- Email: [email protected]
Desalination of water with high conductivity using membrane-free electrodeionization
Separation and Purification Technology 128 (2014) 39–44
Contents lists available at ScienceDirect
Separation and Purification Technology journal homepage: www.elsevier.com/locate/seppur
Desalination of water with high conductivity using membrane-free electrodeionization Xiaolan Shen, Tianjun Li, Xiaping Jiang, Xueming Chen ⇑ Environmental Engineering Department, Zhejiang University, 388 Yuhangtang Road, Hangzhou 310058, China
a r t i c l e
i n f o
Article history: Received 6 February 2014 Received in revised form 14 March 2014 Accepted 15 March 2014 Available online 24 March 2014 Keywords: Membrane-free electrodeionization Purification High purity water Electrical regeneration Resins
a b s t r a c t Our previous work has shown that membrane-free electrodeionization (MFEDI) is a promising process for high purity water (HPW) production. However, this technique was only suitable to feed water with low conductivity, usually below 20 ls/cm. To extend the applicable range, type I strong base resin was used together with type II strong base resin in this work, due to the high capacity and easy regeneration of type II strong base resin and the excellent purification performance of type I strong base resin. It was demonstrated that although the conductivity of the feed water increased to 50 lS/cm, effective purification was still achieved, with effluent conductivity being 0.056–0.066 lS/cm only. The average conductivity of the concentrate collected during regeneration was over 380 lS/cm, indicating high regenerating efficiency. The water recovery reached 86% and the power consumption was 1.5 kW h/m3 water. In addition, repetitive experimental results showed that the MFEDI system could work stably. Ó 2014 Elsevier B.V. All rights reserved.
1. Introduction Nowadays, HPW is widely used in semiconductors and pharmaceuticals production, electric power generation, and chemical and biochemical laboratories [1–4]. With the rapid development of relevant industries, the demand for HPW is increasing greatly. Generally, HPW production involves in a series of operations. Among them, the mixed bed ion exchange (MBIE), which works as a polishing step, plays a very important role in ensuring the quality of HPW [5]. In despite of high efficiency [6], MBIE requires periodical regeneration using alkali and acid. This produces concentrated waste stream containing high concentration of strong acid, strong alkali and salt. Furthermore, the operations of resins separation, respective regeneration of anion and cation exchange resins and remix of the resins are laborious [7–11]. An alternative technique to MBIE is electrodeionization (EDI), which is a hybrid system combining ion exchange (IE) and electrodialysis (ED). EDI has many advantages including no need of chemicals to regenerate the ion exchange resins, environmental friendliness and easy operation [12–15]. Nevertheless, EDI needs selective ion exchange membranes, and thus a series of membrane-associated problems, such as concentrating polarity, chemical precipitation, and membrane fouling, may occur [16–19].
⇑ Corresponding author. Tel.:+86 57187951239; fax: +86 57187952771. E-mail address: [email protected] (X. Chen). http://dx.doi.org/10.1016/j.seppur.2014.03.011 1383-5866/Ó 2014 Elsevier B.V. All rights reserved.
In our previous work, a novel MFEDI process was investigated for high purity water production [20]. This process is different from the electrically regenerated ion exchange invented by Davis and Hills [21] that used ion exchange fabrics and had a complex configuration. It is also different from the one reported before by our group [22] that required at least a pair of ion exchange membranes. MFEDI is operated in batch, that is, the purification and electrical regeneration are alternated. The purification process is illustrated below:
R H þ Naþ ! R Na þ Hþ
R OH þ Cl ! R Cl þ OH
ð1Þ ð2Þ
where R–H represents H-type cation exchange resin and R–OH represents OH-type anion exchange resin. After the effluent conductivity is over a specific value, electricity is supplied to restore the exchange capacity of the resin. Hereafter, high concentration of hydrogen ions and hydroxide ions are produced by water dissociation and water electrolysis. These ions would exchange with salt ions in resin phase to regenerate the resin. The electrical regeneration process is described below:
H2 O ! Hþ þ OH
electrically enhanced water dissociation
ð3Þ
2H2 O 4e ! 4Hþ þ O2 " at the anode
ð4Þ
2H2 O þ 2e ! 2OH þ H2 " at the cathode
ð5Þ
40
X. Shen et al. / Separation and Purification Technology 128 (2014) 39–44
OH þ R Cl ! R OH þ Cl
Hþ þ R Na ! R H þ Naþ
ð6Þ
decomposition method. The detailed fabrication procedures can be found elsewhere [24].
ð7Þ
The MFEDI process shows many advantages. Firstly, membranes are not used in MFEDI system and problems associated with membranes can be avoided. Secondly, MFEDI combines the advantages of MBIE and EDI: on the one hand, the conductivity of the MFEDI effluent is close to pure water; on the other hand, ion exchange resins in MFEDI column can be regenerated electrically in situ and the concentrates produced during regeneration can be recycled to RO. Furthermore, the operation and configuration of MFEDI are simpler than MBIE and EDI. However, MFEDI was only suitable to feed water with low conductivity, usually below 20 lS/cm. To extend the applicable range, type I strong base resin was used together with type II strong base resin in this work. As we know, type I strong base resin is superior in purification. Nonetheless, this type of resin is low in capacity and difficult to be regenerated. On contrast, type II strong base resin is high in capacity and easy in regeneration, but it is poor in purification. The combination of the two types of strong base resins allows the full use of their advantages. In the purification stage, the large capacity of type II strong base resin helps to absorb more ions, and the excellent purification performance of type I strong base resin assists to guarantee the HPW quality; in the electrical regeneration stage, the good regeneration property of type II strong base resin facilitates restoring more capacity. The major objectives of this study are to evaluate the performance of the resins selected for producing HPW from the water with high conductivity, to examine the efficiency of the electrical regeneration, and to investigate the resins stability. 2. Materials and methods 2.1. Resins pretreatment Main physiochemical properties of the gel-type strong-base cation exchange resin (Monosphere 550A, Dow Co., Shanghai, China), gel-type strong-base II cation exchange resin (Monosphere 202, Zhengguang Co., Hangzhou, China) and macroporous-type weakacid anion exchange resin (Monosphere D113, Zhengguang Co., Hangzhou, China) are presented in Table 1. Before use, type I strong base resin and type II strong base resin were washed with deionized water until neutral pH, and weak acid resin was first soaked in 4% NaOH for 6 h, then washed with deionized water, regenerated fully with 5% HCl, and finally washed with deionized water again.
2.3. Experimental setup for equilibrium investigation and resistivity measurement The experimental setup for equilibrium investigation and resistivity measurement is shown in Fig. 1. A pair of Ti/Pt electrodes was installed in a column, which had a diameter of 3 cm. The net spacing of the electrodes was 3 cm, and about 30 mL of mixed ion exchange resins including 10 mL of cation exchange resin and 20 mL of anion exchange resin, were packed between the electrodes. A spring was used to compress the resin layer. A reservoir was used to store the circulation solution. The solution temperature was maintained using a water bath. Nitrogen gas was purged into the solution to eliminate the interference of the carbon dioxide. The equilibrium investigation was conducted in a cyclic manner including an adsorbing stage and a releasing stage. Initially, 600 mL of 0.5 M NaCl solution was circulated to the system and no electricity was supplied. After the ion exchange equilibrium reached, that is, the solution conductivity was unchanged any more, DC electricity was supplied to the resin layer at a current density of 280 A/m2 until a new equilibrium reached. The resistivity was measured by examining the variation of voltage with current density. 2.4. MFEDI system The MFEDI system for HPW production is schematically shown in Fig. 2. The inner diameter of the MFEDI column was also 3 cm. The whole column was divided into two parts: the lower part, 70 cm in effective height, was packed with the mixed weak acid and type II strong base resins with a ratio of 1:2; while the upper part, 30 cm in effective height, was packed with the mixed weak acid and type I strong base resins also with a ratio of 1:2. The MFEDI system was operated in a batch mode including a purification stage and a regeneration stage. In the purification stage, no electricity was supplied, and 50 lS/cm NaCl feed solution passed the MFEDI column upward with a flow velocity of 10 m/h; in the regeneration stage, electricity was supplied at a constant voltage of 1400 V, and deionized water passed the MFEDI column downward with a flow velocity of 15 m/h. To improve regeneration efficiency and to neutralize the concentrate, the electrodes were reversed during regeneration. In the first 20 min of the regeneration, the top electrode worked as a cathode and the bottom electrode worked as an anode. In the latter 40 min of the regeneration, the top electrode worked as an anode and the bottom electrode worked as a cathode.
2.2. Electrodes preparation 2.5. Analysis Mesh electrodes used in MFEDI system were prepared as following: H2PtCl6H2O (99%, Adamas Reagent Co., Shanghai, China) was dissolved in isopropanol (99.5%, Adamas Reagent Co., Shanghai, China) with a concentration of 0.5 M [23]. Titanium meshes, 3 cm in diameter, with an effective area of 7.1 cm2, were used as substrates. The electrodes were fabricated using a thermal
Conductivity was measured using a conductivity meter (Orion 3 Star, Thermo Scientific, Singapore, for MFEDI effluent, resolution: 0.001 lS/cm; and Sension 5, Hach, OH, for the others, resolution: 0.01 lS/cm). pH values and Na+ concentrations were measured using a pH/ISE meter (Orion Dual StarTM, Thermo Scientific, Singa-
Table 1 Properties of ion exchange resins.a
a
Designation
Type
Matrix structure
Functional group
Capacity (mol/L)
D113 550A 202
Weak-acid Type I strong-base Type II strong-base
Polyacrylic acid Polystyrene Polystyrene
Carboxyl Quaternary amine Quaternary amine
4.3 1.4 1.3
Obtained from manufacturers.
41
X. Shen et al. / Separation and Purification Technology 128 (2014) 39–44
by the mixed weak acid and type I strong base resins, indicating better purification performance of the weak acid and type I strong base resins. It can also be seen that, the solution conductivity of mixed weak acid and type II strong base resins increased sharply after DC power was supplied, and finally reached 384 lS/cm, while that of the mixed weak acid and type I strong base resins increased gradually, and finally reached 244 lS/cm. Additionally, it took only 30 min for the mixed weak acid and type II strong base resins to reach a steady state, while the time needed for the mixed weak acid and type I strong base resins was as long as 90 min. Both the faster speed of ions release and the higher conductivity in the steady state suggested better regeneration performance of the mixed weak acid and type II strong base resins. The results above indicate that a combination of two types of strong base resins may give a boost on performance. The mixed weak acid and type II strong base resins played the role of pretreatment of feed water, due to its high regeneration efficiency; while the mixed weak acid and type I strong base resins played the role of further purification, attributing to its high purification efficiency.
Fig. 1. Experimental set-up for equilibrium investigation.
3.2. Purification performance of new resins
(b) Regeneration period Fig. 2. MFEDI system.
pore, resolution: 0.01 for pH and 0.01 mg/L for Na+). Chlorine was measured using the iodometric titration method [25].
3. Results and discussion 3.1. Main property of different mixed resins Fig. 3 shows the conductivity variation of the circulating solution with time after DC electricity was supplied to different mixed resins. Before the electricity was supplied, the solution had been circulated for a sufficiently long time, usually about one and half an hour, and therefore the initial conductivity was essentially the equilibrium conductivity of the solution in cases without electricity imposed. It was found that the equilibrium solution conductivity of the mixed weak acid and type I strong base resins was 2.41 lS/cm, whereas that of the mixed weak acid and type II strong base resins was 10.3 lS/cm, which was almost 3.3 times higher than that of the former. In other words, more ions were adsorbed
The purification performance of the new resins that were fully regenerated with chemicals was investigated first, and the results are shown in Fig. 4. It can be seen that, with an influent conductivity of 50 lS/cm, the effluent conductivity of the lower layer was lower than 2 lS/cm within the initial 29, while that of the upper layer kept below 0.056 lS/cm. The effluent conductivity of the lower layer then increased gradually to 13.2 lS/cm at 45 h and that of the upper layer still kept below 0.056 lS/cm. Finally, the effluent conductivity of the lower layer rose promptly and reached 47.5 lS/ cm at 59 h, whereas that of the upper layer increased gradually and reached 0.066 lS/cm at the same time. In the whole purification process, the ions absorbed by the lower layer amounted to be 81%, and those removed by the upper layer were calculated to be 19%. The experimental results above further verified our original prediction, that is, the lower mixed weak acid and type II strong base resins layer would play a predominant role in removing saline ions from the feed water whereas the upper mixed weak acid and type I strong base resins layer would be responsible to guarantee the effluent quality. 3.3. Electro-regeneration performance 3.3.1. Concentrating efficiency After adsorption, the resins were regenerated electrically at a constant voltage of 1400 V. Fig. 5 shows the variation of the con-
Effluent conductivity of L (µ s/cm)
400
Conductivity (µ s/cm)
320
240
160
D113:550A=1:2 D113:202=1:2
80
0
0
20
40
60
80
100
120
Time (min) Fig. 3. Equilibrium conductivity variations of different mixed resins at 25 °C.
50
0.075
40
0.070
Effluent of lower layer Effluent of upper layer
30
0.065
20
0.060
10
0.055
0
0
10
20
30
40
50
60
0.050
Time (h) Fig. 4. Purification performance of the new resins at 25 °C.
Effluent conductivity of U (µ s/cm)
(a) Purification period
42
X. Shen et al. / Separation and Purification Technology 128 (2014) 39–44
cathode was on the top indicate a constant flux of Na+. As the anode and cathode had different effect on the regeneration of weak acid resin, the reversing of the electrode polarity gave a temporary change in the variations of Na+ concentration. Then, the Na+ concentration decreased at a constant rate too. The Na+ concentration of the mixing concentrate was 75.6 mg/L and the total amount of the NaCl exchanged during 1 h electrical regeneration was 2.048 g.
700
Conductivity (µ s/cm)
600
Concentrate of lower layer Concentrate of upper layer
500 400 300 200 100 0
0
10
20
30
40
50
60
70
Time (min) Fig. 5. Conductivity variations of different resin layers during regeneration at 25 °C.
centrate conductivity with regeneration time. It can be seen from the figure that the concentrate conductivity increased fast initially, and reached a maximum value of 172 lS/cm for the upper layer, and 624 lS/cm for the lower layer at 5 min. The fast increase in the concentrate conductivity reveals the fast ions release from the resin phase to the solution phase. The conductivity then reduced gradually due to a decrease in the concentration of the ions adsorbed in resins and thereby a decrease in the speed of ions release. There was a sudden drop in the conductivity when the electrode polarity was reversed at 20 min. The reason for this phenomenon is that the concentrate was temporarily neutralized after reversing, and the conductivity of the salt solution was lower than that of the acid solution at the same molar concentration. Thereafter, the conductivity increased gradually due to the increase in pH value. It was then reduced again as a result of the decreased speed of ions release. By mixing all the concentrate solution of the lower layer, the conductivity of the concentrate was over 380 lS/cm. It can also be found that, the concentrate conductivity value of the lower resin was about 2.4 times higher than that of the upper resin. This further demonstrates the higher regeneration efficiency of the lower mixed weak acid resin and type II strong base resin. Fig. 6 shows the variation of Na+ concentration of the concentrate with regeneration time. The Na+ concentration in the concentrate followed a predictable trend over the experiment. The linear variations of the Na+ concentration in the first 20 min when the
3.3.2. Current variation Fig. 7 presents the current density variation during regeneration at a constant voltage of 1400 V. It was found that the current density increased gradually at the beginning and achieved a maximum value of 137 A/m2 at 5 min. This should be attributed to an increase in resins conductivity owing to the changes of the resin forms and an increase in the solution conductivity. The current density increase also occurred in the initial 5 min after the electrodes were reversed. Besides, the tendency was that the current density reduced with the regeneration time due to the change of the resin resistance and the decrease of the solution conductivity. To investigate resin resistivity of different resin forms, the correlation of voltage with current density for different resin forms was examined, and the results are shown in Fig. 8. It can be calculated that the resistivity for sodium form of weak acid resin, hydrogen form of weak acid resin, chloride form of type II strong base resin, hydroxide form of type II strong base resin, chloride form of type I strong base resin and hydroxide form of type I strong base resin were 3.0 Xm, 18.7 Xm, 1.9 Xm, 2.3 Xm, 2.0 Xm and 0.8 Xm, respectively. The resistivity difference of different resin forms can be explained by the different binding force between the ions adsorbed and the fixed sites of the resin. The greater the binding force is, the harder the ions can be dissociated from the fixed sites, and thus the larger the resistivity is. Chloride ion has a greater binding force with a fixed site of type I strong base resin than hydroxide ion. As a result, chloride form type I strong base resin has a larger resistivity than hydroxide form of type I strong base resin. As the upper layer packed with weak acid and type I strong base resins was only 30 cm, the general resistivity was determined by weak acid and type II strong base resins. With the proceeding of the electro-regeneration, sodium form of weak acid resin and chloride form of type II strong base resin changed into hydrogen form and hydroxide form. Since the hydrogen form weak acid resin has much higher resistivity than the sodium form weak acid resin, as a whole, the resistivity of the resin layers rose gradually. This explains why the current density tended to reduce during regeneration.
150
Concentrate of lower layer Concentrate of upper layer
100
125 2
Current density (A/m )
Concentration of Na+ (ppm)
125
75
50
25
0
100 75 50 25
0
10
20
30
40
50
60
70
Time (min) Fig. 6. Na+ concentration variations of different resin layers during regeneration at 25 °C.
0
0
10
20
30
40
50
60
70
Time (min) Fig. 7. Current density variation during regeneration at a constant voltage of 1400 V at 25 °C.
43
X. Shen et al. / Separation and Purification Technology 128 (2014) 39–44
300
14
(A) D113-Na D113-H
200
Concentrate of lower layer Concentrate of upper layer
12 10 8
pH
Voltage (V)
250
150
6 100 4 50 0
2
0
100
200
300
400
2
10
20
30
40
50
60
70
Fig. 9. pH variations of different resin layers during regeneration at 25 °C.
30
(B)
25
Voltage (V)
0
Time (min)
Current density (A/m )
550A-OH 550A-Cl
20 15 10 5 0
0
0
100
200
300
400
ions produced by electrolysis in the anode would drain away with the concentrate before they regenerated the cation exchange resin. As a result, the regeneration of anion exchange resin was more available. However, the concentrate of the upper layer kept alkali in the first 20 min. This was because weak acid resin was much more easily to be regenerated than type I strong base resin regardless of the polarity. When the polarity was reversed, the pH value of the concentrate turned into alkali due to the favorable regeneration of cation exchange resin. The pH value of the whole concentrate was 6.9, indicating it can be returned to a pretreatment unit, such as RO, for recovery without the need of neutralization in real applications.
2
Current density (A/m )
3.3.4. Chlorine analysis During regeneration, hydrogen gas was produced at the cathode, while oxygen and chlorine gases were generated at the anode. Since the generated chlorine in the gaseous phase was too low to detect, the concentration of chlorine in the concentrate was analyzed. By mixing all the concentrate collected, the average concentration of the chlorine, which is a good disinfectant for use in RO, was measured to be 2.1 mg/L.
40
(C) Voltage (V)
30
202-OH 202-Cl
20
3.4. Purification performance of the resins after electro-regeneration 10
0
0
100
200
300
400
2
Current density (A/m ) Fig. 8. Voltage–current correlations of resins in different forms: (A) D113, (B) 550A, and (C) 202.
3.3.3. pH Variation Fig. 9 shows the pH variation of the concentrate with regeneration time. It was found that the concentrate of the lower layer was acidic during the first 20 min when the cathode was on the top of the resin layers. This was because such polarity was favorable for the electro-regeneration of anion exchange resin. The hydrogen and hydroxide ions for regeneration can be from electrolysis in the electrodes or electrically enhanced water dissociation. When the cathode was on the top, anion exchange resin can be regenerated by hydroxide ions either from electrolysis or from water dissociation, whereas cation exchange resin can be regenerated by hydrogen ions from electrolysis only. This was because hydrogen
To test the performance of the double layer MFEDI after electroregeneration, a purification experiment of the resins after regenerated electrically was conducted under the same condition as the new resins, and the results are demonstrated in Fig. 10. It can be observed that the effluent conductivity of the lower layer rose gradually and reached 42 lS/cm at 11 h, while that of the upper layer kept constant at 0.056 lS/cm in the initial 8 h and then increased slightly to 0.066 lS/cm in the following 3 h, indicating that the removal performance of the MFEDI system after regeneration was as good as before. The total amount of NaCl removed by ion exchange resins was 2.055 g which could be calculated by flow rate, removal time and concentration of feed water. This value is in good agreement with that of NaCl released from the resins during the regeneration, implying that the capacity of ion exchange resin was unchanged after regeneration. In addition, the water recovery rate was calculated to be 86%. A repetitive experiment of 10 times was carried out on the hybrid ion exchange resins to evaluate the stability of the MFEDI process for production of high-purity water. As shown in Fig. 10, the effluent conductivity of the upper and lower layer kept very stable after 10 times of experiments. Therefore, it is demonstrated that the purification performance of the MFEDI process was stable with
44
X. Shen et al. / Separation and Purification Technology 128 (2014) 39–44
Effluent conductivity of L (µ s/cm)
40
0.070
Effluent of Lower layer Effluent of upper layer
30
0.065
20
0.060
10
0.055
0
0
2
4
6
8
10
0.050 12
Effluent conductivity of U (µ s/cm)
0.075
50
Time (h) Fig. 10. Purification performance of the resins after electrical regeneration at 25 °C.
a cycle time of 12 h, including 11 h of high-purity water production and 1 h of electrical regeneration. 3.5. Energy consumption The energy consumption can be calculated according to the following formula.
E¼
IUt1 ðQ 2 t 2 Q 1 t1 Þ
ð8Þ
anywhere E is the energy consumption, kW h/m3 water; I is the average current during the whole regeneration period, A; U is the voltage, V; Q2 is the flow rate during purification period, m3/h; t2 is the purification time, h; Q1 is the flow rate during regeneration period, m3/h; t1 is the regeneration time, h. The energy consumption was calculated to be 1.5 kW h/m3 water. It was acceptable considering the high concentration of influent. 4. Conclusions By combining type I strong base resin and type II strong base resin, MFEDI were used to produce HPW successfully from the feed water with an conductivity as high as 50 lS/cm. The effluent conductivity was 0.056–0.066 lS/cm, indicating good purification performance. The conductivity of the concentrate collected during electrical regeneration was over 382 lS/cm, suggesting high electrical regeneration efficiency. The concentrate was neutral and therefore could be recycled to RO directly. The energy consumption and the water recovery were 1.5 kW h/m3 water and 86%, respectively. References [1] H.D. Willauer, F.D. Mascio, D.R. Hardy, M.K. Lewis, F.W. Williams, Development of an electrochemical acidification cell for the recovery of CO2 and H2 from seawater, Ind. Eng. Chem. Res. 50 (2011) 9876–9882.
[2] A. Grabowski, G.Q. Zhang, H. Strathmann, G. Eigenberger, Production of highpurity water by continuous electrodeionization with bipolar membranes: influence of concentrate and protection compartment, Sep. Purif. Technol. 60 (2008) 86–95. [3] J. Lu, Y.X. Wang, Y.Y. Lu, G.L. Wang, L. Kong, J. Zhu, Numerical simulation of the electrodeionization (EDI) process for producing ultrapure water, Electrochim. Acta. 55 (2010) 7188–7198. [4] J. Lu, Y.X. Wang, J. Zhu, Numerical simulation of the electrodeionization (EDI) process accounting for water dissociation, Electrochim. Acta. 55 (2010) 2673– 2686. [5] J.H. Song, K.H. Yeon, J. Cho, S.H. Moon, Effects of the operating parameters on the reverse osmosis-electrodeionization performance in the production of high purity water, Korean J. Chem. Eng. 22 (1) (2005) 108–114. [6] T.H. Boyer, P.C. Singer, Bench-scale testing of a magnetic ion exchange resin for removal of disinfection by-product precursors, Water Res. 39 (2005) 1265– 1276. [7] Y.Q. Xing, X.M. Chen, P.D. Yao, D.H. Wang, Continuous electrodeionization for removal and recovery of Cr(VI) from wastewater, Sep. Purif. Technol. 67 (2009) 123–126. [8] A. Mahamoud, A.F.A. Hoadley, An evaluation of a hybrid ion exchange electrodialysis process in the recovery of heavy metals from simulated dilute industrial wastewater, Water Res. 46 (2012) 3364–3376. [9] H.X. Lu, J.Y. Wang, S.F. Bu, L. Fu, Influence of resin particle size distribution on the performance of electrodeionization process for Ni2+ removal from synthetic wastewater, Sep. Sci. Technol. 46 (2011) 404–408. [10] L. Fu, J.Y. Wang, Y.L. Su, Removal of low concentrations of hardness ions from aqueous solutions using electrodeionization process, Sep. Purif. Technol. 68 (2009) 390–396. [11] K.H. Yeon, J.H. Song, S.H. Moon, A study on stack configuration of continuous electrodeionization for removal of heavy metal ions from the primary coolant of a nuclear power plant, Water Res. 38 (2004) 1911–1921. [12] Ö. Arar, Ü. Yüksel, N. Kabay, Mithat Yüksel, Removal of Cu2+ ions by a microflow electrodeionization (EDI) system, Desalination 277 (2011) 296–300. [13] A. Grabowski, G.Q. Zhang, H. Strathmann, G. Eigenberger, The production of high purity water by continuous electrodeionization with bipolar membranes: influence of the anion-exchange, J. Membr. Sci. 281 (2006) 297–306. [14] K.H. Yeon, J.H. Seong, S. Rengaraj, S.H. Moon, Electrochemical characterization of ion-exchange resin beds and removal of cobalt by electrodeionization for high purity water production, Sep. Sci. Technol. 38 (2) (2003) 443–462. [15] J.Y. Wang, S.C. Wang, M.R. Jin, A study of the electrodeionization process: highpurity water production with a RO/EDI system, Desalination 132 (2000) 349– 352. [16] J.H. Song, K.H. Yeon, S.H. Moon, Effect of current density on ionic transport and water dissociation phenomena in a continuous electrodeionization (CEDI), J. Membr. Sci. 291 (2007) 165–171. [17] J.S. Park, J.H. Song, K.H. Yeon, S.H. Moon, Removal of hardness ions from tap water using electromembrane processes, Desalination 202 (2007) 1–8. [18] J.W. Lee, K.H. Yeon, J.H. Song, S.H. Moon, Characterization of electroregeneration and determination of optimal current density in continuous electrodeionization, Desalination 207 (2007) 276–285. [19] J. Wood, J. Gifford, J. Arba, M. Shaw, Production of ultrapure water by continuous electrodeionization, Desalination 250 (2010) 973–976. [20] W.Q. Su, R.Y. Pan, X.M. Chen, Membrane-free electrodeionization for high purity water production, Desalination 15 (2013) 29–86. [21] T.A. Davis, V. Hills, Electrically regenerated ion exchange system, US Patent: 4,032,452, June 1977. [22] Y.Q. Xing, X.M. Chen, D.H. Wang, Electrically regenerated ion exchange for removal and recovery of Cr(VI) from wastewater, Environ. Sci. Technol. 41 (2007) 1439–1443. [23] C. Comninellis, G.P. Vercesi, Characterization of DSA-type oxygen evolving electrodes: choice of a coating, J. Appl. Electrochem. 21 (1991) 335–345. [24] X.M. Chen, G.H. Chen, Investigation of Ti/IrO2–Sb2O5–SnO2 electrodes for O2 evolution: calcination temperature and precursor composition effects, J. Electrochem. Soc. 152 (7) (2005) J59–J64. [25] APHA, Standard Methods for the Examination of Water and Wastewater, 21st ed., American Public Health Association, Port City, Maryland, 2005.
Contents lists available at ScienceDirect
Separation and Purification Technology journal homepage: www.elsevier.com/locate/seppur
Desalination of water with high conductivity using membrane-free electrodeionization Xiaolan Shen, Tianjun Li, Xiaping Jiang, Xueming Chen ⇑ Environmental Engineering Department, Zhejiang University, 388 Yuhangtang Road, Hangzhou 310058, China
a r t i c l e
i n f o
Article history: Received 6 February 2014 Received in revised form 14 March 2014 Accepted 15 March 2014 Available online 24 March 2014 Keywords: Membrane-free electrodeionization Purification High purity water Electrical regeneration Resins
a b s t r a c t Our previous work has shown that membrane-free electrodeionization (MFEDI) is a promising process for high purity water (HPW) production. However, this technique was only suitable to feed water with low conductivity, usually below 20 ls/cm. To extend the applicable range, type I strong base resin was used together with type II strong base resin in this work, due to the high capacity and easy regeneration of type II strong base resin and the excellent purification performance of type I strong base resin. It was demonstrated that although the conductivity of the feed water increased to 50 lS/cm, effective purification was still achieved, with effluent conductivity being 0.056–0.066 lS/cm only. The average conductivity of the concentrate collected during regeneration was over 380 lS/cm, indicating high regenerating efficiency. The water recovery reached 86% and the power consumption was 1.5 kW h/m3 water. In addition, repetitive experimental results showed that the MFEDI system could work stably. Ó 2014 Elsevier B.V. All rights reserved.
1. Introduction Nowadays, HPW is widely used in semiconductors and pharmaceuticals production, electric power generation, and chemical and biochemical laboratories [1–4]. With the rapid development of relevant industries, the demand for HPW is increasing greatly. Generally, HPW production involves in a series of operations. Among them, the mixed bed ion exchange (MBIE), which works as a polishing step, plays a very important role in ensuring the quality of HPW [5]. In despite of high efficiency [6], MBIE requires periodical regeneration using alkali and acid. This produces concentrated waste stream containing high concentration of strong acid, strong alkali and salt. Furthermore, the operations of resins separation, respective regeneration of anion and cation exchange resins and remix of the resins are laborious [7–11]. An alternative technique to MBIE is electrodeionization (EDI), which is a hybrid system combining ion exchange (IE) and electrodialysis (ED). EDI has many advantages including no need of chemicals to regenerate the ion exchange resins, environmental friendliness and easy operation [12–15]. Nevertheless, EDI needs selective ion exchange membranes, and thus a series of membrane-associated problems, such as concentrating polarity, chemical precipitation, and membrane fouling, may occur [16–19].
⇑ Corresponding author. Tel.:+86 57187951239; fax: +86 57187952771. E-mail address: [email protected] (X. Chen). http://dx.doi.org/10.1016/j.seppur.2014.03.011 1383-5866/Ó 2014 Elsevier B.V. All rights reserved.
In our previous work, a novel MFEDI process was investigated for high purity water production [20]. This process is different from the electrically regenerated ion exchange invented by Davis and Hills [21] that used ion exchange fabrics and had a complex configuration. It is also different from the one reported before by our group [22] that required at least a pair of ion exchange membranes. MFEDI is operated in batch, that is, the purification and electrical regeneration are alternated. The purification process is illustrated below:
R H þ Naþ ! R Na þ Hþ
R OH þ Cl ! R Cl þ OH
ð1Þ ð2Þ
where R–H represents H-type cation exchange resin and R–OH represents OH-type anion exchange resin. After the effluent conductivity is over a specific value, electricity is supplied to restore the exchange capacity of the resin. Hereafter, high concentration of hydrogen ions and hydroxide ions are produced by water dissociation and water electrolysis. These ions would exchange with salt ions in resin phase to regenerate the resin. The electrical regeneration process is described below:
H2 O ! Hþ þ OH
electrically enhanced water dissociation
ð3Þ
2H2 O 4e ! 4Hþ þ O2 " at the anode
ð4Þ
2H2 O þ 2e ! 2OH þ H2 " at the cathode
ð5Þ
40
X. Shen et al. / Separation and Purification Technology 128 (2014) 39–44
OH þ R Cl ! R OH þ Cl
Hþ þ R Na ! R H þ Naþ
ð6Þ
decomposition method. The detailed fabrication procedures can be found elsewhere [24].
ð7Þ
The MFEDI process shows many advantages. Firstly, membranes are not used in MFEDI system and problems associated with membranes can be avoided. Secondly, MFEDI combines the advantages of MBIE and EDI: on the one hand, the conductivity of the MFEDI effluent is close to pure water; on the other hand, ion exchange resins in MFEDI column can be regenerated electrically in situ and the concentrates produced during regeneration can be recycled to RO. Furthermore, the operation and configuration of MFEDI are simpler than MBIE and EDI. However, MFEDI was only suitable to feed water with low conductivity, usually below 20 lS/cm. To extend the applicable range, type I strong base resin was used together with type II strong base resin in this work. As we know, type I strong base resin is superior in purification. Nonetheless, this type of resin is low in capacity and difficult to be regenerated. On contrast, type II strong base resin is high in capacity and easy in regeneration, but it is poor in purification. The combination of the two types of strong base resins allows the full use of their advantages. In the purification stage, the large capacity of type II strong base resin helps to absorb more ions, and the excellent purification performance of type I strong base resin assists to guarantee the HPW quality; in the electrical regeneration stage, the good regeneration property of type II strong base resin facilitates restoring more capacity. The major objectives of this study are to evaluate the performance of the resins selected for producing HPW from the water with high conductivity, to examine the efficiency of the electrical regeneration, and to investigate the resins stability. 2. Materials and methods 2.1. Resins pretreatment Main physiochemical properties of the gel-type strong-base cation exchange resin (Monosphere 550A, Dow Co., Shanghai, China), gel-type strong-base II cation exchange resin (Monosphere 202, Zhengguang Co., Hangzhou, China) and macroporous-type weakacid anion exchange resin (Monosphere D113, Zhengguang Co., Hangzhou, China) are presented in Table 1. Before use, type I strong base resin and type II strong base resin were washed with deionized water until neutral pH, and weak acid resin was first soaked in 4% NaOH for 6 h, then washed with deionized water, regenerated fully with 5% HCl, and finally washed with deionized water again.
2.3. Experimental setup for equilibrium investigation and resistivity measurement The experimental setup for equilibrium investigation and resistivity measurement is shown in Fig. 1. A pair of Ti/Pt electrodes was installed in a column, which had a diameter of 3 cm. The net spacing of the electrodes was 3 cm, and about 30 mL of mixed ion exchange resins including 10 mL of cation exchange resin and 20 mL of anion exchange resin, were packed between the electrodes. A spring was used to compress the resin layer. A reservoir was used to store the circulation solution. The solution temperature was maintained using a water bath. Nitrogen gas was purged into the solution to eliminate the interference of the carbon dioxide. The equilibrium investigation was conducted in a cyclic manner including an adsorbing stage and a releasing stage. Initially, 600 mL of 0.5 M NaCl solution was circulated to the system and no electricity was supplied. After the ion exchange equilibrium reached, that is, the solution conductivity was unchanged any more, DC electricity was supplied to the resin layer at a current density of 280 A/m2 until a new equilibrium reached. The resistivity was measured by examining the variation of voltage with current density. 2.4. MFEDI system The MFEDI system for HPW production is schematically shown in Fig. 2. The inner diameter of the MFEDI column was also 3 cm. The whole column was divided into two parts: the lower part, 70 cm in effective height, was packed with the mixed weak acid and type II strong base resins with a ratio of 1:2; while the upper part, 30 cm in effective height, was packed with the mixed weak acid and type I strong base resins also with a ratio of 1:2. The MFEDI system was operated in a batch mode including a purification stage and a regeneration stage. In the purification stage, no electricity was supplied, and 50 lS/cm NaCl feed solution passed the MFEDI column upward with a flow velocity of 10 m/h; in the regeneration stage, electricity was supplied at a constant voltage of 1400 V, and deionized water passed the MFEDI column downward with a flow velocity of 15 m/h. To improve regeneration efficiency and to neutralize the concentrate, the electrodes were reversed during regeneration. In the first 20 min of the regeneration, the top electrode worked as a cathode and the bottom electrode worked as an anode. In the latter 40 min of the regeneration, the top electrode worked as an anode and the bottom electrode worked as a cathode.
2.2. Electrodes preparation 2.5. Analysis Mesh electrodes used in MFEDI system were prepared as following: H2PtCl6H2O (99%, Adamas Reagent Co., Shanghai, China) was dissolved in isopropanol (99.5%, Adamas Reagent Co., Shanghai, China) with a concentration of 0.5 M [23]. Titanium meshes, 3 cm in diameter, with an effective area of 7.1 cm2, were used as substrates. The electrodes were fabricated using a thermal
Conductivity was measured using a conductivity meter (Orion 3 Star, Thermo Scientific, Singapore, for MFEDI effluent, resolution: 0.001 lS/cm; and Sension 5, Hach, OH, for the others, resolution: 0.01 lS/cm). pH values and Na+ concentrations were measured using a pH/ISE meter (Orion Dual StarTM, Thermo Scientific, Singa-
Table 1 Properties of ion exchange resins.a
a
Designation
Type
Matrix structure
Functional group
Capacity (mol/L)
D113 550A 202
Weak-acid Type I strong-base Type II strong-base
Polyacrylic acid Polystyrene Polystyrene
Carboxyl Quaternary amine Quaternary amine
4.3 1.4 1.3
Obtained from manufacturers.
41
X. Shen et al. / Separation and Purification Technology 128 (2014) 39–44
by the mixed weak acid and type I strong base resins, indicating better purification performance of the weak acid and type I strong base resins. It can also be seen that, the solution conductivity of mixed weak acid and type II strong base resins increased sharply after DC power was supplied, and finally reached 384 lS/cm, while that of the mixed weak acid and type I strong base resins increased gradually, and finally reached 244 lS/cm. Additionally, it took only 30 min for the mixed weak acid and type II strong base resins to reach a steady state, while the time needed for the mixed weak acid and type I strong base resins was as long as 90 min. Both the faster speed of ions release and the higher conductivity in the steady state suggested better regeneration performance of the mixed weak acid and type II strong base resins. The results above indicate that a combination of two types of strong base resins may give a boost on performance. The mixed weak acid and type II strong base resins played the role of pretreatment of feed water, due to its high regeneration efficiency; while the mixed weak acid and type I strong base resins played the role of further purification, attributing to its high purification efficiency.
Fig. 1. Experimental set-up for equilibrium investigation.
3.2. Purification performance of new resins
(b) Regeneration period Fig. 2. MFEDI system.
pore, resolution: 0.01 for pH and 0.01 mg/L for Na+). Chlorine was measured using the iodometric titration method [25].
3. Results and discussion 3.1. Main property of different mixed resins Fig. 3 shows the conductivity variation of the circulating solution with time after DC electricity was supplied to different mixed resins. Before the electricity was supplied, the solution had been circulated for a sufficiently long time, usually about one and half an hour, and therefore the initial conductivity was essentially the equilibrium conductivity of the solution in cases without electricity imposed. It was found that the equilibrium solution conductivity of the mixed weak acid and type I strong base resins was 2.41 lS/cm, whereas that of the mixed weak acid and type II strong base resins was 10.3 lS/cm, which was almost 3.3 times higher than that of the former. In other words, more ions were adsorbed
The purification performance of the new resins that were fully regenerated with chemicals was investigated first, and the results are shown in Fig. 4. It can be seen that, with an influent conductivity of 50 lS/cm, the effluent conductivity of the lower layer was lower than 2 lS/cm within the initial 29, while that of the upper layer kept below 0.056 lS/cm. The effluent conductivity of the lower layer then increased gradually to 13.2 lS/cm at 45 h and that of the upper layer still kept below 0.056 lS/cm. Finally, the effluent conductivity of the lower layer rose promptly and reached 47.5 lS/ cm at 59 h, whereas that of the upper layer increased gradually and reached 0.066 lS/cm at the same time. In the whole purification process, the ions absorbed by the lower layer amounted to be 81%, and those removed by the upper layer were calculated to be 19%. The experimental results above further verified our original prediction, that is, the lower mixed weak acid and type II strong base resins layer would play a predominant role in removing saline ions from the feed water whereas the upper mixed weak acid and type I strong base resins layer would be responsible to guarantee the effluent quality. 3.3. Electro-regeneration performance 3.3.1. Concentrating efficiency After adsorption, the resins were regenerated electrically at a constant voltage of 1400 V. Fig. 5 shows the variation of the con-
Effluent conductivity of L (µ s/cm)
400
Conductivity (µ s/cm)
320
240
160
D113:550A=1:2 D113:202=1:2
80
0
0
20
40
60
80
100
120
Time (min) Fig. 3. Equilibrium conductivity variations of different mixed resins at 25 °C.
50
0.075
40
0.070
Effluent of lower layer Effluent of upper layer
30
0.065
20
0.060
10
0.055
0
0
10
20
30
40
50
60
0.050
Time (h) Fig. 4. Purification performance of the new resins at 25 °C.
Effluent conductivity of U (µ s/cm)
(a) Purification period
42
X. Shen et al. / Separation and Purification Technology 128 (2014) 39–44
cathode was on the top indicate a constant flux of Na+. As the anode and cathode had different effect on the regeneration of weak acid resin, the reversing of the electrode polarity gave a temporary change in the variations of Na+ concentration. Then, the Na+ concentration decreased at a constant rate too. The Na+ concentration of the mixing concentrate was 75.6 mg/L and the total amount of the NaCl exchanged during 1 h electrical regeneration was 2.048 g.
700
Conductivity (µ s/cm)
600
Concentrate of lower layer Concentrate of upper layer
500 400 300 200 100 0
0
10
20
30
40
50
60
70
Time (min) Fig. 5. Conductivity variations of different resin layers during regeneration at 25 °C.
centrate conductivity with regeneration time. It can be seen from the figure that the concentrate conductivity increased fast initially, and reached a maximum value of 172 lS/cm for the upper layer, and 624 lS/cm for the lower layer at 5 min. The fast increase in the concentrate conductivity reveals the fast ions release from the resin phase to the solution phase. The conductivity then reduced gradually due to a decrease in the concentration of the ions adsorbed in resins and thereby a decrease in the speed of ions release. There was a sudden drop in the conductivity when the electrode polarity was reversed at 20 min. The reason for this phenomenon is that the concentrate was temporarily neutralized after reversing, and the conductivity of the salt solution was lower than that of the acid solution at the same molar concentration. Thereafter, the conductivity increased gradually due to the increase in pH value. It was then reduced again as a result of the decreased speed of ions release. By mixing all the concentrate solution of the lower layer, the conductivity of the concentrate was over 380 lS/cm. It can also be found that, the concentrate conductivity value of the lower resin was about 2.4 times higher than that of the upper resin. This further demonstrates the higher regeneration efficiency of the lower mixed weak acid resin and type II strong base resin. Fig. 6 shows the variation of Na+ concentration of the concentrate with regeneration time. The Na+ concentration in the concentrate followed a predictable trend over the experiment. The linear variations of the Na+ concentration in the first 20 min when the
3.3.2. Current variation Fig. 7 presents the current density variation during regeneration at a constant voltage of 1400 V. It was found that the current density increased gradually at the beginning and achieved a maximum value of 137 A/m2 at 5 min. This should be attributed to an increase in resins conductivity owing to the changes of the resin forms and an increase in the solution conductivity. The current density increase also occurred in the initial 5 min after the electrodes were reversed. Besides, the tendency was that the current density reduced with the regeneration time due to the change of the resin resistance and the decrease of the solution conductivity. To investigate resin resistivity of different resin forms, the correlation of voltage with current density for different resin forms was examined, and the results are shown in Fig. 8. It can be calculated that the resistivity for sodium form of weak acid resin, hydrogen form of weak acid resin, chloride form of type II strong base resin, hydroxide form of type II strong base resin, chloride form of type I strong base resin and hydroxide form of type I strong base resin were 3.0 Xm, 18.7 Xm, 1.9 Xm, 2.3 Xm, 2.0 Xm and 0.8 Xm, respectively. The resistivity difference of different resin forms can be explained by the different binding force between the ions adsorbed and the fixed sites of the resin. The greater the binding force is, the harder the ions can be dissociated from the fixed sites, and thus the larger the resistivity is. Chloride ion has a greater binding force with a fixed site of type I strong base resin than hydroxide ion. As a result, chloride form type I strong base resin has a larger resistivity than hydroxide form of type I strong base resin. As the upper layer packed with weak acid and type I strong base resins was only 30 cm, the general resistivity was determined by weak acid and type II strong base resins. With the proceeding of the electro-regeneration, sodium form of weak acid resin and chloride form of type II strong base resin changed into hydrogen form and hydroxide form. Since the hydrogen form weak acid resin has much higher resistivity than the sodium form weak acid resin, as a whole, the resistivity of the resin layers rose gradually. This explains why the current density tended to reduce during regeneration.
150
Concentrate of lower layer Concentrate of upper layer
100
125 2
Current density (A/m )
Concentration of Na+ (ppm)
125
75
50
25
0
100 75 50 25
0
10
20
30
40
50
60
70
Time (min) Fig. 6. Na+ concentration variations of different resin layers during regeneration at 25 °C.
0
0
10
20
30
40
50
60
70
Time (min) Fig. 7. Current density variation during regeneration at a constant voltage of 1400 V at 25 °C.
43
X. Shen et al. / Separation and Purification Technology 128 (2014) 39–44
300
14
(A) D113-Na D113-H
200
Concentrate of lower layer Concentrate of upper layer
12 10 8
pH
Voltage (V)
250
150
6 100 4 50 0
2
0
100
200
300
400
2
10
20
30
40
50
60
70
Fig. 9. pH variations of different resin layers during regeneration at 25 °C.
30
(B)
25
Voltage (V)
0
Time (min)
Current density (A/m )
550A-OH 550A-Cl
20 15 10 5 0
0
0
100
200
300
400
ions produced by electrolysis in the anode would drain away with the concentrate before they regenerated the cation exchange resin. As a result, the regeneration of anion exchange resin was more available. However, the concentrate of the upper layer kept alkali in the first 20 min. This was because weak acid resin was much more easily to be regenerated than type I strong base resin regardless of the polarity. When the polarity was reversed, the pH value of the concentrate turned into alkali due to the favorable regeneration of cation exchange resin. The pH value of the whole concentrate was 6.9, indicating it can be returned to a pretreatment unit, such as RO, for recovery without the need of neutralization in real applications.
2
Current density (A/m )
3.3.4. Chlorine analysis During regeneration, hydrogen gas was produced at the cathode, while oxygen and chlorine gases were generated at the anode. Since the generated chlorine in the gaseous phase was too low to detect, the concentration of chlorine in the concentrate was analyzed. By mixing all the concentrate collected, the average concentration of the chlorine, which is a good disinfectant for use in RO, was measured to be 2.1 mg/L.
40
(C) Voltage (V)
30
202-OH 202-Cl
20
3.4. Purification performance of the resins after electro-regeneration 10
0
0
100
200
300
400
2
Current density (A/m ) Fig. 8. Voltage–current correlations of resins in different forms: (A) D113, (B) 550A, and (C) 202.
3.3.3. pH Variation Fig. 9 shows the pH variation of the concentrate with regeneration time. It was found that the concentrate of the lower layer was acidic during the first 20 min when the cathode was on the top of the resin layers. This was because such polarity was favorable for the electro-regeneration of anion exchange resin. The hydrogen and hydroxide ions for regeneration can be from electrolysis in the electrodes or electrically enhanced water dissociation. When the cathode was on the top, anion exchange resin can be regenerated by hydroxide ions either from electrolysis or from water dissociation, whereas cation exchange resin can be regenerated by hydrogen ions from electrolysis only. This was because hydrogen
To test the performance of the double layer MFEDI after electroregeneration, a purification experiment of the resins after regenerated electrically was conducted under the same condition as the new resins, and the results are demonstrated in Fig. 10. It can be observed that the effluent conductivity of the lower layer rose gradually and reached 42 lS/cm at 11 h, while that of the upper layer kept constant at 0.056 lS/cm in the initial 8 h and then increased slightly to 0.066 lS/cm in the following 3 h, indicating that the removal performance of the MFEDI system after regeneration was as good as before. The total amount of NaCl removed by ion exchange resins was 2.055 g which could be calculated by flow rate, removal time and concentration of feed water. This value is in good agreement with that of NaCl released from the resins during the regeneration, implying that the capacity of ion exchange resin was unchanged after regeneration. In addition, the water recovery rate was calculated to be 86%. A repetitive experiment of 10 times was carried out on the hybrid ion exchange resins to evaluate the stability of the MFEDI process for production of high-purity water. As shown in Fig. 10, the effluent conductivity of the upper and lower layer kept very stable after 10 times of experiments. Therefore, it is demonstrated that the purification performance of the MFEDI process was stable with
44
X. Shen et al. / Separation and Purification Technology 128 (2014) 39–44
Effluent conductivity of L (µ s/cm)
40
0.070
Effluent of Lower layer Effluent of upper layer
30
0.065
20
0.060
10
0.055
0
0
2
4
6
8
10
0.050 12
Effluent conductivity of U (µ s/cm)
0.075
50
Time (h) Fig. 10. Purification performance of the resins after electrical regeneration at 25 °C.
a cycle time of 12 h, including 11 h of high-purity water production and 1 h of electrical regeneration. 3.5. Energy consumption The energy consumption can be calculated according to the following formula.
E¼
IUt1 ðQ 2 t 2 Q 1 t1 Þ
ð8Þ
anywhere E is the energy consumption, kW h/m3 water; I is the average current during the whole regeneration period, A; U is the voltage, V; Q2 is the flow rate during purification period, m3/h; t2 is the purification time, h; Q1 is the flow rate during regeneration period, m3/h; t1 is the regeneration time, h. The energy consumption was calculated to be 1.5 kW h/m3 water. It was acceptable considering the high concentration of influent. 4. Conclusions By combining type I strong base resin and type II strong base resin, MFEDI were used to produce HPW successfully from the feed water with an conductivity as high as 50 lS/cm. The effluent conductivity was 0.056–0.066 lS/cm, indicating good purification performance. The conductivity of the concentrate collected during electrical regeneration was over 382 lS/cm, suggesting high electrical regeneration efficiency. The concentrate was neutral and therefore could be recycled to RO directly. The energy consumption and the water recovery were 1.5 kW h/m3 water and 86%, respectively. References [1] H.D. Willauer, F.D. Mascio, D.R. Hardy, M.K. Lewis, F.W. Williams, Development of an electrochemical acidification cell for the recovery of CO2 and H2 from seawater, Ind. Eng. Chem. Res. 50 (2011) 9876–9882.
[2] A. Grabowski, G.Q. Zhang, H. Strathmann, G. Eigenberger, Production of highpurity water by continuous electrodeionization with bipolar membranes: influence of concentrate and protection compartment, Sep. Purif. Technol. 60 (2008) 86–95. [3] J. Lu, Y.X. Wang, Y.Y. Lu, G.L. Wang, L. Kong, J. Zhu, Numerical simulation of the electrodeionization (EDI) process for producing ultrapure water, Electrochim. Acta. 55 (2010) 7188–7198. [4] J. Lu, Y.X. Wang, J. Zhu, Numerical simulation of the electrodeionization (EDI) process accounting for water dissociation, Electrochim. Acta. 55 (2010) 2673– 2686. [5] J.H. Song, K.H. Yeon, J. Cho, S.H. Moon, Effects of the operating parameters on the reverse osmosis-electrodeionization performance in the production of high purity water, Korean J. Chem. Eng. 22 (1) (2005) 108–114. [6] T.H. Boyer, P.C. Singer, Bench-scale testing of a magnetic ion exchange resin for removal of disinfection by-product precursors, Water Res. 39 (2005) 1265– 1276. [7] Y.Q. Xing, X.M. Chen, P.D. Yao, D.H. Wang, Continuous electrodeionization for removal and recovery of Cr(VI) from wastewater, Sep. Purif. Technol. 67 (2009) 123–126. [8] A. Mahamoud, A.F.A. Hoadley, An evaluation of a hybrid ion exchange electrodialysis process in the recovery of heavy metals from simulated dilute industrial wastewater, Water Res. 46 (2012) 3364–3376. [9] H.X. Lu, J.Y. Wang, S.F. Bu, L. Fu, Influence of resin particle size distribution on the performance of electrodeionization process for Ni2+ removal from synthetic wastewater, Sep. Sci. Technol. 46 (2011) 404–408. [10] L. Fu, J.Y. Wang, Y.L. Su, Removal of low concentrations of hardness ions from aqueous solutions using electrodeionization process, Sep. Purif. Technol. 68 (2009) 390–396. [11] K.H. Yeon, J.H. Song, S.H. Moon, A study on stack configuration of continuous electrodeionization for removal of heavy metal ions from the primary coolant of a nuclear power plant, Water Res. 38 (2004) 1911–1921. [12] Ö. Arar, Ü. Yüksel, N. Kabay, Mithat Yüksel, Removal of Cu2+ ions by a microflow electrodeionization (EDI) system, Desalination 277 (2011) 296–300. [13] A. Grabowski, G.Q. Zhang, H. Strathmann, G. Eigenberger, The production of high purity water by continuous electrodeionization with bipolar membranes: influence of the anion-exchange, J. Membr. Sci. 281 (2006) 297–306. [14] K.H. Yeon, J.H. Seong, S. Rengaraj, S.H. Moon, Electrochemical characterization of ion-exchange resin beds and removal of cobalt by electrodeionization for high purity water production, Sep. Sci. Technol. 38 (2) (2003) 443–462. [15] J.Y. Wang, S.C. Wang, M.R. Jin, A study of the electrodeionization process: highpurity water production with a RO/EDI system, Desalination 132 (2000) 349– 352. [16] J.H. Song, K.H. Yeon, S.H. Moon, Effect of current density on ionic transport and water dissociation phenomena in a continuous electrodeionization (CEDI), J. Membr. Sci. 291 (2007) 165–171. [17] J.S. Park, J.H. Song, K.H. Yeon, S.H. Moon, Removal of hardness ions from tap water using electromembrane processes, Desalination 202 (2007) 1–8. [18] J.W. Lee, K.H. Yeon, J.H. Song, S.H. Moon, Characterization of electroregeneration and determination of optimal current density in continuous electrodeionization, Desalination 207 (2007) 276–285. [19] J. Wood, J. Gifford, J. Arba, M. Shaw, Production of ultrapure water by continuous electrodeionization, Desalination 250 (2010) 973–976. [20] W.Q. Su, R.Y. Pan, X.M. Chen, Membrane-free electrodeionization for high purity water production, Desalination 15 (2013) 29–86. [21] T.A. Davis, V. Hills, Electrically regenerated ion exchange system, US Patent: 4,032,452, June 1977. [22] Y.Q. Xing, X.M. Chen, D.H. Wang, Electrically regenerated ion exchange for removal and recovery of Cr(VI) from wastewater, Environ. Sci. Technol. 41 (2007) 1439–1443. [23] C. Comninellis, G.P. Vercesi, Characterization of DSA-type oxygen evolving electrodes: choice of a coating, J. Appl. Electrochem. 21 (1991) 335–345. [24] X.M. Chen, G.H. Chen, Investigation of Ti/IrO2–Sb2O5–SnO2 electrodes for O2 evolution: calcination temperature and precursor composition effects, J. Electrochem. Soc. 152 (7) (2005) J59–J64. [25] APHA, Standard Methods for the Examination of Water and Wastewater, 21st ed., American Public Health Association, Port City, Maryland, 2005.