Brackish Water Desalination by Electrodialysis

8 downloads 64551 Views 2MB Size Report
CaCO3 Scaling Monitoring During Batch ... 273 Soliman 8020 Tunisie, E-mail: [email protected] .... Experiments are performed in batch recirculation.
doi 10.1515/ijcre-2013-0082

International Journal of Chemical Reactor Engineering 2013; 11(1): 1–9

Ilhem Ben Salah Sayadi, Philippe Sistat, Mohamed Ben Amor, and Mohamed Tlili*

Brackish Water Desalination by Electrodialysis: CaCO3 Scaling Monitoring During Batch Recirculation Operation Abstract: The composition of feed water used in electrodialysis (ED) desalination causes the risk of CaCO3 scale crystallization in ED equipments, which leads to the decrease of process efficiency. To control scaling in water systems, several scaling indices and tests, permitting to predict the scaling occurrence, have been devised. In this study, an accelerated scaling test allowing to follow CaCO3 formation in water desalination plant using ED process was proposed. Tests were performed using a pilot unit as a conventional ED in batch recirculation mode. By simple in situ measurements of pH and flow rate of the concentrate, the crystallization onset and growth of CaCO3 were followed during the pilot operation. This leads to precise determination of the number of concentrate recirculation times and therefore to the highest recovery rate without scaling risk as a function of the inlet water quality and the antiscale pretreatment. Keywords: electrodialysis, desalination, calcium carbonate, scaling

*Corresponding author: Mohamed Tlili, Laboratoire de Traitement des Eaux Naturelles, Centre de Recherches et de Technologies des Eaux, BP 273 Soliman 8020 Tunisie, E-mail: mohamed.tlili@certe. rnrt.tn, [email protected] Ilhem Ben Salah Sayadi, Laboratoire de Traitement des Eaux Naturelles, Centre de Recherches et de Technologies des Eaux, BP 273 Soliman 8020 Tunisie, E-mail: [email protected] Philippe Sistat, Institut Européen des Membranes, Montpellier, France, E-mail: [email protected] Mohamed Ben Amor, Laboratoire de Traitement des Eaux Naturelles, Centre de Recherches et de Technologies des Eaux, BP 273 Soliman 8020 Tunisie, E-mail: [email protected]

1 Introduction CaCO3 scaling is a common and a complex phenomenon encountered in desalination, industrial and domestic water plants [1–4]. Aware of its economic and technical impacts, various models and indices have been introduced since the beginning of the nineteenth century to predict water scaling power [5–9]. These attempts to characterize the scaling

potentiality, based on thermodynamic studies of the calcocarbonic system equilibria, have remained insufficient because they are unable to foresee the scaling kinetics and to distinguish the specific role of each foreign dissolved salt [10–12]. Therefore, several precipitation methods called “accelerated scaling techniques”, clearly summarized in the paper of Hui and Lédion [13], were developed to redefine “water scaling power”, to estimate the effect of water composition on the scaling rate and to evaluate the efficiency of antiscale treatments. Their common principle is to induce precipitation by raising the water-pH either by CO2 degasification [14–16] or by OH— generation [17, 18]. The set of these analytical methods, which permit to estimate the scaling risk in circuits under specific conditions, does not allow to define precisely the scaling risk in more complicated systems of water treatment. Thus, in electromembrane desalination plants, several phenomena can coexist leading to the scale formation. Indeed, the polarization concentration, resulting from a concentration gradient established nearby the membrane, increases the scaling probability at the interface membrane solution. Besides, water dissociation, which can occur on the membrane-solution interface [19–21], provides hydroxyl ions responsible for the scale formation inside and/or on the membrane [22]. Over the past 10 years, ED has advanced rapidly because of improved ion exchange membrane properties and advances in technology as the introduction of electrodialysis reversal (EDR) and bipolar membranes. Indeed, ED/EDR has earned a reputation as a membrane desalination process that works economically and reliably on surface water supplies [23–25], reuse water [26–28] and some specific industrial applications [29, 30] when designed and operated properly [31]. Thus, several studies were interested, especially on CaCO3 and/or CaSO4 scale formation risk on the common ED stacks and membranes [32–39]. Recently, Turek et al. [39] have proposed an algorithm for scaling prediction during EDR. The presented algorithm turns out to be successful in case of CaSO4 scale, but not in case of CaCO3 scale prediction because of the complexity of the calcocarbonic system and its dependence on several parameters. This study is a contribution to understand CaCO3 scaling mechanisms during ED desalination. The main

Bereitgestellt von | De Gruyter / TCS Angemeldet | 46.30.84.116 Heruntergeladen am | 14.12.13 14:13

2

I. Ben Salah Sayadi et al.: Brackish Water Desalination by Electrodialysis

purpose is to develop an accelerated scaling test permitting to detect the crystallization onset and growth of CaCO3 during water desalination using ED process by simple in situ measurements of pH and flow rate.

2 Materials and methods 2.1 Work solutions Tested solutions are synthetic brackish water prepared by adding CaCl2 to NaHCO3 solution under CO2 bubbling. The carbonate ions concentration was maintained at 8  10–3 M. However, two different concentrations of calcium were used: 7.5  10−3 and 25  10−3 M. The solutions ionic strength was kept at 845  10−4 M, adjusted by adding NaCl. All reagents are from analytical grade.

2.2 Electrodialysis equipment The ED setup consists of a power DC, a concentrate reservoir, a diluate reservoir, a rinsing electrode reservoir and three pumps (Heidolph D-93309) equipped each with a flow-meter (PCCell GmbH). Figure 1 shows a simplified scheme of ED setup.

The ED cell was a PCCell ED 64-004 (Germany) used as a conventional ED unit with two compartments: the diluate and the concentrate. ED cell is made of two polypropylene blocks supporting electrodes. One electrode is made of Pt/Ir-coated Ti stretched (anode) and the other of Ti stretched metal (cathode). The membranes and spacers are stacked between the two electrode-end blocks. ED stack is formed by 10 repeating sections called cell pairs. A cell pair consists of the following: – cation exchange membrane (PCA-SK) – diluate flow spacer (0.5 mm) – anion exchange membrane (PCA-SA) – concentrate flow spacer (0.5 mm) Spacers are made in plastic and are placed between the membranes to form the flow paths of the diluate and concentrate streams. These spacers are arranged in the stack so that all the diluate and concentrate streams are manifolded separately. For each membrane, the active surface area is 64 cm2. The flow channel width between two membranes is 0.5 mm determined by the thickness of intermembrane spacers. The stack is equipped with three separate external plastic reservoirs: the first serves to concentrate solution, the second to diluate solution and the third to rinse electrode solution. The fluid circulation is achieved using three pumps equipped with flowmeters. Experiments are performed in batch recirculation mode at ambient temperature.

(7)

2.3 Experimental procedure (6)

(5)

(5)

(4)

(1)

(5)

(4)

(2)

(1) Concentrate tank (2) Diluate tank (3) Electrode rinse tank (4) Circulation pumps

(4)

(3)

(5) Flow-meter (6) Electrodialysis cell (7) Current generator

Figure 1 Scheme of the ED installation

During all experiments, the volume of diluate, concentrate and rinsing electrode solution was 1 L each. Rinse solution was 0.1 M Na2SO4 in order to prevent generation of toxic gas. Before the onset of the desalination test, the same solution of brackish water was introduced in diluate and concentrate compartments. Flow rate of electrode rinse solution was fixed at 100 L h−1 throughout the test and for all experiments. For concentrate and diluate solutions, only the initial values of flow rates were fixed (75 L h−1). The experiment starts at time of the potential application (10 V; ~ 1 V/cell). For this potential, the system operates under the limiting current (maximum value fixed by membrane constructor and verified at laboratory scale was 2 V/cell). The solutions pH, conductivity and flow rate were recorded in time. pH was measured with a pH meter (model pH 320/ SET, WTW, Arles, France). Conductivity was measured using conductivity meter (model LF320/SET, WTW, Arles, France). The pH and conductivity electrodes were immersed in each tank (concentrate and diluate).

Bereitgestellt von | De Gruyter / TCS Angemeldet | 46.30.84.116 Heruntergeladen am | 14.12.13 14:13

3

I. Ben Salah Sayadi et al.: Brackish Water Desalination by Electrodialysis

When the conductivity of the product water reaches 0.5  0.1 mS cm−1, a value of high-quality drinking water, voltage is cut and only the dilute solution is replaced by a new working solution. This corresponds to batch duration of 18 min. This operation was repeated several times until detecting stack scaling. After every experiment, ED cell was cleaned with circulation of 0.1 M HCl solution during 30 min followed by three successive rinsing of 15 min each with ultra-pure water.

3 Results and discussions 3.1 Accelerated scaling test For this test, feed solution was brackish water containing 7.5  10−3 M of calcium ions and 8  10−3 M of HCO3—. The pH was fixed at 7.16, corresponding to a supersaturation coefficient (determined using eq. 1) superior to the unity (Ω ≈ 4), for which spontaneous precipitation of calcium carbonate cannot be induced [8, 40–43]. ΩCaCO3 ¼

   γCa2þ  γCO2— ½Ca2þ   CO2— 3 3

ð1Þ

KSPcalcite

where γ, [i] and KSP calcite are the activity coefficients, the ions concentrations and the calcite solubility product, respectively. Figure 2 shows the conductivity values in concentrate and diluate compartments during all desalination runs of

the scaling test. As expected, conductivity values increase with time in the concentrate and decrease in the diluate. During the first desalination run, the conductivity evolution of the concentrate follows a linear pace during the first 10 min with a slope of 0.474 mS cm−1 min−1. The desalination rate is less important during the run remainder and the conductivity curve slope becomes equal to 0.195 mS cm−1 min−1. This behaviour could be explained by high current density at the beginning of the runs when the diluate compartment is still rich in charged species. Note that diluate compartment is emptied and fed after every run. The same behaviour is observed for all runs; however, curve slopes decrease gradually with number of runs: slope values from 0.474 to 0.367 mS cm−1 min−1 for the runs at the beginning of the experiment and from 0.195 to 0.120 mS cm−1 min−1 at the end. This could be due to the hard ionic transit from the diluate to the concentrate when concentrate conductivity progressively increases [44, 45] and/or to the probable precipitation occurrence during the desalination test. The superposition of diluate conductivity curves in Figure 3 shows a slight difference between the first run and the others. This can be explained by the difference between the initial conditions and those reached for the ulterior runs such as the hydrodynamic conditions and the counterion composition of membranes. The perfect superposition of diluate conductivity curves from the second run indicates that ionic species transit does not vary during the different batches of desalination test.

10

30 8 σ Diluate/mS cm–1

σ/ mS cm–1

25 σ Diluate σ Concentrate

20 15 10

1st run 2nd run 3rd run 4th run

6

4

2

5 0

0

0

10

20

30

40 50 Time/min

60

70

80

Figure 2 Conductivity (σ) profiles in the concentrate and the diluate during desalination of synthetic brackish water ([Ca2 þ ] ¼ 7.5  10−3 M; [HCO3–] ¼ 8  10−3 M)

0

2

4

6

8

10 12 Time/min

14

16

18

20

Figure 3 Conductivity (σ) profiles in the diluate during desalination of synthetic brackish water ([Ca2 þ ] ¼ 7.5  10−3 M; [HCO3–] ¼ 8  10−3 M)

Bereitgestellt von | De Gruyter / TCS Angemeldet | 46.30.84.116 Heruntergeladen am | 14.12.13 14:13

4

I. Ben Salah Sayadi et al.: Brackish Water Desalination by Electrodialysis

From these conductivity curves, the scale occurrence in ED pilot unit is not evident in spite of the conductivity curves’ slope decrease of the concentrate versus runs. Generally, precipitation is detected through conductivity decrease. However, in this process, precipitation in concentrate compartment cannot be detected by conductivity decrease since the compartment is enhanced with ions during desalination. Detection is possible only if the rate of precipitation is higher than the rate of desalination. Figure 4 shows experimental pH and flow-rate values recorded in the concentrate, compartment where CaCO3 precipitation could occur with the progress of desalination due to the increase of calcium and hydrogenocarbonate ions. As shown in the graph, pH rises quasi-linearly during the first three runs, reaches a maximum at 7.46 before its continuous drop during the fourth run. This pH reduction is accompanied by a net variation of flow rate which decreases from 75 at the beginning to 67 L h−1 at the end of the test. After four consecutive desalination batches, tanks were emptied and ED stack was dismounted.

78

7.5 pH Flow rate

76 74

7.3 72 7.2 70 7.1

Flow-rate/L h–1

pH Concentrate

7.4

68

1st run

2nd run

3rd run

4th run

7.0

Concentrate Output Concentrate input

Black precipitates

Figure 5 Photo of IEM surface on concentrate side after stack disassembly

from concentrate tank and elementary concentrates. The identification of crystal structure was confirmed by X-ray diffraction. CaCO3 identification permits to explain the pH and flow-rate decrease. At the beginning of the test, pH increase in the concentrate is mainly caused by the rise of HCO3– concentration, since dissolved CO2 concentration is constant: escape or dissolution rate of CO2 at the interface air/solution is insignificant because of a small surface of the tank (78.5 cm2) and a short length of the desalination run (18 min). Considering the second dissociation reaction of carbonic gas (2), it can be observed that both pH and HCO3— rises induce CO32— increase. In addition, calcium ions concentration augments during desalination as well as supersaturation coefficient (eq. 1).

66 0

10

20

30

40 50 Time/min

60

70

80

2— þ HCO— 3 $ CO3 þ H

90

Figure 4 pH and flow-rate profiles in concentrate compartment during desalination of synthetic brackish water ([Ca2 þ ] ¼ 7.5  10−3 M; [HCO3—] ¼ 8  10−3 M)

¼ K2 aCO2— 3

K2

aHCO—3 aHþ

ð2Þ ð3Þ

It is well known that the formed (from Ca2 þ and CO3 ) stable nucleus of CaCO3 can grow in the supersaturated solution following the surface reaction (4) leading to the proton production [40, 43, 46]. At pH levels between 7 and 8, HCO3— represents more than 80% of the total carbon Ct (Ct ¼ [CO2] þ [HCO3—] þ [CO32—]). pH decreases when transformation according to (4) reaction in the concentrate is faster than HCO3– transfer from diluate to concentrate compartments. 2—

Figure 5 shows ionic exchange membrane surface on concentrate side after stack disassembly: blank precipitates are clearly apparent on the concentrate input–output of circulated solutions. Concentrate solutions extracted from concentrate tank (bulk solution) and concentrate side of membranes (membranes surface and gratings of concentrate spacers) were separately filtered through a 0.45 μm filter. SEM photos of precipitates show calcium carbonate under its vaterite (dominant) and calcite shapes (Figure 6). No significant difference was detected between precipitates respectively obtained

2þ $ CaCO3 þ Hþ HCO— 3 þ Ca

ð4Þ

The rapid flow-rate diminution indicates that crystals have reached a size large enough to decrease fluid

Bereitgestellt von | De Gruyter / TCS Angemeldet | 46.30.84.116 Heruntergeladen am | 14.12.13 14:13

5

I. Ben Salah Sayadi et al.: Brackish Water Desalination by Electrodialysis

(a)

S-4800 X 1.50 K

(b)

20.0 um

20.0 um

S-4800 X 1.50 K

Figure 6 SEM photos of precipitates collected at the end of the scaling test (a) from the bulk solution and (b) the membranes surface

3.2 Evaluation tests of water “scaling power” 3.2.1 Effect of calcium content Two brackish waters with the same HCO3— concentration ([HCO3—] ¼ 8  10−3 M) and different calcium concentrations ([Ca2 þ ] ¼ 25  10−3 and 7.5  10−3 M for S1 and S2

7.4 7.2 pH concentrate

circulation in elementary concentrates. This is confirmed by Figure 5 which shows crystal adhesion and/or formation on the input–output of circulated solutions. Precipitation is rapidly detected thanks to the low thickness of elementary concentrates (0.5 mm): crystals are captured on gratings, which form the surface of spacers, and flow-rate decrease is observed. However, formed precipitates do not obstruct ions transfer from diluate to concentrate as previously discussed and shown in Figure 3 depicting diluate conductivity curves. Therefore, we have to distinguish precipitation leading to flow-rate decrease due to crystal formation and/or adherence in input–output of circulated solutions from precipitation leading to ion transit decrease due to membranes scaling. In this study, only the first one is detected and consequently, integrity of PCA-SK and PC-SA membranes is preserved. The similarity of crystals extracted from the concentrate tank and elementary concentrates suggests that nucleation starts in the bulk solution (homogeneous nucleation). As has been shown in the previous study [10] since their formation, calcium carbonate nuclei, since their formation, are shared (in function of the saturation state of the solution) between walls and bulk solution. The crystals formed are then kept by the gratings spacers.

7.0 S2 S1

6.8 6.6 6.4 0

10

20

30

40

50

60

70

80

Time/min Figure 7 pH profiles in concentrate compartment during desalination tests of S1 ([Ca2 þ ] ¼ 25  10−3 M) and S2 ([Ca2 þ ] ¼ 7.5  10−3 M)

solutions, respectively) were tested. The initial pH of feed water was fixed at 7.2 ± 0.05. Figures 7 and 8 describe pH and flow-rate variation during the desalination tests. Experimental data show similar pH variation throughout the first run for S1 and S2 tests. Thus, as previously mentioned, pH behaviour is mainly dependent on carbonate species since the studied solutions have the same carbonate concentration. The pH variation and flow-rate curves’ slopes indicate the scaling occurrence from the second run for S1 and from the fourth one for S2 for maximum pH values of 7.34 and 7.46, respectively. The solution S1 is therefore more scalant than S2 as provided by a prospective calculation of the supersaturation coefficient (eq. 1). Indeed, with less calcium quantity, the supersaturation coefficient must reach a higher value to induce CaCO3 nucleation.

Bereitgestellt von | De Gruyter / TCS Angemeldet | 46.30.84.116 Heruntergeladen am | 14.12.13 14:13

6

I. Ben Salah Sayadi et al.: Brackish Water Desalination by Electrodialysis

7.8

76 74

pH concentrate

7.4

70 68 66

7.2 7.0 6.8

64

6.6

62

6.4

60

6.2

0

10

20

30

40 50 Time/min

60

70

Figure 8 Flow-rate profiles in concentrate compartment during desalination tests of S1 ([Ca2 þ ] ¼ 25  10−3 M) and S2 ([Ca2 þ ] ¼ 7.5  10−3 M)

The pH decline in S1 test is faster than that in S2 with negative curve slopes 3.35  10−2 and 1.88  10−2 pH min−1 for S1 and S2 tests, respectively. This indicates that precipitation rate is higher in the test S1 than that in the test S2. Figure 8 confirms that S1 is more scalant than S2 since their flow rates reach values of 61 and 67 L h−1, respectively. Therefore, the number of desalination batches leading to the complete obstruction of the stack will be more important in the case of the test S2.

3.2.2 Effect of feed water pH Desalination tests were performed using the same solution S1 ([HCO3—] ¼ 8  10−3 M and [Ca2 þ ] ¼ 25  10−3 M) at different initial pH values (pHi): 7.8, 7.2 and 6.2 for which supersaturation coefficients were about 70, 20 and 2, respectively. Note that pH was adjusted by bubbling or CO2 escape and that solution at the highest initial pH was filtered through 0.45 μm few seconds before the launch of the test in order to prevent any CaCO3 nuclei formation in the solution. Figure 9 shows pH profiles during three desalination tests using the same solution at different initial pH. For the pHi 6.2 test, pH gradually increases during the desalination test and reaches the value of 6.9 at the end. For the pHi 7.2 test, pH slowly increases during the first run. At the beginning of the second run, pH values are nearly constant (7.34–7.36) and decrease during the last 4 min of the run. pH decrease is clearly observed during the third run when it reaches a value of 6.72 at its end. During the last

1rst run 0

80

10

3rd run

2nd run 20

30

40 50 Time/min

4th run 60

70

80

Figure 9 Variation of concentrate pH during desalination tests of S1 at different initial pH values (7.8, 7.2 and 6.2)

run, pH decrease is also observed but it is slow. For the last test (pHi 7.8 test), pH decrease is observed since the first run. The decline is very pronounced during the second run and less pronounced during the third and fourth runs. Thus, no scaling was formed during pHi 6.2 test as can also be confirmed by the flow-rate measurements in Figure 10 where we see that the flow rate remains constant. Indeed, in this case, the pH reached at the end of test is not sufficient to initiate CaCO3 nucleation since pH depends on carbonate species in the solution as mentioned previously. However, for pHi 7.2 and pHi 7.8 tests, pH decrease proves CaCO3 formation. This is also supported by the flow-rate measurements in Figure 10. Having the same calcium concentration, solution with the highest pH value is the most scalant

76 74 72 Flow rate/ L h–1

Flow rate/L h–1

72

pHi = 7.8 pHi = 7.2 pHi = 6.2

7.6

S2 S1

pHi = 7.8 pHi = 7.2

70 68

pHi = 6.2

66 64 62 60 58

3rd run

2nd run

1rst run

4th run

56 0

10

20

30

40

50

60

70

80

Time/min Figure 10 Variation of concentrate flow rate during desalination tests of S1 at different initial pH values (7.8, 7.2 and 6.2)

Bereitgestellt von | De Gruyter / TCS Angemeldet | 46.30.84.116 Heruntergeladen am | 14.12.13 14:13

7

I. Ben Salah Sayadi et al.: Brackish Water Desalination by Electrodialysis

S1 S1+0.5mg L–1 STP S1+1mg L–1 STP

7.8 7.6 pH concentrate

one. Flow rate and pH curves are in perfect accordance with these conclusions and confirm the relation between pH and flow-rate decrease with the precipitation occurrence during desalination runs. At the end of each scaling test, concentrate solutions are filtered through a 0.45 μm filter. As expected, precipitates are present and recuperated only for pHi 7.8 and pHi 7.2 tests. Therefore, the CO2 bubbling could be an efficient way to delay calcium carbonate precipitation. As shown in Figure 11, the conductivity profiles indicate precipitation during pHi 7.8 and 7.2 tests while during pHi 6.2 test, the conductivity curve is significantly above the two other. However, conductivity profile does not allow to determine precisely when precipitation is triggered and to differentiate solutions according to their precipitation rate.

7.4 7.2 7.0 6.8 6.6 6.4 0

10

20

30

40

50

60

70

80

Time/min Figure 12 Variation of pH concentrate during desalination tests of S1 (pHi ¼ 7.8) with and without addition of STP (S1; S1 þ 0.5 mg L−1; S1 þ 1 mg L−1)

26 24

22

22

σ/ mS cm–1

20

76

20

Flow rate/L h–1

σ/ms cm

–1

24

18

18

16 40

16

50

60 Time/min

70

80

14 12

pHi = 7.8

10

pHi = 7.2

8

pHi = 6.2

74

S1

72

S1+0.5mg L–1 STP S1+1mg L–1 STP

70 68 66 64 62 60 58 56 0

6 0

10

20

30

40

50

60

70

10

20

30

40 50 Time/min

60

70

80

80

Time/min Figure 11 Variation of concentrate conductivity (σ) during desalination tests of S1 at different initial pH values (7.8, 7.2 and 6.2)

3.2.3 Scale prevention by mineral phosphate additive Figures 12 and 13 illustrate pH and flow-rate progressions during desalination of S1 ([HCO3—] ¼ 8  10−3 M; [Ca2 þ ] ¼ 25  10−3 M and pHi ¼ 7.8) without and with 0.5 or 1 mg L−1 of sodium tripolyphosphate inhibitor (STP: Na5P3O10), respectively. In absence of STP additive, pH measurements show that the CaCO3 nucleation has started since the beginning of the test. This precipitation becomes pronounced from the beginning of the second run where a substantial decrease of flow rate is registered. In presence of STP (S1 þ 0.5 mg L−1 and S1 þ 1 mg L−1), pH variations during desalination tests are similar during the first three runs. This reveals the reproducibility of desalination

Figure 13 Variation of the flow-rate concentrate during desalination tests of S1 (pHi ¼ 7.8) with and without addition of STP (S1; S1 þ 0.5 mg L−1; S1 þ 1 mg L−1)

tests. Nevertheless, in presence of 0.5 mg L−1 of STP, pH remains quasi-constant until the end of the third run and considerably decreases thereafter reflecting CaCO3 precipitation (nucleation and growth). Addition of 0.5 mg L−1 of STP has delayed the nucleation step but did not affect the growth as indicated by flow-rate decrease (Figure 12). The addition of 1 mg L−1 of STP causes a slight and progressive decline of pH throughout the four runs; pH goes from 7.75 to 7.48. This can be attributed to the weak liberation of proton and then to the slight precipitation rate. CaCO3 precipitation is detected through pH evolution but does not affect flow rate which remains constant throughout the desalination as indicated in Figure 12. Further, addition of 1 mg L−1 of STP has not only delayed the nucleation step but also affected the growth of CaCO3 nuclei, avoiding the stack scaling.

Bereitgestellt von | De Gruyter / TCS Angemeldet | 46.30.84.116 Heruntergeladen am | 14.12.13 14:13

8

I. Ben Salah Sayadi et al.: Brackish Water Desalination by Electrodialysis

4 Conclusion In this study, we present a new scaling test, easy to scaleup, leading to identify the CaCO3 scale formation onset conditions and to follow its growth during ED desalination. Conductivity measurements show that ionic species transit was not significantly influenced by precipitation occurrence. Further, membranes were slightly fouled in the experimental conditions of the present investigation. Stack scaling is mainly located on the input–output of circulated solutions in concentrate side leading to flowrate decrease. It has been shown that the presented method is able to determine the number of desalination runs without scaling occurrence and then the highest

recovery ratio from which an antiscale treatment must be imposed. By using this scaling method, it was also possible to compare the scaling power of different brackish waters and to optimize an antiscale treatment. It has been clearly shown that for the same water, a high dissolved CO2 content (lower pH) can play a crucial role in reducing the probability of CaCO3 scale formation. Moreover, results of the chemical inhibition tests when increasing amounts of STP were added into synthetic brackish waters have been very useful to accurately determine the inhibitor content. Indeed, in addition to their detrimental effects on the environment, these chemicals can negatively affect the cost of the produced water if they are not estimated correctly.

References 1. Hawthorn D. Heat transfer: solving scaling problems at the design stage. Chem Eng Res Des 2009;87:193–9. 2. Tlili MM, Korchef A, Ben Amor M. Effect of scalant and antiscalant concentrations on fouling in a solar desalination unit. Chem Eng Process 2007;46:1243–50. 3. Xiaoyan L, Jungang L, Qianya Z, Jinlai F, Yingli L, Jingxin S. The analysis and prediction of scale accumulation for water-injection pipelines in the Daqing Oilfield. J Pet Sci Eng 2009;66:161–4. 4. Yang Q, Liu Y, Gu A, Ding J, Shenz Z. Investigation of calcium carbonate scaling inhibition and scale morphology by AFM. J Colloid Interface Sci 2001;240:608–21. 5. Langelier WF. The analytical control of anticorrosion water treatment. J Am Water Works Assoc 1936;28:1500–23. 6. Ryznar JW. A new index for determining amount of calcium carbonate scale formed by a water. J Am Water Works Assoc 1944;36:472–86. 7. Elfil H, Hannachi A. Reconsidering water scaling tendency assessment. AIChE J 2006;52:3583–91. 8. Gal JY, Fovet Y, Gache N. Mechanisms of scale formation and carbon dioxide partial pressure influence: Part I. Elaboration of an experimental method and a scaling model. Water Res 2002;36:755–63. 9. Poirier G, Legrand L. Théorie des eaux naturelles. Eyrolles, Paris, 1972. 10. Ben Salah I, Tlili MM, Ben Amor M. Influence of foreign salts to the CaCO3-CO2-H2O system and antiscalants on the adherence of calcium carbonate on the stainless steel. Eur J Water Qual 2010;41:51–66. 11. Chen T, Neville A, Yuan M. Assessing the effect of Mg2 þ on CaCO3 scale formation–bulk precipitation and surface deposition. J Cryst Growth 2005;275:1341–7. 12. Meldruma FC, Hyde ST. Morphological influence of magnesium and organic additives on the precipitation of calcite. J Cryst Growth 2001;231:544–58. 13. Hui F, Lédion J. Evaluation methods for the scaling power of water. Eur J Water Qual 2002;33:55–74.

14. Ben Amor M, Zgolli D, Tlili M, Manzola AS. Influence of water hardness, substrate nature and temperature on heterogeneous calcium carbonate nucleation. Desalination 2004;166:79–84. 15. Dedieu L, Hort C, Martin-Dominguez A, Rola M, Roques H. Contribution à l’étude des phénomènes d’entartrage. 1ère partie: Généralités et méthodes d’étude LCGE. Tribune de l’eau 1994;48:3–19. 16. Lédion J, François B, Vienne J. Caractérisation du pouvoir entartrant de l’eau par précipitation contrôlée rapide. J Eur d’hydrol 1997;28:1. 17. Devos O, Gabrielli C, Tlili M, Tribollet B. Nucleation-growth process of scale electrodeposition, influence of the supersaturation. J Electrochem Soc 2003;150:C494–501. 18. Tlili M, Ben Amor M, Gabrielli C, Joiret S, Maurin G, Rousseaua P. Study of electrochemical deposition of CaCO3 by in situ raman spectroscopy: II. Influence of the solution composition. J Electrochem Soc 2003;150:C485–493. 19. Frilette VJ. Preparation and characterization of bipolar ionexchange membranes. J Phys Chem 1956;60:435–9. 20. Kressman TR, Tye FL. The effect of current density on the transport of ions through ions-selective membranes. Discuss Faraday Soc 1956;21:185–292. 21. Tanaka Y. Ion exchange membranes: fundamentals and applications. Membrane Science and Technology Series, vol. 12. Netherlands: Elsevier, 2007. 22. Krol J. Monopolar and bipolar ion exchange membranes. Mass Transport Limitations, Enschede, Netherlands, 1997. 23. Ryabtsev AD, Kotsupalo NP, Titarenko VI, Igumenov IK, Gelfond NV, Fedotova NE, et al. Electrodialysis for production of drinking quality water from artesian waters with salt content up to 8 kg/m3 with productivity up to 1 m3/h. Desalination 2001;136:333–6. 24. Valero F, Arbos R. Desalination of brackish river water using Electrodialysis Reversal (EDR) control of the THMs formation in the Barcelona (NE Spain) area. Desalination 2010;253:170–4.

Bereitgestellt von | De Gruyter / TCS Angemeldet | 46.30.84.116 Heruntergeladen am | 14.12.13 14:13

I. Ben Salah Sayadi et al.: Brackish Water Desalination by Electrodialysis

25. Van Der Hoek JP, Rijnbende DO, Lokin CJ, Bonné PA, Loonen MT, Hofman JA. Electrodialysis as an alternative for reverse osmosis in an integrated membrane system. Desalination 1998;117:159–72. 26. Broens L, Liebrand N, Futselaar H, de Armas JC. Effluent reuse at Barranco Seco (Spain): a 1,000 m3/h case study. Desalination 2004;167:13–16. 27. Chao YM, Liang TM. A feasibility study of industrial wastewater recovery using electrodialysis reversal. Desalination 2008;221:433–9. 28. Zhang Y, Ghyselbrecht K, Meesschaertb B, Pinoy L, Van der Bruggen B. Electrodialysis on RO concentrate to improve water recovery in wastewater reclamation. J Membr Sci 2011;378:101–10. 29. Dalla Costa RF, Klein CW, Bernades AM, Ferreira JZ. Evaluation of the electrodialysis process for the treatment of metal finishing wastewater. J Braz Chem Soc 2002;13:540–7. 30. Pilat BV. Industrial application of electrodialysis reversal systems. Desalination 2003;158:87–9. 31. Valero F, Barceló A, Arbos R. Electrodialysis technology – theory and applications, desalination, trends and technologies. Michael Schorr (Ed.), ISBN: 978-953-307-311-8. Croatia: InTech, 2011. 32. Ayala-Bribiesca E, Araya-Farias M, Pourcelly G, Bazinet L. Effect of concentrate solution pH and mineral composition of a whey protein diluate solution on membrane fouling formation during conventional electrodialysis. J Membr Sci 2006;280:790–801. 33. Ayala-Bribiesca E, Pourcelly G, Bazinet L. Nature identification and morphology characterization of anion-exchange membrane fouling during conventional electrodialysis. J Colloid Interface Sci 2007;308:182–90. 34. Bazinet L, Araya-Farias M. Effect of calcium and carbonate concentrations on cationic membrane fouling during electrodialysis. J Colloid Interface Sci 2005;281:188–96. 35. Bazinet L, Montpetit D, Ippersiel D, Amiot J, Lamarche F. Identification of skim milk electroacidification fouling: a microscopic approach. J Colloid Interface Sci 2001;237:62–9.

9

36. Shaposhnik VA, Zubets NN, Strygina IP, Mill BE. High demineralization of drinking water by electrodialysis without scaling on the membranes. Desalination 2002;145:329–32. 37. Turek M, Dydo P. Electrodialysis reversal of calcium sulphate and calcium carbonate supersaturated solution. Desalination 2003;158:91–4. 38. Turek M, Dydo P, Was J. High efficiency electrodialysis reversal of concentrated calcium sulfate and calcium carbonate solutions. Desalination 2007;205:62–6. 39. Turek M, Waś J, Mitko K. Scaling prediction in electrodialytic desalination. Desalination Water Treat 2012;44:255–60. 40. Elfil H, Roques H. Contribution à l’étude des mécanismes de l’entartrage. Étude de la signification physique de la droite limite de précipitation. Tribune de l’eau 1999;599:3. 41. Gal JY, Bollinger JC, Tolosa H, Gache N. Calcium carbonate solubility: a reappraisal of scale formation and inhibition. Talanta 1996;43:1497–509. 42. Legrand L, Poirier G, Leroy P. Les équilibres carboniques et l’équilibre calcocarbonique dans les eaux naturelles. Eyrolles, Paris, 1981. 43. Tarits C. Étude des variations de composition isotopique du carbone et de l’oxygène dans le domaine de la précipitation spontanée homogène du carbonate de calcium. Thèse, Université René Descartes, Paris V, 1990. 44. Bourdillon C. Transport d’eau dans les membranes échangeuses de cations II. Influence des couches de polarisation en convection naturelle. J Chem Phys 1974;71:819–27. 45. Ripoll C, Demarty M. Transport d’eau dans les membranes échangeuses de cations III, Modèle pour l’interprétation de l’influence des couches de polarisation sur les nombres de transport d’eau. J Chem Phys 1974;71:828–35. 46. Tlili M, Ben Amor M, Gabrielli C, Joiret S, Maurin G. On the initial stages of calcium carbonate precipitation. Eur J Water Qual 2006;37:89–108.

Bereitgestellt von | De Gruyter / TCS Angemeldet | 46.30.84.116 Heruntergeladen am | 14.12.13 14:13