Desalination 275 (2011) 62–66
Contents lists available at ScienceDirect
Desalination j o u r n a l h o m e p a g e : w w w. e l s ev i e r. c o m / l o c a t e / d e s a l
Ion-exchange membrane capacitive deionization: A new strategy for brackish water desalination Haibo Li, Linda Zou ⁎ SA Water Centre for Water Management and Reuse, University of South Australia, Adelaide, SA 5095, Australia
a r t i c l e
i n f o
Article history: Received 13 October 2010 Received in revised form 11 February 2011 Accepted 11 February 2011 Available online 14 May 2011 Keywords: Ion-exchange membrane Membrane capacitive deionization (MCDI) Single walled carbon nanotube (SWCNTs) Salt removal efﬁciency Sorption capacity
a b s t r a c t Membrane capacitive deionization (MCDI) integrates the advantages of capacitive deionization (CDI) and ionexchange membrane technology and has shown great potential to improve the desalting efﬁciency. MCDI works based on the same working principle of CDI. In addition, ion-exchange membranes are introduced in front of the electrodes so that the charged ions can be selectively passed through the membrane layer and are subsequently adsorbed by the oppositely charged electrodes without interference of co-ions and therefore improve the salt removal efﬁciency as well as strengthen the regeneration. In this research, electrodes made from single walled carbon nanotubes (SWCNTs) were used together with cation- and anion-exchange membranes. The correlation between solution concentration and conductivity was calibrated prior to the experiment. Through bench scale batch mode desalination experiments, a salt removal efﬁciency of as high as 97% was achieved with initial conductivity of 110 μs/cm and electrical voltage of 1.2 V. This efﬁciency is much higher than the corresponding CDI without membrane whose salt removal efﬁciency is only about 60%. Further, the obtained adsorption rate constant as a result of adsorption kinetics clearly demonstrated that the ion-exchange membrane can help to achieve a faster ion transfer rate in the electrosorption process due to low co-ions expulsion effect. © 2011 Published by Elsevier B.V.
1. Introduction Ion-exchange membranes are introduced in capacitive deionization (CDI), which is an electrically induced alternative approach for removing salt ions from concentrated aqueous solutions. In an electrical ﬁeld, charged ions are forced into the electrical double layer at an electrode-solution interface when the electrodes are connected to an external power supply. Fig. 1(a) shows a typical CDI unit. For membrane capacitive deionization (MCDI), cation- and anion-exchange membranes used and separated by an insulator to avoid shorting the circuit. The inlet stream contains a large number of salt ions that ﬂow through the insulator compartment, in which the ions were permeated through the ion-exchange membrane and then adsorbed into internal porous structure of the electrodes. As a result, the salt concentration of outlet stream would be reduced. In a reversal process, a reverse voltage is applied to the two porous electrodes, the adsorbed ions can be released back into the bulk solution and therefore the regeneration is achieved. Different from the conventional CDI, the introduced ion-exchange membranes can directly restrict co-ions from accessing the electrodes so the counter-ions can be easily absorbed by the electrodes and therefore the salt removal efﬁciency is increased. A brief schematic diagram of adsorption
⁎ Corresponding author. Tel.: + 61 8 830 25489; fax.: + 61 8 830 23386. E-mail address: [email protected]
(L. Zou). 0011-9164/$ – see front matter © 2011 Published by Elsevier B.V. doi:10.1016/j.desal.2011.02.027
mechanism associated with MCDI is depicted in Fig. 1(b). Speciﬁcally, a cation-exchange membrane is placed in front of the porous electrode that is deﬁned as cathode while an anion-exchange membrane is placed in front of the porous electrode that is positively biased as anode. As a result, the counter-ions can move freely in and out of the ion-exchange membranes and porous electrode, while the co-ions activity is avoided. The state-of-art theory relates to MCDI that was proposed and developed by Biesheuvel , which showed that the co-ion expulsion effect would be strictly blocked and for each electron transferred between the electrodes, ideally one salt molecule will be removed from the bulk solution. In addition to the minimization of co-ions expulsion effect, the regeneration of MCDI would be more efﬁcient than CDI because the adsorbed counter-ions can be fully released. From the aspect of practical application, the MCDI would be more promising in energy efﬁciency as it removes the salt ions, which are only a small percentage of the feed solution, as compared to most other desalination technologies that shift water, which accounts of 90% of the feed solution. Conversely, MCDI is also an environmentally friendly process because no contaminants or harmful by-products are produced during the desalination and regeneration of electrodes. In the last few decades, extensive studies associated with CDI electrodes have concentrated on the preparation of carbon electrodes with higher speciﬁc surface area and superior conductivity, such as carbon cloth, carbon aerogels, ordered mesoporous carbons, and carbon nanotubes (CNTs) [2–7]. Among these advanced carbon materials,
H. Li, L. Zou / Desalination 275 (2011) 62–66
pod Ralskem, Czech). The detailed descriptions related to the speciﬁcations of the membranes can be found elsewhere . Ultrapure Milli-Q water (18.2 MΩ cm, Millipore Corporation, France) was used in all of the experiments. Each ion-exchange membrane with dimension of 140 mm long × 70 mm wide was immersed in Milli-Q water for 2 days to make it fully swelled before use. 2.2. Electrode fabrication To fabricate MCDI electrodes, the SWCNTs, graphite powder as conductive material and Polytetraﬂuoroethylene (PTFE) as binder were used [6,7,9]. Their respective percentages in the ﬁnal electrodes are 72%, 20% and 8%. Each electrode was 70 mm wide × 140 mm long × 0.3 mm thick and ﬂow-through hole with a diameter of 4 mm. Ethanol (10–20 ml) was added dropwise to the mixture to make it moist and it was then pressed onto a graphite sheet. A typical MCDI unit consists of one pair of ion-exchange membranes, one pair of SWCNTs electrodes and a spacer. Taking one electrode, it was assembled in the sequence of retaining plate, carbon electrode, ionexchange membrane and nylon spacer, as shown in Fig. 1(b). 2.3. Characterization
Fig. 1. Schematic diagrams of (a) CDI and MCDI unit (b) working principle of CDI compared with MCDI.
CNTs have been considered a good candidate for being used as CDI electrode due to its unique physical and chemical properties since it was ﬁrst observed in 1991 . Particularly in developing MCDI technology, carbon nanotube and nanoﬁber composites have been pioneered to be used as sorption electrodes and good desalination results are achieved through coupling with ion-exchange membranes . In this work, single walled carbon nanotubes (SWCNTs) incorporating with ion-exchange membranes were employed as electrodes for MCDI. The surface morphology and porous structure were characterized by means of scanning electron microscopy and N2 adsorption system, respectively. Further, batch mode experiments were performed to investigate several aspects relating to SWCNTbased MCDI. Finally, the desalination performance of MCDI was compared with CDI under the same experimental condition, and the results were presented.
Scanning electron microscope (SEM, XL-30 Netherland) was employed to observe the morphology and ﬁne structure of SWCNTs. The N2 adsorption–desorption isotherms used to determine the porosity of the SWCNTs were performed at − 196°C on Belsorp system (BEL JAPAN, INC). The pore size distribution curves were calculated by the Barrett–Joyner–Halenda (BJH) method from the desorption branch. The speciﬁc surface area was calculated from the adsorption data in the relative pressure interval from 0.04 to 0.2 using the BET method. The total volume (V) was estimated from the amount adsorbed at a relative pressure of 0.98. The Dubinin–Radushkevich (DR) theory was employed for estimating the micropore volume (Vmi), and the as-plot method was used for the external surface area (Se) and the micropore surface area (Smi). The mesopore fraction was obtained from (V Vmi). The surface morphology and porosity property of SWCNTs are presented in Fig. 2. It can be observed that the buddled CNTs are entangled and show a network structure that is expected to increase the speciﬁc surface area and thereby beneﬁcial to adsorption. The speciﬁc surface area of SWCNTs resulted from N2 adsorption– desorption curve is 455 m2/g and the dominant peak appeared in pore size distribution focused on 1.88 nm, indicating that the average
2. Materials and methods 2.1. Materials The SWCNTs (length, 5–15 μm; ash percentage, ≤2 wt.%; surface area, N400 m2/g; diameter, b2 nm, respectively; purity, ≥90%) were supplied by Nanotech Port Co., Ltd. (Shenzhen, China). The ionexchange membranes including cation-exchange membrane and anion-exchange membrane were provided by MemBrain s.r.o. (Straz
Fig. 2. N2 adsorption–desorption curve, inset in bottom right is pore size distribution and another one in top left shows the SEM image of SWCNTs.
H. Li, L. Zou / Desalination 275 (2011) 62–66
diameter of SWCNTs are smaller than 2 nm, which is accordance with product speciﬁcations. To fully utilize the inner space of SWCNTs, the SWCNTs were treated by HNO3 to remove the residual catalyst  and therefore open up the tip of the tubes and at the same time attach some useful function groups on the surface, such as –OH and –COOH, to further increase the sorption capacity of SWCNTs. 2.4. Desalination experiment In the desalting experiments, the solution was pumped into the unit cell by a peristaltic pump and the efﬂuent was returned to the unit cell. The solution volume was maintained at 50 ml and the ambient temperature was kept at 298 K, respectively. Meanwhile, the applied voltage was adjusted from 0.8 V to 1.6 V to determine the optimum working voltage. The above-mentioned experiments were performed using a synthetic NaCl solution, which had an initial conductivity of around 110 μS/cm. The relationship between conductivity and concentration was established according to a calibration table made prior to the experiment. The NaCl conductivity changes of the solution were continuously monitored at the outlet of the unit cell by using conductivity meter. In our experiment, the salt removal efﬁciency and sorption capacity were deﬁned as follows: Salt removal efficiency ¼
Sorption capacity ¼
C0 Cf 100% C0
ð 0:91878 þ 0:48188 CÞ 0:05 1000 ðμmol=gÞ 58:5 M ð2Þ
Where C0 and Cf are the initial and ﬁnal concentrations (μs/cm), M is the mass of both electrodes (g). The equation (−0.91878+0.48188×C)× 1000 presents numerical correlation between the concentration unit μmol/g and conductivity μs/cm that are derived by calibrating the NaCl solutions prior to the experiment. 3. Results and discussion
Fig. 3. SWCNTs based MCDI processes (a) at different working voltages (inset shows sorption performance from 30 to 60 min) and (b) the corresponding salt removal efﬁciency as well as sorption capacity. The initial conductivity was 110 μs/cm.
3.1. Desalination experiments 3.1.1. Desalination performance of SWCNTs based MCDI Fig. 3 depicts the typical SWCNT-based MCDI process at different working voltages and the corresponding salt removal efﬁciency as well as sorption capacity. As expected, once the electrical voltage was applied, the conductivity dramatically decreased until it would not change any further, indicating that saturation was achieved. Further, when we removed the voltage and made two electrodes short circuit, the electrode can be quickly regenerated, that is, the adsorbed ions are desorbed from the electrodes due to the disappearance of electrostatic force. The salt removal efﬁciency increased from 91.3 to 98.1% by varying the polarized bias from 0.8 to 1.4 V and the sorption capacity has the same trend as well. However, when the voltage was ﬁxed at 1.6 V, the sorption capacity slightly decreased compared to corresponding sorption capacity at 1.4 V, indicating that water electrodialysis was taking place (the voltage for water electrodialysis is about 1.23 V). Thus, by taking into account both energy consumption and salt removal efﬁciency as well as sorption capacity, the optimum working voltage for SWCNTs based MCDI is recommended as 1.2 V, which is lower than the voltage of water electrodialysis. It is observed that the slope of the adsorption curve is very steep, which indicates a high adsorption rate constant (as shown in Fig. 6(b)) and therefore conﬁrms the theory that the co-ion effects can be signiﬁcantly reduced by incorporation of ion-exchange membrane [12,13]. Further, the high adsorption rate constant also indicates that the ions access the electrode faster and thus the desalination process can be completed in a very short time. In addition, several experiments have been performed
to check the regeneration property of SWCNT-based MCDI. Fig. 4 implies that the salt ions in the testing solution can be almost completely depleted during the adsorption process and the electrodes are easily regenerated in the desorption process for three adsorption– desorption cycles.
Fig. 4. Regeneration property of SWCNTs based MCDI process at an initial conductivity of 130 μs/cm.
H. Li, L. Zou / Desalination 275 (2011) 62–66
3.1.2. Regeneration of SWCNTs based MCDI In order to further investigate the desalination performance of SWCNT-based MCDI at much higher conductivity, several experiments were conducted at an initial conductivity of 500 μs/cm and 1125 μs/cm respectively. Fig. 5(a) shows the conductivity reduction versus adsorption time at an initial salt conductivity of 500 μs/cm and inset gives the corresponding salt removal efﬁciency as well as sorption capacity. It is noted that the salt removal efﬁciency increased from 55.4 to 88.2% and the sorption capacity increase from 80.9 to 122.5 μmol/g with the increase of applied voltage from 0.8 to 1.6 V. This salt removal efﬁciency is still lower than the corresponding salt removal efﬁciency obtained when the initial conductivity is lower, while the sorption capacity was larger than it was when tested with lower conductivity, for instance, 110 μs/cm. This could imply that the electrode has not saturated yet in the lower conductivity experiment. According to previous research , the sorption behaviour of MCDI would follow the Langmuir isotherm so that the maximum sorption capacity could be predicted by the corresponding isotherm but herein it is not investigated. Further, when the initial conductivity was increased to 1125 μs/cm, the salt removal efﬁciency continued to decrease but the sorption capacity is nearly equal to that of sorption capacity at the initial conductivity of 500 μs/cm using the same electrical voltage, as shown in Fig. 5(b). However, the regeneration
Fig. 5. Desalination performance of SWCNTs based MCDI process at (a) an initial conductivity of 500 μs/cm by varying electrical voltage from 0.8 to 1.6 V, inset depicts the corresponding salt removal efﬁciency as well as sorption capacity as a function of electrical voltage (b) an initial conductivity of 1000 μs/cm with applied voltage of 1.2 V.
curve still showed good outcomes, which are similar to the results obtained from the experiments conducted at a lower conductivity. 3.2. Comparison of MCDI with CDI The desalting performance of the SWCNTs based MCDI with ionexchange membranes was compared with the CDI without the membranes under the same experimental conditions. The result showed in Fig. 6 demonstrated the typical adsorption–desorption process based on two types of devices. In Fig. 6(a), both MCDI and CDI processes showed a dramatic decrease in salt conductivity within ﬁrst 30 min, gradually approached saturation, and then increased to the initial level once the electricity was cut off. Apparently, the salt removal efﬁciency of MCDI with the ion-exchange membranes (about 97%) is much higher than that of CDI without the membranes (about 60%). As discussed in the previous section, selective shielding of the membrane, the salt removal efﬁciency is less affected by co-ions effect. It can be seen that the presence of the ion-exchange membrane MCDI is very beneﬁcial for electrosorption salt ions by making the counter-ions transfer faster and be adsorbed by electrode without obstruction of the co-ions. It is conﬁrmed by the steeper slope of the adsorption curve for MCDI. Moreover, to further prove our explanation, the sorption kinetics, implying the sorption rate, was simulated and calculated from Fig. 6(b). The adsorption rate constant of salt ions adsorbed onto electrodes was determined by the Lagergren equation,
Fig. 6. (a) Comparative desalination result by SWCNTs based CDI and MCDI. (the initial conductivity is around 100 μs/cm, the electrical voltage is of 1.2 V, the rest experimental conditions are according to the Desalination experiments section) (b) the sorption kinetics of SWCNTs based CDI and MCDI in the same experimental condition.
H. Li, L. Zou / Desalination 275 (2011) 62–66
which is often so-called pseudo-ﬁrst-order adsorption kinetics . It was formulated as: logðqe qÞ¼ log qe
play a useful role in brackish water desalination due to its notable advantages comparing with CDI without membrane and other conventional desalting technologies.
Where k is the adsorption rate constant (min−1), qe and q are the adsorption capacity at equilibrium (μmol/g) and time t (min), respectively. The calculated data, including rate constant k as well as regression coefﬁcient R2, which can be utilized to determine the error between experimental data and theory simulation, are considered. The coefﬁcient R2 for SWCNTs based MCDI and CDI are 0.9985 and 0.9986, respectively, which is very close to 1, indicating the experimental data correlated with the adsorption equations very well. Further, the rate constant of MCDI is 0.3995, which is much higher than that of CDI whose rate constant is of 0.1933. This again conﬁrmed that the ion transferred in MCDI is much faster than in corresponding CDI. 4. Conclusion In this paper, the performance of MCDI using SWCNT as electrodes was compared with the CDI without membranes. Through bench scale batch mode desalination experiments, it is found that the salt removal efﬁciency of SWCNTs based MCDI is proportional to electrical voltage in different ionic strength. Speciﬁcally, a salt removal efﬁciency of as high as 97% was achieved with initial conductivity of 110 μs/cm and electrical voltage of 1.2 V. This efﬁciency is much higher than the corresponding CDI without membrane whose salt removal efﬁciency is only about 60%. The obtained adsorption rate constant as a result of adsorption kinetics clearly demonstrated that the ion-exchange membrane can help to achieve faster ion transferring rate in the electrosorption process due to low co-ions expulsion effect, and this is evidenced by comparing the salt removal efﬁciencies of MCDI with ion-exchange membrane and CDI without the membrane. Further, the regeneration property of SWCNTs based MCDI was investigated and was found that it can be regenerated very well by imposing a reverse equal voltage. It is believed that the membrane-based CDI process will
Acknowledgements The authors acknowledge the ﬁnancial support of WQRA research grant 1025-09. The author H.B Li acknowledges the China Scholarship Council (CSC, File No. 2009614107) for the ﬁnancial support. References  P.M. Biesheuvel, A. van der Wal, Membrane capacitive deionization, J. Membr. Sci. 346 (2010) 256–262.  Y. Oren, Capacitive deionization (CDI) for desalination and water treatment-past, present and future (a review, Desalination 228 (2008) 10–29.  J.H. Lee, W. Bae, J. Choi, Electrode reactions and adsorption/desorption performance related to the applied potential in a capacitive deionization process, Desalination 258 (2010) 159–163.  X.Z. Wang, M.G. Li, Y.W. Chen, R.M. Cheng, S.M. Huang, L.K. Pan, Z. Sun, Electrosorption of ions from aqueous solutions with carbon nanotubes and nanoﬁbers composite ﬁlm electrodes, Appl. Phys. Lett. 89 (2006)8 053127–3.  C. Gabelich, T. Tran, A.H. Melsuffet, Electrosorption of inorganic salts from aqueous solution using carbon aerogels, Environ. Sci. Technol. 36 (2002) 3010–3019.  L. Zou, G. Morris, D.D. Qi, Using activated carbon electrode in electrosorptive deionisation of brackish water, Desalination 225 (2008) 329–340.  L. Zou, L. Li, H. Song, G. Morris, Using mesoporous carbon electrodes for brackish water desalination, Water Res. 42 (2008) 2340–2348.  S. Iijima, Helical microtubules of graphitic carbon, Nature 354 (1991) 56–58.  H.B. Li, Y. Gao, L.K. Pan, Y.P. Zhang, Y.W. Chen, Z. Sun, Electrosorptive desalination by carbon nanotubes and nanoﬁbres electrodes and ion-exchange membranes, Water Res. 42 (2008) 4923–4928.  http://www.membrain.cz/en/specialization/processes-and-technologies/ion-exchange-membrane/.  Q. Liao, J. Sun, L. Gao, Adsorption of chlorophenols by multi-walled carbon nanotubes treated with HNO3 and NH3, Carbon 46 (2008) 553–555.  Y.J. Kim, J.H. Choi, Improvement of desalination efﬁciency in capacitive deionization using a carbon electrode coated with an ion-exchange polymer, Water Res. 44 (2010) 990–996.  J.S. Kim, J.H. Choi, Fabrication and characterization of a carbon electrode coated with cation-exchange polymer for the membrane capacitive deionization applications, J. Membr. Sci. 355 (2010) 85–90.  H.B. Li, L.K. Pan, Y.P. Zhang, L.D. Zou, C.Q. Sun, Y.K. Zhan, Z. Sun, Kinetics and thermodynamics study for electrosorption of NaCl onto carbon nanotubes and carbon nanoﬁbers electrodes, Chem. Phys. Lett. 485 (2010) 161–166.