Abstract
A method combining adsorption with electric field-driven ion enrichment to remove the trace metal ions in wastewater was investigated. Composite adsorbent prepared from chitosan (CS) and sodium phytate (SP) supported into polyethylene glycol terephthalate (PET) nonwoven fabric by nonsolvent induced phase separation was employed and located near the cathode in a pair of titanium plate electrodes. Results showed the removal rate of copper ions (Cu(II)) adsorbed onto CS-SP/PET adsorbent was increased from 56% to 88% for 10 mg L−1 Cu(II) solution when the applied voltage was from 0 to 1.2 V. The adsorption behavior was well correlated with the Langmuir isotherm model. And adsorption process fitted well for pseudo-second-order kinetic equation, suggesting that chemical adsorption was rate-limiting step. And the energy consumption was economical, about 4.35 × 10−3 kW·h for 1 m3 of water with 10 mg L−1 Cu(II). It was suggested that the adsorption performance for Cu(II) with CS-SP/PET adsorbent was enhanced due to the enrichment of Cu(II) under electric field. This work provides a new way to remove trace heavy metal ions from wastewater.
Introduction
Heavy metal pollution is a common hazardous pollutant for the environment at present. Heavy metal ions could be accumulated in living organisms, causing several disorders and diseases to human beings (Ding et al., 2016; Sellaoui et al., 2018; Wang et al., 2017). Therefore, the treatment of heavy metal ions from polluted water is necessary and of great significance. A variety of techniques had been applied to the removal of heavy metal ions from polluted water, such as chemical precipitation (Verbych et al., 2005), ion exchange (Inglezakis and Loizidou, 2007), membrane filtration (Lam et al., 2018), reverse osmosis (Abdullah et al., 2009), adsorption (Liang et al., 2017; Rathinam et al., 2018; Singh et al., 2018), and electric field-driven technology (Chen et al., 2011; Pan et al., 2009). Among these technologies, it is very simple and effective to remove the heavy metal ions from polluted water by adsorption. Various absorbents were studied and applied in water treatment, such as zeolites, activated carbon, clay, chitosan (CS), and so on (Mfoumou et al., 2018; Öztürk and Malkoc, 2014; Rathinam et al., 2017; Wu et al., 2018). Among these adsorbent materials, CS, a partially acetylated glucosamine and low-cost natural biopolymer, shows high adsorption ability for heavy metal ions and is widely used in many wastewater treatment areas due to rich amino and hydroxyl functional groups on its structure. Evans and coworkers studied the adsorption of CS from crab shells for Cd(II) and the adsorption capacity was 105.0 mg g−1 (Evans et al., 2002). Salehi prepared CS/PVA membranes by adding MWCNT-NH2 and PEG additives. The result showed that adsorption of the CS/PVA membrane for Cu(II) was 35 mg g−1 (Salehi et al., 2012). Swayampakula et al. (2009) prepared CS-coated perlite as a adsorbent to adsorb copper and nickel ions and found that the adsorption quantities were 147.1 and 38.9 mg g−1, respectively. In addition, PEG/CS membranes were prepared by Reaid and applied to iron and manganese ions removal. The maximum adsorption capacities were 80.0 and 50.0 mg g−1 for iron and manganese ions, respectively (Reiad et al., 2012). It was noted that the adsorption performance of heavy metal ions by CS was still insufficient when the concentration of heavy metals was very low, such as less than 10.0 mg L−1 levels, due to adsorption dynamic limit (Shaker and Taiwan, 2015). In order to overcome the drawbacks of CS, we prepared modified CS and studied its adsorption on Cu(II). Although the adsorption performance (88.3 mg g−1) for CuII) was improved, the adsorption equilibrium time (420 min) was still long compared with the pure CS (Kong et al., 2018). When the rest of Cu(II) concentration in aqueous solution was relatively low, Cu(II) could not be further adsorbed even if further prolonging adsorption time or increasing adsorbent amount (Kong et al., 2018; Saleh et al., 2016).
Electric field-driven separation technologies for removing heavy metal ions including electrodialysis (ED) named electrically driven membrane process and capacitive deionization (CDI), can be also used to remove dissolved ions from aqueous solutions based on directional shift of ions under electric field (Fan et al., 2016; Wang et al., 2015). In ED, the positively and negatively charged ions can be easily enriched, pass through ion exchange membrane, and be transferred to the opposite electrodes under an applied electric field. But the energy cost is relatively high, and the high performance ion-exchange membrane is the key in the ED and often has high requirements (Paul and Kollsman, 1957). In CDI, the positively and negatively charged ions can be enriched in opposite directions and then adsorbed onto the cathode and anode under an applied electric field, respectively. The adsorbed ions can be easily released back to the bulk solution when the electric field is canceled (Subramani et al., 2011; Zafra et al., 2014). But the electrode materials in CDI must be often conductive, high surface area for adsorbing more ions, which limited the widely application of CDI technology up to now (Pan et al., 2009).
In this study, a method combining adsorption technology and electric field-driven to remove the trace metal ions in wastewater was firstly investigated. CS composite adsorbent prepared from CS and sodium phytate (SP) supported into PET nonwoven fabric by nonsolvent induced phase separation was used as adsorbent. The adsorption unit was composed of a pair of titanium plate electrodes, which were separated by nylon spacer at a certain distance. CS-SP/PET composite adsorbent was placed near the cathode. Trace heavy metal ions in solution were migrated to cathode, enriched near the cathode and adsorbed onto CS-SP/PET adsorbent under electric field. In this method, the role of electrodes was only to provide electric field to promote ion direct migration, without considering surface area and conductivity of electrode that required in EDI. The efficiency of the adsorption process for Cu(II) was optimized by regulating voltage, flow rate, and electrode spacing.
Experiment
Chemicals
CS (deacetylation degrees 85%) and SP (95% solution) were purchased from Shandong Laizhou Haili Biotechnology Co. (China) and Henan Shengzhide Trading Co. (China), respectively. Copper sulfate (CuSO4) was obtained from Sigma–Aldrich. Polyethylene glycol terephthalate (PET) nonwoven fabric (NWF) was obtained from Tianjin Yuyuan Technology Co. (China). Before the experiment, pretreatment of PET NWF was required by being immersed into distilled water at 100°C for 1 h to remove impurities and dried in an oven to remove distilled water.
Preparation and characterization of CS-SP/PET composite adsorbent
A flat sheet CS-SP/PET composite adsorbent was prepared by nonsolvent induce phase separation (NIPS). Firstly, CS (3.0 g) was dissolved in 97.0 g acetic acid solution (2.0 wt%) to form CS solution. Then, the pretreated PET NWF was immersed into the CS solution for 2 h in vacuum degassing. Finally, the CS/PET NWF was immediately transferred to the SP solution (9.0 wt%) to form CS-SP/PET composite adsorbent at room temperature. During the coagulation, CS can be simultaneously ionically crosslinked by SP to improve the stability of the CS in aqueous solution (Kong et al., 2018). The CS-SP/PET composite adsorbent was obtained and kept in distilled water as standby. Scanning electron microscope (SEM) (Hitachi-TM3030) was utilized to observe the morphology of CS-SP/PET composite adsorbent.
Removal of Cu(II) onto CS-SP/PET composite adsorbent under electric field
To investigate the total adsorption capacity of CS-SP/PET composite adsorbent under an applied electric field, the experiments were conducted in a continuously recycling mode. The experiment schematic diagram was shown in Figure 1. A potentiostat (TD1717) was operated to provide a certain external voltage (0.0–2.0 V). The adsorption unit cell was specifically composed of a pair of titanium plate electrodes with effective size of 2 cm × 6 cm, as anode and cathode pairs, respectively. And the electrodes were separated by nylon spacer at a certain distance. CS-SP/PET composite adsorbent with effective size of 2 cm × 6 cm was located near the cathode. Initial Cu(II) concentration was 10 mg L−1. The experiments were conducted at pH of 6.0 and 30°C. The Cu(II) concentration before and after adsorption was determined by inductively coupled plasma-atomic emission spectrometry (ICP-715-ES instrument, Varian). The removal efficiency of Cu(II) with CS-SP/PET composite adsorbent, S (%), was defined as follows

Schematic diagram of the CS-SP/PET adsorption system enhanced by electric field.
The electrical energy consumption during adsorption under an applied electric field was expressed as follows
Results and discussion
Characterization of CS-SP/PET adsorbent
The structure of the PET NWF and the CS-SP/PET adsorbent were observed as shown in Figure 2. PET NWF had excellent properties of high strength and good compatibility with CS, and was an ideal supporting material. From Figure 2(a), PET NWF was composed of plenty of disordered fiber stacked and CS solution was easily adsorbed on the surface and inside of PET NWF (Luo et al., 2017). From Figure 2(b), it could be seen that the surface of the CS-SP/PET adsorbent was covered with a layer of CS, and the surface of the CS had some small pores. And the CS was also attached among the fibers with some small pores as shown in Figure 2(c). The structure of the CS-SP/PET adsorbent had a high specific area 8.9 m2 g−1, which was obtained by nitrogen adsorption isotherms at 77 K with automated physical and chemical adsorption measurement (Nova 4200e). The higher the specific surface area, the more reactive groups such as hydroxyl groups and amino groups would be exposed, which were beneficial to Cu(II) adsorption (Kong et al., 2018).

SEM of (a) PET NWF, (b) surface, and (c) cross-section of CS-SP/PET adsorbent.
Electrochemical characteristics
In order to ensure that the Cu(II) was not reduced during the adsorption process, cyclic voltammetry measurements were measured in a two-electrode electrochemical unit by electrochemical workstation. From Figure 3(a), the cyclic voltammograms of the 10.0 mg L−1 Cu(II) solution was performed with a scan rate of 50 mV s−1 at various applied voltages ranging from 0.0 to 2.0 V. It could be seen that when the voltage changed between 0.0 and 1.2 V, the corresponding current also steadily increased. And the cyclic voltammetries of copper ion solution were measured at various sweep rates, and the results were shown in Figure 3(b). There was no obvious redox peak during the adsorption process, suggesting that no clear oxidation/reduction reactions occurred. However, when the applied voltage was further increased to the range of 1.5 to 2.0 V, redox reactions clearly occurred. Therefore, the adsorption was carried out in the range between 0.0 and 1.2 V in order to avoid the oxidation/reduction reactions and decrease energy consumption.

Cyclic voltammograms for titanium plate electrode in 10 mg L−1 Cu(II) solution (a) at various sweep potentials and (b) at various sweep rate.
Effect of applied voltage on Cu(II) removal
Figure 4 showed the residual Cu(II) concentration in the solution after adsorption with CS-SP/PET adsorbent at various voltages (0.0–1.2 V). When the voltage was 0, after adsorption, Cu(II) concentration decreased from 10.0 to 4.4 mg L−1 at 7 h. However, once the electric field was applied, residual Cu(II) concentration further dropped because more Cu(II) were attracted by oppositely charged electrode to enrich near the electrode and adsorbed by CS-SP/PET adsorbent. When the voltage was progressively increased from 0 to 0.3, 0.5, 0.8, 1.0, and 1.2 V, the residual Cu(II) concentration decreased from 4.4 to 3.6, 3.2, 2.8, 2.2, and 1.2 mg L−1 at 7 h, respectively. The removal efficiency at 1.2 V was 88.0%, 32% higher than the removal efficiency (56%) at open circuit (0.0 V). This was because a higher electric field drove more Cu(II) to enrich around the cathode, which was favorable for the adsorption of CS-SP/PET adsorbent due to higher concentration Cu(II) to overcome the dynamic limit.

The residual of Cu(II) concentration after adsorption with CS-SP/PET adsorbent under various applied potentials (initial Cu(II): 10 mg L−1; pH: 6.0; temperature: 30°C).
Effect of flow rate on Cu(II) removal
In order to determine optimum operation conditions for CS-SP/PET adsorbent under applied potential, the effect of flow rate of the solution on Cu(II) removal was investigated as shown in Figure 5. The adsorption efficiency of Cu(II) increased from 48% to 88% with increasing the flow rate from 0.5 to 5 mL min−1, and then slowed down gradually to 50% with further increasing the flow rate to 20 mL min−1. This phenomena could be explained as follows. When flow rate is slow, the mass transfer resistance of Cu(II) at the surface of adsorbent is high, resulting in a low removal efficiency of Cu(II). But when flow rate is high, Cu(II) flow directly out of the device before Cu(II) laterally transferred to the surface of the adsorbent, which also lead to a low removal efficiency of Cu(II) (Mossad et al., 2013). From the experiments, Cu(II) removal efficiency of 88% was the highest and the residual concentration of Cu(II) in aqueous solution was about 1.2 mg L−1 at a flow rate of 5 mL min−1. It could be allowed to discharge into surface water according to the maximum permissible limit concentration in water (2.0 mg L−1) specified by the World Health Organization (WHO) (Raval et al., 2016).

Effect of different flow rate on the adsorption of Cu(II) with CS-SP/PET adsorbent (initial Cu(II): 10 mg L−1; pH: 6.0; temperature: 30°C).
Effect of electrode spacing on Cu(II) removal
The adsorption ability for Cu(II) was also studied by adjusting the distance between the two electrodes. Figure 6 showed that Cu(II) removal efficiency was decreased when the distance of two electrodes increased. Cu(II) removal efficiency was 87%, 69%, and 61% with a plate distance of 2, 3, and 4 mm, respectively. It was because the smaller the spacing between the electrodes, the stronger the electric field. Small spacing was helpful to shorten the migration distance of Cu(II) and increased migration rate of Cu(II) (Wang et al., 2015), which would benefit for enrichment around the electrode and adsorption of Cu(II) onto CS-SP/PET adsorbent. However, the distance between the electrodes was too short, which was not conducive to the entire adsorption process, and sometimes a short circuit problem would occur. Therefore, 2 mm could be considered to be more appropriate electrode spacing in our experiment.

Effect of electrode plate distance on the adsorption of Cu(II) with CS-SP/PET adsorbent (initial Cu(II): 10 mg L−1; pH: 6.0; temperature: 30°C).
Power consumption
To calculate the energy requirement for the process of the removal of Cu(II), the energy consumption was calculated based on Cu(II) adsorption process. As shown in Figure 7, there was an increasing trend for the electric current with the applied voltage increasing during the whole adsorption process. As the voltage was progressively increased to 1.2 V, the corresponding current was 5.19 × 10−5 A. According to equation (2), it could be calculated that the energy consumed for processing 100 mL of Cu(II) removing from 10 to 1.2 mg L−1 was about 4.35 × 10−7 kW·h. When one cubic meter of waste water with the same Cu(II) concentration was treated with the CS-SP/PET composite adsorbent, the energy consumption would be only 4.35 × 10−3 kW·h. Therefore, it would be a promising and cost-effective method to remove Cu(II) with CS-SP/PET composite adsorbent under an electric field.

Changes of electric current with the prolongation of the treatment time.
Adsorption kinetics
Adsorption kinetics can not only show valuable insights into the reaction pathway, but also evaluate the mechanism of adsorption process. In order to analyze the adsorption kinetics in this experiment, the pseudo-first-order and pseudo-second-order model are employed to fit experimental data.
The form of pseudo-first-order model equation is represented by Albadarin et al. (2017)
The form of pseudo-second-order model equation is formulated as Campos et al. (2018)
Electric field can not only influence the adsorption rate but also drive ions near the electrode. Through the adsorption data of Cu(II) on the CS-SP/PET composite adsorbent under an electric field, the nonlinear pseudo-first-order kinetic and pseudo-second-order kinetic curves and parameters were shown in Figure 8 and Table 1, respectively. As shown in Figure 8 and Table 1, the equilibrium adsorption capacity of Cu(II) increased as the supplied voltage increased from 0.0 V to 1.2 V due to the increasing enrichment concentration of Cu(II) around the CS-SP/PET adsorbent under higher electrostatic force. According to the determination coefficient (R2), the pseudo-second-order kinetic was more in line with the adsorption process of the CS-SP/PET composite adsorbent under an electric field. It indicated that the adsorption mechanism of the process for Cu(II) with CS-SP/PET composite adsorbent was mainly controlled by chemical adsorption.

The adsorption kinetics for Cu(II) with CS-SP/PET composite adsorbent at different supplied voltage (initial Cu(II): 10 mg L−1; pH: 6.0; temperature: 30°C).
The kinetic parameters for adsorption of Cu(II) onto CS-SP/PET composite adsorbent at different supplied voltage (initial Cu(II): 10 mg L−1; pH: 6.0; temperature: 30°C).
Adsorption isotherms
Adsorption isotherms can express the relationship between adsorbate molecules and the liquid phase concentration. In this section, Langmuir and Freundlich models are used to analyze the Cu(II) adsorption isotherms onto the CS-SP/PET composite adsorbent.
The nonlinear equation for Langmuir isotherm is represented by the following equation (Foo and Hameed, 2010; Peng et al., 2018)
The nonlinear equation for Freudlich isotherm is an empirical equation and given below (Foo and Hameed, 2010; Peng et al., 2018)
Langmuir and Freundlich models can be used to describe the monolayer adsorption on homogeneous and multilayer adsorption on heterogeneous, respectively. The initial Cu(II) concentration varied from 5.0 to 30.0 mg L−1 at 30°C. The Langmuir and Freundlich isotherm for Cu(II) onto the CS-SP/PET composite adsorbent and parameters were shown in Figure 9 and Table 2, respectively. According to the determination coefficients, it could be concluded that the Langmuir isotherm model (R2 = 0.999) fitted better than the Freundlich model (R2 = 0.980). It could be suggested that monolayer adsorption was more reasonable to explain the adsorption behavior.

The adsorption isotherm for Cu(II) with CS-SP/PET composite adsorbent at 1.2 supplied voltage (pH: 6; temperature: 30°C).
The isotherm model parameters for adsorption of Cu(II) onto CS-SP/PET composite adsorbent at 1.2 supplied voltage (pH: 6; temperature: 30°C).
Conclusions
A novel method combining adsorption technology with electric field-driven to remove the trace metal ions in wastewater was investigated. Comparing with only CS-SP/PET adsorbent with a removal rate of 56%, the removal efficiency of Cu(II) adsorbed with CS-SP/PET adsorbent was increased to 88% when the applied voltage was 1.2 V, the electrode spacing was 2 mm, and the flow rate was 5 mL min−1. The residual concentration of Cu(II) in aqueous solution reached to 1.2 mg L−1 from the initial concentration of 10 mg L−1, which was lower than the international organization emission standards. The adsorption fitted well with the pseudo-second-order kinetic equation, suggesting that chemical adsorption was rate-limiting step. The equilibrium isotherm fitted better represented by the Langmuir model. And the energy consumption was very economical, about 4.35 × 10−3 kW·h for 1 m3 of water with 10 mg L−1 Cu(II). The results demonstrated that the adsorption of Cu(II) with CS-SP/PET composite adsorbent enhanced by electric field was effective for removing trace heavy metal ions.
Footnotes
Declaration of Conflicting Interests
The author(s) declared no potential conflicts of interest with respect to the research, authorship, and/or publication of this article.
Funding
The author(s) disclosed receipt of the following financial support for the research, authorship, and/or publication of this article: This work was financially supported by the National Natural Science Foundation of China (grant no. 21776218), the Tianjin Natural Science Foundation of China (grant no. 17PTSYJC00040), and the Program for Innovative Research Team in University of Ministry of Education of China (grant no. IRT-17R80).
