Carbon Material-Based Flow-Electrode Capacitive Deionization for Continuous Water Desalination

Abstract

Flow-electrode capacitive deionization (FCDI) offers an electrochemical, energy-efficient technique for water desalination. In this work, we report the study of carbon-based FCDI, which consists of one desalination chamber and one salination chamber and applies a carbon nanomaterials-based flow electrode that circulates between the cell anode and cathode, to achieve a fast, continuous desalination process. Five different carbon nanomaterials were used for preparing the flow electrode and were studied for the desalination performance, with properties including average salt removal rate (ASRR), salt removal efficiency (SRE), energy consumption (EC) and charge efficiency (CE) being quantitatively determined for comparation. Different FCDI parameters, including carbon concentration and flow rate of the flow electrode and cell voltage, were investigated to examine the influences on the desalination. Long-term operation of the carbon-based FCDI was evaluated using the optimal results found in the conditions of 1.5 M concentration, 1.5 V cell voltage, and 20 mL min⁻¹ flow rate of electrode and water streams. The results showed an ASRR of 63.7 µg cm⁻² min⁻¹, EC of 162 kJ mol⁻¹, and CE of 89.3%. The research findings validate a good efficiency of this new carbon-based FCDI technology in continuous water desalination and suggest its good potential for real, long-term application.

1. Introduction

Water is a basic need and crucial resource for the survival and growth of life. The increasing requests for freshwater resources create an urgent need to develop ways and solutions for alternative and healthy water resources to enhance this situation [1,2]. Capacitive deionization (CDI) is one of the techniques that was developed to help overcome this problem. It utilizes the electrosorption mechanism to remove ions and charged particles between two carbon-based electrodes sandwiching a saline water stream with a cell voltage applied [3]. This method gained increasing interest owing to its environmental friendliness, low energy consumption, and simple cell structural design [3,4]. In addition, CDI does not require high pressure or temperature to operate, unlike many other methods that are being used for water desalination [3,4,5]. Over the years, different techniques based on the CDI mechanism have been developed to achieve better performance. One of the developed techniques is membrane CDI (MCDI), which uses an ion exchange membrane around the water stream to allow specified ions to pass through, resulting i ion adsorption by the oppositely charged electrodes to enhance the salt removal efficiency [4,6,7,8]. Hybrid CDI (HCDI) is another example of such a technique that uses faradaic material electrodes to increase the salt removal capacity [9,10,11,12,13].

Flow-electrode CDI (FCDI) is an advanced technique that allows an electrode solution or slurry to flow around the water stream. It became popular based on its outstanding advantages, including excellent ion-adsorption capacity, and allowing for continuous desalination operation [4,14]. Flow-through CDI is a common FCDI configuration [15,16]. In comparison, flow-by CDI shows a better long-term stability, attributed to a lower diffusion rate of dissolved oxygen [17]. There are different operation modes for the flow-by CDI, such as, isolated closed-cycle (ICC) mode, short-circuited closed-cycle (SCC) mode, and open cycle (OC) mode. Previous comparisons have been studied regarding these modes, with SCC being found to be the most effective mode [18,19,20]. Different electrode materials were used in the FCDI system, such as sulfonated nickel phthalocyanine redox [21], nanoporous carbon polyhedron [22], mixed vanadium redox couples and carbon nanotubes [23], mixed electrode solvent of water and ethanol [24], and other materials [25,26,27,28,29]. In general, carbon-based materials, such as activated carbon, carbon nanotubes, and carbon black, are most commonly used due to their decent adsorption capacity, large specific surface, good stability, high bulk conductivity, low cost and environmental friendliness. However, their reported performances in different literature studies were often not consistent with each other, likely caused by inconsistent FCDI parameters and testing conditions [30,31,32,33,34,35,36,37].

In this work, different carbon-based materials, including BLACK PEARLS^® 200 carbon black (CB BP), VULCAN^® XC-72 carbon black (CB V), carbon nanotubes (CNT), mesoporous carbon (MC), and carbon nanofibers (CNF), were systematically investigated for a detailed, comparative study to examine the influences of carbon-based materials on FCDI performance. The effects of FCDI operation parameters, including carbon concentration and flow rate of the flow electrode and cell voltage, on the desalination performance, including average salt removal rate (ASRR), salt removal efficiency (SRE), energy consumption (EC), and charge efficiency (CE) properties, were investigated. Long-term operation of carbon-based FCDI was evaluated by desalinating saline water all the way to a drinkable level. The results show promising features of the carbon-based FCDI technology and its good potential for practical, long-term desalination applications.

2. Experimental Section

2.1. Materials

BLACK PEARLS^® 2000 carbon black (CB BP) was purchased from Cabot. VULCAN^® XC-72 carbon black (CB V) was purchased from Cabot. Multi-walled carbon nanotubes (CNT), mesoporous carbon (MC), and carbon nanofibers (CNF) were purchased from Aldrich. The size and morphology of these carbon-based materials were characterized using transmission electron microscopy (TEM) on a JEOL JEM-1230 microscope prior to use. Sodium chloride (NaCl) was purchased from Merck kGaA. Anion exchange membrane (AEM, Fumasep FAS-PET-130) and cation exchange membrane (CEM, Fumasep FKS-PET-130) were from the online Fuel Cell Store.

2.2. Preparation of Carbon-Based Flow Electrode

The flow electrodes were prepared by dispersing 0.5 M (0.6 wt.%) corresponding carbon material and 0.1 M NaCl in DI water (20 mL), unless otherwise specified. The resultant slurry was sonicated for 30 min to help stabilize carbon particles dispersion and form a uniform ink. The ink was further stirred overnight at room temperature using a magnetic stirrer to improve the slurry stability for longer than the duration of the experimental test.

2.3. FCDI Configuration and Desalination Mechanism

The schematic, mechanism, and configuration of the FCDI cell in this study are displayed in Figure 1a,b. The FCDI cell comprises the following components in sequence: cathode (graphite plate), chamber 1, CEM, chamber 2, AEM, chamber 3, CEM, chamber 4, and anode (graphite plate). The carbon flow electrode is circulated between chambers 1 and 4 with a peristaltic pump (Golander, BT101L Pump). Chambers 2 and 3 are separated by AEM, with each having its own private circulation saline water stream around the chamber and the stream having an initial 50 mM NaCl concentration and a 15 mL volume. The desalination mechanism of the FCDI cell is that when the carbon particles flowing in chamber 1 touch the negatively charged graphite plate (cathode), they become negatively charged that electrostatically drive Na⁺ ions migration from chamber 2 through CEM to chamber 1 and adsorb the migrated Na⁺ ions. These negatively charged carbon particles with adsorbed Na⁺ become discharged and desorb Na⁺ when they flow to chamber 4 and touch the positively charged graphite plate (anode). This generates excessive positive charges that repel the Na⁺ ions away from chamber 4 through CEM to chamber 3. As the Na⁺ ions are transferred from chamber 2 to chamber 1 and from chamber 4 to chamber 3; chamber 2 becomes imbalanced with excess Cl^- ions and chamber 3 becomes imbalances with excess Na⁺ ions. This charge imbalance thus drives migration of Cl^- ions from chamber 2 through AEM to chamber 3 to neutralize both streams. As such, the stream in chamber 2 becomes desalinated and chamber 3 becomes salinized. The salt concentration of the stream in chamber 2 was determined by measuring the conductivity every 10 s during the entire experiment using a conductivity meter. One big advantage of using 2 chambers of water streams instead of only 1 chamber is that the 2 water chamber design allows for a continuous desalination process, with no need to stop and reverse the process to remove the saturated salt ions in the flow electrode as in the 1 water chamber design, because the salt ions can be continuously concentrated and removed out of the system through chamber 3.

2.4. Electrochemical Measurements

The electrochemical measurements of assembled FCDI cells were conducted using CHI 760D electrochemical workstation (CH Instruments, Inc., Bee Cave, TX, USA). Cyclic voltammograms (CV) were obtained with a two-electrode mode and a voltage sweeping range of −1.0 to 1.0 V at a scan rate of 100 mV s⁻¹. In a typical desalination experiment, FCDI with a flow electrode consisting of designated type and concentration of carbon-based material was operated at a designated cell voltage. The cell current as a function of time was recorded, together with measuring the conductivity change of the saline water stream in chamber 2 every 10 s. The conductivity measurements were translated to salt concentration data by establishing a concentration–conductivity calibration curve.

2.5. Evaluation of Desalination Performance

The desalination performance of the FCDI cell was evaluated using experimental measurements and the determination of relevant metrics. Important metric parameters, including average salt removal rate (ASRR, µg_NaCl cm⁻² min⁻¹), salt removal efficiency (SRE, %), energy consumption of NaCl removal (EC, kJ mol_NaCl⁻¹), and charge efficiency (CE, %), were calculated using the following equations:

$$\mathrm{ASRR}=\frac{\left(C_o-C_t\right)\times Mw\times V_s}{S\times t}$$

(1)

$$\mathrm{SRE}=\frac{\left(C_o-C_t\right) }{C_o}\times 100\%$$

(2)

where C_o and C_t are the initial and final concentrations of NaCl from chamber 2, Mw is the molecular weight of NaCl, V_s is the saline water stream volume, S is the effective contact area between flow electrode and ion-exchange membrane, and t is the time of operation. SRE for all samples was calculated based on the values in the 30th min of the experiment. The EC is the energy consumption for the process of removing NaCl from the desalinating stream. The pumping energy is not included in this calculation.

$$\mathrm{EC}=\frac{V\times \int Idt}{\left(C_o-C_t\right)\times V_s}$$

(3)

$$\mathrm{CE}=\frac{F\times \left(C_o-C_t\right)\times V_s}{\int Idt}\times 100\%$$

(4)

where V is the applied cell voltage, I is the current, and F is the Faraday constant (96,485 C mol⁻¹).

3. Results and Discussion

3.1. Characterization of Carbon Materials

The commercial CB BP, CB V, CNT, MC, and CNF samples were characterized using TEM for the carbon particles’ size and morphology information (Figure 2). CB BP and CB V show a spheroidal shape of their carbon particles (Figure 2a,b). The average particle size of the CB BP is significantly smaller than that of the CB V sample (about 10 nm versus about 50 nm), suggesting a larger specific surface area. The CNT sample consists of multi-walled carbon nanotubes of about 10 nm in diameter and micrometers in length (Figure 2c). The MC sample consists of micrometer-sized particles that contain meso-sized pores within individual particles (Figure 2d). The CNF sample is made of micro-sized carbon fibers, with average fiber diameter of over 100 nm and length of hundreds of micrometers (Figure 2e). Considering the general trend of the charging/discharging capacity of a material being proportionally correlated with its specific surface area [38], these carbon-based materials would be expected with different desalination performances due to their significant differences in particle size and morphology.

3.2. Effect of Carbon Material on FCDI Performance

A FCDI cell was assembled and tested with different carbon material-based flow electrodes, including CB BP, CB V, CNT, MC, and CNF, to investigate the effect of carbon material on the desalination performance. Figure 3a shows the scanned CV curves with all five samples in the voltage range of −1.0 to 1.0 V. All carbon samples show a quasi-rectangular shape, confirming that there is effective capacitive charge transferring between the carbon flow electrode and graphite current collectors. The FCDI with MC showed the highest current of 8.3 mA at 1.0 V, followed by 7.9, 7.9, 7.1, and 6.1 mA with CB BP, CB V, CNT, and CNF, respectively. Figure 3b shows the measured I-t curves using these prepared flow electrodes, with the carbon concentration and cell voltage being fixed at 0.5 M and 1 V, and flow rate for all streams set at 20 mL min⁻¹ for a fair comparison. For all five samples, the current exhibits a rapid decay within the first few tens of seconds, which can be attributed to the instantaneous charging effect resultant of a sudden voltage change at the beginning. The current then quickly becomes stabilized, suggesting a steady-state operation. The steady current varies with the carbon materials, indicating their different desalination performances. Figure 3c displays the conductivity and concentration changes in the desalinated stream over time. The CB BP-based FCDI cell showed the highest average current value and desalination performance. This could attribute to its high surface area and small particle sizes that create larger dynamic interface contacts with the graphite electrode and allows the ions to transfer in between the particles faster without being blocked, which enhance the charging/discharging process. On the other hand, the performance of the MC-based cell was poor despite the fact that it shows high CV signals. This can be ascribed to the fact that it has poor ion transport through mesopores within individual MC particles that slows the charging/discharging process [39,40]. The desalination performance was quantitatively evaluated by calculating the ASRR and SRE valued (Figure 3d). The CB BP shows the highest ASRR of 21.4 µg cm⁻² min⁻¹ and SRE of 11.3% among all samples, in agreement with the I-t measurements. The CB V and CNT show about the same performance, with ASRR and SRE of around 7.6 µg cm⁻² min⁻¹ and 7.7%, respectively. The MC and CNF show significantly lower values, indicating a poor desalination performance and are thus dropped from further investigation. The energy consumption (EC) and charge efficiency (CE) data of the three top performed samples are displayed in Figure 3e, which exhibits some but not very significant differences. Considering the CB BP exhibiting significantly better ASRR and SRE and comparable EC and CE compared to other carbon materials, it was selected for more detailed investigations to examine the influence of other operation parameters on the desalination performance.

3.3. Effect of Carbon Flow Electrode Concentration on FCDI Performance

To investigate the concentration effect of the CB BP on the FCDI performance, three flow electrodes containing 0.1, 1.0, and 1.5 M (0.12, 1.2, and 1.8 wt.%, respectively) CB BP were prepared. The experiments were run for 30 min, with the cell voltage set at 1.0 V and the flow rate for all streams set at 20 mL min⁻¹. Figure 4a presents the measured I-t curves, from which it can be observed that the 1.5 M CB BP cell resulted in the highest current and the 0.1 M one had the lowest. Figure 4b reflects the current trend, such that the higher the current that the electrode gives, the faster the desalination that results. The conductivity and concentration over time using 1.5 M CB BP showed a significantly deeper drop after the experiment started and kept decreasing over time, with 12.4 mM of NaCl concentration drop in the desalination chamber after 30 min. The 1.0 and 0.1 M cells showed a slower decrease in the NaCl concentration, with concentration drops of 4.9 and 0.6 mM, respectively. This proves that the higher the concentration of the CB BP in the flow electrode, the faster that the FCDI cell desalinates. This is reasonable, as a higher concentration of CB BP particles in the flow electrode would generate more frequent contact between CB BP and graphite electrode, which allows for more change transfer between them, i.e., faster charging/discharging process, and thus faster salt ions electrostatic adsorption/desorption. This eventually drives a faster migration of Cl^- and Na⁺ ions out of the desalination chamber. Figure 4c. displays the calculated ASRR and SRE. The 1.5 M CB BP cell showed the highest ASRR and SRE values of 48.3 µg cm⁻² min⁻¹ and 24.8%, respectively, among the three examined concentrations, confirming it has the fastest desalination performance. The 1.5 M CB BP cell also has the lowest energy consumption and highest charge efficiency as presented in Figure 4d. The charge efficiency for the 1.5 M cell reached 86% while keeping the energy consumption at 112 kJ mol⁻¹. This result is ascribed to the fact that a higher carbon particle concentration in the flow electrode promotes the charge transfer rate, which lowers the required cell overvoltage and thus improves the energy utilization efficiency [35].

3.4. Effect of Cell Voltage on FCDI Performance

In order to determine the effect of cell voltage on the FCDI performance, the desalination experiments were conducted with different applied voltages of 0.5, 1.0, 1.2 and 1.5 V, with the CB BP concentration kept at 0.5 M in the flow electrode and the flow rate set constant at 20 mL min⁻¹. The obtained I-t data are plotted in Figure 5a, in which the 1.5 V run showed the highest current while the current became lower when the applied voltage was reduced. In a same trend, the 1.5 V run showed the biggest NaCl concentration drop of 5.21 mM within 30 min of the experiment (Figure 5b), In comparison, the other three runs at 1.2, 1.0 and 0.5 V showed smaller concentration drops of 4.52, 3.47 and 1.67 mM, respectively. The ASRR and SRE were determined to be 20.3 µg cm⁻² min⁻¹ and 10.4% for the 1.5 V run (Figure 5c), whereas the values were 17.6 µg cm⁻² min⁻¹ and 9.0% for the 1.2 V run, 13.5 µg cm⁻² min⁻¹ and 6.9% for the 1.0 V run and 6.5 µg cm⁻² min⁻¹ and 3.3% for the 0.5 V run. The faster desalination with a higher cell voltage ascribes to a larger overvoltage, which generates a stronger driving force for charge transfer at the carbon particle-graphite plate interface and ion transport from the desalination chamber. However, applying a higher voltage significantly raised the energy consumption as displayed in Figure 5d, where it increased from 62.8 kJ mol⁻¹ at 0.5 V to 170.5 kJ mol⁻¹ at 1.5 V. The charge efficiency increased from 76.9% at 0.5 V to 91% for both applied voltages of 1.0 and 1.2 V. However, it decreased slightly to 84.9% when the applied voltage increased to 1.5 V. It was reported in the literature studies that exceeding 1.2 V cell voltage would lead to unwanted faradaic reactions, such as water electrolysis and carbon electrode oxidation, which cause a decline in the charge efficiency [8,36].

3.5. Effect of Carbon Electrode Flow Rate on FCDI Performance

Another factor that was investigated in this study is the effect of flow rate of the flow electrode on FCDI performance. The desalination experiments were performed by varying the flow electrode flow rate to be 5, 15, 20 mL min⁻¹ while keeping the flow rate of water streams at 20 mL min⁻¹, maintaining a constant CB BP concentration in the flow electrode of 0.5 M and a constant cell voltage of 1.0 V. Figure 6a,b show that the 5 mL min⁻¹ run resulted in a lower cell current and a smaller concentration drop of a mere 0.3 mM in the desalination chamber after 30 min of operation. In comparison, the concentration drops increased to about 2.2 mM at 15 mL min⁻¹ and 3.5 mM at 20 mL min⁻¹. Correspondingly, the ASRR of 13.5 µg cm⁻² min⁻¹ and SRE of 6.9% were the highest at 20 mL min⁻¹ (Figure 6c), while the ASRR were 8.7 and 1.2 µg cm⁻² min⁻¹ at 15 and 5 mL min⁻¹, respectively. Figure 6d displays the energy consumption and charge efficiency for the three runs. The energy consumption decreased from 149.4 kJ mol⁻¹ at 5 mL min⁻¹ to 105.5 kJ mol⁻¹ at 20 mL min⁻¹, showing that the higher the flow electrode flow rate, the less energy the system consumes. Meanwhile, the charge efficiency increased with the flow rate, with 91.4% charge efficiency at 20 mL min⁻¹. These results indicate that increasing the flow rate of the flow electrode helps to improve the ASRR, SRE and charge efficiency and to lower the energy consumption. This can be attribute to the fact that a higher flow rate of the flow electrode accelerates the contact frequency of the carbon particles with the graphite electrode, which promotes the charging–discharging process of the flow electrode that results in faster and more efficient desalination.

3.6. Long Run FCDI to Reach Drinkable NaCl Level

To evaluate the applicability of the FCDI technology, a long run was performed to desalinate 50 mM NaCl saline water all the way to a drinkable level, i.e., <4.6 mM (or <270 mg L⁻¹) [41]. The test was run with a flow electrode containing 1.5 M (1.8 wt.%) CB BP, a cell voltage of 1.5 V, and 20 mL min⁻¹ flow rate of the flow electrode and water streams. Figure 7a shows the current change as a function of time. It can be observed that the current dropped quickly in the first 10 s of the test, after which it started to decline gradually and did not reach a steady state. This can be attributed to the fact that the CB BP particles gradually agglomerated, particularly at 1.5 M high concentration and over a long period, which destabilized the flow electrode and caused a decay in the cell current. Nevertheless, the salt concentration in the desalination chamber kept decreasing and that in the salination chamber kept increasing in a near symmetric pattern (Figure 7b), confirming that the salt ions were essentially transferred from the desalination chamber to the salination chamber during the FCDI cell operation. The drinkable level of water was reached successfully after about 5000 s (<1.5 h), with the overall ASRR and SRE determined to be 63.7 µg cm⁻² min⁻¹ and 42.7% and the energy consumption and charge efficiency calculated to be 162 kJ mol⁻¹ and 89.3%, respectively. These results demonstrated the feasibility of a carbon-based FCDI cell for water desalination to produce drinkable water and a good potential of this technology for long-term, practical application.

4. Conclusions

This work reports the study of a carbon-based flow-electrode capacitive deionization (FCDI) cell consisting of one desalination chamber and one salination chamber for continuous desalination application and the desalination performance in SCC mode. Different carbon materials, including CB BP, CB V, CNT, MC, and CNF samples, were investigated for use as flow electrodes and were compared to understand their characteristics and performances in the FCDI. The CB BP sample showed the best desalination performance among all five examined samples, which was attributed to its large surface area and small particle sizes that allowed for more charge transfer at the carbon particle–graphite electrode interface and for more efficient ion transfer between the carbon particles to facilitate the charging/discharging process. Effects of carbon electrode concentration, cell voltage, and flow rate of the flow electrode on the FCDI performance were investigated, which showed different influences on the desalination performances. An increase both in the carbon electrode concentration and in the flow rate of the flow electrode led to increases in the ASRR, SRE, and CE and a decrease in the EC, which was attributed to more frequent and more efficient charge transfer between the carbon particles and current collectors. A larger cell voltage resulted in an increase in the ASRR and SRE, benefiting from a larger overvoltage driving force, at a cost of a higher EC. Long-term FCDI operation was demonstrated for successfully desalinating 50 mM saline water all the way to the drinkable level, with ASRR of 63.7 µg cm⁻² min⁻¹, SRE of 42.7%, EC of 162 kJ mol⁻¹, and CE of 89.3% being determined. The research findings confirm the feasibility of carbon-based FCDI for efficient water purification and show a good potential of this technology for long-term, practical desalination applications.