The MCDI stack comprised 4 parallel cells that were sandwiched in an acrylic housing.
Two types of electrode, a film electrode casted with activated carbon particles, PACMM (PACMM 203, Material Methods, Irvine, CA, USA, δe1=290 µm) and an activated carbon cloth FM10K (Zorflex®, Pittsburg, PA, USA, δe2=350 µm), were used in this study. Each cell consists of a pair of graphite foil as current collectors (Alfa Aesar, thickness δ=130 µm), a pair of porous electrodes (PACMM or FM 10K), a pair of anion- and cation-exchange membranes (Neosepta AMX, δmem=140 µm, and Neosepta CMX, δmem=170 µm, Tokuyama Co., Japan), and a glass fiber spacer
52
(Whatman, δ=250 µm). Each electrode was cut to a 6×6 cm2 square with a 1.5×1.5 cm2 square hole at the center, yielding a total active electrode area of 270 cm2 for the entire stack. A peristaltic pump drives the water to enter from the periphery of the square MCDI stack, flow along the spacers sandwiched between the ion exchange membrane/electrode assemblies, and exit from the center square hole. A schematic showing the MCDI structure and flow direction is provided in the Supplementary Information. The total mass of the 4 pairs of electrodes were 2.56 g and 4.29 g for PACMM and FM 10K, respectively.
4.2.2. Experimental methods and design.
NaCl solutions were used as the model feed solution throughout the study. The feed solution was stored in a 10L feed tank with constant nitrogen purging to minimize the oxygen content for mitigating electrode oxidation. The feed solution was pumped through the MCDI stack by a peristaltic pump and the effluent of MCDI stack was sent back to the feed tank. The conductivity of the effluent was measured by a flow-through conductivity meter (isoPod EP357, eDAQ, Australia) installed right at the exit of the stack, which was further converted into salt concentration according to a pre-established calibration. Although pH fluctuation has been observed in previous CDI experiments due to possible oxidation/reduction of the carbon electrodes140, 141, the measured concentrations of H+or OH- in the effluent is orders of magnitude lower than the NaCl concentration used in our experiments. It is therefore reasonable to assume that the removal of NaCl by a pair of electrodes is asymmetric, i.e., equal amount of Na+ and Cl- ions are removed in the CDI system. Desalination performance was evaluated with constant current charging and zero-voltage discharge as controlled using a potentiostat (SP 150, Bio-Logic, France) that recorded the real-time current and cell voltage.
To obtain a kinetics-energetics trade-off curve, we conducted MCDI experiments with different current densities and evaluated the ASAR (mg g min-1), which represents the kinetic efficiency, and SEC (J mg-1), which represents the energy efficiency. Previous studies using different current densities adopted an operation protocol that used a constant flow rate and terminated the charging when the cell voltage reached a pre-determined value 11, 139. However, such a protocol led to different effluent salinities and thus achieved different underlying
53
separations. In this study, we choose to use a different approach by adjusting the flow rate to the various charging currents to maintain the same effluent salinity, and thereby achieving the same underlying separation with different current densities. The flow rates leading to the target effluent salinity were obtained using trial-and-error method. The impacts of feed concentration (𝑐0), diluted water concentration (𝑐𝐷), volume of diluted water(𝑉𝐷), i.e. volume of the effluent in the charge stage, and electrode materials on the kinetics-energetics tradeoff were systematically evaluated, with detailed conditions for different sets of experiments summarized in Table 4.1.
4.2.3. Data analysis
The kinetic efficiency of MCDI operations were quantified using ASAR defined as the mass of salt removed per gram of electrode per time. We only evaluated the ASAR for the charging half- cycle following convention, even though ASAR can also be defined based on the full charging and discharge cycle. ASAR (mg g min-1) is calculated according to the following equation:
𝐴𝑆𝐴𝑅 = 𝑀𝑊𝑁𝑎𝐶𝑙𝑄 ∫ (𝑐0𝑡𝐶 0− 𝑐(𝑡))𝑑𝑡
𝑤𝑒𝑡𝐶 (4.1)
where 𝑀𝑊𝑁𝑎𝐶𝑙 is the molecular weight of NaCl, 𝑐0 is the salt concentration (mM) of the feed (or influent) stream, 𝑐(𝑡) is the effluent salt concentration (mM) and should be constant short after charging starts, 𝑄 is the flowrate (L min-1), 𝑤𝑒 is the total mass of electrodes (g), and 𝑡𝐶 is the charging time (min).
The energetic efficiency of the process is quantified using SEC defined as the energy consumed (J) to remove a unit mass of salt (mg). SEC is calculated using equation 2 from experimental data:
𝑆𝐸𝐶 = 𝐼 ∫ 𝑉(𝑡)𝑑𝑡0𝑡𝐶
𝑀𝑊𝑁𝑎𝐶𝑙𝑄 ∫ (𝑐0𝑡𝐶 0− 𝑐(𝑡))𝑑𝑡 (4.2) where 𝐼 is the applied current (A), and 𝑉(𝑡) is the time dependent cell voltage (V).
54 Table 4.1 Experimental conditions in MCDI tests
Flow rate (mL/min)
Applied Current (mA) 120 100 80 60 40
Current Density (mA/cm2) 0.89 0.74 0.59 0.44 0.30
Series 1: Effluent concentration, 𝑐𝐷 ( 𝑐0 = 20 mM, 𝑉𝐷 = 28.5 mL, PACMM)
10 mM 7 5.6 4.2 3.9 2.7
13 mM 9.5 8.1 6.4 4.8 3.4
16 mM 13.0 11.8 9.5 7.7 4.5
Series 2: feed concentration, 𝑐0 ( 𝑐𝐷 = 13 mM, 𝑉𝐷 = 28.5 mL, PACMM)
16 mM 15.2 12.5 10.9 8.3 6.0
20 mM 9.5 8.1 6.4 4.8 3.4
24.2 mM 6.7 5.5 4.4 3.3 2.2
Series 3: diluted water volume, 𝑉𝐷 ( 𝑐0 = 20 mM, 𝑐𝐷 = 13 mM, PACMM)
19 mL 9.5 7.8 6.4 5.0 3.3
28.5 mL 9.5 8.1 6.4 4.8 3.4
33 mL 9.5 7.8 6.2 5.0 3.3
Series 4: Electrode Material (𝑐0 =20 mM, 𝑐𝐷 = 16 mM, 𝑉𝐷 = 33 mL)
PACMM 13.0 11.8 9.5 7.7 4.5
FM10K 16.8 14.3 11.0 8.3 5.4
In this study, we will quantify the kinetics-energetics tradeoff using the relationship between ASAR and the inverse of SEC, i.e., SEC-1 (unit: mg J-1). The interpretation of SEC-1 is the mass of salt removed per amount of energy spent, which also serves equally well, if not even better, as an intuitive metric for energy efficiency. The charge efficiency, defined as the ratio between the salt removal and the electrical charge transferred between the electrodes, is calculated following equation 4.3:
55 𝛬 = 𝑀𝑊𝑁𝑎𝐶𝑙𝑄∫ (𝑐0𝑡𝐶 0− 𝑐(𝑡))𝑑𝑡
𝐼𝑡𝑐 (4.3)