hydroxylamine (NH 2 OH). Khalid et al. prepared the AO-containing PAN nanofibers and found that the flexibility of the nanofiber mats decreased as the nitrile conversion increased (Saeed et al. 2008). Similar finding was reported in some other works (Neghlani et al. 2011). Mats contraction in size with high degree of AO formation (> 45%) was also observed (Neghlani et al. 2011). The reason for the above structural change or even deterioration of nanofiber is that the various chemical reactions originally targeting surface adjustment may cause irreversible change to the overall fiber if the reaction agents penetrates into the deeper part (Sagitha et al. 2018). Consequently, there is a compromise between separation performance and mechanical stability upon chemical modification.
In recent years, coaxial electrospinning to prepare fibers with composite core-shell structure receives much attention in catalysis, energy, filtration, tissue engineering, pharmaceutical, etc. . As compared with integral structure, the composite one can fulfill more diversified purpose by material selection and additional control. In Li et al.'s work, ultra-thin shells were designed for enhancing the fast dissolution of poorly water-soluble drugs . Removing the core layer by an extraction process delivered hollow nanofibers exhibiting good performance in filtration and lithium ion batteries (Aslan et al. 2016;Li et al. 2017). The coaxial electrospinning was used to protect glucosamine sulfate (GAS) from environment by the outer polycaprolactone (PCL) layer, so as to control drug release over time . The technology has also been adopted for tackling the typical conductivity-strength trade-off of proton exchange membrane (PEM). Key to the breakthrough is the enhanced fiber strength by extra electrostatic force provided by the inner nozzle wall . As far as coaxial operation is concerned, it presents as a more convenient alternative to multi-step fabrication targeting similarly the composite structure .
In current work, flexible PAN nanofibrous membranes with AO functionality were prepared for treating heavy metals by adsorption. The fibers were designed with core-shell structure and fabricated by coaxial electrospinning. AO groups were introduced by reaction of PAN in shell layer with NH 2 OH, as mentioned above (Saeed et al. 2008). The desired flexibility relies on the core layer made from PVDF. This is a polymer that has excellent chemical stability in many environments and is expected to maintain structural integrity during PAN chemical modification (Graphical Abstract). The nanofibrous membranes fabricated were tested in adsorption and filtration for Pb 2+ , a possibly carcinogenic chemical. The tensile characterization was conducted to confirm to what degree the PAN/PVDF composite nanofibrous membrane can maintain the flexibility and strength during chemical modification and regeneration.

Preparation of membranes with core-shell nanofibers
PAN and PVDF powders were vacuum-dried at 120 °C overnight before dope preparations. They were, respectively, added into DMAc solvent and mixed by magnetic stirrer at 60 °C for 4 h, until a homogeneous solution is formed. Place the solution at room temperature until the bubbles are removed; PAN and PVDF solutions of different concentrations were transferred into two 10-mL syringes, respectively, which were then connected to coaxial needles through a polytetrafluoroethylene (PTFE) tube. The nanofibers were fabricated by electrospinning device (JDF04, Changsha Nano Instrument Technology Co., Ltd, China) at a flow rate of 0.5/0.5 mL/h for 2 h. The needle is fixed 15 cm above the roller; the diameter and length of the roller are 92 mm and 19.5 cm, respectively, and the rotation speed is 300 rpm. The temperature and relative humidity during electrospinning were 25 °C and 55%, respectively. After fabrication, the membrane was placed in oven at 60 °C for removing the residual solvent.

Chemical modification
The outer PAN was chemically modified to generate amidoxime chelating group, and the reaction mechanism is given in Fig. 1a. The modification was carried out by hydrothermal method. Solutions with different ratio of anhydrous sodium carbonate (Na 2 CO 3 ) and hydroxylamine hydrochloride (NH 2 OH·HCl) (prepared by dissolving NH 2 OH in acid) were added to 100 mL deionized (DI) water at 70 °C, followed with immersing the membrane into the solution for the modification. The membranes attached to non-woven support temporarily were fixed by clamps to prevent shrinkage. The membranes after modification were symbolized AOPAN/PVDF and rinsed by deionized water for three times to removal residual chemicals. The AOPAN/PVDF membrane was then freeze-dried under vacuum.

Adsorption and regeneration experiments
The static adsorption experiments were carried out in well-sealed blue cap bottles, each of which contained 50 mL aqueous solution of Pb(NO 3 ) 2 at 200 mg/L with the membrane dosage of 100 mg/L and was shaken for 24 h. To analyze the pH effect, solutions with concentrations of 200 mg/L and the pH values of the solution were adjusted from 2 to 5 by HCl and NaOH. For regeneration study, the AOPAN/PVDF nanofibrous membranes were first equilibrated with Pb 2+ in solution at concentration of 200 mg/L for XX hours; the membrane was then taken out and removed of residual liquid. Thereafter, the AOPAN/PVDF membranes was immersed into 50 mL 0.1 M HCl for 1 h.
The Pb 2+ concentration of solutions before and after adsorption were determined. The amount of Pb 2+ ions adsorbed on the membrane (mg/g) was calculated by the following equation (Habiba et al. 2017;Dognani et al. 2019): where q e is the amount of Pb 2+ adsorbed (mg/g), C 0 and C e is the initial and the equilibrium Pb 2+ ions concentration (mg/L), respectively, V is the solution volume (L), and m is the amount of membrane used (g).
In the flow-through test, the dynamic adsorption was performed on a dead-end filtration setup using a single-pass flow with pH = 5 (Fig. 1b). The setup consists of a peristaltic pump (BT-600EA, JIHPUMP, China) and a circular sand core support with an active diameter of 4 cm. The filtrates were sampled at a definite interval. The dynamic adsorption where v t (mg/min) was the adsorption rate at time t (min); C 0 (mg/mL) was the initial concentrations of Pb 2+ solution; C t (mg/mL) was the concentration of the filtrate at time t; v (mL/min) was the flow rate. The dynamic adsorption capacity per unit membrane area (q d ) was calculated by was calculated by following equation (Fan et al. 2019): where q d (mg/cm) was the adsorption capacity per unit area of the membrane; q e ' (mg) was the total adsorption capacity of the membrane, which is equal to the area under v t -t curve; A (cm 2 ) is effective area of the membrane.
The concentration of Pb 2+ in aqueous solution was measured by inductively coupled plasma-atomic emission spectroscopy (Agilent Technologies, ICP-5100-VDV, Malaysia). The data were used for the calculation in Eqs. (1-3).

Other characterization
The morphology of the membranes was analyzed by double beam electron microscope (FESEM, Helios Nanolab G3 UC, Czech). The diameter of fibers was measured by software ImageJ. Chemical structure of the membranes was characterized by Fourier transform infrared spectrometer (FTIR) (Bruker, TENSOR27, USA). The Mechanical strength test of the membrane was performed with a commercial test machine (WDW-100 N, Jilin Tanhor Testing Machine). The TEM images were observed by spherical aberration transmission electron microscope (FEI, Titan G2 60-300, USA). The specific surface area was measured at 77 K with a liquid nitrogen bath and analyzed using Quadrasorb SI-3MP (Quantachrome, Quadrasorb SI-3MP, USA). The wide-angle X-ray diffraction (WAXD) patterns were detected by an X-ray diffractometer (West Germany's Siemens Company, D500, Germany). Membrane thermal stability was tested by thermogravimetry-differential scanning calorimetry (TG-DSC) (Netzsch Instruments Manufacturing Co., Ltd., NETZSCH STA 449F3, Germany).

Preparation of core-shell fiber membrane
In electrospinning, the polymer solution is pumped out of the needle at a constant speed, overcoming the surface tension by internal charge repulsion and external electrostatic force and stretching to form nanofibers [39]. The physical properties of the nanofibers are influenced by several factors, such as polymer concentration, operation voltage, temperature and humidity (Drosou et al. 2018).
In electrospinning process, the voltage adopted is very critical for achieving solution stretching to form fibers. Only when attaining threshold and higher voltage can the fibers be developed. Researchers have some different views about the influence of the voltage in electrospinning. Zhang et al. suggested that the volume ejection rate for polymer solution was higher at high voltage, which promotes the formation of fibers with larger diameter (Zhang et al. 2005). More works found that a higher voltage generated higher degree of polymer solution stretching due to a stronger electric field and hence delivered smaller fiber diameter (Bhardwaj and Kundu 2010). We used a camera to observe the process characteristics of the coaxial electrospinning (vedio1-3). Confirmed by TEM pictures is that both voltages of 13 and 14 kV can lead to the formation of a core-shell structure. As can be seen from Fig. 2 a1-a3 of the pictures captured from the videos, the linear regions of the ejected fibers were stable and regular at 13 and 14 kV. When the voltage reached 15 kV, two jet streams appeared between the tip and collector, as shown in the red circles in Fig. 2-a3. Multi-ejection at high voltage has been reported by some researches (Zanjani and J, Saner Okan B, Letofsky-Papst I, Yildiz M, Menceloglu YZ 2015). Another possible reason in current work is that the two solutions respond to charge and electrostatic field in different manner. When voltage is high (i.e., 15 kV), the difference might be big enough to allow their peeling off from each other after extrusion. As result, approximately singlelayer fiber structure was formed, as shown by TEM (Fig. 2-c3). The problem with 13 kV is that some bead-on-string phenomenon occurs, as shown in Fig. 2-c1. The fibers connected with beads were thinner, which might collapse easily when subject to mechanical stress . Lower applied voltage results in weaker electrostatic attraction between the needle tip and the collector, generating lower drawing stress in the jet, which leads to the polymer solution aggregation in the needle tip ( Fig. 2-a1). This is the major mechanism for the bead development (Lee et al. 2003). It has been revealed in some researches that higher voltage brought about unstable and unpredictable, which formed multiple jet flow during electrospinning process. Therefore, 14 kV was the most suitable voltage for the coaxial electrospinning.
The effect of polymers concentration on core-shell fibers morphology were studied in this work and different combinations of PAN concentration in outer layer and PVDF concentration in inner layer were selected, as shown in Table 1. It was found that when the concentrations of solutions for both inner and outer layers were lower than 10 wt%, there existed fiber disruption and bead formation. Raising polymer concentration increased the spinnability; larger and more uniform diameter of fibers and reduced number of bead-onstring phenomenon were resulted with higher concentration. This observation is consistent with that reported in Fong et al.'s and Chen and Jiang's research (Fong et al. 1999;Chen and Jiang 2020). The diameter change was believed to be related to the increase in the solution viscosity (Dong et al. 2004). Nonetheless, higher viscosity will enhance the propensity of needle clogging during electrospinning. Observed was also the dependence of the integrity of the dual-layer structure on solution concentration. As shown in Fig. 3 a2-a4, some fibers exhibit a cross section with an incomplete outer layer (C-shaped). Additional phases introduce an interface interaction (miscibility and stress due to the viscosity difference), and the individual physical behavior of different phases (solidi cation and conductivity) also complicates the coaxial electrospinning . When the concentration of the outer solution is low, the polymer chain entanglement was not sufficient to generate continuous thinning and deposition that can cover fully the core layer. On the other hand, an insufficient core solution cannot support a continuous smooth core phase (Ou et al. 2011). In other words, the concentrations of both layers are critical for obtaining a uniform and complete morphology.
The mechanical strength of different fiber membrane was characterized by a commercial test machine (WDW-100 N, Jilin Tanhor Testing Machine), and the results are displayed in Fig. 4. It was shown that increasing the concentration of PVDF in the inner layer improved the mechanical strength of the nanofibrous membrane, and the relationship is approximately linear. As discussed in the above section, the core-shell fibers attained bigger diameter with higher   (Table 1). The change in the diameter is one of the factors underlying the variation in mechanical strength . For membranes using a concentration in the range of 10-6%, the strength of the membrane is too low to be accurately analyzed. Also revealed in Fig. 4 is the change in mechanical strength with PAN concentration at fixed PVDF concentration. When PVDF concentration was equal and lower than 8 wt% and PAN is 10%, the fibers could not be formed. For PVDF at 10 and 12%, an abrupt enhancement followed by a mild increase as PAN content changed from 10 to 14%. The mechanical strength of the core-shell fiber depends on both layers. When PAN is at low level, the outer layer is not complete (i.e., C-shape indicated in Fig. 2), which presents as defects for the overall structure. For fibers with completeness in outer layer, the slight increase in the mechanical strength with PAN content might be due to the minor proportion of the outer layer in the overall structure. Taking structural integrity and mechanical strength as the major standard, the 12 wt% (outer)-12% (inner) combination was selected for the following chemical modification.

The effect of chemical modification on core-shell fibers
The PVDF/PAN core-shell fibers were treated by hydrothermal method to obtain amidoxime functional groups (the reaction route is shown in Fig. 1a). Our preliminary experiment showed that 70 °C was appropriate for the amidoxime formation. The reaction rate slows down with the temperature drops, while at higher temperature the membrane will shrink seriously. The reactants played an important role as well in the conversion rate of nitrile groups and we investigated three different ratios of hydroxylamine hydrochloride (NH 2 OH·HCl) and Na 2 CO 3 . The concentration of amidoxime group in the core-shell fibers is calculated by the following formula (Saeed et al. 2008): (4) where W 0 (g) is the mass of nanofiber mat before reaction; W 1 (g) is the mass of nanofiber mat after reaction; M 1 is the molecular weight of hydroxylamine (33 g/mol).
The reaction rate constant can be obtained by using the content of amidoxime group at different hydroxylamine concentrations. The equation is as follows (Alakhras et al. 2005): where V (mmol/s) is the reaction rate; C i is the component concentration at time i (mol/L); t is time; n is the reaction order; K is the reaction rate constant.
The value of K is estimated from the intercept of the straight curve shown in Fig. 5, and the other parameters in Eq. (5) are given in Table 2. The excess of both reactants hindered amidoxime formation. When Na 2 CO 3 was excessive, the reaction rate decreased obviously. Higher content of Na 2 CO 3 increases the pH of the solution and reduces the amount of free hydroxylamine ions (NH 2 OH).
For the pure PAN fibrous membranes, the fibers became very brittle when the conversion rate of nitrile group is over 35% (Saeed et al. 2008). The conversion rate of amino indicates the degree of amidoxime formation reaction and is calculated by following equation (Saeed et al. 2008): where C n (%) is the conversion rate of nitrile group in PAN into amidoxime group; ΔM (g) is the mass of core-shell nanofiber mat difference before and after reaction; M 0 (g) is the mass of the core-shell nanofiber mat before reaction; W 0 is the molecular weight of acrylonitrile monomer (53 g/ mol); and W 1 is the molecular weight of hydroxylamine (33 g/mol).
The different conversion degrees for nitrile group were further confirmed by FTIR (Fig. 6a). The adsorption peak at 2250 cm −1 was assigned to the -CN group stretching vibration, and the peak intensity decreased as the conversion of the nitrile to the amidoxime group increased (Puspitasari et al. 2018); the peak of C = O in ester at 1730 cm −1 was reduced due to the introduction of hydroxamic acid groups (Alakhras et al. 2005); the observed absorption band at 1650 cm −1 in curses were attributed to the stretching of C = N (Hu et al. 2016).
The XRD patterns of AOPAN/PVDF nanofibers are shown in Fig. 6b. There had a sharp peak at around 21.4°, which corresponds to the PVDF molecular chains (Sun et al. 2020). The characteristic peak at 23.8° and 26.5° is commonly observed with PAN fibers (Feng et al. 2020). The result shows that the crystallinity of the core-shell nanofibers had been largely retained with different AO conversion. This provides a validation for the strength retention of core-shell fibers.
The thermal stability of PAN/PVDF and AOPAN/PVDF core-shell nanofibers is displayed in Fig. 6c-d. The curve of PAN/PVDF had three step changes in mass with temperature increase, which occurred at 70, 250 and 400 °C, respectively. The curves for AO modified fibers also exhibited three major step change, but with the second being obviously lower. The weight loss happening at the temperature range from 50 °C to around 150 °C are generally ascribed to evaporation of moisture. And the change at temperature higher than 320 °C are supposed to be related with original polymer chain including both PAN and PVDF. Therefore,  this indicated AOPAN/PVDF fibers still retained part of its structure unchanged in amidoxime modification (Blendelectrospun poly(vinylidene fluoride)). The difference in the second inflection point implies that the thermal stability of the modified fibers degraded. This is thought due to the decomposition of the amidoxime groups. The mechanical strength of neat PVDF and AOPAN/ PVDF fiber mats with different conversion rates are shown in Table 3. The neat PVDF nanofiber membrane was prepared by 12% PVDF dope solution at 14 kV voltage for 2 h. After being treated by hydroxilamine at 70 °C, the mechanical strength of neat PVDF fiber is slightly improved, which could be used as a proof for the chemical stability for PVDF in the reaction. The reason behind the improvement may be the thermal annealing that rearrange the polymer chain packing. Both tensile strength and tensile elongation decreased  with increase in conversion degree, but core-shell nanofibers with a conversion rate of 72.8% maintained much of its original morphology according to SEM image (Fig. 7). Bending tests also showed that the nanofibrous membrane have durable flexibility. The mechanical strength of nanofiber reported in other literature is shown in Table 4, and the property of the fibers by current work is better than that in most of the other researches. The specific surface area was 5.813 m 2 /g. Therefore, AOPAN/PVDF obtained from core-shell electrospinning is a promising material for water treatment.

Static adsorption
The reaction mechanism for amidoxine group is given in Fig. 8. The effect of the initial pH on the adsorption capacity of the AOPAN/PVDF nanofibers mats for the same Pb 2+ is summarized in Fig. 9a. The time duration for all the tests was set at 24 h, to ensure the adsorption equilibrium was attained. The results showed that the adsorption capacity  of Pb 2+ on AOPAN/PVDF core-shell fibers increased with the increase of pH value. And the adsorption capacity was very low at pH = 2. This was first attributed to the competition between Pb 2+ and H + toward the adsorption site. Additionally, the electrostatic repulsion from protonated amino was higher for Pb 2+ than H + ; the former has larger diameter and higher valence. Therefore, it is difficult for Pb 2+ to approach and adhere to the functional group of AOPAN/ PVDF core-shell fibers (Hong et al. 2015). When the pH is above 5, the Pb 2+ would precipitate in the form of Pb(OH) 2 and PbCO 3 , which affects greatly its adsorption behavior.
The Pb 2+ adsorption performance compared to other adsorbents is summarized in Table 5. The maximum adsorption capacity of AOPAN/PVDF for Pb 2+ is 89.29 mg/g, which is higher than part of the other nanomaterials. It is expected to be a new material for Pb 2+ removal from water. Figure 9b shows the adsorption amount of Pb 2+ on the AOPAN/PVDF core-shell fibers as a function of the contact time (pH = 5). The capacity increased with an increase of the contact time. The equilibrium adsorption capacity of Pb 2+ on the AOPAN/PVDF core-shell fibers were reached within 12 h. Adsorption kinetics can be employed to describe the  adsorption rate and possible adsorption mechanism. Pseudo first-order and pseudo second-order models were used to analyze the nature of the kinetics and the rate of Pb 2+ adsorption by the AOPAN/PVDF membrane. The linear forms of these two models are fitted by following equations, respectively (Zhao et al. 2011).
where q e (g·g −1 ) is the equilibrium adsorption capacity, q t (g·g −1 ) is the amount of metal ions adsorbed at time t, t (min) is the duration of the adsorption, k 1 (min −1 ) and k 2 (g·g −1 min −1 ) are the first-and second-order rate constants, respectively. The best fit model was selected based on linear regression correlation coefficient (R 2 ). According to Fig. 9c, pseudo second-order model (R 2 = 0.9921) showed a better fit with the experimental data than pseudo first-order model (R 2 = 0.9557). The basic assumption of pseudo second-order model is that the adsorption process is major in chemical adsorption. Many researchers have reported the similar mechanism (Lu et al. 2013;Zhu et al. 2016). The deprotonation of -NH 2 and -OH on amidoxime group can be chelated with Pb 2+ in liquid, which is the main mechanism of Pb 2+ adsorption by AOPAN/PVDF (Zhuang et al. 2018).
The FTIR spectra of AOPAN/PVDF and after adsorption are shown in Fig. 10a; it is found that new peaks appear at 1598 and 1528 cm −1 after adsorption, which is related to the affinity binding with Pb 2+ and also confirms that Pb 2+ is adsorbed on the surface of AOPAN/PVDF (Hong et al. 2015). The membrane adsorption property and regeneration efficiency in five consecutive cycles is shown in Fig. 10b. It is revealed that the adsorption capacity of AOPAN/PVDF membranes decreased with each adsorption-desorption cycle. And the capacity became 44 mg/g in the adsorption at the 5th cycle. The Pb 2+ on membrane could not be completely removed from the membrane during the desorption 1 q e t step; some lead ions permanently occupied the adsorption sites and the adsorption sites on membrane decreased with increasing repetitions (Ren et al. 2018). It should be noted that the mechanical strength of the AOPAN/PVDF membrane did not change significantly in the reuse process. This was attributed to the core-shell structure. The inner support layer avoids the collapse of the membrane during the modification process, so that the mechanical properties of the membrane can be maintained.

Flow-through adsorption
Dynamic adsorption in membrane filtration mode was performed on the dead-end filtration setup (Fig. 1b) using a single-pass flow. The filtrates were sampled at the set time, and the concentrations of Pb 2+ ions in the samples were determined by ICP. The Pb 2+ adsorption for the AOPAN/PVDF membranes obtained at different filtration rates is presented in Fig. 11a. Obviously, the filtration rate had a significant influence on Pb 2+ capture. For all the three tests, Pb 2+ breakthrough happened almost immediately after the starting of the test. When the filtration rate was 2.66 mL/min, the concentration of Pb 2+ in effluent increased sharply within 20 min. Higher filtration rate signifies that the residence time of the feed solution in the AOPAN/PVDF membrane may be too short, hence resulting in faster breakthrough. With higher initial concentration, the time required for Pb 2+ to break through the membrane was very short, as shown in Fig. 11b. When the concentration was at lower level of 10 mg/L, Pb 2+ concentration increase at the initial stage was milder (Fig. 11b). The dynamic adsorption rate v t at different filtration rate and initial Pb 2+ concentrations are summarized in Fig. 12. Table 6 displays the q d of conditions which calculated by Fig. 12. It was observed that q d increased and then decreased with the filtration rate. With a higher filtration rate, more amount of Pb 2+ ions will pass through the membrane within the same period of filtration time; but ions may not have sufficient time to transport to the adsorption sites. If the latter factor overtakes the former, q d will be reduced.

Conclusions
The core-shell PAN/PVDF nanofiber membrane was fabricated by coaxial electrospinning, and then modified by amidoxime. The applied voltage has a great influence on the process of electrospinning and the morphology of the fiber. The polymer solution cannot be fully stretched under 13KV, which caused bead-on-string phenomenon, while 15KV causes instability in the electrospinning process. Through fiber morphology and mechanical strength test, 12%PAN-12%PVDF is the most suitable polymer concentration ratio. The core-shell structure greatly enhances the stability of the fiber in amidoxime modification. The morphology and strength of the membrane are still preserved when the conversion of nitrile group in PAN was over 70%. The membrane has excellent removal performance on Pb 2+ , and more importantly, its mechanical properties are stable during the regeneration process. The AOPAN/ PVDF membrane can effectively retain Pb 2+ in dynamic filtration, indicating the potential of the AOPAN/PVDF for industrial applications.