Ureido-functionalized mesoporous polyvinyl alcohol/silica composite nanofibre membranes were prepared by electrospinning technology and their application for removal of Pb2+ and Cu2+ from wastewater was discussed. The characteristics of the membranes were investigated by scanning electron microscopy, transmission electron microscopy, X-ray diffraction, and N2 adsorption-desorption analysis. Results show that the membranes have long ﬁbrous shapes and worm-like mesoporous micromorphologies. Fourier transform infrared spectroscopy confirmed the membranes were successfully functionalized with ureido groups. Pb2+ and Cu2+ adsorption behavior on the membranes followed a pseudo-second-order nonlinear kinetic model with approximately 30 minutes to equilibrium. Pb2+ adsorption was modelled using a Langmuir isotherm model with maximum adsorption capacity of 26.96 mg g−1. However, Cu2+ adsorption was well described by a Freundlich isotherm model with poor adsorption potential due to the tendency to form chelating complexes with several ureido groups. Notably, the membranes were easily regenerated through acid treatment, and maintained adsorption capacity of 91.87% after five regeneration cycles, showing potential for applications in controlling heavy metals-related pollution and metals reuse.
Heavy metals, including copper (Cu), lead (Pb), cadmium (Cd), mercury (Hg), chromium (Cr), and arsenic (As), are used in diverse industrial products and are commonly found in high concentrations in industrial wastewater and landfill leachate (Sang et al. 2008). Due to recalcitrance, toxicity and tendency toward bioaccumulation, high levels of heavy metals are a major cause for public health concern, causing disorders such as kidney disease, skin inflammation and nerve paralysis (Singh & Prasad 2015). Developing methods for effective removal of heavy metals from wastewater is therefore extremely important.
Many methods have been developed for the removal of heavy metals, including coagulation, chemical precipitation, membrane separation, ion exchange and adsorption (Zhao et al. 2016). Among these techniques, adsorption is the most common technology used in heavy metal removal due to its simplicity, ease of operation, cost-effectiveness and environmental friendliness. The adsorbents that have been studied and used to separate heavy metals include zeolites (Zhao 2016), clay minerals (Uddin 2017), activated carbon (Zaini et al. 2010), nanoscale zero valent iron (Zou et al. 2016) and graphene oxides (Yuan et al. 2013), among others. A great deal of recent research has focused on silica-based mesoporous materials with or without functionalized organic groups as heavy metal adsorbents because they have a uniquely large specific area, regular pore structure, and highly adsorptive properties (Jiang et al. 2013; Lin et al. 2015; Hao et al. 2016). However, these powdered porous adsorbents are difficult to recycle and separate after application to heavy metal-contaminated water. Thus, adsorbents based on nanofibre membrane materials are gaining popularity as an alternative.
Owing to the interconnected porous structure, the high surface area and extremely long length, the versatility of design, relatively low cost and high production rate, fibre membranes fabricated by electrospinning are promising candidates for filtration and adsorption in wastewater treatment (Sahay et al. 2012). Many electrospun polymer fibres have recently been evaluated for heavy metal pollution control (Huang et al. 2014; Ma et al. 2016). Nevertheless, most of the polymer nanofibres suffer from weak physical–chemical stability and thermal resistance in application (Zhang et al. 2012). In order to overcome these obstacles, organic-inorganic composite nanofibre membranes have been developed by electrospinning. Polyvinyl alcohol (PVA)/silica composite fibre membranes, among the most commonly studied organic-inorganic composite materials, have been successfully fabricated (Shao et al. 2003) and explored as important candidates for enhancement of adsorption performance by introducing functional groups such as mercapto (Wu et al. 2010), cyclodextrin (Islam et al. 2015), and amino (Keshtkar et al. 2013). For example, Islam et al. (2015) reported the electrospinning preparation of phosphine grafted PVA/silica composite nanofibre adsorbents and their application for the highly effective removal of Mn(II) and Ni(II). However, few studies have investigated the preparation and application of PVA/silica nanofibre membranes functionalized with ureido (-NHCONH2) groups.
We present fabricated ureido-functionalized PVA/silica mesoporous fibre membranes which possess both high specific surface area and long fibrous morphology by electrospinning PVA, tetraethyl orthosilicate (TEOS) and ureidopropyltriethoxysilane sol-gel precursor with cetyltrimethyl ammonium bromide (CTAB) as the template (Figure 1). Additionally, we discuss the potential of these membranes in their Pb2+ and Cu2+ adsorption abilities in terms of effect of pH, the adsorption kinetics, the adsorption isotherms and regeneration. Results show that the materials displayed excellent adsorption rates and removal efficiencies for Pb2+ and Cu2+. Moreover, a high regeneration capacity of the membranes was obtained by acid treatment. Our work has important applications for fabrication of functionalized environmental materials and their application in water pollution control and treatment.
Polyvinyl alcohol (PVA, Mw = 1,750), TEOS, CTAB, lead (II) nitrate (Pb(NO3)2), copper (II) nitrate trihydrate (Cu(NO3)2·3H2O), hydrochloric acid (HCl, 36%–38%) and absolute ethanol were provided from Sinapharm Chemical Reagent Co., Ltd, China. Ureidopropyltriethoxysilane was purchased from Jingzhou Jianghan Fine Chemical Co., Ltd, China.
Fabrication of PVA/silica fibre membranes
First, (Sol I) CTAB (0.74 g) was dissolved in absolute ethanol (2.50 g) and magnetically stirred for 0.5 hours at 40 °C. Meanwhile, (Sol II) a certain amount of ureidopropyltriethoxysilane was mixed with deionized water (2.92 g) and 2 mol L−1 HCl solution (0.17 g) and stirred for 1 hour at room temperature. TEOS (2.25 g) was then slowly added into the mixture of Sol I and Sol II and further stirred for 2 hours. The molar ratios of ureidopropyltriethoxysilane/(ureidopropyltriethoxysilane + TEOS) in the mixture were 0, 0.2, 0.3, and 0.4. Finally, the above solution was dropped into 10% PVA solution (3.8 g) and stirred vigorously for 5 hours at 40 °C. After 2 days aging at room temperature, the precursor solutions were put into a plastic syringe with a stainless-steel needle. The needle was connected to a high-voltage generator as the anode, while the cathode was connected to a fibre collector covered with a piece of aluminium foil. A voltage of 15–18 kV and a constant flow rate of 0.5 mL h−1 were applied to the solution with a tip-to-collector distance of 20 cm. After electrospinning for several hours, the collected PVA/silica fibre membranes were refluxed in ethanol/HCl solution (volume ratios of 10:1) for 24 hours at 70 °C to extract the template, and finally dried for 6 hours at 80 °C under vacuum. After cooling to room temperature, the functionalized PVA/silica mesoporous fibre membranes were obtained and denoted as m-0, m-2, m-3, and m-4, respectively, corresponding to the ureidopropyltriethoxysilane molar percentage in the precursor sol.
The structure and morphology of the prepared fibre membranes were examined by scanning electron microscopy (SEM) on an XL-30 scanning electron microscope, transmission electron microscopy (TEM, JEM-2011) operating on a JEM-2011 electron microscope at 200 kV, and X-ray diffraction recorded with a Bruker D8 Advance X-ray diffractometer using a secondary monochromator and Cu Ka radiation (40 kV, 100 mA, h = 5–70, and k = 0.15406 nm). The grafted organic moieties were characterized by Fourier transform infrared spectroscopy (FTIR, Nicolet 5700) of KBr powder pressed pellets carried out on a Nicolet 5700 spectrometer. N2 adsorption/desorption isotherms were taken at 77 K on a Micrometrics ASAP-2020 adsorption meter. The Brunauer–Emmett–Teller (BET) method and Barrett–Joyner–Halenda (BJH) model were employed to calculate the specific surface area and the pore volume and pore size distributions, respectively. The concentrations of the Pb2+ and Cu2+ were determined by an Inductively Coupled Plasma Spectrometer (ICP, Optima 2100 DV, USA).
Pb(NO3)2 and Cu(NO3)2·3H2O were used for the preparation of metal aqueous solutions. Adsorption experiments were conducted as follows. In the adsorption kinetics test, 20 mg membranes and 20 mL Pb2+ or Cu2+ solution of 20 mg L−1 were added into a batch of stoppered vials and shaken at a speed of 200 rpm for different time intervals (0.5 to 180 minutes). Adsorption isotherms were obtained by immersing 20 mg membranes into 20 mL Pb2+ or Cu2+ solution with gradient concentrations (ranging from 5 mg L−1 to 100 mg L−1) and shaken for 3 hours. The influence of pH (range 1–6) on adsorption of heavy metals was estimated and pH was adjusted with HCl or NaOH. After reaching adsorption equilibrium, the suspension was filtered with a 0.45 μm cellulose acetate filter. The Pb2+ and Cu2+ concentration in the filtrate was determined by ICP/OES spectroscopy.
The regenerability of the adsorbents was evaluated by immersing the membranes full of Pb2+ into 1 mol L−1 HCl solution with oscillation treatment at 25 °C for 2 hours. Subsequently, the adsorbents were filtered and washed with 1 mol L−1 NaHCO3 solution for 12 hours and washed with deionized water and dried in a vacuum oven at 60 °C.
RESULT AND DISCUSSION
Characterization of ureido-functionalized PVA/silica nanofibre membranes
Figure S1(a) shows the SEM image of the originally fabricated ureido-functionalized electrospun PVA/silica fibre membranes. The membranes are composed of rather uniform and smooth fibres, of which the diameter ranged between 600–900 nm. After removal of the CTAB template in ethanol/HCl solution, the surface of the fibres became rough but without any serious cracks in morphology (Figure S1(b)). (Figure S1 is available with the online version of this paper.)
The TEM image shown in Figure 2(b) illustrates that the surface of the obtained mesoporous PVA/silica fibres is wormlike and mesoporous after extraction of template compared with that of the original precursor fibres (Figure 2(a)), which might be attributed to the removal of CTAB, interference of electrostatic force in electrospinning as well as surface tension of the precursor gel solution. Moreover, the structure of these pores shows no long-range periodicity.
Figure 3 shows the FTIR spectra of the ureido-functionalized PVA/silica mesoporous fibre membranes. The broad band at 3,000–3,600 cm−1 is compared to C–OH and NH/O-H stretching vibration. The peaks appear at around 1,600 cm−1 and 1,557 cm−1 due to the bending vibration band of the amine N–H group. The peaks at 1,646 cm−1 are assigned to the amide C = O stretching vibration. The results provide direct evidence for the successful modification of the ureido group onto the PVA/silica fibre membrane skeleton by hydrolytic poly-condensation.
The corresponding XRD patterns of the fabricated membranes are shown in Figure S2 (available online). As illustrated in the wide-angle WXRD pattern, there is only a broad band at around 2θ = 23.90, indicating amorphous PVA/silica. The insert of the small-angle XRD pattern contains a dominant d100 weak diffraction peak at around 2θ = 2.57, which reflects a wormhole channel mesoporous structure and disordered pore channel structure. The result is similar to those reported previously (Wu et al. 2010) and corresponded well with the result obtained from TEM.
The N2 adsorption/desorption isotherms of the mesoporous fibre membranes are shown in Figure 4. The presence of a typical type IV isotherm with an N2 hysteresis loop is indicative of the presence of mesostructure in the samples. The calculated pore size distribution further confirms the presence of mesoporous channels with diameters greater than 2 nm. However, after adding ureidopropyltriethoxysilane, the shape of the isotherms changed slightly and the BET surface area, pore diameter and pore volume (shown in Table S1, available with the online version of this paper) decreased with the increasing amount of ureidopropyltriethoxysilane. This can be attributed to grafting of the functional groups on the surface of the channel walls.
Effect of pH on adsorption
The initial solution pH is one of the factors influencing the adsorption of heavy metals in that it is related to the forms of metal ions in aqueous phases and ligand properties of the adsorbents. The effect of pH on the adsorption of Pb2+ and Cu2+ on the ureido-functioned PVA/silica fibre membranes was investigated and the results are shown in Figure 5. The pH had a marked influence on Pb2+ and Cu2+ removal efficiency. The removal efficiency of Pb2+ and Cu2+ by the adsorbents was very low at pH 2, only 1.28% for Pb and 5.26% for Cu, which might result from electrostatic repulsion between Pb2+ and Cu2+ and the protonated amine group on the ureido at low pH (Deng et al. 2003; Puanngam & Unob 2008). Removal efficiency of Pb2+ and Cu2+ increased consistently with increasing pH. At higher pH (pH < 7), the lead and copper in aqueous solution exist as Pb(OH)+, Pb(OH), or Pb2(OH)22+, and Cu(OH)+, Cu(OH) or Cu2(OH)22+ (Neghlani et al. 2011; Benettayeb et al. 2017). The strong binding ability of chelation between amine group on the ureido group and metal ions is responsible for the adsorption of Pb2+ and Cu2+. With greater pH, precipitation of the metal hydroxide was observed. The maximum removal efficiency emerged at pH 6 for Pb2+ and pH 5–6 for Cu2+ respectively. In this pH range, neither precipitation of the metal hydroxide nor protonation of the nitrogen atom on the ureido groups is expected.
Effect of contact time and adsorption kinetics
Adsorption rates were determined by examining the removal efficiency of Pb2+ and Cu2+ on adsorbents over contact time. The results suggested that the adsorption processes of Pb2+ and Cu2+ on the ureido-functioned PVA/silica fibre membranes were similar at different temperatures (Figures S3–S6, available with the online version of this paper). In the initial adsorption stage, the adsorbents exhibited a rapid adsorption to Pb2+ and Cu2+. Adsorption rate increased modestly with the extension of contact time. Adsorption equilibrium was reached at approximately 30 min with removal efficiency of Pb2+ up to 94.55% for m-4. Additionally, the adsorption performance of the membranes for Pb2+ and Cu2+ was better at higher temperatures as the increase in temperature enhanced molecular thermal motion to promote contact frequency of metal ions with adsorbents. Moreover, the removal efficiency of Pb2+ was greater than that for Cu2+, as Pb2+ occupied one reaction site with the amine group, while Cu2+ was likely to form chelating complexes with several amine groups. This mechanism explains the Pb2+ and Cu2+ adsorption isotherm models as well (Figure 7).
The fitting curves generated from the kinetic data of Pb2+ and Cu2+ adsorption on m-4 are depicted in Figure 6(a) and corresponding kinetic parameters are calculated in Table S2. The kinetic parameters of m-0, m-2 and m-3 are calculated in Tables S3–S5. (Tables S2–S5 are available with the online version of this paper.) Per the fitting results, the correlation coefficient (R2) for the pseudo-second-order model is higher than that for the pseudo-first-order model for the four adsorbents, indicating that the pseudo-second-order model is more suitable to describe the adsorption kinetics of Pb2+ and Cu2+ and the concentration of metal ions and adsorbent performance exercise some effect on the adsorption rate. The saturated adsorption capacity of Pb2+ and Cu2+ on m-4 were estimated to be 19.73 mg g−1 and 5.31 mg g−1 respectively, and the adsorption rate constants for Pb2+ and Cu2+ were 0.28 g mg−1 min−1 and 0.17 g mg−1 min−1, respectively.
The inter-particle diffusion model fits are shown in Figure 6(b). The adsorption data for both Pb2+ and Cu2+ are fitted to two straight lines. The first one passes through the origin and indicates rapid adsorption at this adsorption stage and intra-particle diffusion is the only rate-limiting factor. The second line is straight with low slope, showing that the adsorption rate starts to decrease due to reduction of metal ion concentration in the aqueous solution. Meanwhile, the second line that does not pass through the origin (representing C ≠ 0) demonstrates that boundary layer diffusion effects the diffusion process.
The adsorption isotherms for Pb2+ and Cu2+ on m-4 at 298 K and the simulating curves of the above two models are depicted in Figure 7. The corresponding parameters are listed in Table S2. The adsorption capacities of the ureido-functioned PVA/silica fibre membranes for both Pb2+ and Cu2+ increase dramatically from initial concentration and finally reach equilibrium. The Pb2+ adsorption isotherm fits well with the Langmuir model with an R2 value of 0.9735 while Cu2+ adsorption data are very well described by the Freundlich isotherm equation with R2 value of 0.9622. The isotherm parameters of m-0, m-2 and m-3 listed in Tables S3–S5 suggest similar fitting results. For Pb2+ adsorption, the monolayer coverage of Pb2+ on the membranes predominates with a maximum adsorption capacity of 26.96 mg g−1. In contrast, Cu2+ adsorption shows multilayer adsorption. However, the calculated value of 1/n (>1) represents poor adsorption potential of the absorbents for Cu2+, demonstrated by higher adsorption of Pb2+ than that of Cu2+.
Compared to various adsorbents reported in the literature (Bassi et al. 2000; Çekiç et al. 2004; Sprynskyy et al. 2006; Chaisuwan et al. 2010; Prakash et al. 2012) (Table S7, available online), the adsorptive capacities of clinoptilolite (Sprynskyy et al. 2006) for Pb2+ and Cu2+, found to be 27.7 mg g−1 and 25.76 mg g−1 respectively, are higher than those of m-4. However, the time to adsorption equilibrium of ureido-functionalized PVA/silica mesoporous fibre membranes is far shorter (30 min) than clinoptilolite (>5 hours), which might be related to the structure of the nanopores in the fibres (Yang et al. 2010) which may improve the diffusion of metal ions in the pores. The reported adsorption data of other adsorbents listed in Table S7 are lower than the value in our study. Thus, the ureido-functionalized PVA/silica mesoporous fibre membranes appear to be an effective adsorbent for removal of Cu2+ and Pb2+ from aqueous solution.
The adsorption mechanism of the ureido-functionalized PVA/silica mesoporous fibre membranes for Pb2+ and Cu2+ can be explained by Pearson's hard and soft acids and bases (HSAB) principle. Pearson (1966) divided every conceivable Lewis base into hard, soft, and borderline base. Correspondingly, Lewis acids are classified into hard, soft, and borderline acids. According to the HSAB principle, hard acids prefer to combine with hard bases through coulombic forces and soft acids prefer to bind soft bases with covalent bonds. In this work, ureido is considered as a hard base with respect to the adsorbents. Cu2+ and Pb2+ are considered as borderline metal cations, which can bind both hard bases and soft bases. Thus, ureido groups can complex Pb2+ and Cu2+. It can also be seen from the FTIR patterns (Figure 3) that the membrane surface has many hydroxyl groups, which also play a role in removal of Pb2+ and Cu2+ by coordinating with metal ions. In addition, Yang et al. (2010) proposed that the interrelated voids formed from entanglement of fibers can influence adsorption kinetics positively by making contact between the functional groups and the adsorbates easier and the flow of the solution smoother.
Regeneration of adsorbents
Desorption investigation aimed at recovery of metal as well as regeneration of adsorbent. The extraction of Pb2+ from Pb2+-loaded fibre membranes was performed using 1 mol L−1 HCl solution. As illustrated in Figure 8(a), up to 70% of adsorbed Pb2+ was desorbed within 5 min, and the desorption efficiency was over 95% after 2 h, which suggests that HCl is an ideal stripping agent for regeneration of the ureido-functionalized adsorbents. The protonation of ureido groups after treatment with HCl solution is the main reason for the regeneration of adsorbents. Pb2+ and Cu2+ are desorbed from ureido groups due to electrostatic repulsion, re-releasing adsorption sites. Moreover, Cl− and heavy metals can form stable complexes, which attenuates the adsorption effect of ureido functional groups for Pb2+ and Cu2+. Figure 8(b) graphically depicts the removal efficiency of Pb2+ on regenerated adsorbents at different regeneration cycles. After acid treatment, the adsorption capacity of Pb2+ decreases from 98.96% to 91.87% after five regeneration cycles. This is attributed to the physical loss of functional groups and irreversible binding of some ligands. Consequently, these results strongly suggest that ureido-functionalized PVA/silica fibre membranes have great potential for removal and recovery of heavy metal ions in aqueous solution.
In conclusion, ureido-functionalized PVA/silica mesoporous fibre membranes were fabricated using an electrospinning and hydrolytic poly-condensation technique. The characterization of this material indicated that the ureido groups were successfully added to the PVA/silica fibre membrane skeleton, possessing a continuous long and ﬁbrous morphology, mesoporous structures, and high BET surface area of 459.45 m2 g−1. The adsorption arrays of Pb2+ and Cu2+ indicated that the optimum pH for the removal of Pb2+ was 6 and 5–6 for Cu2+. The pseudo-second-order nonlinear kinetic model was more suitable for describing the adsorption kinetics of both Pb2+ and Cu2+, and adsorption equilibrium was achieved within 30 min. According to the R2 values, the Langmuir isotherm model effectively described Pb2+ adsorption by the membranes with monolayer adsorption and maximum adsorption capacity of 26.96 mg/g (R2 = 0.9735). Cu2+ adsorption data were best fitted to the Freundlich isotherm equation (R2 = 0.9622) with multilayer adsorption and poor adsorption potential. Meanwhile, the amount of adsorption of Pb2+ was superior to that of Cu2+, which can be explained by the capture of Cu2+ by several ureido groups. Furthermore, the experimental materials were easily regenerated through acid treatment, and adsorption capacity was maintained even after five regeneration cycles. Consequently, ureido-functionalized PVA/SiO2 mesoporous fibre membranes show strong potential for use as efficient adsorbents of heavy metal ions.
This work was funded by Natural Science Basic Research Plan in Shaanxi Province of China (2017JQ5008) and the Fundamental Research Funds for the Central Universities of China (310829161017; 310829173602; 310829161006; 310829172002), Special funds of education and teaching reform for the Central Universities of China (310629172112). One Hundred Talent Plan of Shaanxi Province and the Innovation Training Program for Undergraduate Students of Chang'an University (201610710079, 201710710096).