Enhanced desalination performance utilizing sulfonated carbon nanotube in the flow-electrode capacitive deionization process

Enhanced desalination performance utilizing sulfonated carbon nanotube in the flow-electrode capacitive deionization process

Separation and Purification Technology xxx (xxxx) xxxx Contents lists available at ScienceDirect Separation and Purification Technology journal homep...

3MB Sizes 0 Downloads 10 Views

Separation and Purification Technology xxx (xxxx) xxxx

Contents lists available at ScienceDirect

Separation and Purification Technology journal homepage: www.elsevier.com/locate/seppur

Enhanced desalination performance utilizing sulfonated carbon nanotube in the flow-electrode capacitive deionization process ⁎

Yanmeng Caia,b,c,d, Xiaotong Zhaoa,b,c,d, Yue Wanga,b,c,d, , Dongya Maa,b,c,d, Shichang Xua,c,d a

Chemical Engineering Research Center, School of Chemical Engineering and Technology, Tianjin University, Tianjin 300072, PR China State Key Laboratory of Chemical Engineering, Tianjin 300072, PR China c Tianjin Key Laboratory of Membrane Science and Desalination Technology, Tianjin 300072, PR China d Collaborative Innovation Center of Chemical Science and Engineering (Tianjin), Tianjin 300072, PR China b

A R T I C LE I N FO

A B S T R A C T

Keywords: Flow-electrode capacitive deionization Sulfonated carbon nanotubes Suspension stability Salt removal efficiency Electrode regeneration efficiency

The novel sulfonated carbon nanotubes (CNT-S) material was successfully introduced into the flow-electrode capacitive deionization (FCDI) process, which could reduce inter-tube aggregation and promote good contact between the active material and the current collector by forming the conductive network. The physical characteristic and electrochemical tests were carried out to analyze the effects of surface properties of CNT-S on the suspension stability and electrochemical performance of the flow-electrode, respectively. The results indicated that the CNT-S flow-electrode slurry displayed better dispersible property and suspension stability, higher charging capacity and lower ohmic impedance than that of the CNT flow-electrode slurry. Moreover, the desalination performance of the CNT-S flow-electrode was investigated based on the symmetric FCDI cell. For the 2.0 wt% CNT-S flow-electrode, the mean electrosorption rate of 4.85 mg/(g·min) and the salt removal efficiency of 45.8% were achieved in 1.0 g/L NaCl solution, both of which were all about 1.6 times that of the CNT flowelectrode. It was also found that the CNT-S FCDI system coupled with the adsorption and desorption process simultaneously can achieve high electrode regeneration efficiency by controlling the flux ratio of the washing fluid to the feed solution. Therefore, the CNT-S flow-electrode was a promising active material in the FCDI process and displayed prominent advantage in improving the desalination performance of the FCDI cell.

1. Introduction The shortage of freshwater resources has become a globalization burning question along with the aggravation of water pollution and the increase of population. The seawater desalination technology attracts considerable attention as an alternative to address the water shortage issue [1–3]. In recent years, with the advantages of energy efficient, low cost, easy to regenerate and environmental-friendliness, capacitive deionization (CDI) has been considered as a promising technology for desalting salt-water with low or moderate concentration [4–8]. The CDI technology is an electrochemical deionization technique based on the theory of the electrical double layers (EDL), which the adsorption and desorption process occur at the surface of active materials under a low direct current potential. The CDI performance is strongly determined by the properties of the electrode materials and tremendous efforts have been devoted to the development of advanced carbon-based materials with the high accessible surface area for ion accumulation, the suitable pore size for ion mobility within the poreand network, the low inner



resistivity for effective charge holding, and the high hydrophilicity for ion diffusion [9–13]. However, the application of the CDI technology is challenging because the fixed amount of active material on the solid-electrode restricts the improvement of desalting capacity and efficiency of the CDI cell and its scale-up application [14–20]. Also, the adsorption and desorption processes are carried out in the same CDI cell, which results in the discontinous CDI process and reduces the desalination efficiency. To overcome the limitations of the solid-electrode in the CDI technology, flow-electrode capacitive deionization (FCDI), which utilizes flowable electrodes instead of fixed electrodes, has aroused extensive attention and made significant improvements in the desalting performance [21–23]. Recent researches [24–32] have proved that the FCDI system can achieve higher salt removal capacity owing to the theoretically unlimited supply of the flowable electrode suspension. What’s more, instead of the discharge step taking place in the same cell, the flowelectrode slurry can easily be regenerated outside the FCDI cell by directly mixing the two slurry streams together, which makes the FCDI

Corresponding author at: Chemical Engineering Research Center, School of Chemical Engineering and Technology, Tianjin University, Tianjin 300072, China. E-mail address: [email protected] (Y. Wang).

https://doi.org/10.1016/j.seppur.2019.116381 Received 29 July 2019; Received in revised form 16 November 2019; Accepted 1 December 2019 1383-5866/ © 2019 Elsevier B.V. All rights reserved.

Please cite this article as: Yanmeng Cai, et al., Separation and Purification Technology, https://doi.org/10.1016/j.seppur.2019.116381

Separation and Purification Technology xxx (xxxx) xxxx

Y. Cai, et al.

and the operation parameters including the addition of the dispersant, the working voltage, the feed concentration, the flow rate ratio of the feed solution to the flow-electrode and the flux ratio of the washing fluid to the feed solution were also optimized in the symmetric FCDI cell. The desalination performance results were greatly enhanced by the formation of the conducting network owing to the addition of CNT-S. Therefore, the sulfonation method can be regarded as an effective way to solve the problem of scale-up operation of the FCDI process.

system scale-up production possible by increasing the amount of the flow-electrode feed continuously. Thus, the FCDI system can provide infinite ion adsorption capacity and continuous desalination operation. Although the FCDI technology has displayed a promising potential for the treatment of high concentration salt solution, the desalting performance largely suffers from the limitation of the flow-electrode materials. In previous works, due to the discontinuous conductive electrical network of the active materials in the flow-electrode slurry, the conductivity of the commonly utilized activated carbon (AC) flowelectrode slurry is a few orders of magnitude lower than that of the conventionally AC solid electrode, which results in lower desalination capacity [26,33]. Increasing the loading of the active material in the flow-electrode slurry might be an effective method to improve the flowelectrode conductivity and enhance the salt removal performance, but the viscosity of the flow-electrode slurry significantly increases at the same time, leading to the aggregation of the active material in the flowelectrode slurry and the clogging of the flow-electrode channels. Hatzell et al. [29] fabricated a flow-electrode with a high loading of AC particles up to 28 wt% by chemical oxidation method. The results demonstrated that the salt removal efficiency of the FCDI cell is improved, but the problems of serious clogging of the flow-channels are urgent to be addressed [34,35]. Therefore, it is highly desired to find a suitable material for the FCDI system to prepare flow-electrode with the properties of the excellent conductive electrical network, good suspension stability, low flow viscosity and high material-loading [36]. Carbon nanotubes (CNT) display extraordinary electrochemical and desalination advantages due to its unique hollow tubular structures and intertwined nanoscale network structures. Also, the CNT with larger aspect ratio has been proved to be beneficial to facilitate the conductivity of the flow-electrode in the FCDI process [37–39]. However, the agglomeration of carbon nanotubes in the electrolyte occurs easily due to the strong van der Waals forces and poor hydrophilicity. Especially, this phenomenon can be even worse when CNT is used as the active material of the flow-electrode, which can cause the clogging of flow-channels during the deionization process. In order to address this problem, the surface modification of CNT has been adopted as an effective method to enhance the hydrophilicity of CNT and improve the affinity between the CNT surface and the electrolyte [28,32]. And the modified surface characteristic of materials plays an important role in affecting the ion removal efficiency because good connecting network and bridge between material particles can accelerate the charge transportation and storage in the percolating networks [40–42], which will also enhance the conductivity of flow-electrode and the desalination performance of the FCDI process. In addition, previous studies always ignore the fact that the effects of flow-electrode on the deionization performance is the comprehensive interaction between the flow factors and the natural properties of the flow-electrode. Dennison et al. [43] investigated the effects of the flow factors (e.g. flow rate, channel depth and rheological properties) on the capacitance and conductivity of flow-electrode, and the results demonstrated that the conductivity significantly varied with changing the flow factors and a rapid decay of the capacitance was also observed with increasing channel depth. Thus, the integrated effects of the natural properties of the flow-electrode (suspension stability and capacitance) and operation parameters (including the working voltage, the feed concentration and the flow rate of the flow-electrode) need to be optimized for the FCDI system during the desalting process to achieve the best desalination properties. Herein, the sulfonation treatment of CNT was prepared by using the sulfonating reagents and the sulfonated CNT (CNT-S) was used as the active material of the flow-electrode. To investigate the effects of CNT-S material’s surface characteristic on the suspension stability and capacitance of the flow-electrode, a series of dispersible property experiments and electrochemical tests for the CNT-S flow-electrode slurry were implemented. Meanwhile, the desalination performance of the CNT-S flow-electrode was confirmed through the desalting experiments

2. Experimental 2.1. Sulfonation treatment of CNT The sulfonated carbon nanotube (CNT-S) was synthesized by using chemical oxidation method with nitric acid and p-aminobenzenesulfonic acid [52,53]. The specification information of CNT has been presented in Table S1. Before the sulfonation process, CNT was pretreated by the HNO3 solution to remove the amorphous carbon and metal impurities. In brief, CNT was mixed with 10 mol/L HNO3 solution and refluxed in the water bath at 90 °C for 4 h, and the purified CNT was obtained by rinsing with deionized water for several times. After that, 10.4 g of p-aminobenzenesulfonic acid was reacted in 1 mol/L HCl solution with slowly adding 1 mol/L NaNO2 under an ice bath for 1 h and the p-diazonium benzenesulfonate was obtained. Then, the p-diazonium benzenesulfonate reacted with the pretreated CNT with dropwise adding 90 mL 50 wt% hypophosphorous acid (H3PO2) solution as the reducing agent. The whole process was carried out under the ice bath. The obtained sulfonated CNT sample was filtered, washed with deionized water and dried in vacuum at 80 °C for 24 h and named as CNT-S. The sulfonic group (eSO3−) was introduced into the surface of CNT to increase the amount of charged functional groups. To make the synthesis process more clearly, the schematic diagram is presented in Fig. S1.

2.2. Characterization of CNT-S materials and CNT-S flow-electrode slurry The morphology and structure of CNT-S and CNT were characterized by field emission scanning electron microscope (SEM, Nanosem 430, Netherlands), field emission transmission electron microscope (TEM, Tecnai G2 F20, Netherlands), Fourier transform infrared (FTIR) spectroscopy (MultiGas 2030, America), and X-ray photoelectron spectroscopy (XPS) (PHI5000 Versa Probe, Japan). The surface wettability of CNT-S and CNT material were tested by the dynamic contact angle analysis method (the two materials are pressed into thin films with the thickness of about 0.5 mm). The flow-electrode slurry was fabricated by mixing sulfonated carbon nanotubes (CNT-S) and Sodium dodecyl sulfate (SDS) with deionized water (DI water). In detail, the CNT-S and DI water was mixed with a mass ratio of 1: 5 to ensure the concentration of the slurry to be 2.0 wt%. Then the dispersant SDS was added into the above slurry. The prepared slurry was treated by ultrasonic oscillation for 1.5 h. The suspension and dispersion stability of the flow-electrode slurry was evaluated by the viscosity test, the centrifugal sedimentation test, the particle size distribution test and the Zeta potential test. The viscosity of the flow-electrode slurry was measured with a digital viscometer (SHP, NDJ-5S Viscometer) at a rotating speed of 60 rpm. The centrifugal sedimentation operation was performed at the rotational speed of 8000 rpm with the centrifugation time of 30 min. The particle size distribution test (Zetasizer Nano ZS) was conducted to measure the distribution of the CNT-S particles in the flow-electrode slurry. And the Zeta potential test (Zetasizer Nano ZS) was employed to evaluate the charged characteristics of CNT-S particles.

2

Separation and Purification Technology xxx (xxxx) xxxx

Y. Cai, et al.

2.3. Electrochemical characterization of the CNT-S flow-electrode

and 0.3 mm depth). This setup contained the anion- and cation-exchange membranes (TRJAM-Ⅱ and TRJCM-Ⅱ, 160 μm thickness) which were used to separate the feed solution channel from the flow-electrode compartments. The ion exchange membrane has an effective contact area of 217 cm2 between the part of the ion exchange membrane and the flow electrode. A silicone gasket and three-layer polyester spacers with the thickness of 0.6 mm for every spacer were placed between cation- and anion-exchange membranes. All parts were held together using phenolic resin end plates with dimensions (length × width × thickness) of 31 cm × 28 cm × 2.5 cm.

To investigate the electrochemical performances of the CNT-S flowelectrode, the electrochemical experiments were performed in terms of the cyclic voltammetry (CV) and electrochemical impedance spectroscopy (EIS) test in a three-electrode system with the test electrode as working electrode, the saturated calomel electrode (SCE, Hg/HgCl) as reference electrode and the platinum foil (Pt) as counter electrode. All electrochemical tests were conducted in 1 M NaCl solution. Cyclic voltammetry was carried out at the scanning potential range of −0.2 ~ 0.6 V vs OC and the scan rate of 5 mV/s. The specific capacitance (Cs , F/g) was calculated from the CV curves according to the following formula (1):

Cs =

2.4.2. Operation and desalination characteristics of the FCDI process The FCDI process used in this research was depicted in Fig. 2, which consisted of the measurement section and two fluid circulation units (the flow electrode-loop and feed solution flow channel). The feed solution with a concentration of 1.0 g/L NaCl solution (160 mL) was pumped into the flow channel at a flow rate of 20 mL/ min and was operated under close cycle condition with one reservoir. The flow rate of the flow-electrode, operated in a closed cycle, was maintained constant at 25 mL/min. The fresh cathode and anode flowelectrode were pumped from two flow-electrode reservoirs to the FCDI cell. After the desalting process, the used cathode and anode flowelectrode were regenerated in another FCDI cell with washing fluid and the opposite voltage. The cathode and anode CNT-S flow-electrodes were adopted to assemble the symmetric FCDI cell. For comparison, the symmetric FCDI cell was assembled by using the CNT flow-electrodes. A constant voltage of 1.3 V was applied to the FCDI cell through a DC power supply for desalting experiments. The variation of NaCl concentration was monitored at the outlet of the FCDI cell using a conductivity meter. During the desalination test, the salt removal efficiency (η, %) and adsorption rate (νA , mg/min/cm2) can be calculated in the following equations (3) and (4), respectively:

∫ idV (1)

2mv ΔV

where ∫ idV is the integral area of the CV curves; m is the active material mass of one electrode (g); ΔV is the window voltage (V); and v is the scan rate (mV/s). The impedance characteristics of the flow-electrode have great influences on the capacitance and desalting performance. In order to study the impedance characteristics of the flow-electrode and the diffusion process of ions, EIS test was employed on the Ametek PARSTAT4000 electrochemical workstation at a scanning frequency range of 0.01 Hz to 100 kHz. The capacitance in the EIS test was calculated by the following equation:

C=

1 ωZ′ ′

(2) ′

where C is the specific capacitance (F/g); Z′ is the imaginary resistance of the electrode impedance (Ω); ω is the angular frequency of the applied ac signal, ω = 2πf ; f is the scanning frequency (Hz). 2.4. Desalination performance test of the FCDI cell

η (%) = To evaluate the desalting performance of the CNT-S flow-electrode, the self-designed FCDI system was assembled and operated.

νA = 2.4.1. Assembly of the FCDI cell The FCDI cell (Fig. 1) is composed of a pair of titanium current collectors with the size of 200 mm × 172 mm × 0.4 mm (length × width × thickness) with carved flow channels (4 mm width

k 0 − k1 × 100% k0

α (k 0 − k1 ) V At A

(3)

(4)

where k 0 , k1are the initial and final conductivity of NaCl solution (μS/ cm), respectively; α is the fitting coefficient between the conductivity and concentration of NaCl solution (5.79 × 10−4 mg/(μS·cm2); V is the

Fig. 1. The assembly of the FCDI setup. 3

Separation and Purification Technology xxx (xxxx) xxxx

Y. Cai, et al.

Fig. 2. The schematic diagram of the FCDI desalination system.

openings appear at the end of CNT-S nanotubes, which may increase the adsorption sites in the surface and tube end of CNT-S. What’s more, the length of CNT-S is largely shorten to little nanotubes, which is beneficial to reduce the interaction between nanotubes and promote the dispersion property in the flow-electrode. The FT-IR test was performed to characterize the chemical structures of the material surface by qualitatively analyzing the bonding types and the interactions between the grafted functional groups and CNT. It can be seen from Fig. 4(a), compared with CNT, the main peaks at 3428.42 cm−1, 1642.05 cm−1 and 1086.83 cm−1 which are attributed to the OeH, C]C and CeH stretching vibration bonds respectively remain in the structure of CNT-S, indicating that the sulfonation modification process doesn’t change the main structure of CNT. The peak locating at 547.68 cm−1 is ascribed to the stretching vibration of SeO and the absorption band at 628.68 cm−1 belongs to the symmetric stretching of S]O. And the absorption peaks at 1168.59 cm−1, 1114.08 cm−1 and 1005.06 cm−1 are corresponding to the S]O stretching vibration peaks on the sulfonic acid group, which demonstrates the successful preparation of CNT-S [52]. In order to further determine the modification degree, the XPS was carried out by quantitatively analyzing the contents and types of functional groups on the material surface. Fig. 4(b) shows the XPS spectra of the CNT and CNT-S materials, and the elemental contents of two materials are presented in Table 1. It can be found that after the sulfonation treatment of CNT, the carbon content decreases from 95.83% to 93.93%, and the oxygen content increases from 4.17% to 4.88%. The proportion of sulfur in CNT-S indicates the degree of sulfonation. The high-resolution spectrum for the S 2p in Fig. S2 shows that the peak appearing at 168.1 eV is ascribed to the oxidized S (SeO/ S]O), which comes from the grafted sulfonic acid groups [12,54]. In addition, compared with CNT, a certain amount of nitrogen (0.36%) appears in the XPS spectra of CNT-S. The nitrogen element mainly comes from the reactant p-aminobenzenesulfonic acid. During the chemical reaction between the p-diazonium benzenesulfonate and the pretreated CNT, a certain amount of nitrogen would attach on the surface of CNT or is doped on the CNT. The nitrogen that attaches on the surface of CNT with the sulfonic group would reduce the interaction between nanotubes and promote the dispersible property of the flow-

total volume of NaCl solution (mL); and tA is the adsorption time (min). A represents the effective contact area between the ion-exchange membrane and the flow-electrode. To evaluate the desalination capacity of the flow-electrode, the specific adsorption capacity (Γ, mg/g), the mean electrosorption rate (v, mg/(g·min)), charge efficiency (Λ ) and electrode regeneration efficiency (ηr) of the flow-electrode were calculated based on the following equations (5)–(8):

Γ=

α ΔkV m

(5)

ν=

α ΔkV mt

(6)

Λ=

mF Γ × 100% MQ

(7)

ηr =

Γn × 100% Γ1

(8)

whereΔk is the difference between the initial and final conductivity of NaCl solution (μS/cm); m is the mass of flow-electrode slurry in the electrode chamber during the desalination process (g); t is the electrosorption time (min); F is the Faraday’s constant (96,485C/mol); M is the molecular weight of NaCl (58.5 g/mol); and Q is the total charge in the adsorption process (C); Γ (mg/g) is the specific adsorption capacity of the FCDI cell. Γ1 and Γn are the specific adsorption capacity of the first cycle and the nth cycle, respectively. 3. Results and discussion 3.1. Characterization of CNT-S material The morphology and structure of CNT and CNT-S were investigated by SEM and TEM, as illustrated in Fig. 3(a)-(d). From the SEM images in Fig. 3(a) and (b), it can be clearly observed that the CNT-S still retains the tubular feature of CNT after the modification treatment, whereas the surface roughness of CNT-S is larger than that of CNT. Compared with CNT, the TEM image of CNT-S in Fig. 3(d) clearly shows that more 4

Separation and Purification Technology xxx (xxxx) xxxx

Y. Cai, et al.

Fig. 3. SEM and TEM images of CNT (a), (c) and CNT-S (b), (d).

electrode. And the doped nitrogen in the CNT-S would improve the hydrophilicity of the CNT-S, which is beneficial to ion transportation and enhance the desalination performance. The content changes of carbon and oxygen elements and the presence of sulfur and nitrogen elements effectively demonstrated the successful grafting of the sulfonic acid groups on the surface of CNT. The dynamic contact angle analysis was used to confirm the good hydrophilicity of the CNT-S. Fig. 5 presents the changes of dynamic contact angle for CNT and CNT-S materials. It clearly shows that the initial contact angle of CNT-S (63.2°) is smaller than that of CNT (96.2°). What’s more, the time of the dynamic contact angle change from the initial value to zero for CNT-S is much shorter than that of

Table 1 Elements content of CNT and CNT-S. Sample

CNT CNT-S

Content percentage (%) C1s

O1s

N1s

S2p

95.83 93.93

4.17 4.88

0 0.36

0 0.83

Fig. 4. The FT-IR spectra (a) and XPS spectra (b) of CNT and CNT-S. 5

Separation and Purification Technology xxx (xxxx) xxxx

Y. Cai, et al.

Table 2 The centrifugal separation test of the CNT and CNT-S flow-electrode slurry. Sample

CNT

CNT-S

Precipitate /g

0.450

0.102

the CNT-S flow-electrode slurry and improve the dispersion property of the CNT-S flow-electrode slurry. The excellent suspension stability of CNT-S material over that of CNT material can be ascribed to the fact that the steric hindrance effect of sulfonic groups on the surface of CNTS material is stronger than that of hydroxyl groups on the surface of CNT material. The particle size distribution is adopted to measure the aggregation effect of the flow-electrode slurry. Fig. 6(b) exhibits the particle size distribution of CNT and CNT-S flow-electrode slurry. It clearly shows that the average particle size of the CNT flow-electrode slurry is about 220 nm, while the average particle size of CNT-S flow-electrode slurry is only 106 nm. The results indicate that the sulfonic groups on the surface of the CNT-S material, which can form electrostatic repulsion between particles, can effectively improve the hydrophilicity and weaken the aggregation of CNT-S material in the flow-electrode slurry. The centrifugal sedimentation experiments of the CNT and CNT-S flow-electrode slurry are performed and the relevant results are listed in Table 2. After the centrifugal sedimentation operation, the CNT-S flowelectrode slurry produces 0.102g precipitate, which is less than that of the CNT flow-electrode slurry (0.450g). This test demonstrates that the suspension stability of the CNT-S flow-electrode slurry is better than that of the CNT flow-electrode slurry, which can be attributed to the presence of the sulfonic functional groups on the surface of the CNT-S materials. These results indicate that the functional modification is beneficial for the preparation of the flow-electrode slurry. The Zeta potential test can be performed to evaluate the interaction between the dispersible particles and the dispersant. The Zeta potential of the CNT and CNT-S flow-electrode slurry is presented in Table 3. It clearly displays that the Zeta potentials of the two flow-electrode slurries all present negative value and the Zeta potential of CNT-S flowelectrode slurry (−18.4 mV) is much lower than that of the CNT flowelectrode slurry (−5.45 mV). And this phenomenon can be ascribed to the fact that the grafted sulfonic groups on the surface of the CNT-S material are more than that of the carboxyl and hydroxy groups on the surface of the CNT. These sulfonic groups can promote the good dispersible property and reduce the average particle size of the CNT-S material. The smaller the particle size is, the higher the absolute value of the Zeta potential is, which makes the flow-electrode slurry more stable. Therefore, the Zeta potential value of the CNT-S flow-electrode slurry becomes more negative and its absolute value is higher than that of the CNT flow-electrode slurry, indicating the good suspension

Fig. 5. The dynamic contact angle analysis of CNT and CNT-S.

CNT, which indicates the good hydrophilicity of CNT-S. 3.2. Characterization of CNT-S flow-electrode slurry The flowability and rheologic properties of the flow-electrode are important for the desalting performance and system stability. Thus, the viscosity measurements, the centrifugal sedimentation experiment, the particles size distribution test and the Zeta potential test were implemented to investigate the dispersible property of the flow-electrode slurry. The viscosity is an important index to evaluate the rheologic and dispersible property of the flow-electrode slurry. The poor dispersion of the flow-electrode slurry would exacerbate the aggregation of particles and increase the viscosity of the flow-electrode slurry. Thus, the viscosity is negatively correlated with the dispersion ability of the flowelectrode slurry. The viscosity results of CNT and CNT-S flow-electrode slurry are displayed in Fig. 6(a). It can be seen from Fig. 6(a) that the viscosity of CNT-S flow-electrode slurry is close to that of DI water and obviously lower than that of the CNT flow-electrode slurry. Namely, CNT-S flow-electrode slurry possesses similar flow behavior with the DI water and superior to the CNT flow-electrode slurry. The above results can be attributed to the following reasons. The presence of sulfonic functional groups can make the surface of the CNT-S material form an effective electrostatic layer. The effective electrostatic layer overcomes the interaction between the CNT-S nanotubes. So the hydrophilicity and dispersible stability of CNT-S material are enhanced. Meantime, the functional groups forms a steric hindrance effect [39–42] and prevents the aggregation of CNT-S material, which can reduce the viscosity of

Fig. 6. The viscosity (a) and the particle size distribution (b) of the CNT and CNT-S flow-electrode slurry. 6

Separation and Purification Technology xxx (xxxx) xxxx

Y. Cai, et al.

charging capacity of the CNT-S flow-electrode is higher than that of the CNT flow-electrode. CV curves obtained at the scan rate of 1 mV/s in Fig. S3 maintain the quasi-rectangular shapes and no evident oxidation/reduction peaks are observed in the scope of the given potential window, implying that the adsorption of ions on the flow-electrode follows the ideal capacitive electrical double layer mechanism. And this results can be explained by the following two reasons: (i) After the modification treatment, CNT-S presents several charge characteristics due to the grafted sulfonic functional groups on the surface, which accelerates the selective migration of counter-ions in the feed solution. So the fast ions migration rate can promote the specific capacitance of the CNT-S flow-electrode. (ii) The hydrophilicity of CNT-S material is enhanced by the modification treatment, which results in the decline of charge resistance in the charging and discharging process. The charge characteristic plays a dominant role in the application of the flowelectrode. And the improvement of materials’ hydrophilicity has a prominent effect on the increase of slurry conductivity. Therefore, the synergistic effect of two types of effects (the promotion effect of charge characteristics and the improvement effect of hydrophilicity) makes the specific capacitance of the CNT-S flow-electrode superior to that of the CNT flow-electrode. Fig. 8 gives the Nyquist plots and the specific capacitance as a function of the frequency of the CNT and CNT-S flow-electrode slurries. Typically, the Nyquist plots consist of two distinct regions: a linear region at low frequency, which is associated with the diffusion of the salt ions and capacitive behavior derived from EDL, and an arc semicircle region at high frequency, which reveals information about the charge transfer rate of ions into the CNT and CNT-S materials. The charge resistance is a comprehensive effect of the ohmic resistance and the ions diffusion resistance. The low ohmic resistance facilitates the fast ion transport and charge percolation [36,55,56]. The measured EIS plots are fitted using an equivalent circuit diagram to achieve the values of the circuit components, which can be qualitatively associated with process transport parameters. As shown in Fig. 8(a), the ohmic impedance (Rs) of the CNT-S flow-electrode slurry obtained from the Nyquist plots at high frequency is 2.11 Ω, which is lower than that of the CNT flow-electrode slurry (2.92 Ω). Due to the same amount of the active material in all tested flow-electrode slurries, this difference is mainly attributed to the hydrophilicity of the active materials. The good hydrophilicity of the active material is beneficial to decrease the ohmic resistance of the flow-electrode, which can facilitate the conductivity of flow-electrodes. Thus, the conductivity of CNT-S flow-electrode is better than that of the CNT flow-electrode. Simultaneously, in the Nyquist plots, the ion diffusion behavior of the two flow-electrode slurries is obtained by the sloped region at low frequency. It is found that the CNT-S flow-electrode slurry exhibits a slightly larger inclination,

Table 3 The Zeta potential test of CNT and CNT-S flow-electrode slurry. Sample

CNT

CNT-S

Zeta potential (mV)

−5.45

−18.4

Fig. 7. Cyclic voltammetry of the CNT and CNT-S flow-electrode at the scan rate of 5 mV/s.

stability of CNT-S flow-electrode slurry. 3.3. Electrochemical performances of the flow-electrode slurry The electrochemical performances of the CNT and CNT-S flowelectrode slurry were evaluated in terms of the specific capacitance through cyclic voltammetry (CV) test and ohmic resistance using electrochemical impedance spectroscopy (EIS). During the electrochemical performance test process, the flow-electrode slurry was uniformly loaded into the channel of a static cell and keep the still state. The CV experiment is used to test the specific capacitance for the CNT and CNT-S flow-electrode slurries. The CV curves of the CNT and CNT-S flow-electrode slurries are shown in Fig. 7. As we can see, with the scan rate of 5 mV/s, the CV curves of the two flow-electrode slurries show near-rectangular shapes without the obvious redox peak, indicating that the modified treatment does not change the ions storage mechanism of the electrical double layers. The specific capacitances of the CNT-S and CNT flow-electrode calculated according to the formula (1) are 33.70F/g and 23.77F/g respectively, indicating that the

Fig. 8. The Nyquist plots (a) (the inset is the local magnification and the equivalent circuit) and specific capacitance (b) of the CNT and CNT-S flow-electrode slurries. 7

Separation and Purification Technology xxx (xxxx) xxxx

Y. Cai, et al.

has been compared with previous results listed in Table S2. From Table S2, it clearly shows that compared with other cells [49–51], the CNT-SSDS FCDI cell reveals the dominant advantage in the salt removal efficiency. All the above good results can be ascribed to the addition of the dispersant SDS and the good hydrophilicity of CNT-S material. The dispersant SDS consists of the hydrophilic groups and the hydrophobic groups. When it is added into the CNT-S flow-electrode slurry, the hydrophobic groups connect with the CNT-S surface by the interaction to form a steric hindrance layer and the hydrophilic groups with the electronegativity will form the charge layer to adsorb the counter-ions [47]. And when it interacts with the negative charged CNT-S material, the steric hindrance layer formed by the hydrophobic groups and the charge layer formed by the hydrophilic groups can produce the synergistic effect to promote the adsorption of counter-ions, which would improve the salt removal capacity of the CNT-S flow-electrode. In addition, the salt removal efficiency of 45% for the CNT-S-SDS FCDI cell is much higher than that of the CNT-SDS FCDI cell (26%). This can be ascribed to the grafted sulfonic functional groups. On the one hand, the sulfonated CNT with the functional groups present good dispersible property and conducting network, which can significantly promote the stability and conductivity of the FCDI cell. On the other hand, these grafted functional groups on the surface of CNT can reduce the co-ion effect [42–46,48] inside the flow-electrode slurry in the desalting process and enhance the adsorption capacity and desalination efficiency of the CNT-S flow electrode. Therefore, the dispersant plays an important role in improving the desalination performance of the FCDI cell. And the following desalination experiments would use SDS as the dispersant in CNT-S flowelectrode and investigate the effects of the operating parameters on the CNT-S FCDI cell.

indicating its good ion diffusion ability. Thus, the charge resistance of the CNT-S flow-electrode slurry is far lower than that of the CNT flowelectrode slurry. In Fig. 8(b), the specific capacitances of the two flowelectrode slurries all have a sharp decrease at the low frequency range of 0.01–10 Hz and reach a plateau at the high frequency region. The specific capacitance of the CNT-S flow-electrode is higher than that of the CNT flow-electrode in the low frequency range, and this result is consistent with the CV test. According to the above tests, the dispersible property and electrochemical performance of the CNT-S flow-electrode slurry are superior to that of the CNT flow-electrode slurry. Therefore, the CNT-S material is more suitable as the flow-electrode in the FCDI cell considering its excellent dispersible property and electrochemical performance. 3.4. Desalination performance tests of the FCDI system 3.4.1. The effect of the dispersant on the desalination performance of the FCDI system The suitable dispersant can significantly improve the dispersible property and the desalting performance of the flow-electrode. Here the anionic dispersant sodium dodecyl sulfate (SDS) is added into the prepared flow-electrode. And four symmetric FCDI cells assembled with the CNT, CNT-S, CNT-SDS and CNT-S-SDS flow-electrodes are used to perform the saturated desalination experiments. CNT-SDS and CNT-SSDS flow-electrodes denotes that the anionic dispersant SDS is added into the two flow-electrodes respectively. Fig. 9(a) displays the change of the normalized solution conductivity. It clearly shows that the normalized conductivities of the four FCDI cells declines sharply at the beginning and then reach to the stable state at about 100 min. The salt removal efficiencies of the CNT, CNT-S, CNT-SDS and CNT-S-SDS FCDI cells are 20%, 29%, 26% and 45% respectively. According to the equation (5), the corresponding specific adsorption capacities of the four FCDI cells are 3.44 mmol/g, 6.0 mmol/g, 5.25 mmol/g and 8.29 mmol/g respectively. And the mean electrosorption rates of the four FCDI cells are 2.01 mg/(g·min), 3.51 mg/(g·min), 3.07 mg/(g·min) and 4.85 mg/(g·min) shown in Fig. 9(b). Compared with the CNT FCDI cell, the CNT-S, CNT-SDS and CNT-S-SDS FCDI cells display better performance in the salt removal efficiency and mean electrosorption rate, and especially the CNT-S-SDS FCDI cell presents the best advantages with the salt removal efficiency of 45% and mean electrosorption rate of 4.85 mg/(g·min). And the CNT-SDS FCDI cell with the addition of the dispersant SDS also shows higher salt removal efficiency (26%) and faster electrosorption rate (3.07 mg/(g·min)) than that of CNT FCDI cell (20%, 2.01 mg/(g·min)) without SDS. The same trend also appears between the CNT-S FCDI cell and the CNT-S-SDS FCDI cell. What’s more, the desalination performance of the CNT-S-SDS FCDI cell

3.4.2. The effect of the operating voltage on the desalination performance of the FCDI system In order to evaluate the effect of the operating voltage on the desalination performance of the CNT-S FCDI cell, the saturated adsorption experiments were implemented in the 1.0 g/L NaCl solution at different operating voltages from 0.8 V to 1.5 V. Fig. 10 depicts the saturated adsorption curves of the CNT-S FCDI cell at different operating voltages. In Fig. 10, when the working voltage is applied, the decrease of conductivity for the CNT-S FCDI cell at different operating voltages can be observed at first and the adsorption process becomes saturated at about 100 min. As can be seen from Table 4, with increasing the operating voltages from 0.8 V to 1.5 V, the desalination efficiency, adsorption rate and charge efficiency of the CNT-S FCDI cell all display the same trend, and the results are calculated by the equation (3), (4) and (7). Moreover, the maximum desalination efficiency (45.8%), the fastest

Fig. 9. The normalized conductivity (a) and mean electrosorption rates (b) of the four FCDI cells. 8

Separation and Purification Technology xxx (xxxx) xxxx

Y. Cai, et al.

removal efficiency is the smallest (38.6%) and the saturated adsorption time is the longest (200 min), which means that the FCDI cell at this condition presents the worst desalination performance. The reason can be explained by the fact that as the flow rate of the feed solution is set as the same with that of the flow-electrode, the driving force between the feed solution and the flow-electrode is the weakest and the adsorption process can be regarded as the relatively static state, which will result in the slow adsorption rate, low desalination efficiency and long saturated adsorption time. As the flow rate of the feed solution is lower than that of the flow-electrode (the ratio of 0.6 and 0.8), the relative flow rate difference between the feed solution and the flow-electrode would promote ions transportation and adsorption rate, resulting in the high desalination efficiency. And the highest desalination efficiency of 45.0% is obtained at the flow rate ratio of 0.8. The desalination efficiency at the flow rate ratio of 0.6 (41.4%) is lower than that at the flow rate ratio of 0.8, which is mainly due to the low flow rate leading to the inadequate adsorption of ions. When the flow rate ratio is higher than 1.0, the relative flow rate difference still exists between the feed solution and the flow-electrode, but the ion residence time would be shortened. At the high flow rate ratio (1.2, 1.4 and 1.6), salt ions in the feed solution are not in full contact with the flow-electrode, the new feed solution would take salt ions out of the cell, leading to the inadequate ions adsorption and low salt removal efficiency. More time will be needed to achieve the saturated adsorption state of the flowelectrode. Thus, considering the salt removal efficiency and the saturated adsorption time, the optimized flow rate ratio of the feed solution to the flow-electrode is set as 0.8 in the whole desalination process.

Fig. 10. The saturated adsorption curves of the CNT-S FCDI cell at different operating voltages (0.8–1.5 V).

adsorption rate (0.0039 mg/min/cm2) and the highest charge efficiency (59.1%) are all achieved at the operating voltage of 1.3 V. These results are attributed to the fact that the operating voltage can provide the effective driving force of the ions migration and ensure the ions adsorption in the solution. When the operating voltage is lower than 1.3 V, the adsorption rate, the desalination efficiency and the charge efficiency decrease firstly and then increase with increasing the operating voltage. In contrast to the optimal operating voltage of 1.3 V, when the voltage is up to 1.4 V and 1.5 V, the decreasing trend of the desalination efficiency and adsorption rate occurs. What’s more, the charge efficiency of the CNT-S FCDI cell at 1.3 V is about 59.1% and obviously higher than that at 1.4 V (44.8%) and 1.5 V (38.4%). The above results indicate that when the voltage is lower than 1.3 V, the driving force of the ions migration is not enough for high desalination performance, while the higher voltage (1.4 V and 1.5 V) would accelerate water decomposition degree and decrease the charge efficiency, leading to the attenuation of desalination performance of the FCDI cell. Therefore, it is taking into account that the optimal operating voltage in the FCDI process is 1.3 V and it is adopted in the subsequent desalting experiment.

3.4.4. Adsorption experiments of FCDI system with different NaCl concentrations To investigate the desalination performance of the FCDI process at high concentration salt solution, the CNT-S FCDI cell was operated at the concentrations of NaCl solution from 1.0 g/L to 35 g/L, as shown in Fig. 12(a). The desalination performance is evaluated in terms of the conductivity variations in the effluent water concentration. In Fig. 12(a), the conductivities of all the NaCl solution in the effluent stream decrease rapidly within the first 30 min and then reaches to a plateau as the ions adsorption on the flow-electrode reaches the equilibrium state. The steady concentrations of the effluent stream are calculated as 31.35, 17.64, 8.92, 4.09, 1.35 and 0.55 g/L for the initial NaCl solution with concentration of 35.0, 20.0, 10.0, 5.0, 2.0 and 1.0 g/ L, respectively. It can be seen that the CNT-S FCDI cell can effectively adsorb ions in the high concentration of salt water, which indicates that the CNT-S FCDI cell can be applied in the high concentration NaCl solution as it overcomes many limitations of the typical CDI systems. In addition, the salt removal efficiency of the CNT-S FCDI cell is also investigated at various NaCl solutions, as provided in Fig. 12(b). The calculated salt removal efficiencies with initial NaCl salt concentrations of 1.0, 2.0, 5.0, 10.0, 12.0, 20.0 and 35.0 g/L are 45.2%, 32.5%, 18.3%, 10.8%, 11.3%, 11.8% and 10.4%, respectively. With the increase of the initial salt concentration, the removal efficiency decreases. The decreasing trend of salt removal efficiency can be explained by the fact that the amount of the fresh salt ions introduced into the system is greater than that removed from the system. Thus, the CNT-S FCDI cell can also be used to treat high concentration NaCl solution and the adsorption capacity of the flow-electrode is influenced deeply by the initial NaCl concentration.

3.4.3. The optimization of the flow rate ratio of the feed solution to the flowelectrode in the FCDI process The flow rate ratio of the feed solution to the flow-electrode is an important parameter in the continuous FCDI process because the relative flow rate of the two fluids determines the desalination efficiency in a large degree. So it is necessary to optimize the relative flow rate ratio of the two fluids. The flow rate ratio of the feed solution to the CNT-S flow-electrode is set as 0.6, 0.8, 1.0, 1.2, 1.4 and 1.6 respectively. Fig. 11(a) presents the conductivity changes with time for the CNT-S FCDI cell at different flow rate ratios. The salt removal efficiencies of the FCDI cell at these six flow rate ratios (0.6, 0.8, 1.0, 1.2, 1.4 and 1.6) are 41.4%, 45.0%, 38.6%, 42.5%, 41.7% and 40.7% respectively and the corresponding saturated adsorption time are 160 min, 140 min, 200 min, 150 min, 155 min and 180 min respectively, as shown in Fig. 11(b). The above results indicate that when the flow rate ratio reaches to 1.0 (the same flow rate of the feed solution and flow-electrode), the salt

Table 4 The desalination efficiency, adsorption rate and charge efficiency of the CNT-S FCDI cell. Voltage /(V)

0.8

0.9

1.0

1.1

1.2

1.3

1.4

1.5

Desalination efficiency/(%) Adsorption rate /(mg/min/cm2) Charge efficiency/(%)

43.0 0.0037 57.7

42.9 0.0036 56.6

41.5 0.0035 54.6

41.1 0.0037 53.9

44.7 0.0038 58.6

45.8 0.0039 59.1

35.1 0.0032 44.8

30.3 0.0028 38.4

9

Separation and Purification Technology xxx (xxxx) xxxx

Y. Cai, et al.

Fig. 11. The conductivity vs. time (a) and the salt removal efficiency and the saturated adsorption time (b) for different flow rate ratios.

3.4.5. The influence of the flux of washing fluid on the flow-electrode regeneration process In order to study the continuous desalting process for the FCDI system, this work adopts the continuous CNT-S FCDI system coupled with the adsorption and desorption process simultaneously as shown in Fig. S4. The influence of the flux of washing fluid on the flow-electrode regeneration efficiency is investigated by controlling the flow rate of the washing fluid in the desorption process. In this part, the CNT-S flowelectrode is continuously recirculated in the FCDI desalination cell at a flow rate of 25 mL/min. The feed NaCl solution flows into the FCDI cell with the flow rate of 20 mL/min and the flow rate of the washing fluid was controlled to make sure the flux ratio of the washing fluid to the feed NaCl solution as 5/5, 4/6, 3/7 and 2/8 respectively. Fig. 13 displays the conductivity changes of the washing fluid and feed solution at different flux ratios of the washing fluid to the feed NaCl solution. It clearly shows that the conductivity of the feed solution declines fast while the washing fluid presents the opposite trend at the initial stage of the desalination process. Then, the conductivity changes of the feed solution and washing fluid reach the stable state at about 200 min, which means that the desalination and desorption process reaches to the balanced state. According to the Fig. 13, the effluent concentration of the feed solution and the washing fluid at different flux ratios are listed in Table 5. From Table 5, it clearly displays that with the decrease of the flow rate of the washing fluid, the electrode generation efficiency declines seriously from 100% to 83.1%. And this phenomenon can be ascribed to the following reasons. On the one hand, the declining flow rate of the washing fluid can make the regeneration process of the fresh flow-

Fig. 13. The conductivity of the washing fluid and feed solution vs. time at different flux ratios.

electrode slow, which in turn accelerates the rate of electrode aging process and reduces the desalination performance. On the other hand, as the flow rate of the washing fluid becomes slow, the time for the regeneration process of the flow-electrode would be prolonged. The stability of the flow-electrode after the long-term operation will decay and the flow-electrode particles may aggregate and clog the flow channel, which will lead to the aggregation of the flow-electrode and influence the desalination performance and electrode regeneration

Fig. 12. Conductivity variations in the effluent stream (a) and NaCl removal efficiencies (b) for the CNT-S FCDI cell at different NaCl concentrations. 10

Separation and Purification Technology xxx (xxxx) xxxx

Y. Cai, et al.

Appendix A. Supplementary material

Table 5 The effluent concentration of washing fluid and feed solution at different flux ratios. Ratio

5/5

4/6

3/7

2/8

Flow rate of the feed solution (ml/min) Flow rate of the washing fluid (ml/min) Initial concentration (g/L) The concentration of the feed solution (g/L) The concentration of the washing fluid (g/L) Electrode generation efficiency (%)

25 25 1.019 0.164 1.879 100

25 17 1.017 0.185 2.239 98.4

25 11 1.019 0.180 2.715 86.6

25 6 1.018 0.228 3.644 83.1

Supplementary data to this article can be found online at https:// doi.org/10.1016/j.seppur.2019.116381. References [1] M. Elimelech, W.A. Phillip, The future of seawater desalination: energy, technology, and the environment, Science 333 (2011) 712–717. [2] A. Subramani, J.G. Jacangelo, Emerging desalination technologies for water treatment: a critical review, Water Res. 75 (2015) 164–187. [3] M.E. Suss, S. Porada, X. Sun, P.M. Biesheuvel, J. Yoon, V. Presser, Water desalination via capacitive deionization: what is it and what can we expect from it? Energy Environ. Sci. 8 (2015) 2296–2319. [4] S. Porada, R. Zhao, A. van der Wal, V. Presser, P.M. Biesheuvel, Review on the science and technology of water desalination by capacitive deionization, Prog. Mater Sci. 58 (2013) 1388–1442. [5] Y. Liu, B. Soucaze-Guillous, P.-L. Taberna, P. Simon, Understanding of carbon-based supercapacitors ageing mechanisms by electrochemical and analytical methods, J. Power Sources 366 (2017) 123–130. [6] L. Han, K.G. Karthikeyan, M.A. Anderson, K.B. Gregory, Exploring the impact of pore size distribution on the performance of carbon electrodes for capacitive deionization, J. Colloid Interface Sci. 430 (2014) 93–99. [7] J. Zhang, J. Fang, J. Han, T. Yan, L. Shi, D. Zhang, N, P, S co-doped hollow carbon polyhedra derived from MOF-based core–shell nanocomposites for capacitive deionization, J. Mater. Chem. A 6 (2018) 15245–15252. [8] J. Han, L. Shi, T. Yan, J. Zhang, D. Zhang, Removal of ions from saline water using N, P co-doped 3D hierarchical carbon architectures via capacitive deionization, Environ. Sci. Nano 5 (2018) 2337–2345. [9] J. Lee, S. Kim, C. Kim, J. Yoon, Hybrid capacitive deionization to enhance the desalination performance of capacitive techniques, Energy Environ. Sci. 7 (2014) 3683–3689. [10] M.E. Suss, T.F. Baumann, W.L. Bourcier, C.M. Spadaccini, K.A. Rose, J.G. Santiago, M. Stadermann, Capacitive desalination with flow-through electrodes, Energy Environ. Sci. 5 (2012) 9511. [11] T. Yan, J. Liu, H. Lei, L. Shi, Z. An, H.S. Park, D. Zhang, Capacitive deionization of saline water using sandwich-like nitrogen-doped graphene composites via a selfassembling strategy, Environ. Sci. Nano 5 (2018) 2722–2730. [12] G. Divyapriya, K.K. Vijayakumar, I. Nambi, Development of a novel graphene/ Co3O4 composite for hybrid capacitive deionization system, Desalination 451 (2019) 102–110. [13] Z.U. Khan, T. Yan, L. Shi, D. Zhang, Improved capacitive deionization by using 3D intercalated graphene sheet–sphere nanocomposite architectures, Environ. Sci. Nano 5 (2018) 980–991. [14] M.E. Suss, P.M. Biesheuvel, T.F. Baumann, M. Stadermann, J.G. Santiago, In situ spatially and temporally resolved measurements of salt concentration between charging porous electrodes for desalination by capacitive deionization, Environ. Sci. Technol. 48 (2014) 2008–2015. [15] L. Han, K.G. Karthikeyan, M.A. Anderson, J.J. Wouters, K.B. Gregory, Mechanistic insights into the use of oxide nanoparticles coated asymmetric electrodes for capacitive deionization, Electrochim. Acta 90 (2013) 573–581. [16] H. Wang, L. Shi, T. Yan, J. Zhang, Q. Zhong, D. Zhang, Design of graphene-coated hollow mesoporous carbon spheres as high performance electrodes for capacitive deionization, J. Mater. Chem. A 2 (2014) 4739–4750. [17] Y. Bai, Z.-H. Huang, X.-L. Yu, F. Kang, Graphene oxide-embedded porous carbon nanofiber webs by electrospinning for capacitive deionization, Colloids Surf. A 444 (2014) 153–158. [18] J. Yang, L. Zou, N.R. Choudhury, Ion-selective carbon nanotube electrodes in capacitive deionisation, Electrochim. Acta 91 (2013) 11–19. [19] C.-H. Hou, N.-L. Liu, H.-L. Hsu, W. Den, Development of multi-walled carbon nanotube/poly(vinyl alcohol) composite as electrode for capacitive deionization, Sep. Purif. Technol. 130 (2014) 7–14. [20] G. Rasines, P. Lavela, C. Macías, M.C. Zafra, J.L. Tirado, C.O. Ania, On the use of carbon black loaded nitrogen-doped carbon aerogel for the electrosorption of sodium chloride from saline water, Electrochim. Acta 170 (2015) 154–163. [21] Y. Cho, K.S. Lee, S. Yang, J. Choi, H.-R. Park, D.K. Kim, A novel three-dimensional desalination system utilizing honeycomb-shaped lattice structures for flow-electrode capacitive deionization, Energy Environ. Sci. 10 (2017) 1746–1750. [22] S.-I. Jeon, H.-R. Park, J.-G. Yeo, S. Yang, C.H. Cho, M.H. Han, D.K. Kim, Desalination via a new membrane capacitive deionization process utilizing flowelectrodes, Energy Environ. Sci. 6 (2013) 1471. [23] A. Rommerskirchen, Y. Gendel, M. Wessling, Single module flow-electrode capacitive deionization for continuous water desalination, Electrochem. Commun. 60 (2015) 34–37. [24] K.B. Hatzell, M. Beidaghi, J.W. Campos, C.R. Dennison, E.C. Kumbur, Y. Gogotsi, A high performance pseudocapacitive suspension electrode for the electrochemical flow capacitor, Electrochim. Acta 111 (2013) 888–897. [25] K.B. Hatzell, E. Iwama, A. Ferris, B. Daffos, K. Urita, T. Tzedakis, F. Chauvet, P.L. Taberna, Y. Gogotsi, P. Simon, Capacitive deionization concept based on suspension electrodes without ion exchange membranes, Electrochem. Commun. 43 (2014) 18–21. [26] S. Porada, D. Weingarth, H.V.M. Hamelers, M. Bryjak, V. Presser, P.M. Biesheuvel, Carbon flow electrodes for continuous operation of capacitive deionization and capacitive mixing energy generation, J. Mater. Chem. A 2 (2014) 9313.

efficiency. Thus, the flux ratio of washing fluid to feed solution can affect the electrode regeneration in a large degree. To achieve the optimal desalination performance and electrode regeneration efficiency, it is proposed to take the flux ratio of 5/5 as the best choice. 4. Conclusions In this study, the CNT-S was adopted as the active material to prepare the flow-electrode in the FCDI process. The enhanced wettability of the CNT-S material was confirmed by the functionalization method. And the excellent desalination performance was also achieved by the symmetric CNT-S FCDI cell. Some specific conclusions above the paper are listed as follows: (1) The introduction of the CNT-S material significantly improves the dispersible property and suspension stability of the flow-electrode slurry. (2) The CNT-S flow-electrode displays excellent electrochemical properties with a high specific capacitance of 33.7F/g and low ohmic resistance of 2.11 Ω. (3) For the symmetric FCDI cell assembled with only 2.0 wt% CNT-S flow-electrode with the SDS as the dispersant, the mean electrosorption rate of 4.85 mg/(g·min) and the high salt removal efficiency of 45.8% were achieved in 1.0 g/L NaCl solution, both of which were much higher than those of the CNT FCDI flow-electrode (3.07 mg/(g·min), 28.8%). (4) The FCDI system could achieve the best desalination performance at the voltage of 1.3 V and the flow rate ratio of 0.8 for the feed solution to the flow-electrode. What’s more, the continuous desalination and desorption processes can be realized in the FCDI cell and the flow-electrode can be well regenerated with the flux ratio of 5/5 for the washing fluid to the feed NaCl solution. Therefore, the modification method of material and the operation parameters of the FCDI system could provide some valuable references for the future development of the FCDI process. CRediT authorship contribution statement Yanmeng Cai: Data curation, Writing - original draft. Xiaotong Zhao: Conceptualization, Methodology, Software. Yue Wang: Writing review & editing. Dongya Ma: Visualization, Investigation. Shichang Xu: Supervision. Declaration of Competing Interest The authors declare that they have no known competing financial interests or personal relationships that could have appeared to influence the work reported in this paper. Acknowledgments This research is supported by the National Natural Science Foundation of China (No. 21576190). 11

Separation and Purification Technology xxx (xxxx) xxxx

Y. Cai, et al.

[42] K.B. Hatzell, M. Boota, E.C. Kumbur, Y. Gogotsi, Flowable conducting particle networks in redox-active electrolytes for grid energy storage, J. Electrochem. Soc. 162 (2015) A5007–A5012. [43] C.R. Dennison, M. Beidaghi, K.B. Hatzell, J.W. Campos, Y. Gogotsi, E.C. Kumbur, Effects of flow cell design on charge percolation and storage in the carbon slurry electrodes of electrochemical flow capacitors, J. Power Sources 247 (2014) 489–496. [44] N. Hauptman, M. Klanjšek Gunde, M. Kunaver, M. Bešter-Rogač, Influence of dispersing additives on the conductivity of carbon black pigment dispersion, J. Coat. Technol. Res. 8 (2011) 553–561. [45] V. Rezazadeh, M.R. Pourhossaini, A. Salimi, Effect of amine-functionalized dispersant on cure and electrical properties of carbon nanotube/epoxy nanocomposites, Prog. Org. Coat. 111 (2017) 389–394. [46] J. Yu, N. Grossiord, C.E. Koning, J. Loos, Controlling the dispersion of multi-wall carbon nanotubes in aqueous surfactant solution, Carbon 45 (2007) 618–623. [47] X.J. Wang, X.F. Li, S. Yang, Influence of pH and SDBS on the stability and thermal conductivity of nanofluids, Energy Fuels 23 (2009) 2684–2689. [48] Y. Ge, Z. Li, D. Xiao, P. Xiong, N. Ye, Sulfonated multi-walled carbon nanotubes for the removal of copper (II) from aqueous solutions, J. Ind. Eng. Chem. 20 (2014) 1765–1771. [49] W. Huang, Y. Zhang, S. Bao, R. Cruz, S. Song, Desalination by capacitive deionization process using nitric acid-modified activated carbon as the electrodes, Desalination 340 (2014) 67–72. [50] X. Gao, A. Omosebi, J. Landon, K. Liu, Enhancement of charge efficiency for a capacitive deionization cell using carbon xerogel with modified potential of zero charge, Electrochem. Commun. 39 (2014) 22–25. [51] H.-R. Park, J. Choi, S. Yang, S.J. Kwak, S.-I. Jeon, M.H. Han, D.K. Kim, Surfacemodified spherical activated carbon for high carbon loading and its desalting performance in flow-electrode capacitive deionization, RSC Adv. 6 (2016) 69720–69727. [52] P. Liu, H. Wang, T. Yan, J. Zhang, L. Shi, D. Zhang, Grafting sulfonic and amine functional groups on 3D graphene for improved capacitive deionization, J. Mater. Chem. A 4 (2016) 5303–5313. [53] D. Ma, Y. Cai, Y. Wang, S. Xu, J. Wang, M.U. Khan, Grafting the charged functional groups on carbon nanotubes for improving the efficiency and stability of capacitive deionization process, ACS Appl. Mater. Interfaces 11 (2019) 17617–17628. [54] S. Yang, L. Peng, P. Huang, X. Wang, Y. Sun, C. Cao, W. Song, Nitrogen, phosphorus, and sulfur co-doped hollow carbon shell as superior metal-free catalyst for selective oxidation of aromatic alkanes, Angew. Chem. Int. Ed. Engl. 55 (2016) 4016–4020. [55] K. Tang, S. Yiacoumi, Y. Li, C. Tsouris, Enhanced water desalination by increasing the electroconductivity of carbon powders for high-performance flow-electrode capacitive deionization, ACS Sustain. Chem. Eng. 7 (2019) 1085–1094. [56] T.N. Baroud, E.P. Giannelis, High salt capacity and high removal rate capacitive deionization enabled by hierarchical porous carbons, Carbon 139 (2018) 614–625.

[27] S.-I. Jeon, J.-G. Yeo, S. Yang, J. Choi, D.K. Kim, Ion storage and energy recovery of a flow-electrode capacitive deionization process, J. Mater. Chem. A 2 (2014) 6378. [28] Y. Gendel, A.K.E. Rommerskirchen, O. David, M. Wessling, Batch mode and continuous desalination of water using flowing carbon deionization (FCDI) technology, Electrochem. Commun. 46 (2014) 152–156. [29] K.B. Hatzell, M.C. Hatzell, K.M. Cook, M. Boota, G.M. Housel, A. McBride, E.C. Kumbur, Y. Gogotsi, Effect of oxidation of carbon material on suspension electrodes for flow electrode capacitive deionization, Environ. Sci. Technol. 49 (2015) 3040–3047. [30] S. Yang, J. Choi, J.G. Yeo, S.I. Jeon, H.R. Park, D.K. Kim, Flow-electrode capacitive deionization using an aqueous electrolyte with a high salt concentration, Environ. Sci. Technol. 50 (2016) 5892–5899. [31] C. Zhang, K.B. Hatzell, M. Boota, B. Dyatkin, M. Beidaghi, D. Long, W. Qiao, E.C. Kumbur, Y. Gogotsi, Highly porous carbon spheres for electrochemical capacitors and capacitive flowable suspension electrodes, Carbon 77 (2014) 155–164. [32] S. Yang, S.-I. Jeon, H. Kim, J. Choi, J.-G. Yeo, H.-R. Park, D.K. Kim, Stack design and operation for scaling up the capacity of flow-electrode capacitive deionization technology, ACS Sustain. Chem. Eng. 4 (2016) 4174–4180. [33] G.J. Doornbusch, J.E. Dykstra, P.M. Biesheuvel, M.E. Suss, Fluidized bed electrodes with high carbon loading for water desalination by capacitive deionization, J. Mater. Chem. A 4 (2016) 3642–3647. [34] A. Rommerskirchen, C.J. Linnartz, D. Müller, L.K. Willenberg, M. Wessling, Energy recovery and process design in continuous flow-electrode capacitive deionization processes, ACS Sustain. Chem. Eng. 6 (2018) 13007–13015. [35] L. Wang, J.E. Dykstra, S. Lin, Energy efficiency of capacitive deionization, Environ. Sci. Technol. 53 (2019) 3366–3378. [36] Y. Cho, C.Y. Yoo, S.W. Lee, H. Yoon, K.S. Lee, S. Yang, D.K. Kim, Flow-electrode capacitive deionization with highly enhanced salt removal performance utilizing high-aspect ratio functionalized carbon nanotubes, Water Res. 151 (2019) 252–259. [37] J.W. Campos, M. Beidaghi, K.B. Hatzell, C.R. Dennison, B. Musci, V. Presser, E.C. Kumbur, Y. Gogotsi, Investigation of carbon materials for use as a flowable electrode in electrochemical flow capacitors, Electrochim. Acta 98 (2013) 123–130. [38] M. Karami, M.A. Akhavan Bahabadi, S. Delfani, A. Ghozatloo, A new application of carbon nanotubes nanofluid as working fluid of low-temperature direct absorption solar collector, Sol. Energy Mater. Sol. Cells 121 (2014) 114–118. [39] S.S.J. Aravind, P. Baskar, T.T. Baby, R.K. Sabareesh, S. Das, S. Ramaprabhu, Investigation of structural stability, dispersion, viscosity, and conductive heat transfer properties of functionalized carbon nanotube based nanofluids, J. Phys. Chem. C 115 (2011) 16737–16744. [40] M.A. Sabiha, R. Saidur, S. Hassani, Z. Said, S. Mekhilef, Energy performance of an evacuated tube solar collector using single walled carbon nanotubes nanofluids, Energy Convers. Manage. 105 (2015) 1377–1388. [41] F.Y. Fan, W.H. Woodford, Z. Li, N. Baram, K.C. Smith, A. Helal, G.H. McKinley, W.C. Carter, Y.-M. Chiang, Polysulfide flow batteries enabled by percolating nanoscale conductor networks, NanoLetters 14 (2014) 2210–2218.

12