- Email: [email protected]
Boron removal in water using a hybrid membrane process of ion exchange resin and microfiltration without continuous resin addition
Journal of Water Process Engineering 17 (2017) 32–39
Contents lists available at ScienceDirect
Journal of Water Process Engineering journal homepage: www.elsevier.com/locate/jwpe
Boron removal in water using a hybrid membrane process of ion exchange resin and microfiltration without continuous resin addition Assma Alharati, Yousef Swesi, Koffi Fiaty, Catherine Charcosset ∗ Laboratoire d’Automatique et de Génie des Procédés (LAGEP), Université Claude Bernard Lyon 1, 43 bd du 11 Novembre 1918, Bâtiment CPE, 69622 Villeurbanne Cedex, France
a r t i c l e
i n f o
Article history: Received 13 October 2016 Received in revised form 27 February 2017 Accepted 6 March 2017 Keywords: Boron removal Ion exchange Amberlite IRA743 resin Microfiltration Hybrid membrane process Regeneration
a b s t r a c t Boron contained in seawater and some natural ground waters is harmful for plants and humans, especially water obtained by reverse osmosis desalination can contain a high level of boron. In this study, we investigated a hybrid process for boron removal from water which associates sorption on ion exchange resin in batch and microfiltration. In this approach, the ion exchange resin Amberlite IRA 743 was first ground to a mean particle size of 40–60 m to increase its kinetics. A ceramic microfiltration membrane was used to retain the ion exchange resin in the feed tank and the circulation loop while the model solution of boron was continuously added and the permeate collected for analysis. The effect of resin dosage, boron initial concentration, transmembrane pressure and membrane pore size was studied. A concentration below 0.3 mg/L was obtained during a long time at sufficient resin dosage. The results were analyzed in terms of volume treated at breakthrough and permeate flux. In particular, it was shown that the transmembrane pressure and membrane pore size increased the permeate flux, but also decreased the volume treated at breakthrough, probably due to insufficient residence time. Moreover, the ion exchange resin and the microfiltration membrane were efficiently regenerated with HCl (0.37%) followed by NaOH (1%) and reused. Overall, it is suggested that the hybrid process of ion exchange resin in batch and microfiltration may be a possible technique for boron removal. © 2017 Elsevier Ltd. All rights reserved.
1. Introduction In seawater, boron is present mainly as boric acid B(OH)3 [1]. Boron is an essential micronutrient for humans, animals and plants, however, it can become toxic if the amount is slightly higher than required [2–4]. For humans, boron toxicity depends on the length, frequency, and level of exposure. A chronic exposure of boron may cause adverse effects such as cutaneous disorders, retarded growth and negative impact on reproduction [1,5]. The level set by the World Health Organization (WHO) was 0.3 mg/L until 1998, which was increased to 0.5 mg/L and then to 2.4 mg/L in 2011 [6]. For plants, the degree of damage depends upon time, concentration, crop sensitivity and crop water use, and if damage is severe enough, crop yield is reduced. It is usually admitted that boron concentration in irrigation water should not exceed 0.3–4 mg/L depending on the plant and soil characteristics [5]. Due to the growing demand of drinking and irrigation water over the world, associated to population increase, the production
∗ Corresponding author. E-mail address: [email protected] (C. Charcosset). http://dx.doi.org/10.1016/j.jwpe.2017.03.002 2214-7144/© 2017 Elsevier Ltd. All rights reserved.
of high quality water at low cost is highly needed. For that purpose, reverse osmosis (RO) seawater desalination is increasingly implemented throughout the world especially in the Middle East and North Africa regions [1]. The average boron concentration in the oceans and seas is around 5 mg/L. However, boron removal by RO is often insufficient under common RO conditions due to the small size and lack of charge of the boric acid molecule that may diffuse through the RO membrane. The membranes available at present for seawater desalination reject about 60–90% of boron [7]. Higher rejections as high as 96% have been reported for specific conditions such as a two-pass configuration and a new generation of RO membrane with higher boron rejection. However, there is still a strong need to improve the efficiency of conventional RO and to develop alternative techniques such as ion exchange and hybrid membrane systems [8]. A major technique to remove boron makes use of boron specific ion exchange resins [9]. Commercially available boron selective resins are based on a macroporous polystyrene matrix treated by chloromethylation and amination with N-methyl glucamine. The ion exchange resins exhibit good removal efficiency of boron, even at very low concentrations. They are specifically designed and used to remove boric acid and borate from water, magnesium brine or
A. Alharati et al. / Journal of Water Process Engineering 17 (2017) 32–39
other solutions under a variety of conditions. Most studies on boron removal using ion exchange resins were performed in batch or fixed bed column to measure the kinetic and capacity for boron removal. Several kinetic equations have been proposed, for example the Ho pseudo-second-order kinetic model [10,11], a second-order pseudo-homogeneous reaction model [12], and a combination of film and particle diffusion [13]. In recent years, hybrid membrane processes have been increasingly reported for water treatment [14]. These processes associate membrane filtration (microfiltration (MF), ultrafiltration (UF), nanofiltration (NF) or (RO)) to other techniques such as precipitation, ion exchange, sorption, etc, with the purpose to increase the size of the species retained by the membrane. For boron removal, hybrid membrane processes include polymer flocculation associated to UF or MF, activated carbon associated to UF or MF, and specific ion exchange resin combined to UF or MF [15]. In this last process, fine ion exchange resins are used to increase the binding area, leading to a higher sorption capacity and improved kinetics. The advantages of the hybrid membrane technique are then related to its high efficiency and low pressure drop [16–18]. Several configurations of specific ion exchange resins associated to MF have been investigated [19]. In a first configuration, a fresh relatively diluted suspension of ion exchange resin was fed into the feed tank to which the solution to be treated was continuously added; simultaneously partly saturated resin was withdrawn for regeneration by UF or MF to keep the feed volume constant [18,20–22]. UF or MF were operated in a submerged configuration [20,21] or in a circulation loop [18,22]. Different parameters have been investigated such as resin concentration in the feed suspension, flowrates of fresh and saturated resin and permeate flowrate [20,21]. At optimum conditions, boron concentration was found below the required level (0.3 mg/L or 1 mg/L) during a long period of operation. From results obtained with model solutions and sea waters, the process was applied to the removal of boron from geothermal waters. For example, the boron concentration was reduced from 11.0 mg/L to ≤1 mg/L in 20 min by using 2 g resin/L geothermal water with the ion exchange resin Dowex-XUS 43594.00 ground to an average particle size of 20 m [22]. Two other configurations have been tested with less attention. In the first configuration, the feed was mixed with a concentrated suspension of ion exchange resin and flew through a MF loop [23]. The permeate obtained by MF was recovered and the ion exchange suspension entered the regeneration cycle with desorption and removal of boron. Regenerated adsorbent was then returned back to the sorption process. In another approach, the feed solution flew through a well-mixed tank containing the ion exchange resin and the permeate obtained by MF was recovered [24,25]. However, the boron concentration was found much higher than required because large ion exchange resin (diameter 190 and 530 m) was used.
33
In our study, boron removal is evaluated using this last approach with fine ion exchange resin. The effects of several parameters (boron concentration, transmembrane pressure, resin dosage, and membrane pore size) are investigated on the variation of boron concentration in the permeate and permeate flux versus time. In a second part, the simultaneous regeneration of resin and membrane is performed and evaluated in terms of boron concentration in the permeate and permeate flux. The main purpose of this study is to evaluate the hybrid process using microfiltration and ion exchange resin, without continuous addition of resin. With no resin addition, the process is simple to perform and the resin can be used until a high saturation is reached. Another advantage is that the resin and the membrane can be regenerated simultaneously in the same experimental set-up. 2. Materials and methods 2.1. Chemicals The chemicals used in this work were boric acid (99.97%) and Amberlite IRA 743 boron selective ion exchange resin supplied by Sigma-Aldrich (France). Boric acid solutions were prepared by dissolving the appropriate amount of boric acid (0.751 g) in 1 L of demineralized water, the concentration of this solution being 100 mg/L. Appropriate concentrations were obtained from this stock solution. For the regeneration of resin and membrane, hydrochloride acid (HCl 37%) and sodium hydroxide (NaOH) were supplied by Sigma-Aldrich (France). For the analysis of boron concentration, azomethine-H and other reactants were supplied by Sigma-Aldrich (France). 2.2. Preparation of fine ion exchange resins The Amberlite IRA743 resin has a size between 300 and 720 m. From this resin, a fine ground resin with an average particle size between 40 and 60 m was obtained using a planetary ball mill (PM 100, Retsch, France), followed by sieving on a vibratory sieve shaker (AS 200, Retsch, France). The size distribution of the resin was obtained using a Mastersizer 2000 (Malvern, France) which measures the intensity of scattered light as a laser beam passes through a sample of dispersed particles. 2.3. Sorption/MF set-up ®
The experimental set-up included a Micro Kerasep membrane device (Novasep, France) as shown in Fig. 1. The boron solution was continuously added to the 3 L reactor using a Quattroflow 150S pump (Pall, France). The resin suspension was recirculated in a
Fig. 1. Experimental set-up of the ion exchange/MF hybrid system, P: pressure gauge, Qp : permeate flowrate, Qi : inlet flowrate of the boron solution.
34
A. Alharati et al. / Journal of Water Process Engineering 17 (2017) 32–39
Table 1 Summary of the parameters used and results obtained. Resin dosage (g/L)
Boron concentration (mg/L)
Transmembrane pressure (bar)
Membrane pore size (m)
Limiting level (mg/L)
Time of treatment (min)
Volume of water treated at breakthrough (L)
Volume of water treated at breakthrough L/g resin
1 1.66 2.33 3.33 2.33 2.33 3.33 3.33 3.33 2.33 2.33 3.33 2.33
5 5 5 5 5 5 3 7 10 5 5 5 5
1 1 1 1 1 1 1 1 1 2 3 1 1
0.1 0.1 0.1 0.1 0.1 0.1 0.1 0.1 0.1 0.1 0.1 0.45 0.45
0.3 0.3 0.3 0.3 1 2.4 0.3 0.3 0.3 0.3 0.3 0.3 0.3
– – 105 >150 150 >150 »130 135 – 60 30 >75 40
– – 3.5 >5.0 5.0 »5.0 »5.0 3.8 – 3.2 2.25 >5.0 2.6
– – 0.5 >0.5 0.7 »0.7 »0.5 0.4 – 0.45 0.3 >0.5 0.4
closed loop using a Quattroflow 1000S pump (Pall, France). Two pressure gauges were placed at the inlet and outlet of the module, and a valve at the outlet for increasing the transmembrane pressure. A mechanic stirrer (RW 20, IKA-Werke, France) was used to stir continuously the feed suspension of ion exchange resin in the reactor (200 rpm). ® The Kerasep ceramic membrane is tubular, with an outside diameter of 10 mm, an inner diameter of 6 mm and a length of 40 cm; the active membrane area is therefore 0.0075 m2 . The active layer is made of ZrO2 -TiO2 deposited on a monolithic TiO2 /Al2 O3 support. Two membranes were tested with respectively 0.1 and 0.45 m pore size.
2.4. Sorption/MF experiment The boron solution was delivered to the reactor at a flowrate equal to the permeate flowrate in order to maintain a constant volume in the reactor. The process was operated in a closed loop configuration, the suspension being continuously returned to the feed tank. The feed flowrate was set to 5.45 L/min, corresponding to a mean tangential velocity of 3.2 m/s in the tubular membrane.The pH in the reactor was continuously measured (SevenMulti pH meter, Mettler Toledo, France) and drops of NaOH 1 M solution were added to maintain a pH value around 8.2. The temperature of the suspension in the reactor was set to 20–22 ◦ C. At regular time intervals, the permeate flowrate was measured. The permeate flux was obtained by dividing the flowrate by the membrane area. Permeate samples were collected at regular time intervals for analysis of boron concentration by the azomethine-H method [26]. After adding appropriate reactants, the adsorbance was measured at 420 nm on a visible–UV spectrophotometer Cary 50 Probe (Agilent Technologies, France). To evaluate the process performance, the volume treated was evaluated at the breakthrough point, i.e. when the boron concentration in the permeate reached 0.3 mg/L. To investigate the effect of resin dosage, the volume treated at breakthrough was divided by the amount of resin used. Parameters and results are summarized in Table 1. After each experiment, the ion exchange resin was removed by flushing the experimental set-up with water in an open loop configuration. The membrane and the experimental set-up were then cleaned by flushing successively with acid, water, base, water in a closed loop configuration where the permeate was removed. The protocol can be summarized as follows: (1) Open loop configuration (to remove the ion exchange resin): • 5 L deionized water at T = 25 ◦ C
(2) Closed loop configuration: • • • •
2 L 1.85% HCl during 15 min at T = 50 ◦ C 5 L deionized water during 15 min at T = 45 ◦ C 2 L 5% NaOH during 15 min at T = 50◦ C 5 L deionized water during 15 min at T = 45 ◦ C
After each cleaning cycle, the membrane permeability was checked to be close to its initial value (more than 95%). 2.5. Simultaneous regeneration of ion exchange resin and MF membrane In the last set of experiments, the ion exchange resin and the MF membrane were regenerated simultaneously. For that purpose, the ion exchange resin was kept in the reactor. The resin and membrane were then cleaned by flushing successively with water, acid, water, base, and water in a closed loop configuration where the permeate was removed. At each step, MF was run until the suspension volume in the reactor was around 1 L. The following protocol was used: Closed loop configuration: • • • • •
V1 V2 V1 V2 V1
(L) deionized water during 15 min at T = 25 ◦ C (L) 0.37% HCl during 15 min at T = 50 ◦ C (L) deionized water during 15 min at T = 45 ◦ C (L) 1% NaOH (ou 5%) during 15 min at T = 50◦ C (L) deionized water during 15 min at T = 45 ◦ C
where V1 is the volume of deionized water and V2 the volume of acid or base. Different regeneration cycles were tested (with different values of V1 and V2 ) to evaluate the regeneration and reuse of the boron selective ion exchange resin and the MF membrane. 3. Results First, the effect of several parameters including the resin dosage, initial boron concentration, transmembrane pressure, and membrane pore size were investigated. For each experiment, the boron concentration in the permeate and the permeate flux were measured versus time. Then, the regeneration of the ion exchange resin and MF membrane were performed simultaneously and optimized. 3.1. Effect of resin dosage The effect of resin dosage in the reactor was studied for values between 1 and 3.33 g/L. The variation of boron concentration in the permeate and permeate flux are shown in Fig. 2a and b, respec-
A. Alharati et al. / Journal of Water Process Engineering 17 (2017) 32–39
35
Fig. 2. Effect of resin dosage. Experimental conditions: initial boron concentration 5 mg/L, transmembrane pressure 1 bar, membrane pore diameter 0.1 m.
tively. The concentration of boron in the permeate decreased as boron was retained by the ion exchange resin and then increased progressively as the resin binding sites became more saturated. From Fig. 2a, it can be observed that sorption took place very fast. For the higher resin dosages of 2.33 and 3.33 g/L, the concentration in the permeate decreased from 5 mg/L to 0.35 and 0.19 mg/L, respectively, after 3 min. As reported previously (i.e. Darwish et al. [27]), the fast kinetics of boron removal obtained with fine ground resin suggests that intra-particle diffusion is almost negligible. As expected, an increase in resin dosage in the feed tank reduced the boron concentration in the permeate. The increase in boron uptake with the resin dosage is explained by the increase in the surface area and thus the number of binding sites on the resin surface [13,27]. The resin dosage had a major effect on the volume of boron solution that could be treated before reaching the limit level (breakthrough). Volumes treated at breakthrough are summarized in Table 1. For the experimental conditions used, a resin dosage below 1.66 g/L was not sufficient to reach the concentration of 0.3 mg/L (C/C0 = 0.6). With the resin dosages of 2.33 g/L and 3.33 g/L, the treatment of 3.5 L and >5.0 L of boron solution was possible, respectively, corresponding to 0.5 and » 0.5 L of boron solution/g resin. Therefore, the higher resin dosage gave a higher volume treated at breakthrough per mass of resin and was then more favorable. It can be noted that the choice of a limiting concentration greatly influenced the volume treated (Fig. 2a and Table 1). For example, at a resin dosage of 2.33 g/L, volumes to be treated were respectively equal to 3.5, 5.0 and »5.0 L before a boron level of 0.3, 1 and 2.4 mg/L, respectively. As expected, a higher limiting level implies a higher volume treated at breakthrough. This suggests that high resin dosage will be adequate for the treatment of large volumes of water. The process investigated is similar to that proposed by Holdich et al. [25] and called seeded MF. In this study, the resin was put in a stirred cell device equipped with a flat micro engineered membrane with 8 m pore size, the boron solution was continuously added to the stirred cell whereas the permeate was eliminated. However, the boron concentration in the permeate remained high and the concentration did not reach the required limit of 0.3 mg/L, which was explained by to the large size of the resin (530 and 190 m). Our study shows that the seeded adsorption/MF method with resins with small sizes of 40–60 m gave low boron concentration and could be then an alternative to other hybrid/sorption processes.
Breakthrough occurs at a certain time in batch operations as observed by Jia et al. [28] using powdered active carbon for trace organics removal. A new batch of ion exchange resin or activated carbon is required to be added into the system to restore the water treatment capacity. While for continuous addition of resin or active carbon, breakthrough can be avoided at optimal conditions, and once the water quality is reached, the operation can proceed on without changing system parameter. Another difference observed by the same authors [28] is that for the batch mode, the trace organic concentration in the product water reaches the lowest value very soon after the dosing of activated carbon. This was also observed in the present study using ion exchange resin. While for the continuous dosing mode, the lowest target organic concentration was approached gradually because of the less carbon mass at the beginning of the operation. The filtration flux reached a steady state value after a few minutes (Fig. 2b), the transitory regime corresponding to the development of a cake layer at the membrane surface [29]. The steady state obtained suggests low internal membrane fouling. Indeed, ion exchange particles are much larger than the membrane pore size (40–60 m compared to 0.1 m) which prevents them for being trapped by the membrane pores; also the relative low resin dosages used are favorable to low membrane fouling. In another set of experiments (data not shown), we checked that boron alone did not cause fouling and was not retained by the MF membrane in the absence of ion exchange resin. It may be also noted that the permeate flux obtained at the lowest resin dosage (1 g/L) was slightly lower than those obtained at higher resin dosages (240 instead of 265 L/h m2 ). During crossflow microfiltration, a layer of particles accumulates on top of the membrane forming a cake layer [30,31]. The properties of this cake layer are modified by phenomena such as particles compaction and/or agglomeration, which can change the permeate flux data. The slightly lower permeate flux observed at 1 g/L of resin may be due to the change in the cake layer formed at the membrane surface. Indeed, at this lower resin dosage, the resin becomes more saturated and this could make the cake layer more compact leading to a decrease in permeate flux.
3.2. Effect of boron concentration The effect of boron concentration was investigated between 3 and 10 mg/L. Indeed, boron concentration in seawater is around
36
A. Alharati et al. / Journal of Water Process Engineering 17 (2017) 32–39
Fig. 3. Effect of the initial boron concentration. Experimental conditions: resin dosage 3.33 g/L, transmembrane pressure 1 bar, membrane pore diameter 0.1 m.
5 mg/L, higher concentrations are found in groundwater and geothermal waters, and lower boron concentrations are obtained in seawaters treated by RO. The variation versus time are shown in Fig. 3a and b, respectively for boron concentration in the permeate and permeate flux. At the different initial concentration, the variation of boron concentration in the permeate showed similar trend (Fig. 3a). At 10 mg/L, the concentration was higher than 0.3 mg/L during almost all the experiment, while this concentration was reached after »5.0 L, >5.0 L, and 3.8 L, respectively at 3, 5 and 7 mg/L. Therefore, the volume treated at breakthrough was higher at low boron concentration. At all boron concentrations, the permeate flux reached a stable value after a few minutes suggesting again low internal membrane fouling. Moreover, a decrease in permeate flux was observed at higher boron concentration. As previously mentioned, a cake layer is formed at the membrane surface during crossflow microfiltration. At higher boron concentration of the feed solution, the resin becomes more saturated which may lead to changes in the cake layer properties compaction and to lower the permeate flux. It can be also mentioned that boron alone does not cause a decrease in permeate flux, that is to say membrane fouling (data not shown). This may be explained by the size of the borate ion B(OH)−4 which is such smaller than the membrane pore size.
3.3. Effect of transmembrane pressure The effect of transmembrane pressure was studied for 1, 2 and 3 bar. The variation of boron concentration in the permeate and permeate flux are shown respectively in Fig. 4a and b. At 1, 2 and 3 bar, the concentration of 0.3 mg/L was reached for permeate volumes of 3.5 L, 3.2 and 2.25 L, respectively (Table 1). This suggests that a high transmembrane pressure (leading to a high permeate flux) may be inadequate for the hybrid membrane process. Indeed, the boron solution is introduced in the reactor at the same flowrate as the permeate flowrate; the residence time in the reactor may be then not sufficient for boron removal by the ion exchange resin, and leads to a lower volume treated at breakthrough. Fig. 4b shows the effect of transmembrane pressure on the permeate flux variation versus time. At all pressures, the permeate flux reached a steady state value after a few minutes. As mentioned above, internal membrane fouling was rather low. As expected, the permeate flux increased with pressure: 265 L/h/m2 at 1 bar and 597 L/h/m2 at 3 bar. Higher permeate fluxes are preferable as they decrease processing time. However, as mentioned above, a higher permeate flux was associated to a decrease in the volume treated at breakthrough.
Fig. 4. Effect of transmembrane pressure. Experimental conditions: initial boron concentration 5 mg/L, membrane pore diameter 0.1 m, of resin dosage 2.33 g/L.
A. Alharati et al. / Journal of Water Process Engineering 17 (2017) 32–39
37
Fig. 5. Effect of the membrane pore size. Experimental conditions: initial boron concentration 5 mg/L, transmembrane pressure 1 bar, dosage of resin 2.33 g/L.
3.4. Effect of membrane pore size
3.5. Regeneration of the hybrid system (resin + membrane)
In order to see the effect of membrane pore size, a MF membrane with pore size 0.45 m was tested and the results compared to those obtained with the 0.1 m pore size membrane. The boron concentration and permeate flux variation versus time are shown respectively in Fig. 5a and b. For the 0.1 m and 0.45 m pore size membranes, the concentration of 0.3 mg/L was reached after 3.5 and 2.6 L of permeate (Table 1). The decrease in volume treated at breakthrough for the 0.45 m pore size membrane may due to the low residence time, as the permeate flux was much higher for this membrane (Fig. 5b). This effect was similar to the influence of transmembrane pressure, where the higher transmembrane pressure led to lower volumes treated at breakthrough. The comparison between both membranes was also performed at a higher resin dosage (3.33 g/L instead of 2.33 g/L). The boron concentration and permeate flux variation versus time are then shown respectively in Fig. 6a and b. For this higher resin dosage, the boron concentration in the permeate did not rise sufficiently to see an effect of membrane pore size. For both membranes, the concentration of 0.3 mg/L was reached after >5.0 L of permeate (Table 1). This confirms the effect of resin dosage, where the volume treated at breakthrough was found to increase with an increase in resin dosage.
Due to the high price of ion exchange resin, their regeneration is a mandatory step in water treatment. Generally, the regeneration process consists of two main steps, removal of boron from the resin using acid (HCl) followed by neutralization of the resins with a basic solution (NaOH) [32]. The regeneration is reported to be very effective. For example, Darwish et al. [27] used the ion exchange resin during 5 cycles of sorption-regeneration and showed that the resin was efficiently reused even if it was previously ground by mechanical grinding. In this study, we make use of the fact that similar regenerations (water, acid, water, base, water) are used for both the ion exchange resin and the membrane. Both resin and membrane were regenerated simultaneously in the experimental set-up shown in Fig. 1. Fig. 7a–c present data on boron sorption by the ion exchange resin in the hybrid process, before and after regeneration with HCl and NaOH solutions at different volumes. The acid and alkali solutions were heated as usually done for membrane regeneration. A first sorption of boron solution was performed, followed by regeneration, and finally a new sorption of boron solution. Fig. 7a–c show respectively boron concentration in the permeate before and after regeneration for decreasing volumes of acid and base, respectively V = 1 L, 0.5 L, and 0.2 L. Higher volumes and
Fig. 6. Effect of the membrane pore size. Experimental conditions: initial boron concentration 5 mg/L, transmembrane pressure 1 bar, dosage of resin 3.33 g/L.
38
A. Alharati et al. / Journal of Water Process Engineering 17 (2017) 32–39
Fig. 7. Effect of volumes of acid and base (V1 ) and water (V2 ) used for regeneration (a) V1 = 1 L, V2 = 3 L, (b) V1 = 0.5 L, V2 = 2 L, (c) V1 = 0.2 L, V2 = 1 L. Other conditions for regeneration are specified in the Materials and Methods section. For boron sorption: initial boron concentration 5 mg/L, transmembrane pressure 1 bar, membrane pore diameter 0.1 m, dosage of resin 3.33 g/L.
higher concentrations of acid and base were also tested, but much lower boron removal was obtained after regeneration suggesting that these high amounts of acid and base may lead to the decrease in resin capacity. As shown in Fig. 7, it can be seen that the regeneration was efficient for both V1 = 1 L and 0.5 L, whereas the capacity of the ion exchange resin was found lower at volumes of 0.2 L of acid and base. The volume 0.5 L was then chosen as it was the lower volume found to regenerate efficiently the resin. In addition, the permeate flux was checked to be the same before and after regeneration (data not shown). In addition, the regeneration was checked to be effective when the resin has reached completed saturation (Fig. 8). For that, an experiment was conducted by passing 25 L of boron solution, until the boron concentration in the permeate was equal to its initial concentration of 5 mg/L. The regeneration was then performed and a new experiment of boron removal was conducted. The permeate flux and boron concentration in the permeate were found very similar before and after regeneration. This suggests that the regeneration protocol, especially the acid and base volumes (V = 0.5 L), could be used for the treatment of larger water volumes. In a last study, the sorption-regeneration cycle was performed 3 times (Fig. 9). For that, sorption and regeneration were run at identical conditions, and the permeate flux and boron concentration were measured after each cycle. The results showed that membrane flux (data not shown) and boron concentration were the same before
and after regeneration for the 3 cycles. Overall, the obtained results proved that the resin and the membrane can be efficiently reused.
Fig. 8. Regeneration after complete saturation of the resin by 25 L of boron solution. Conditions for regeneration are specified in the Materials and Methods section, with: V1 = 0.5 L, V2 = 2 L. For boron sorption: initial boron concentration 5 mg/L, transmembrane pressure 1 bar, membrane pore diameter 0.1 m, dosage of resin 3.33 g/L.
A. Alharati et al. / Journal of Water Process Engineering 17 (2017) 32–39
Fig. 9. Regeneration after 3 cycles. Conditions for regeneration are specified in the Materials and Methods section, with: V1 = 0.5 L, V2 = 2 L. For boron sorption: initial boron concentration 5 mg/L, transmembrane pressure 1 bar, membrane pore diameter 0.1 m, dosage of resin 3.33 g/L.
4. Conclusion In this study, a hybrid process of ion exchange resin in batch and MF was shown to be effective for boron removal below a level of 0.3 mg/L without adding fresh resin into the system. Several parameters were studied to optimize the process in terms of boron removal and permeate flux. An increase in transmembrane pressure and membrane pore size led to a higher permeate flux. However, this positive effect was at the expense of the volume treated at breakthrough which was found to decrease at higher permeate flux. Moreover, an increase in the resin dosage led to an increase in volume treated at breakthrough; however, this effect has to be evaluated at larger volumes of water. Lastly, the ion exchange resin and the membrane were effectively regenerated and reused at optimum amounts of HCl followed by NaOH for reconditioning. For industrial applications, it is foreseen that large membrane devices and reactors will have to be used. To run continuously the process, two similar plants could be installed, running alternatively for boron removal and ion exchange resin/membrane regeneration. It is suggested that this hybrid membrane process could be a possible alternative to other hybrid ion exchange/membrane technique. However, further studies are needed to investigate boron removal from seawater and for much larger volumes. Acknowledgement We thank the Libyan Embassy in France to provide a grant to Mrs Assma Alharati during her PhD thesis. References [1] N. Hilal, G.J. Kim, C. Somerfield, Boron removal from saline water: a comprehensive review, Desalination 273 (1) (2011) 23–35. [2] F.S. Kot, Chapter 1 − boron in the environment, in: N.K.B. Hilal (Ed.), Boron Separation Processes, Elsevier, Amsterdam, 2015, pp. 1–33.
39
[3] Y. Duydu, N. Bas¸aran, H.M. Bolt, Chapter 3 − risk assessment of borates in occupational settings, in: N.K.B. Hilal (Ed.), Boron Separation Processes, Elsevier, Amsterdam, 2015, pp. 65–105. [4] F.H. Nielsen, Update on human health effects of boron, J. Trace Elem. Med. Biol. 28 (4) (2014) 383–387. [5] J. Wolska, M. Bryjak, Methods for boron removal from aqueous solutions – a review, Desalination 310 (0) (2013) 18–24. [6] World Health Organization, Boron in Drinking-water: Background Document for Development of WHO Guidelines for Drinking-water Quality, 2009. [7] C. Jacob, Seawater desalination: boron removal by ion exchange technology, Desalination 205 (1–3) (2007) 47–52. [8] A.B. Koltuniewicz, A. Witek, K. Bezak, Efficiency of membrane-sorption integrated processes, J. Membr. Sci. 239 (1) (2004) 129–141. [9] E. Güler, et al., Boron removal from seawater: state-of-the-art review, Desalination 356 (2015) 85–93. [10] N. Kabay, et al., Removal of boron from seawater by selective ion exchange resins, React. Funct. Polym. 67 (12) (2007) 1643–1650. [11] N. Kabay, et al., Removal of boron from SWRO permeate by boron selective ion exchange resins containing N-methyl glucamine groups, Desalination 223 (1–3) (2008) 49–56. [12] R. Boncukcuoˇglu, et al., An empirical model for kinetics of boron removal from boron-containing wastewaters by ion exchange in a batch reactor, Desalination 160 (2) (2004) 159–166. [13] N. Öztürk, T.E. Köse, Boron removal from aqueous solutions by ion-exchange resin: batch studies, Desalination 227 (1–3) (2008) 233–240. [14] M. Pontié, C. Charcosset, Seawater, brackish waters and natural waters treatment with hybrid membrane processes, in: A. Basile, C. Charcosset (Eds.), Integrated Membrane Systems and Processes, Wiley, 2016. [15] M. Bryjak, N. Kabay, Chapter 10 – boron removal from water by sorption–membrane filtration hybrid process, in: N.K.B. Hilal (Ed.), Boron Separation Processes, Elsevier, Amsterdam, 2015, pp. 237–248. [16] M. Bryjak, J. Wolska, N. Kabay, Removal of boron from seawater by adsorption–membrane hybrid process: implementation and challenges, Desalination 223 (1–3) (2008) 57–62. [17] N. Kabay, et al., Removal of boron from aqueous solutions by a hybrid ion exchange–membrane process, Desalination 198 (1–3) (2006) 158–165. [18] N. Kabay, et al., Adsorption-membrane filtration (AMF) hybrid process for boron removal from seawater: an overview, Desalination 223 (1–3) (2008) 38–48. ˇ [19] M. Blahuˇsiak, S. Schlosser, N. Kabay, Chapter 16 – hybrid adsorption–microfiltration process with plug flow of microparticulate adsorbent for boron removal, in: N.K.B. Hilal (Ed.), Boron Separation Processes, Elsevier, Amsterdam, 2015, pp. 339–354. [20] N. Kabay, et al., An innovative integrated system for boron removal from geothermal water using RO process and ion exchange-ultrafiltration hybrid method, Desalination 316 (2013) 1–7. [21] N. Kabay, et al., Coupling ion exchange with ultrafiltration for boron removal from geothermal water-investigation of process parameters and recycle tests, Desalination 316 (2013) 17–22. [22] S. Samatya, et al., Utilization of geothermal water as irrigation water after boron removal by monodisperse nanoporous polymers containing NMDG in sorption–ultrafiltration hybrid process, Desalination 364 (2015) 62–67. ˇ Schlosser, Simulation of the adsorption—microfiltration pro[23] M. Blahuˇsiak, S. cess for boron removal from RO permeate, Desalination 241 (1–3) (2009) 156–166. [24] I. Yilmaz Ipek, et al., Kinetic behaviour of boron selective resins for boron removal using seeded microfiltration system, React. Funct. Polym. 67 (12) (2007) 1628–1634. [25] R.G. Holdich, I.W. Cumming, S. Perni, Boron mass transfer during seeded microfiltration, Chem. Eng. Res. Des. 84 (1) (2006) 60–68. [26] AFNOR, Dosage du bore par spectrométrie d’absorption moléculaire, Méthodes à l’azométhine H, in Qualité de l’eau, ANFOR, 2001, 2017, p. T 90–041. [27] N.B. Darwish, V. Kochkodan, N. Hilal, Boron removal from water with fractionized Amberlite IRA743 resin, Desalination 370 (2015) 1–6. [28] Y. Jia, R. Wang, A.G. Fane, Hybrid PAC-submerged membrane system for trace organics removal II: system simulation and application study, Chem. Eng. J. 149 (1–3) (2009) 42–49. [29] B. Onderková, et al., Microfiltration of suspensions of microparticulate boron adsorbent through a ceramic membrane, Desalination 241 (1) (2009) 148–155. [30] I.H. Huisman, G. Trägårdh, C. Trägårdh, Particle transport in crossflow microfiltration – II. Effects of particle–particle interactions, Chem. Eng. Sci. 54 (2) (1999) 281–289. [31] G. Belfort, R.H. Davis, A.L. Zydney, The behavior of suspensions and macromolecular solutions in crossflow microfiltration, J. Membr. Sci. 96 (1–2) (1994) 1–58. [32] Y. Xu, J.-Q. Jiang, Technologies for boron removal, Ind. Eng. Chem. Res. 47 (1) (2008) 16–24.
Contents lists available at ScienceDirect
Journal of Water Process Engineering journal homepage: www.elsevier.com/locate/jwpe
Boron removal in water using a hybrid membrane process of ion exchange resin and microfiltration without continuous resin addition Assma Alharati, Yousef Swesi, Koffi Fiaty, Catherine Charcosset ∗ Laboratoire d’Automatique et de Génie des Procédés (LAGEP), Université Claude Bernard Lyon 1, 43 bd du 11 Novembre 1918, Bâtiment CPE, 69622 Villeurbanne Cedex, France
a r t i c l e
i n f o
Article history: Received 13 October 2016 Received in revised form 27 February 2017 Accepted 6 March 2017 Keywords: Boron removal Ion exchange Amberlite IRA743 resin Microfiltration Hybrid membrane process Regeneration
a b s t r a c t Boron contained in seawater and some natural ground waters is harmful for plants and humans, especially water obtained by reverse osmosis desalination can contain a high level of boron. In this study, we investigated a hybrid process for boron removal from water which associates sorption on ion exchange resin in batch and microfiltration. In this approach, the ion exchange resin Amberlite IRA 743 was first ground to a mean particle size of 40–60 m to increase its kinetics. A ceramic microfiltration membrane was used to retain the ion exchange resin in the feed tank and the circulation loop while the model solution of boron was continuously added and the permeate collected for analysis. The effect of resin dosage, boron initial concentration, transmembrane pressure and membrane pore size was studied. A concentration below 0.3 mg/L was obtained during a long time at sufficient resin dosage. The results were analyzed in terms of volume treated at breakthrough and permeate flux. In particular, it was shown that the transmembrane pressure and membrane pore size increased the permeate flux, but also decreased the volume treated at breakthrough, probably due to insufficient residence time. Moreover, the ion exchange resin and the microfiltration membrane were efficiently regenerated with HCl (0.37%) followed by NaOH (1%) and reused. Overall, it is suggested that the hybrid process of ion exchange resin in batch and microfiltration may be a possible technique for boron removal. © 2017 Elsevier Ltd. All rights reserved.
1. Introduction In seawater, boron is present mainly as boric acid B(OH)3 [1]. Boron is an essential micronutrient for humans, animals and plants, however, it can become toxic if the amount is slightly higher than required [2–4]. For humans, boron toxicity depends on the length, frequency, and level of exposure. A chronic exposure of boron may cause adverse effects such as cutaneous disorders, retarded growth and negative impact on reproduction [1,5]. The level set by the World Health Organization (WHO) was 0.3 mg/L until 1998, which was increased to 0.5 mg/L and then to 2.4 mg/L in 2011 [6]. For plants, the degree of damage depends upon time, concentration, crop sensitivity and crop water use, and if damage is severe enough, crop yield is reduced. It is usually admitted that boron concentration in irrigation water should not exceed 0.3–4 mg/L depending on the plant and soil characteristics [5]. Due to the growing demand of drinking and irrigation water over the world, associated to population increase, the production
∗ Corresponding author. E-mail address: [email protected] (C. Charcosset). http://dx.doi.org/10.1016/j.jwpe.2017.03.002 2214-7144/© 2017 Elsevier Ltd. All rights reserved.
of high quality water at low cost is highly needed. For that purpose, reverse osmosis (RO) seawater desalination is increasingly implemented throughout the world especially in the Middle East and North Africa regions [1]. The average boron concentration in the oceans and seas is around 5 mg/L. However, boron removal by RO is often insufficient under common RO conditions due to the small size and lack of charge of the boric acid molecule that may diffuse through the RO membrane. The membranes available at present for seawater desalination reject about 60–90% of boron [7]. Higher rejections as high as 96% have been reported for specific conditions such as a two-pass configuration and a new generation of RO membrane with higher boron rejection. However, there is still a strong need to improve the efficiency of conventional RO and to develop alternative techniques such as ion exchange and hybrid membrane systems [8]. A major technique to remove boron makes use of boron specific ion exchange resins [9]. Commercially available boron selective resins are based on a macroporous polystyrene matrix treated by chloromethylation and amination with N-methyl glucamine. The ion exchange resins exhibit good removal efficiency of boron, even at very low concentrations. They are specifically designed and used to remove boric acid and borate from water, magnesium brine or
A. Alharati et al. / Journal of Water Process Engineering 17 (2017) 32–39
other solutions under a variety of conditions. Most studies on boron removal using ion exchange resins were performed in batch or fixed bed column to measure the kinetic and capacity for boron removal. Several kinetic equations have been proposed, for example the Ho pseudo-second-order kinetic model [10,11], a second-order pseudo-homogeneous reaction model [12], and a combination of film and particle diffusion [13]. In recent years, hybrid membrane processes have been increasingly reported for water treatment [14]. These processes associate membrane filtration (microfiltration (MF), ultrafiltration (UF), nanofiltration (NF) or (RO)) to other techniques such as precipitation, ion exchange, sorption, etc, with the purpose to increase the size of the species retained by the membrane. For boron removal, hybrid membrane processes include polymer flocculation associated to UF or MF, activated carbon associated to UF or MF, and specific ion exchange resin combined to UF or MF [15]. In this last process, fine ion exchange resins are used to increase the binding area, leading to a higher sorption capacity and improved kinetics. The advantages of the hybrid membrane technique are then related to its high efficiency and low pressure drop [16–18]. Several configurations of specific ion exchange resins associated to MF have been investigated [19]. In a first configuration, a fresh relatively diluted suspension of ion exchange resin was fed into the feed tank to which the solution to be treated was continuously added; simultaneously partly saturated resin was withdrawn for regeneration by UF or MF to keep the feed volume constant [18,20–22]. UF or MF were operated in a submerged configuration [20,21] or in a circulation loop [18,22]. Different parameters have been investigated such as resin concentration in the feed suspension, flowrates of fresh and saturated resin and permeate flowrate [20,21]. At optimum conditions, boron concentration was found below the required level (0.3 mg/L or 1 mg/L) during a long period of operation. From results obtained with model solutions and sea waters, the process was applied to the removal of boron from geothermal waters. For example, the boron concentration was reduced from 11.0 mg/L to ≤1 mg/L in 20 min by using 2 g resin/L geothermal water with the ion exchange resin Dowex-XUS 43594.00 ground to an average particle size of 20 m [22]. Two other configurations have been tested with less attention. In the first configuration, the feed was mixed with a concentrated suspension of ion exchange resin and flew through a MF loop [23]. The permeate obtained by MF was recovered and the ion exchange suspension entered the regeneration cycle with desorption and removal of boron. Regenerated adsorbent was then returned back to the sorption process. In another approach, the feed solution flew through a well-mixed tank containing the ion exchange resin and the permeate obtained by MF was recovered [24,25]. However, the boron concentration was found much higher than required because large ion exchange resin (diameter 190 and 530 m) was used.
33
In our study, boron removal is evaluated using this last approach with fine ion exchange resin. The effects of several parameters (boron concentration, transmembrane pressure, resin dosage, and membrane pore size) are investigated on the variation of boron concentration in the permeate and permeate flux versus time. In a second part, the simultaneous regeneration of resin and membrane is performed and evaluated in terms of boron concentration in the permeate and permeate flux. The main purpose of this study is to evaluate the hybrid process using microfiltration and ion exchange resin, without continuous addition of resin. With no resin addition, the process is simple to perform and the resin can be used until a high saturation is reached. Another advantage is that the resin and the membrane can be regenerated simultaneously in the same experimental set-up. 2. Materials and methods 2.1. Chemicals The chemicals used in this work were boric acid (99.97%) and Amberlite IRA 743 boron selective ion exchange resin supplied by Sigma-Aldrich (France). Boric acid solutions were prepared by dissolving the appropriate amount of boric acid (0.751 g) in 1 L of demineralized water, the concentration of this solution being 100 mg/L. Appropriate concentrations were obtained from this stock solution. For the regeneration of resin and membrane, hydrochloride acid (HCl 37%) and sodium hydroxide (NaOH) were supplied by Sigma-Aldrich (France). For the analysis of boron concentration, azomethine-H and other reactants were supplied by Sigma-Aldrich (France). 2.2. Preparation of fine ion exchange resins The Amberlite IRA743 resin has a size between 300 and 720 m. From this resin, a fine ground resin with an average particle size between 40 and 60 m was obtained using a planetary ball mill (PM 100, Retsch, France), followed by sieving on a vibratory sieve shaker (AS 200, Retsch, France). The size distribution of the resin was obtained using a Mastersizer 2000 (Malvern, France) which measures the intensity of scattered light as a laser beam passes through a sample of dispersed particles. 2.3. Sorption/MF set-up ®
The experimental set-up included a Micro Kerasep membrane device (Novasep, France) as shown in Fig. 1. The boron solution was continuously added to the 3 L reactor using a Quattroflow 150S pump (Pall, France). The resin suspension was recirculated in a
Fig. 1. Experimental set-up of the ion exchange/MF hybrid system, P: pressure gauge, Qp : permeate flowrate, Qi : inlet flowrate of the boron solution.
34
A. Alharati et al. / Journal of Water Process Engineering 17 (2017) 32–39
Table 1 Summary of the parameters used and results obtained. Resin dosage (g/L)
Boron concentration (mg/L)
Transmembrane pressure (bar)
Membrane pore size (m)
Limiting level (mg/L)
Time of treatment (min)
Volume of water treated at breakthrough (L)
Volume of water treated at breakthrough L/g resin
1 1.66 2.33 3.33 2.33 2.33 3.33 3.33 3.33 2.33 2.33 3.33 2.33
5 5 5 5 5 5 3 7 10 5 5 5 5
1 1 1 1 1 1 1 1 1 2 3 1 1
0.1 0.1 0.1 0.1 0.1 0.1 0.1 0.1 0.1 0.1 0.1 0.45 0.45
0.3 0.3 0.3 0.3 1 2.4 0.3 0.3 0.3 0.3 0.3 0.3 0.3
– – 105 >150 150 >150 »130 135 – 60 30 >75 40
– – 3.5 >5.0 5.0 »5.0 »5.0 3.8 – 3.2 2.25 >5.0 2.6
– – 0.5 >0.5 0.7 »0.7 »0.5 0.4 – 0.45 0.3 >0.5 0.4
closed loop using a Quattroflow 1000S pump (Pall, France). Two pressure gauges were placed at the inlet and outlet of the module, and a valve at the outlet for increasing the transmembrane pressure. A mechanic stirrer (RW 20, IKA-Werke, France) was used to stir continuously the feed suspension of ion exchange resin in the reactor (200 rpm). ® The Kerasep ceramic membrane is tubular, with an outside diameter of 10 mm, an inner diameter of 6 mm and a length of 40 cm; the active membrane area is therefore 0.0075 m2 . The active layer is made of ZrO2 -TiO2 deposited on a monolithic TiO2 /Al2 O3 support. Two membranes were tested with respectively 0.1 and 0.45 m pore size.
2.4. Sorption/MF experiment The boron solution was delivered to the reactor at a flowrate equal to the permeate flowrate in order to maintain a constant volume in the reactor. The process was operated in a closed loop configuration, the suspension being continuously returned to the feed tank. The feed flowrate was set to 5.45 L/min, corresponding to a mean tangential velocity of 3.2 m/s in the tubular membrane.The pH in the reactor was continuously measured (SevenMulti pH meter, Mettler Toledo, France) and drops of NaOH 1 M solution were added to maintain a pH value around 8.2. The temperature of the suspension in the reactor was set to 20–22 ◦ C. At regular time intervals, the permeate flowrate was measured. The permeate flux was obtained by dividing the flowrate by the membrane area. Permeate samples were collected at regular time intervals for analysis of boron concentration by the azomethine-H method [26]. After adding appropriate reactants, the adsorbance was measured at 420 nm on a visible–UV spectrophotometer Cary 50 Probe (Agilent Technologies, France). To evaluate the process performance, the volume treated was evaluated at the breakthrough point, i.e. when the boron concentration in the permeate reached 0.3 mg/L. To investigate the effect of resin dosage, the volume treated at breakthrough was divided by the amount of resin used. Parameters and results are summarized in Table 1. After each experiment, the ion exchange resin was removed by flushing the experimental set-up with water in an open loop configuration. The membrane and the experimental set-up were then cleaned by flushing successively with acid, water, base, water in a closed loop configuration where the permeate was removed. The protocol can be summarized as follows: (1) Open loop configuration (to remove the ion exchange resin): • 5 L deionized water at T = 25 ◦ C
(2) Closed loop configuration: • • • •
2 L 1.85% HCl during 15 min at T = 50 ◦ C 5 L deionized water during 15 min at T = 45 ◦ C 2 L 5% NaOH during 15 min at T = 50◦ C 5 L deionized water during 15 min at T = 45 ◦ C
After each cleaning cycle, the membrane permeability was checked to be close to its initial value (more than 95%). 2.5. Simultaneous regeneration of ion exchange resin and MF membrane In the last set of experiments, the ion exchange resin and the MF membrane were regenerated simultaneously. For that purpose, the ion exchange resin was kept in the reactor. The resin and membrane were then cleaned by flushing successively with water, acid, water, base, and water in a closed loop configuration where the permeate was removed. At each step, MF was run until the suspension volume in the reactor was around 1 L. The following protocol was used: Closed loop configuration: • • • • •
V1 V2 V1 V2 V1
(L) deionized water during 15 min at T = 25 ◦ C (L) 0.37% HCl during 15 min at T = 50 ◦ C (L) deionized water during 15 min at T = 45 ◦ C (L) 1% NaOH (ou 5%) during 15 min at T = 50◦ C (L) deionized water during 15 min at T = 45 ◦ C
where V1 is the volume of deionized water and V2 the volume of acid or base. Different regeneration cycles were tested (with different values of V1 and V2 ) to evaluate the regeneration and reuse of the boron selective ion exchange resin and the MF membrane. 3. Results First, the effect of several parameters including the resin dosage, initial boron concentration, transmembrane pressure, and membrane pore size were investigated. For each experiment, the boron concentration in the permeate and the permeate flux were measured versus time. Then, the regeneration of the ion exchange resin and MF membrane were performed simultaneously and optimized. 3.1. Effect of resin dosage The effect of resin dosage in the reactor was studied for values between 1 and 3.33 g/L. The variation of boron concentration in the permeate and permeate flux are shown in Fig. 2a and b, respec-
A. Alharati et al. / Journal of Water Process Engineering 17 (2017) 32–39
35
Fig. 2. Effect of resin dosage. Experimental conditions: initial boron concentration 5 mg/L, transmembrane pressure 1 bar, membrane pore diameter 0.1 m.
tively. The concentration of boron in the permeate decreased as boron was retained by the ion exchange resin and then increased progressively as the resin binding sites became more saturated. From Fig. 2a, it can be observed that sorption took place very fast. For the higher resin dosages of 2.33 and 3.33 g/L, the concentration in the permeate decreased from 5 mg/L to 0.35 and 0.19 mg/L, respectively, after 3 min. As reported previously (i.e. Darwish et al. [27]), the fast kinetics of boron removal obtained with fine ground resin suggests that intra-particle diffusion is almost negligible. As expected, an increase in resin dosage in the feed tank reduced the boron concentration in the permeate. The increase in boron uptake with the resin dosage is explained by the increase in the surface area and thus the number of binding sites on the resin surface [13,27]. The resin dosage had a major effect on the volume of boron solution that could be treated before reaching the limit level (breakthrough). Volumes treated at breakthrough are summarized in Table 1. For the experimental conditions used, a resin dosage below 1.66 g/L was not sufficient to reach the concentration of 0.3 mg/L (C/C0 = 0.6). With the resin dosages of 2.33 g/L and 3.33 g/L, the treatment of 3.5 L and >5.0 L of boron solution was possible, respectively, corresponding to 0.5 and » 0.5 L of boron solution/g resin. Therefore, the higher resin dosage gave a higher volume treated at breakthrough per mass of resin and was then more favorable. It can be noted that the choice of a limiting concentration greatly influenced the volume treated (Fig. 2a and Table 1). For example, at a resin dosage of 2.33 g/L, volumes to be treated were respectively equal to 3.5, 5.0 and »5.0 L before a boron level of 0.3, 1 and 2.4 mg/L, respectively. As expected, a higher limiting level implies a higher volume treated at breakthrough. This suggests that high resin dosage will be adequate for the treatment of large volumes of water. The process investigated is similar to that proposed by Holdich et al. [25] and called seeded MF. In this study, the resin was put in a stirred cell device equipped with a flat micro engineered membrane with 8 m pore size, the boron solution was continuously added to the stirred cell whereas the permeate was eliminated. However, the boron concentration in the permeate remained high and the concentration did not reach the required limit of 0.3 mg/L, which was explained by to the large size of the resin (530 and 190 m). Our study shows that the seeded adsorption/MF method with resins with small sizes of 40–60 m gave low boron concentration and could be then an alternative to other hybrid/sorption processes.
Breakthrough occurs at a certain time in batch operations as observed by Jia et al. [28] using powdered active carbon for trace organics removal. A new batch of ion exchange resin or activated carbon is required to be added into the system to restore the water treatment capacity. While for continuous addition of resin or active carbon, breakthrough can be avoided at optimal conditions, and once the water quality is reached, the operation can proceed on without changing system parameter. Another difference observed by the same authors [28] is that for the batch mode, the trace organic concentration in the product water reaches the lowest value very soon after the dosing of activated carbon. This was also observed in the present study using ion exchange resin. While for the continuous dosing mode, the lowest target organic concentration was approached gradually because of the less carbon mass at the beginning of the operation. The filtration flux reached a steady state value after a few minutes (Fig. 2b), the transitory regime corresponding to the development of a cake layer at the membrane surface [29]. The steady state obtained suggests low internal membrane fouling. Indeed, ion exchange particles are much larger than the membrane pore size (40–60 m compared to 0.1 m) which prevents them for being trapped by the membrane pores; also the relative low resin dosages used are favorable to low membrane fouling. In another set of experiments (data not shown), we checked that boron alone did not cause fouling and was not retained by the MF membrane in the absence of ion exchange resin. It may be also noted that the permeate flux obtained at the lowest resin dosage (1 g/L) was slightly lower than those obtained at higher resin dosages (240 instead of 265 L/h m2 ). During crossflow microfiltration, a layer of particles accumulates on top of the membrane forming a cake layer [30,31]. The properties of this cake layer are modified by phenomena such as particles compaction and/or agglomeration, which can change the permeate flux data. The slightly lower permeate flux observed at 1 g/L of resin may be due to the change in the cake layer formed at the membrane surface. Indeed, at this lower resin dosage, the resin becomes more saturated and this could make the cake layer more compact leading to a decrease in permeate flux.
3.2. Effect of boron concentration The effect of boron concentration was investigated between 3 and 10 mg/L. Indeed, boron concentration in seawater is around
36
A. Alharati et al. / Journal of Water Process Engineering 17 (2017) 32–39
Fig. 3. Effect of the initial boron concentration. Experimental conditions: resin dosage 3.33 g/L, transmembrane pressure 1 bar, membrane pore diameter 0.1 m.
5 mg/L, higher concentrations are found in groundwater and geothermal waters, and lower boron concentrations are obtained in seawaters treated by RO. The variation versus time are shown in Fig. 3a and b, respectively for boron concentration in the permeate and permeate flux. At the different initial concentration, the variation of boron concentration in the permeate showed similar trend (Fig. 3a). At 10 mg/L, the concentration was higher than 0.3 mg/L during almost all the experiment, while this concentration was reached after »5.0 L, >5.0 L, and 3.8 L, respectively at 3, 5 and 7 mg/L. Therefore, the volume treated at breakthrough was higher at low boron concentration. At all boron concentrations, the permeate flux reached a stable value after a few minutes suggesting again low internal membrane fouling. Moreover, a decrease in permeate flux was observed at higher boron concentration. As previously mentioned, a cake layer is formed at the membrane surface during crossflow microfiltration. At higher boron concentration of the feed solution, the resin becomes more saturated which may lead to changes in the cake layer properties compaction and to lower the permeate flux. It can be also mentioned that boron alone does not cause a decrease in permeate flux, that is to say membrane fouling (data not shown). This may be explained by the size of the borate ion B(OH)−4 which is such smaller than the membrane pore size.
3.3. Effect of transmembrane pressure The effect of transmembrane pressure was studied for 1, 2 and 3 bar. The variation of boron concentration in the permeate and permeate flux are shown respectively in Fig. 4a and b. At 1, 2 and 3 bar, the concentration of 0.3 mg/L was reached for permeate volumes of 3.5 L, 3.2 and 2.25 L, respectively (Table 1). This suggests that a high transmembrane pressure (leading to a high permeate flux) may be inadequate for the hybrid membrane process. Indeed, the boron solution is introduced in the reactor at the same flowrate as the permeate flowrate; the residence time in the reactor may be then not sufficient for boron removal by the ion exchange resin, and leads to a lower volume treated at breakthrough. Fig. 4b shows the effect of transmembrane pressure on the permeate flux variation versus time. At all pressures, the permeate flux reached a steady state value after a few minutes. As mentioned above, internal membrane fouling was rather low. As expected, the permeate flux increased with pressure: 265 L/h/m2 at 1 bar and 597 L/h/m2 at 3 bar. Higher permeate fluxes are preferable as they decrease processing time. However, as mentioned above, a higher permeate flux was associated to a decrease in the volume treated at breakthrough.
Fig. 4. Effect of transmembrane pressure. Experimental conditions: initial boron concentration 5 mg/L, membrane pore diameter 0.1 m, of resin dosage 2.33 g/L.
A. Alharati et al. / Journal of Water Process Engineering 17 (2017) 32–39
37
Fig. 5. Effect of the membrane pore size. Experimental conditions: initial boron concentration 5 mg/L, transmembrane pressure 1 bar, dosage of resin 2.33 g/L.
3.4. Effect of membrane pore size
3.5. Regeneration of the hybrid system (resin + membrane)
In order to see the effect of membrane pore size, a MF membrane with pore size 0.45 m was tested and the results compared to those obtained with the 0.1 m pore size membrane. The boron concentration and permeate flux variation versus time are shown respectively in Fig. 5a and b. For the 0.1 m and 0.45 m pore size membranes, the concentration of 0.3 mg/L was reached after 3.5 and 2.6 L of permeate (Table 1). The decrease in volume treated at breakthrough for the 0.45 m pore size membrane may due to the low residence time, as the permeate flux was much higher for this membrane (Fig. 5b). This effect was similar to the influence of transmembrane pressure, where the higher transmembrane pressure led to lower volumes treated at breakthrough. The comparison between both membranes was also performed at a higher resin dosage (3.33 g/L instead of 2.33 g/L). The boron concentration and permeate flux variation versus time are then shown respectively in Fig. 6a and b. For this higher resin dosage, the boron concentration in the permeate did not rise sufficiently to see an effect of membrane pore size. For both membranes, the concentration of 0.3 mg/L was reached after >5.0 L of permeate (Table 1). This confirms the effect of resin dosage, where the volume treated at breakthrough was found to increase with an increase in resin dosage.
Due to the high price of ion exchange resin, their regeneration is a mandatory step in water treatment. Generally, the regeneration process consists of two main steps, removal of boron from the resin using acid (HCl) followed by neutralization of the resins with a basic solution (NaOH) [32]. The regeneration is reported to be very effective. For example, Darwish et al. [27] used the ion exchange resin during 5 cycles of sorption-regeneration and showed that the resin was efficiently reused even if it was previously ground by mechanical grinding. In this study, we make use of the fact that similar regenerations (water, acid, water, base, water) are used for both the ion exchange resin and the membrane. Both resin and membrane were regenerated simultaneously in the experimental set-up shown in Fig. 1. Fig. 7a–c present data on boron sorption by the ion exchange resin in the hybrid process, before and after regeneration with HCl and NaOH solutions at different volumes. The acid and alkali solutions were heated as usually done for membrane regeneration. A first sorption of boron solution was performed, followed by regeneration, and finally a new sorption of boron solution. Fig. 7a–c show respectively boron concentration in the permeate before and after regeneration for decreasing volumes of acid and base, respectively V = 1 L, 0.5 L, and 0.2 L. Higher volumes and
Fig. 6. Effect of the membrane pore size. Experimental conditions: initial boron concentration 5 mg/L, transmembrane pressure 1 bar, dosage of resin 3.33 g/L.
38
A. Alharati et al. / Journal of Water Process Engineering 17 (2017) 32–39
Fig. 7. Effect of volumes of acid and base (V1 ) and water (V2 ) used for regeneration (a) V1 = 1 L, V2 = 3 L, (b) V1 = 0.5 L, V2 = 2 L, (c) V1 = 0.2 L, V2 = 1 L. Other conditions for regeneration are specified in the Materials and Methods section. For boron sorption: initial boron concentration 5 mg/L, transmembrane pressure 1 bar, membrane pore diameter 0.1 m, dosage of resin 3.33 g/L.
higher concentrations of acid and base were also tested, but much lower boron removal was obtained after regeneration suggesting that these high amounts of acid and base may lead to the decrease in resin capacity. As shown in Fig. 7, it can be seen that the regeneration was efficient for both V1 = 1 L and 0.5 L, whereas the capacity of the ion exchange resin was found lower at volumes of 0.2 L of acid and base. The volume 0.5 L was then chosen as it was the lower volume found to regenerate efficiently the resin. In addition, the permeate flux was checked to be the same before and after regeneration (data not shown). In addition, the regeneration was checked to be effective when the resin has reached completed saturation (Fig. 8). For that, an experiment was conducted by passing 25 L of boron solution, until the boron concentration in the permeate was equal to its initial concentration of 5 mg/L. The regeneration was then performed and a new experiment of boron removal was conducted. The permeate flux and boron concentration in the permeate were found very similar before and after regeneration. This suggests that the regeneration protocol, especially the acid and base volumes (V = 0.5 L), could be used for the treatment of larger water volumes. In a last study, the sorption-regeneration cycle was performed 3 times (Fig. 9). For that, sorption and regeneration were run at identical conditions, and the permeate flux and boron concentration were measured after each cycle. The results showed that membrane flux (data not shown) and boron concentration were the same before
and after regeneration for the 3 cycles. Overall, the obtained results proved that the resin and the membrane can be efficiently reused.
Fig. 8. Regeneration after complete saturation of the resin by 25 L of boron solution. Conditions for regeneration are specified in the Materials and Methods section, with: V1 = 0.5 L, V2 = 2 L. For boron sorption: initial boron concentration 5 mg/L, transmembrane pressure 1 bar, membrane pore diameter 0.1 m, dosage of resin 3.33 g/L.
A. Alharati et al. / Journal of Water Process Engineering 17 (2017) 32–39
Fig. 9. Regeneration after 3 cycles. Conditions for regeneration are specified in the Materials and Methods section, with: V1 = 0.5 L, V2 = 2 L. For boron sorption: initial boron concentration 5 mg/L, transmembrane pressure 1 bar, membrane pore diameter 0.1 m, dosage of resin 3.33 g/L.
4. Conclusion In this study, a hybrid process of ion exchange resin in batch and MF was shown to be effective for boron removal below a level of 0.3 mg/L without adding fresh resin into the system. Several parameters were studied to optimize the process in terms of boron removal and permeate flux. An increase in transmembrane pressure and membrane pore size led to a higher permeate flux. However, this positive effect was at the expense of the volume treated at breakthrough which was found to decrease at higher permeate flux. Moreover, an increase in the resin dosage led to an increase in volume treated at breakthrough; however, this effect has to be evaluated at larger volumes of water. Lastly, the ion exchange resin and the membrane were effectively regenerated and reused at optimum amounts of HCl followed by NaOH for reconditioning. For industrial applications, it is foreseen that large membrane devices and reactors will have to be used. To run continuously the process, two similar plants could be installed, running alternatively for boron removal and ion exchange resin/membrane regeneration. It is suggested that this hybrid membrane process could be a possible alternative to other hybrid ion exchange/membrane technique. However, further studies are needed to investigate boron removal from seawater and for much larger volumes. Acknowledgement We thank the Libyan Embassy in France to provide a grant to Mrs Assma Alharati during her PhD thesis. References [1] N. Hilal, G.J. Kim, C. Somerfield, Boron removal from saline water: a comprehensive review, Desalination 273 (1) (2011) 23–35. [2] F.S. Kot, Chapter 1 − boron in the environment, in: N.K.B. Hilal (Ed.), Boron Separation Processes, Elsevier, Amsterdam, 2015, pp. 1–33.
39
[3] Y. Duydu, N. Bas¸aran, H.M. Bolt, Chapter 3 − risk assessment of borates in occupational settings, in: N.K.B. Hilal (Ed.), Boron Separation Processes, Elsevier, Amsterdam, 2015, pp. 65–105. [4] F.H. Nielsen, Update on human health effects of boron, J. Trace Elem. Med. Biol. 28 (4) (2014) 383–387. [5] J. Wolska, M. Bryjak, Methods for boron removal from aqueous solutions – a review, Desalination 310 (0) (2013) 18–24. [6] World Health Organization, Boron in Drinking-water: Background Document for Development of WHO Guidelines for Drinking-water Quality, 2009. [7] C. Jacob, Seawater desalination: boron removal by ion exchange technology, Desalination 205 (1–3) (2007) 47–52. [8] A.B. Koltuniewicz, A. Witek, K. Bezak, Efficiency of membrane-sorption integrated processes, J. Membr. Sci. 239 (1) (2004) 129–141. [9] E. Güler, et al., Boron removal from seawater: state-of-the-art review, Desalination 356 (2015) 85–93. [10] N. Kabay, et al., Removal of boron from seawater by selective ion exchange resins, React. Funct. Polym. 67 (12) (2007) 1643–1650. [11] N. Kabay, et al., Removal of boron from SWRO permeate by boron selective ion exchange resins containing N-methyl glucamine groups, Desalination 223 (1–3) (2008) 49–56. [12] R. Boncukcuoˇglu, et al., An empirical model for kinetics of boron removal from boron-containing wastewaters by ion exchange in a batch reactor, Desalination 160 (2) (2004) 159–166. [13] N. Öztürk, T.E. Köse, Boron removal from aqueous solutions by ion-exchange resin: batch studies, Desalination 227 (1–3) (2008) 233–240. [14] M. Pontié, C. Charcosset, Seawater, brackish waters and natural waters treatment with hybrid membrane processes, in: A. Basile, C. Charcosset (Eds.), Integrated Membrane Systems and Processes, Wiley, 2016. [15] M. Bryjak, N. Kabay, Chapter 10 – boron removal from water by sorption–membrane filtration hybrid process, in: N.K.B. Hilal (Ed.), Boron Separation Processes, Elsevier, Amsterdam, 2015, pp. 237–248. [16] M. Bryjak, J. Wolska, N. Kabay, Removal of boron from seawater by adsorption–membrane hybrid process: implementation and challenges, Desalination 223 (1–3) (2008) 57–62. [17] N. Kabay, et al., Removal of boron from aqueous solutions by a hybrid ion exchange–membrane process, Desalination 198 (1–3) (2006) 158–165. [18] N. Kabay, et al., Adsorption-membrane filtration (AMF) hybrid process for boron removal from seawater: an overview, Desalination 223 (1–3) (2008) 38–48. ˇ [19] M. Blahuˇsiak, S. Schlosser, N. Kabay, Chapter 16 – hybrid adsorption–microfiltration process with plug flow of microparticulate adsorbent for boron removal, in: N.K.B. Hilal (Ed.), Boron Separation Processes, Elsevier, Amsterdam, 2015, pp. 339–354. [20] N. Kabay, et al., An innovative integrated system for boron removal from geothermal water using RO process and ion exchange-ultrafiltration hybrid method, Desalination 316 (2013) 1–7. [21] N. Kabay, et al., Coupling ion exchange with ultrafiltration for boron removal from geothermal water-investigation of process parameters and recycle tests, Desalination 316 (2013) 17–22. [22] S. Samatya, et al., Utilization of geothermal water as irrigation water after boron removal by monodisperse nanoporous polymers containing NMDG in sorption–ultrafiltration hybrid process, Desalination 364 (2015) 62–67. ˇ Schlosser, Simulation of the adsorption—microfiltration pro[23] M. Blahuˇsiak, S. cess for boron removal from RO permeate, Desalination 241 (1–3) (2009) 156–166. [24] I. Yilmaz Ipek, et al., Kinetic behaviour of boron selective resins for boron removal using seeded microfiltration system, React. Funct. Polym. 67 (12) (2007) 1628–1634. [25] R.G. Holdich, I.W. Cumming, S. Perni, Boron mass transfer during seeded microfiltration, Chem. Eng. Res. Des. 84 (1) (2006) 60–68. [26] AFNOR, Dosage du bore par spectrométrie d’absorption moléculaire, Méthodes à l’azométhine H, in Qualité de l’eau, ANFOR, 2001, 2017, p. T 90–041. [27] N.B. Darwish, V. Kochkodan, N. Hilal, Boron removal from water with fractionized Amberlite IRA743 resin, Desalination 370 (2015) 1–6. [28] Y. Jia, R. Wang, A.G. Fane, Hybrid PAC-submerged membrane system for trace organics removal II: system simulation and application study, Chem. Eng. J. 149 (1–3) (2009) 42–49. [29] B. Onderková, et al., Microfiltration of suspensions of microparticulate boron adsorbent through a ceramic membrane, Desalination 241 (1) (2009) 148–155. [30] I.H. Huisman, G. Trägårdh, C. Trägårdh, Particle transport in crossflow microfiltration – II. Effects of particle–particle interactions, Chem. Eng. Sci. 54 (2) (1999) 281–289. [31] G. Belfort, R.H. Davis, A.L. Zydney, The behavior of suspensions and macromolecular solutions in crossflow microfiltration, J. Membr. Sci. 96 (1–2) (1994) 1–58. [32] Y. Xu, J.-Q. Jiang, Technologies for boron removal, Ind. Eng. Chem. Res. 47 (1) (2008) 16–24.