Krishnaveni Venugopal and
Sangeetha Dharmalingam*
Department of Mechanical Engineering, Anna University, Guindy, Chennai 600025, Tamil Nadu, India. E-mail: sangeetha@annauniv.edu; Fax: +91 044 22357744; Tel: +91 044 22357763
First published on 24th August 2015
In this study, polysulfone (PSu) was functionalized and modified using resin and fiber reinforcements for the preparation of mono and bipolar ion exchange membranes. In the case of the bipolar membrane, platinum was used as the intermediate layer. A solution of NaCl in the brackish water concentration range from 5 g L−1 to 25 g L−1 was used as the feed solution for the desalting technique using an electrodialysis process. A commercially procured ion exchange membrane made of polystyrene divinyl benzene was also evaluated for the purpose of comparison. The BPMED performance reached a highest current efficiency of 82.5% and 53.6% with the energy consumption of 0.52 W h and 1.39 W h for the synthesized and commercial membranes respectively. Also the performance of the fabricated unit was assessed in terms of electrical conductivity, salinity and ion (sodium and chloride) concentrations for the feed solution after an 8 hour duration. The reproducibility performance of the membranes was also analyzed for the synthetic salt solution.
The new membrane based electrodialysis (ED) process was commonly used for the production of quality water through the desalination technique. The energy consumption utilized for the process is directly related to the salinity of the feed water. Membrane based ED processes have considerable advantages in desalting brackish water and are used over a wide range of salinity concentration levels from brackish to sea water.2 Many types of membranes exist due to classification under various criteria. Generally ion exchange membranes (IEMs) can be defined as a thin layer of electrolytic barrier which can separate two phases and restrict the transport of various chemical species through it; they are subdivided into cation exchange membranes (CEMs), anion exchange membranes (AEMs) and bipolar membranes (BPMs).3
Monopolar membranes (CEM & AEM) were prepared by functionalization methodologies whereas BPM can be prepared by simply laminating conventional CEM and AEM.4 But laminated BPMs prepared with anion and cation exchange layers (CEL/AEL) alone, often do not perform efficient water dissociation.5 But the presence of heterogeneous materials like ion-exchange resin (IER) particles in non-conducting polymer matrix layers of a BPM and the presence of catalytic intermediates like quaternary6 and non-quaternary amine groups,7 weak acids such as phenolic, carboxylic acid or phosphoric acid groups8 and their corresponding base, inorganic substances such as Mg, Ni, Co, Mn, Cu, Fe and Al,9 metallic compounds such as ruthenium tri-chloride, chromic nitrate, indium sulfate, hydrated zirconium oxide, etc.,10,11 heavy/noble metal ions like Cr3+, Fe2+, Ag, Au, Pt, Pd, Os, Rh, Ti, Sn, Zr, Pa, Ru etc.,12,13 macromolecules such as polyethylene glycol (PEG),14 polyvinyl alcohol (PVA)15 and bovine serum albumin (BSA) containing both carboxylic and amino groups,16 starburst dendrimer polyamidoamine (PAMAM)17 etc.,4,5 in between the CEL and AEL as an intermediate layer (IL), usually results in a BPM with higher mechanical stability18,19 with improved water dissociation effects (the rate was found to be 50 million times faster than the ordinary water dissociation in the presence of a catalyst). Hence in the present study, all the IEMs were prepared with resin and glass fiber reinforcements.
Managing high salinity concentrates generated during membrane water desalination processes was a primary need to be addressed due to its potential threat to the receiving ecosystems regarding brine discharge. Depending on the desired application, bipolar membrane electrodialysis (BPMED) processes can be performed with different stack configurations. Tongwen & Weihua20 evaluated the effect of cell configurations such as A–C (type I), C–BP–C (type II) and BP–A–C–BP (type III) on the production of citric acid using BPMED. From the results, magnitudes of cell voltage, CE and energy consumption followed the analogous order as type II < type I < type III, while the potential drop across a BPM followed the order type II > type I ≈ type III suggesting that type II seemed to be a favorable cell configuration for the production of citric acid. Nagasawa et al.21 used both conventional ED and five compartmental BPMED techniques separately to remove boron from sodium tetra borate solution containing 100 mg L−1 of boron. The results confirmed that with the use of BPMED, more than 90% of the boron was removed under both acidic (pH 2.3) and basic (pH 9.1) conditions, whereas only 35–45% of the boron was removed using conventional ED. Nataraj et al. studied the ED pilot plant unit coupled with a membrane stack containing 11 cation-exchange and 10 anion-exchange membranes for the removal of nitrates and hardness from simulated aqueous mixtures containing salts that are usually encountered in brackish water. More than 94% of the nitrates, 89% of the chlorides, and 86% of the TDS were successfully removed after 150 min of operation of the ED.22 Using the same membrane stack configuration, wastewater effluent from the paper industry could be treated. The obtained results show that 80% of wastewater could be recovered by the use of the MF/ED hybrid process, while the remaining 20% of flow (concentrate) can be used as a biomass.23
The present study highlights the work designed to evaluate the desalination efficiency of prepared monopolar (CEM and AEM) and bipolar (with platinum as IL) IEM with resin and glass fiber reinforcements using polysulfone (PSu). The use of platinum has been given the greatest economic importance in various applied fields such as automobiles, oil refining, glass-making, medical instruments, electronics, dentistry, jewellery, computers, fertilizers, plastics, fuel cells, etc. which has also urged its use in water treatment applications. In addition, its certain properties such as superior conductivity, resistance to oxidation and strong catalytic property compared to other types of intermediates have shown platinum to be one of the best choices to be used as an intermediate for BPM in the present study. The study was conducted for five different synthetic salt water concentrations ranging from 5 g L−1 to 25 g L−1 of NaCl solution (used in feed compartment (FC)) in terms of conductivity, solution pH, transport number of ions (T. No.), feed concentration, current efficiency (CE), energy consumption, water dissociation efficiency (WDE), water dissociation flux and acid–alkali production in the acid compartment (AC) and base compartment (BC) up to 8 h. The stack performance using the synthesized membranes was compared with that of the commercial polystyrene divinyl benzene based (PSDVB) IEM under similar experimental conditions. In addition, the decrease in sodium-chloride ion concentration, salinity and electrical conductivity of the feed water were observed and the reproducibility test for the membrane system with the highest CE was conducted.
The obtained membranes with various resin content were subjected to conductivity studies as described in the ESI S1†; from Table 1it is observed that the conductivity of the prepared membranes increased with the increase in IER loadings. It was optimized at 40% for both resins26 because beyond 50%, the nature of the membranes changed to brittle from their soft nature. Reinforced bipolar membrane (RBPM i.e. RPSu–Pt) was prepared using RCEM (RSPSu) and RAEM (RQPSu) as CEL and AEL, respectively. Then on one side of both the CEL and AEL layers, a solution containing approximately 6 mL of 0.12 g of Pt-containing solution was coated on to form the IL. Finally the IL-coated sides of both layers (CEL and AEL) were sandwiched and subjected to a hot press to finally obtain RBPM represented as RBPM–Pt as shown in Fig. 1.
% of resin added | CER conductivity (S cm−1) | AER conductivity (S cm−1) |
---|---|---|
10 | 1.68 × 10−6 | 8.69 × 10−7 |
20 | 2.78 × 10−6 | 4.78 × 10−6 |
30 | 5.31 × 10−6 | 7.57 × 10−6 |
40 | 4.76 × 10−5 | 3.8 × 10−5 |
50 | 3 × 10−5 | 2.73 × 10−5 |
60 | 2.39 × 10−6 | 1.79 × 10−6 |
70 | 1.7 × 10−6 | 1.2 × 10−6 |
![]() | ||
Fig. 3 pH changes in FC, EC, AC and BC with time for various feed concentrations of the RPSu–Pt IEM system. |
![]() | ||
Fig. 4 pH changes in FC, EC, AC and BC with time for various feed concentrations of the PSDVB IEM system. |
Whereas, in the case of EC, since 0.05 M NaCl was taken as the electrolyte solution each time, the initial pH remained constant for all performances. Though both systems showed the final solution to be acidic in nature, the acidity was greater for the PSDVB IEM system than for the RPSu–Pt IEM system. The difference in pH observed between the two types of IEM systems in FC and EC was mainly attributed to the leakage of ions occurring through membranes arranged between the compartments in a stack. Similarly in the case of AC and BC, the solution pH was found to be approximately around 2.45 and 10.74, respectively, for all performances because of the dilute acid and base solutions that were taken initially.
From Fig. 3 and 4, it was additionally observed that, irrespective of whether it was AC or BC, the acidity or basicity was found to be increased during the first half stage of a performance and the increase was not uniform because of the leakage of some ions into the neighboring compartments. The nature of pH change reflected their acid and base production for both PSDVB and RPSu–Pt IEM systems. This pH change in various compartments clearly suggested that both IEM systems possessed adequate capacity to split water into its co-ions under the influence of an electric field. Also, it is observed that for the reproducibility test, more or less the same result as that of its original feed concentration was noticed.
![]() | ||
Fig. 5 Conductivity changes in AC, BC and FC for different feed concentrations with time for the RPSu–Pt IEM system. |
For the PSDVB IEM system, from Fig. 6, the highest acid and base conductivity values were found to be 1.29 mS cm−1 and 0.7 mS cm−1, respectively. In addition, the lower acid and base conductivity values for 5 g L−1 among other feed concentrations indirectly proved its lower acid and base pH change. In the case of FC the conductivity initially decreased depending upon the feed concentration, which then increased slightly and finally remained constant over time. This increased conductivity was observed to be a little higher than its initial value of that particular feed concentration mainly because of the leakage of ions that occurred from the neighboring compartments through the IEM into FC. It should also be noted that the reproducibility test in both these cases showed more or less the same results as that of its original feed concentration.
![]() | ||
Fig. 6 Conductivity changes in AC, BC and FC for different feed concentrations with time for the PSDVB IEM system. |
However, Fig. 7(a) clearly shows that the RPSu–Pt IEM system produced higher acid and base concentrations of about 0.014 N and 0.006 N for the same feed concentration. It has to be noted that once the higher concentration of acid/base was reached in a particular feed concentration, it remained constant until a certain duration of time after which it decreased with the increase in process time due to a decrease in NaCl concentration or diminished mass transfer of Na+ and Cl− ions in the feed solution.
![]() | ||
Fig. 7 Changes in acid–base yield with time for various feed concentrations for (a) the RPSu–Pt IEM system and (b) the PSDVB IEM system. |
It was observed from Fig. 7(b) that the PSDVB IEM system showed higher acid and base concentrations of about 0.009 N and 0.006 N respectively for various feed concentrations. Though the PSDVB based IEM cell is meant for base production rather than acid production because of higher specific permselectivity of CEM for H+ ions as reported in the literature,26 in the present study, from Fig. 7(b), it was observed that the alkalinity concentration was lower in the case of the PSDVB IEM system. Furthermore, H+ ions in the presence of water have a higher intrinsic mobility than OH− ions thereby resulting in more leakage of H+ ions through the AEM which leads to a decrease in the concentration of acid in AC.27 A similar decreasing trend was also observed in the case of BC for various feed concentrations.
In addition it should be noted that irrespective of the type of system used (either RPSu–Pt IEM or PSDVB IEM system), the produced acid and base values purely represent the acid and base generated during the process because the concentration value obtained from the titration procedure was subtracted from its initial concentration value (0.01 N). Thus the final value is solely due to the produced acid and base. Again, it should be noticed that the reproducibility test in both types of IEM systems showed more or less the same results as that of its original feed concentration.
![]() | ||
Fig. 8 Changes in the T. No. of ions and WDE for various feed concentrations with time for the RPSu–Pt IEM system. |
![]() | ||
Fig. 9 Changes in the T. No. of ions and WDE for various feed concentrations with time for the PSDVB IEM system. |
From Fig. 8, it can be observed that the RPSu–Pt IEM system showed the highest T. No. of 0.57 and 0.33 for chloride and sodium ions, respectively, for a 15 g L−1 feed concentration. Similarly from Fig. 9, it is observed that the PSDVB IEM system showed a greater T. No. of 0.16 and 0.08 of chloride and sodium ions respectively, for the same feed concentration.33 While in the case of WDE, as per Fig. 8, the RPSu–Pt IEM system showed results of about 0.87 finally for different feed concentrations. The reason for this difference was mainly due to the salt concentration in FC and hence, when the feed concentration was lower it was a better opportunity for the transport of water dissociated ions. But when the feed concentration was higher, the competition between these ions would be higher and more preferably the salt ion transport occurred initially followed by the water dissociated ions.34 Whereas, as per Fig. 9, the highest WDE value observed for the PSDVB IEM system was about 0.07. The highest T. No. and steady increase in WDE with time observed for the RPSu–Pt IEM system was mainly due to the increase in electric field, pre-polarization of water molecules at the membrane–solution interface and the presence of a catalytic Pt intermediate in between the two monopolar layers of the BPM.35 In the reproducibility test for the RPSu–Pt IEM systems, a little lesser value was obtained for T. No. of both ions whereas WDE showed a little higher performance when compared with that of its original feed concentration performance. For the PSDVB IEM system, the T. No. of both chloride and sodium ions was observed to be little lesser and the WDE showed a more or less same performance.
![]() | ||
Fig. 10 Variation of CE and energy consumption for various feed concentrations with time for (a) the RPSu–Pt IEM system and (b) the PSDVB IEM system. |
In the case of the RPSu–Pt IEM system, as per Fig. 10(a), the highest CE and energy consumption values observed were about 82.5% and 0.52 W h respectively, for a 15 g L−1 feed concentration. In the case of the PSDVB IEM system the same was observed to be 53% and 1.39 W h respectively, for a 10 g L−1 feed concentration, as shown in Fig. 10(b). From Fig. 10(a), it can be observed that the RPSu–Pt IEM system showed a uniform increase in energy consumption with time until the final stages of experiment without any observation of oscillations with respect to various feed concentrations. This was mainly because of the presence of IER in addition to the fiber reinforcements during the membrane preparation which created an increased electrical resistance. So to overcome this additional resistance, the RPSu–Pt IEM system consumed a little additional energy till the final stage of a performance resulting in the steady increase in value. Whereas from Fig. 10(b), it can be seen that, though PSDVB IEM system displayed oscillations at the initial stages, its increase in energy consumption with time finally depended upon its feed concentration. In terms of the reproducibility test, the CE was observed to be lower and the energy consumption was observed to be higher than its original feed concentration for both types of IEM system because the membrane loses some of its IEC and functional groups depending on its usage time.
![]() | ||
Fig. 11 Variation of current, potential, protons and hydroxyl ion water dissociation fluxes with time for the RPSu–Pt IEM system. |
![]() | ||
Fig. 12 Variation of current, potential, protons and hydroxyl ion water dissociation fluxes with time for the PSDVB IEM system. |
The voltage variation slightly increased with time and attained a maximum value, depending upon feed concentration which then decreased. This is because the production of acid and alkali increased with the increase in voltage to some extent, after which further increase in voltage deteriorated the membrane properties as discussed by Trivedi et al.36 The highest voltage reached by the RPSu–Pt IEM system among the various feed concentrations was about 11.5 V, as per Fig. 11, and it was about 23.1 V for the PSDVB IEM system, as can be seen from Fig. 12. The reason for the observed lower voltage for the RPSu–Pt IEM system can be explained theoretically using a protonation and de-protonation reactions model and the hydrophilicity change in the interface.14
From Fig. 11 and 12, it was clear that for both IEM systems, whether acidic or basic, the water dissociation flux was observed to decrease with time for the various feed concentrations mainly due to the depletion of Na+ ions in the anode compartment and Cl− ions in the cathode compartment. As per Fig. 11, the higher acid and basic fluxes observed were about 37 × 10−6 mol m−2 s−1 and 14.8 × 10−6 mol m−2 s−1 in AC and BC respectively, for the RPSu–Pt IEM system. And for the PSDVB IEM system the same was about 14.8 × 10−6 mol m−2 s−1 and 3.7 × 10−6 mol m−2 s−1, respectively. The greater flux values for the RPSu–Pt IEM system confirmed the catalytic activity of the Pt intermediate layer at the BPM interface by means of hydrogen bonding and polar interaction between Pt and water molecules.
In the reproducibility test in the case of the RPSu–Pt IEM system, both acidic and basic fluxes were observed to be lower than its original feed concentration. Whereas the current change was a higher value and the potential change was also higher at the initial time and then fluctuated around its original feed concentration during the later stage of the performance. In the case of the PSDVB IEM system, the reproducibility test revealed that the dissociation fluxes were lower than the original feed concentration, while both potential and current values were observed to be higher.
Final feed sample solution | Electrical conductivity (mS cm−1) | Salinity (%) | Chloride ion concentration (mg per 100 mL) | Sodium ion concentration (ppm) | ||||
---|---|---|---|---|---|---|---|---|
Initial | Final | Initial | Final | Initial | Final | Initial | Final | |
5 g L−1 | 12.53 | 6.84 | 8.2 | 4.2 | 35.96 | 17.55 | 42.9 | 11.7 |
10 g L−1 | 20.90 | 8.2 | 14.3 | 5.2 | 39.12 | 22.5 | 44.4 | 15.5 |
15 g L−1 | 30.40 | 10.69 | 21.3 | 7.0 | 46.69 | 26.1 | 46.2 | 17.4 |
20 g L−1 | 37.7 | 12.55 | 26.9 | 8.3 | 54.26 | 26.4 | 48.0 | 17.6 |
25 g L−1 | 44.9 | 15.34 | 33.1 | 10.2 | 58.68 | 31.95 | 49.1 | 21.3 |
Reproducibility test (15 g L−1) | 30.40 | 10.68 | 21.3 | 7.1 | 46.69 | 24.3 | 46.2 | 16.2 |
Final feed sample solution | Electrical conductivity (mS cm−1) | Salinity (%) | Chloride ion concentration (mg per 100 mL) | Sodium ion concentration (ppm) | ||||
---|---|---|---|---|---|---|---|---|
Initial | Final | Initial | Final | Initial | Final | Initial | Final | |
5 g L−1 | 12.53 | 11.30 | 8.2 | 7.3 | 35.96 | 20.8 | 42.9 | 24.5 |
10 g L−1 | 20.90 | 19.4 | 14.3 | 13.9 | 39.12 | 27.3 | 44.4 | 38.9 |
15 g L−1 | 30.40 | 27.5 | 21.3 | 19.1 | 46.69 | 30.4 | 46.2 | 44.3 |
20 g L−1 | 37.7 | 31.3 | 26.9 | 22.1 | 54.26 | 37.6 | 48.0 | 45.3 |
25 g L−1 | 44.9 | 36.9 | 33.1 | 26.6 | 58.68 | 43.4 | 49.1 | 46.5 |
Reproducibility test (10 g L−1) | 20.90 | 20.4 | 14.3 | 14.0 | 39.12 | 28.7 | 44.4 | 40.0 |
Footnote |
† Electronic supplementary information (ESI) available. See DOI: 10.1039/c5ra10616h |
This journal is © The Royal Society of Chemistry 2015 |