Enhanced phosphate selectivity from wastewater using copper-loaded chelating resin functionalized with polyethylenimine

Journal of Colloid and Interface Science 409 (2013) 129–134

Contents lists available at ScienceDirect

Journal of Colloid and Interface Science www.elsevier.com/locate/jcis

Enhanced phosphate selectivity from wastewater using copper-loaded chelating resin functionalized with polyethylenimine Byungryul An a, Juhee Nam a, Jae-Woo Choi a, Seok-Won Hong a, Sang-Hyup Lee a,b,⇑ a b

Center for Water Resource Cycle Research, Korea Institute of Science and Technology, Hwarangno 14-gil 5, Seongbuk-gu, Seoul 136-791, Republic of Korea Graduate School of Convergence Green Technology and Policy, Korea University, Anam-ro, Seongbuk-gu, Seoul 136-701, Republic of Korea

a r t i c l e

i n f o

Article history: Received 20 May 2013 Accepted 17 July 2013 Available online 29 July 2013 Keywords: Phosphate Ion exchanger Selectivity Regeneration

a b s t r a c t In water and wastewater, phosphate is considered a critical contaminant due to cause algae blooms and eutrophication. To meet the stringent regulation of phosphate in water, a new commercial chelating resin functionalized with polyethylenimine was tested for phosphate removal by loading Cu2+ and Fe2+/Fe3+ to enhance selectivity for phosphate. Batch and column experiments showed that CR20-Cu exhibited high selectivity for phosphate over other strong anions such as sulfate. The average binary phosphate/nitrate and phosphate/sulfate factors for CR20-Cu were calculated to be 7.3 and 4.8, respectively, which were more than 0.97 and 0.22 for a commercial anion exchanger (AMP16). The optimal pH for the phosphate removal efficiency was determined to be 7. According to the fixed-bed column test, the breakthrough 2   sequence for multiple ions was HPO2 4 > SO4 > NO3 > Cl . Saturated CR20-Cu can be regenerated using 4% NaCl at pH 7. More than 95% of the phosphate from CR20-Cu was recovered, and the phosphate uptake capacity for CR20-Cu was not reduced after 7 regeneration cycles. Ó 2013 Elsevier Inc. All rights reserved.

1. Introduction The discharge of phosphorus as a nutrient in wastewater enters the environment from municipal, industrial, human, detergent, and natural sources. It leads to serious environmental problems, such as algae blooms and eutrophication. Moreover, it has been reported that the concentration of phosphorus has significantly increased [1] and continues to increase in Europe [2]. Although concentrations of phosphorus are currently unregulated in drinking water, phosphate ranging from 0.03 to 0.1 mg/L is associated with algae blooms [3,4]. To prevent algal growth, the USEPA has recommended that the total phosphates which comprise the phosphorus concentration in water should not exceed 25 lg/L in a lake or reservoir [3]. In general, phosphorus is present in various forms of phosphate in solution, including organic phosphate, metaphosphates, orthophosphate, and polyphosphate. Among these four forms, the most predominant species of phosphate in treated municipal and industrial wastewater is orthophosphate which has three acid ionization constant (pKa1 = 2.1, pKa2 = 7.2, and pKa3 = 12.3), indicating that at above pH 7.2, the primary species that exists in the aqueous phase is HPO2 4 [5].

⇑ Corresponding author at: Center for Water Resource Cycle Research, Korea Institute of Science and Technology, Hwarangno 14-gil 5, Seongbuk-gu, Seoul 136791, Republic of Korea. Fax: +82 2 958 5839. E-mail address: [email protected] (S.-H. Lee). 0021-9797/$ - see front matter Ó 2013 Elsevier Inc. All rights reserved. http://dx.doi.org/10.1016/j.jcis.2013.07.038

There are a variety of conventional and advanced technologies used in the treatment of phosphate from wastewater. Biological treatment, chemical precipitation, and adsorption are some of the existing technologies used to remove phosphate. Biological treatment does not require the addition of chemical reagents for high removal efficiency. However, biological treatment processes are sensitive to seasonal and diurnal variations in temperature and to changes in feed composition, which are needed to maintain high-efficiency removal [6]. Chemical precipitation, which involves adding calcium, aluminum and iron, is a traditional and effective method that has long been used as a way to regulate phosphate concentrations. Nevertheless, both chemical and biological treatment methods have difficulty removing phosphorus below 0.1 mg/L [7–9] and require further treatment to dispose of the large amount of sludge that is produced. To overcome these drawbacks, adsorption using ion exchanger (IX) was considered to be an alternative method to effectively remove phosphate from water and wastewater. According to prior research [10], the use of IX in the field of water treatment has proven operational simplicity, adaptability to change flow rates and input compositions, and reusability. With these advantages, IX is currently one of the EPA-identified best available technologies (BAT) for the removal of arsenate [11], which is characterized by similar behavior to phosphate in water. However, the target contaminants can be significantly less efficient when current commercial strong base anion (SBA) resins are used. Clifford [10] and Zhao and SenGupta [12] reported that in using IX for arsenate removal,

130

B. An et al. / Journal of Colloid and Interface Science 409 (2013) 129–134

the removal capacity was highly dependent on the presence of other dissolved anions, such as sulfate, which is commonly present in wastewater at much higher concentrations than phosphate due to the strong competition between phosphate and sulfate on the active sites. The concept of ligand exchange based on organic ion exchanger was first introduced by Helfferich [13] to separate and isolate ammonia and diamine, which can form complexes or adducts with transition metal ions (Cu2+ and Ni2+). Later, a new ligand exchanger was prepared and used to remove arsenate by loading ferric ions onto a weak base chelating resin [14]. Due to the low amount of Fe3+ loaded onto the chelating resin, Ramana and SenGupta [15] and Zhao and SenGupta [12] tried to replace Fe3+ with Cu2+ on the same chelating resins (known as DOW 2N and 3N) with 2-picolylamine and di(2-picolyl)amine groups, respectively. Using a ligand exchanger loaded with Cu2+ resulted in the higher selectivity of phosphate over sulfate and increased the removal efficiency of phosphate despite high concentrations of sulfate. In this paper, organic-based chelating resin functionalized with polyethylenimine was used as the parent resin, and copper and iron were loaded onto the surface of this parent resin as functional groups. The overall goal of this study was to explore new ways to achieve the selective separation of strong ligands such as phosphate. The specific objectives were to (1) prepare a new Polymeric Ligand Exchanger (PLE) using CR20 chelating resin, (2) determine and characterize the phosphate adsorption capacity for the new PLE, (3) perform column tests to obtain breakthrough behavior, and (4) investigate the availability of regeneration for a saturated PLE.

where qe is the equilibrium mass uptake of phosphate by a sorbent (mg/g), V is the solution volume (L), Co and Ce are the initial and final concentrations of phosphate in solution, respectively (mg/L), and M is the mass of a sorbent added (g). Another batch experiments were conducted to determine the removal efficiency of phosphate with competing ions. 250 mL solution containing 100 mg/L of nitrate, phosphate, and sulfate was prepared at 0.25 g of each resin. The mixture samples were shook at 150 rpm on a platform shaker (JSSI-100C) under the same experiment condition of isotherm tests.

2. Materials and methods

2.4. Fixed-bed column test

2.1. Materials

Fixed-bed column tests were performed to test the phosphate breakthrough profiles of both anion exchangers. The column test setup included an acrylic column (10 mm in diameter and 150 mm in length), MasterFlex L/S pump, and GILSON FC203B collector. For each column test, 1.5 g of each resin was packed into column with a 0.45 lm filter at the bottom of column to protect against the loss of resin beads. The influent passed in the downflow mode contained phosphate and other competing ions, such as nitrate, and sulfate, and the pH was initially fixed at 7.5. The physical operating conditions for the flow rate and superficial liquid velocity (SLV) were 1.0 mL/min and 0.76 m/h, respectively. The effluent was automatically collected in 8.5 mL glass tubes and was stored in a refrigerator until analysis.

In this study, two ion exchanger resins were used to investigate phosphate removal. A commercial chelating resin (CR20, Mitsubishi Chemical, Japan) with polyethylenimine groups as the parent resin was compared with a commercial anion exchanger resin (AMP16, Samyang Chemicals, Ulsan, Korea). Three different copper salts, i.e., CuCl22H2O, CuSO4 2H2O, Cu(NO3)23H2O and NaOH and HCl were used, and all were ACS grade. All solutions were prepared with ultrapure, deionized (DI, 18.2 M O) water.

2.2. Synthesis of modified ion exchanger The procedure for loading the copper onto CR20 followed that of An et al. [16], except for the initial copper solution concentration. Both the CR20 and AMP16 resins were first conditioned using 1 M HCl, DI water, and 1 M NaOH for 3 h, sequentially. After rinsing with DI water until the pH reached 7.0, AMP16 resin was dried in air and then tested. For CR20 chelating resin, the process was as follows. First 10 g of CuCl22H2O was dissolved in 500 mL DI in three round-bottom flasks, and then, 100 g of CR20 was added to the 2% copper solution. During the day, the mixed copper solution was kept at 70 °C, which can enhance the copper loading kinetics and stability by swelling the resin, and it was then brought down to ambient temperature at night (21 °C). The solution took two weeks to reach copper equilibrium. During preparation, the pH was kept at 4.0 ± 0.2 with dilute NaOH and HCl, and nitrogen gas was fed into the solution at regular intervals. After two weeks, the resin was rinsed with DI several times and air-dried. The CR20 loaded with Cu was referred to as CR20-Cu. This same procedure was used to prepare CR20-Cu from solutions of CuSO4 and Cu(NO3)2.

2.3. Batch adsorption test Isotherm tests for both resins in 55 mL tubes were conducted by adding different initial concentrations of phosphate ranging from 0 mg/L to 100 mg/L at 0.05 mg of each resin. The mixture was then gently rotated at 30 rpm for 48 h, which is enough time to reach equilibrium, as confirmed by kinetic tests for both resins. After 48 h, 5 mL samples from each tube were taken out, filtered with 2   a 0.2 lm filter, and then analyzed for PO2 4 , SO4 , NO3 and Cl . According to the preliminary tests, the additional process of filtration using syringes did not affect the concentrations of phosphate or other anions. The pH of the solution was initially 7.5–8.0 and was maintained using dilute HCl and NaOH. The amount of uptake for each anion was calculated using the following equation based on mass balance:

qe ¼

VðC o  C e Þ M

ð1Þ

2.5. Regeneration and reuse Regeneration of saturated CR20-Cu was carried out in the batch test with 4% and 8% NaCl for 6 h. During regeneration, the solution pH was kept at 7.0. After regeneration of CR20-Cu, the regenerated resin was reused with the same batch conditions described in Section 2.3. All batch experiments were duplicated.

2.6. Chemical analyses Phosphate, sulfate, nitrate, and chloride were analyzed by Dionex Ion Chromatography (Model: ICS-1000, USA), and Inductively Coupled Plasma (ICP) (model: Agilent 700 series, USA) was used for copper and iron analyses. Fourier transform infrared spectroscopy (FTIR) was used to determine the formation of copper and iron (model: spotlight 200, PerkinElmer, USA). The solution pH was measured using an ORION Star A211 (Thermo scientific, USA).

B. An et al. / Journal of Colloid and Interface Science 409 (2013) 129–134 Table 1 Maximum copper/iron-loading with various metal solutions on CR20 chelating resin. Loading solution

CuCl2

CuSO4

Cu(NO3)2

FeCl3

FeSO4

Maximum loading capacity (mg/g)

50.0

58.3

47.7

2.21

19.3

(0.672)

(2.63)

(1.28)

(0.291)

(0.616)

131

that copper uptake was determined according to the structure of the functional group. Phosphate in water formed tertiary complexes with the central metal ions that were hosted by chelating resins. Accordingly, the removal capacity of the target contaminate depended on the mass of copper attached to the chelating resin [16]. 3.2. Batch experiments

3. Results and discussion 3.1. Characteristics of PLE Three different copper solutions, CuCl2, CuSO4, and Cu(NO3)2, were loaded onto virgin CR20 resin. Other experiments using Fe(II) and Fe(III) instead of Cu(II) have been conducted with the same procedure to determine the removal efficiency of the phosphate. Chanda [14] used iron with two different chelating resins for arsenic removal, both of which showed low removal efficiency for arsenic. In Table 1, the amount of transition metal loaded to the CR20 was obtained after metal desorption with 20% (v/v) H2SO4. The mass of copper loaded ranged from 47.7 mg/g for CuSO4 to 58.3 mg/g for Cu(NO3)2, which are of a slightly higher capacity than the previous result of 46.4 mg/g (0.73 mmol) [17], but the iron uptake only ranged from 0.5 mg/g to 10 mg/g for Fe(III) and Fe(II), respectively. The higher copper loading capacity can be explained by the Irving-Williams [18] rule on the sequence of complex stability: 2þ

Mn2þ < Fe2þ < Co2þ < Ni

< Cu2þ

ð2Þ

Based on the Irving-Williams series, the higher complex stability led to the Lewis acid–base (LAB) interaction-enhanced loading of transition metals, which can be strongly related to phosphate removal. Table 2 lists the five chelating resins based on polystyrene (PS)/divinylbenzene (DVB) that were tested for copper removal abilities and ligand structures reported in the literature indicating

A series of batch tests using CR20 loaded with Cu(II) and Fe(II)/ (III) was performed to determine the phosphate removal efficiency with competing anions. Fig. 1 shows the removal efficiency of phosphate and other anions. In Fig. 1, the phosphate and sulfate removal efficiencies were obtained for all CR20-Cu samples and were approximately 50% and 30%, respectively, indicating that phosphate has a much higher affinity to copper than sulfate. As a result, LAB interactions enhanced phosphate selectivity over sulfate. As observed in Table 1, the phosphate removal efficiency was not proportionally in accordance with the mass of copper on the resin, which might be due to the presence of counter ions. Although loading by CuSO4 showed that the capacity of copper was increased to 16% compared to Cu(NO3)2, the counter ion, SO2 4 , is still strongly bound to copper and is harder to exchange than chloride or nitrate. Another result from using CR20-Fe, which replaced copper with iron, is shown in Fig. 1. The removal efficiency for all anions was reduced compared to CR20-Cu, and the phosphate removal efficiency was approximately 25% and 30% for Fe(II) and Fe(III), respectively. It can be surmised that the amount of Fe on the chelating resin is much less than that of copper, which results in a lower phosphate removal efficiency. Based on past studies of Fe(III) as the central metal by Chanda [14], who tried to remove arsenate using the chelating resin, DOW3 N loaded with Fe(III), arsenate uptake did not compete with DOW3N-Cu [16]. The low uptake capacity for Fe(III) on CR20 can be explained by Fe(III) solubility which is strongly dependent on the solution pH. Under very acidic conditions, dissolved Fe(III) competes with hydrogen ions for bonding with the ligands, whereas at a relatively high pH, Fe(III) is present

Table 2 Summary of chelating resins reported in the literature for Cu(II) binding on the PS/DVB matrix. Functional group

Structure

Di(2-picolyl)amine

CH2

H N

Refs. [12]

0.9

[31]

0.73

[17]

0.83

[17]

0.6–0.7

[17]

CH2

N

N

Iminodiacetate

CH2

P CH2 O C

O-

N

CH2 -

O

C O

Polyethylenimine

PS/DVB HN

NH

NH

Pyridyl imidazolyl

P

Copper-loading capacity (mmol/g) 1.5

NH

N

CH2 N

NH

N Alkyl (2-picolyl) amine

N P

CH2 N CH2 R

R=H, CH3, CH2CH3, or CH2CH2OH

132

B. An et al. / Journal of Colloid and Interface Science 409 (2013) 129–134

50

Conditions Co: 100 mg/L of NO3-, PO43-, SO42pH: 7.5

3383

CR20-Fe(II)

phosphate sulfate

CR20-Fe(III)

Transmittance

Removal efficiency,%

60

40 30 20

CR20 CR20-Cu

464

10

407

0 4000

CR20 CuCl2 CuSO4 Cu(NO3)2 Fe(II) Fe(III)

3500

3000

2500

2000

1500

1000

500

Wavenumbver, cm-1

Loading solution Fig. 1. Removal efficiency for phosphate and sulfate using CR20-Cu and CR20-Fe loaded with different copper solutions and iron forms.

Fig. 2. Comparison of FTIR spectra of CR20-Cu and CR20-Fe (arrow indicates the different peaks).

in its hydroxide form. Accordingly, low mass of Fe(III) results in less phosphate removal efficiency compared with Cu(II). Fig. 2 shows FTIR spectra ranging from 380 cm1 to 4000 cm1 for CR20 and CR20-Cu as well as CR20-Fe(II) and Fe(III); these spectra compare the characteristics of the copper and iron after being loaded onto the CR20. For the CR20-Cu(II) and CR20-Fe(III), the peak locations and intensities were exactly the same, indicating both transition metals were bound to nitrogen in CR20. However, for CR20-Fe(II) resin, different interactions between CR20 and iron were detected for some identified peaks. The peak at 3483 cm1, which was found for all samples, corresponded to O-H stretching in H2O [19], but the peak for CR20-Fe(II) was broader and more intensified. This difference in peaks results from the addition of OH bonds from Fe-OH, which is consistence with Myneni et al. [20]. The very strong and broad band from 545 cm1 to 380 cm1, with peaks at 464 cm1 and 407 cm1, was detected for CR20-Fe(II), which is a typical band for iron(III) hydroxide [21]. As result, phosphate was removed from solution to Cu(II) and Fe(III) coordinated to polyethylenimine via the LAB and electrostatic interactions, while phosphate adsorption by Fe(OH)3 was assumed to a major process for CR20-Fe(II).

was similar at pH 8 and 9. This finding can be explained by the Donnan co-ion exclusion principle [22], which occurs at the surface of an adsorbent. An OH- counter ion is attracted to the fixed metal ion, while H+ ions are excluded from the surface of the IX resin. As a result, due to excessive OH-, the pH at the surface is greater than the pH of the bulk solution, resulting in H2 PO 4 transforming into HPO2 4 . This interfacial pH shift was also observed by Zhao and SenGupta [12] using DOW3N-Cu resin. By increasing the pH to 9, the removal efficiency of phosphate does not change, but that of sulfate decreases. This result can be attributed to the presence of hydroxyl anions, which play a key role in adsorption as competing ions under alkaline conditions. Although phosphate exists as the HPO2 4 species, which is a stronger ligand, the removal efficiency was not increased. Nevertheless, the sulfate removal efficiency decreases with increasing pH, indicating that the reduced sulfate adsorption capacity was due not to phosphate but to hydroxyl anions.

3.3. pH effect on phosphate removal efficiency The species of phosphate is determined by the solution pH, which strongly impacts the selectivity of phosphate in ion exchange processes. Fig. 3 shows the pH effect on phosphate and sulfate uptake with the phosphate species fraction shown as a function of pH. At pH 5, the phosphate and sulfate removal efficiency is 22% and 59%, respectively. Increasing the solution pH to 9 dramatically improved the phosphate removal efficiency to 51% by pH 7, and this efficiency remained at this level through pHs 8 and 9, while the sulfate removal efficiency continuously decreased 2 to 17% at pH 9. At pH 5, H2 PO 4 and SO4 (pKa = 1.9) are the predominant phosphate and sulfate species, respectively. The low ionic strength of H2 PO 4 , which cannot compete with divalently charged sulfate anions, makes it much less adsorbable. As a result, phosphate removal governed by electrostatic interaction is lower than sulfate removal at pH 5. However, above pH 7, the fraction of HPO2 increases and after pH 7.2 (pKa2), this species becomes 4 predominant. At the same time, the phosphate removal efficiency was more than doubled and that of sulfate was halved. Based on both ligand strength and ionic charge, the HPO2 4 species is more adsorbable than HPO2 4 , so the adsorption affinity for total phosphate overcomes the competition from sulfate. Despite the low fraction of HPO2 at pH 7, the capacity of phosphate adsorption 4

3.4. Equilibrium isotherm test One of the critical limitations for commercial standard SBA is lower selectivity for phosphate in the presence of sulfate ions, which is always present in municipal and industrial wastewaters. Thus, the application of SBA requires more frequent regeneration and produces huge volumes of regenerant [23]. In contrast, the use of copper-loaded chelating resin, in which copper complexes with phosphate via LAB interactions, enhances selectivity for phosphate over sulfate [12], so it follows that CR20-Cu should have high phosphate selectivity. Furthermore, according to the model of phosphate and sulfate adsorption, the formation of phosphate and sulfate ligands occurs with inner and outer sphere complexes, respectively [24]. Fig. 4 shows the isotherms of phosphate sorption with 100 mg/L of sulfate and 100 mg/L of nitrate. Based on the experimental isotherm data, the phosphate-sulfate separation factors (ap/s) calculated can be used to compare the relative affinity of the CR20-Cu toward these two competing ligands. The phosphate/sulfate separation factor was defined as follows:

aAs=S ¼

qAs C S C As qS

ð3Þ

where qp and qs are the concentrations of phosphate and sulfate in the solid-phase and Cp and Cs are the concentrations of phosphate and sulfate in the solution. In general, a value of ap/s greater than one generally indicates greater selectivity for phosphate than sulfate.

133

B. An et al. / Journal of Colloid and Interface Science 409 (2013) 129–134

100

3.5

HPO42-

3.0

60

40

40 20 20

Influent

(a)

pH:7.5-8.0 PO43-

Cl-: 52.3mg/L NO3-:11.7mg/L

2.5

NO3-

PO43-:3.4mg/L 2-

SO4 :48.8mg/L

C/C0

80

Phospate fraction,%

60

Removal efficiency,%

Phosphate Sulfate

H2PO4-

2.0

EBCT: 4min SLV:0.76m/hr

Cl--

1.5 1.0

SO42-

0.5 0

0.0

0 4

5

6

7

8

9

0

10

1000

2000

pH

3000

4000

5000

Time, min

Fig. 3. Phosphate and sulfate uptake by CR20-Cu as a function of pH. The solid line 2 shows the calculated speciation curves of H2 PO 4 and HPO4 .

3.5 Influent

60

(b)

pH:7.5-8.0

3.0

Cl-: 52.3mg/L NO3-:11.7mg/L

2.5

PO43-:3.4mg/L SO42-:48.8mg/L

PO4 uptake, mg/g

3-

C/C0

CuCl2 AMP16

50 40

Cl--

EBCT: 4min SLV:0.76m/hr

1.5 1.0

NO3-

30

SO42-

0.5

PO4

20

3-

0.0

0

10

2000

4000

Q=64.4 mg/g for CR20-Cu Q=48.3 mg/g for AMP16

0

20

40

60

Fig. 4. Phosphate isotherms for CR20-Cu and conventional SBA resin in the presence of competing sulfate and nitrate ions (symbols: observed data; lines: Langmuir model fits).

The average values of (ap/s) are shown in Table 3 as 4.8 and 0.22 for CR20-Cu and AMP16, respectively, indicating that the affinity of CR20-Cu for phosphate is higher than that for sulfate. The enhanced selectivity of CR20-Cu for phosphate results from the greater ligand strength between Cu and phosphate; in the commercial anion exchange resin AMP16, phosphate selectivity was only governed by electrostatic interactions, resulting in higher sulfate selectivity than phosphate selectivity. With higher selectivity, CR20-Cu can perform phosphate removal in the presence of sulfate. These experimental isotherm data were interpreted using the classical Langmuir model to determine the maximum phosphate uptake.

bQC e 1 þ bC e

ð4Þ

where qe is the equilibrium phosphate uptake (mg/g), Ce is the equilibrium concentration of phosphate in aqueous phase (mg/L), Q is Table 3 The values of the model-fitted Langmuir parameters (Q, b) and the mean phosphate/ nitrate or sulfate separation factors (ap/n,s). Q (mg/g)

b (L/g)

CR20-Cu AMP16

64.4 48.3

0.052 0.030

8000

10000

Fig. 5. Breakthrough histories of phosphate and competing anions in a multicomponent system using a standard SBA resin, AMP16 (a) and a polymeric ligand exchanger, CR20-Cu (b).

80

Ce, mg/L

Resin

6000

Time, Min

0

qe ¼

2.0

Separation factor

ap/n

ap/s

7.3 0.97

4.8 0.22

the maximum capacity for phosphate, and b is the Langmuir affinity coefficient between the sorbent and sorbate. The values of the model-fit Q values provided in Table 3 were 64.4 and 48.3 mg/g for CR20-Cu and AMP16, respectively. These Q values show that phosphate sorption capacity with CR20-Cu is higher than that with AMP16. 3.5. Breakthrough profile Fixed-bed column tests were conducted to investigate the breakthrough behaviors of nitrate, chloride, sulfate, and phosphate with CR20-Cu. As a comparison, AMP16 was tested under the same operating conditions. The influent concentrations and hydrodynamic conditions (e.g., SLV) are provided in Fig. 5, which shows the ratio of the concentration of the effluent solution (Y axis) after passing through the column versus the influx time (X axis). The breakthrough sequence obtained from Fig. 5a during the fixedbed column run with AMP16 indicates that the resin’s affinity to the anions, as determined by batch tests, corresponds with the breakthrough order. 

2  SO2 4 > HPO4 > NO3 > Cl

ð5Þ

Furthermore, Fig. 5a reveals a sharp chromatographic peaking of nitrate and phosphate, which occurred as a result of competing ion exchanges and corresponded to sulfate’s affinity over phosphate and nitrate [25,26]. In contrast, a completely different breakthrough behavior for sulfate and phosphate can be observed in Fig. 5b. According to the breakthrough sequence, the selectivity order for

134

B. An et al. / Journal of Colloid and Interface Science 409 (2013) 129–134

anions is the following, which is the same as in Zhao and SenGupta [23] : 

2  HPO2 4 > SO4 > NO3 > Cl

ð6Þ

Similar results were shown in fixed-bed column tests at a 10 M ratio of phosphate to sulfate by SenGupta and Pandit [27] using DOW 3N-HFO-Cu. The shape of the breakthrough curves of sulfate and phosphate can be attributed to this difference in selectivity. The breakthrough curve of sulfate quickly reached the initial concentration within 1000 min, but it took phosphate 5 times longer to reach initial concentration. Based on these phenomena, despite the lower concentration of phosphate, the affinity for phosphate using CR20Cu is higher than that for sulfate. 3.6. Regeneration test Regeneration and reuse in ion exchange are appealing because it is cost effective and environmentally friendly [28,29]. The critical parameters used to decide the effectiveness of regeneration include the type, concentration, and pH of the feeding regenerant. Based on previous studies of copper-loaded PLE saturated with arsenic and phosphate, using 4% NaCl [16] and 6% NaCl [12] effectively regenerated the saturated PLE resin and using NaOH for hybrid anion exchange recovered over 90% of the phosphate from the bed [30]. According to these past results, this study opted to use 4% NaCl as a regenerant at pH 7 because NaOH is a strong base and it is difficult to handle. A series of regeneration experiments for phosphate-saturated CR20-Cu were tested in batch tests, and phosphate recovery was calculated using the mass balance from the initial and final phosphate concentrations of the solution. Eq.(7) illustrate the regeneration reaction stoichiometry at the specified pH. At pH 7.0, 

þ xH2 PO4

ð7Þ

Under neutral conditions, phosphate is present as both H2 PO 4 and HPO2 4 species (pKa2 = 7.2). As expected, 4% NaCl at pH 7 recovered almost 100% of phosphate from exhausted CR20-Cu within 6 h. Fig. 6 shows the equilibrium of phosphate uptake by virgin CR20Cu compared to the uptake over 7 consecutive operation cycles. The capacity of phosphate uptake for the regenerated CR20-Cu remained strong and showed no significant drop-off compared to its initial state.

25

phosphate uptake, mg/g

The main findings and conclusions from this study can be summarized as follows:  Based on the batch adsorption tests performed using a variety of CR20 loaded copper resins, the efficiency of phosphate removal depended on the amount of transition metal (Cu(II) and Fe(III)) loaded onto the CR20 resin. The effect of different copper solutions on loading 45 mg/g and 58 mg/g for CuSO4 and CuNO3, respectively, did not significantly affect phosphate removal.  Selectivity for phosphate based on the separation factor was always higher for CR20-Cu, whereas commercial SBA resin showed selectivity for sulfate over phosphate.  During a fixed column run, the breakthrough order using CR20 2   was HPO2 4 > SO4 > NO3 > Cl , and the phosphate did not show any chromatographic elution, which is indicative of phosphate selectivity over that of sulfate.  CR20-Cu was found to be extremely durable, mechanically strong and chemically stable. With 4% NaCl as a regenerant, 7 cycles were successfully achieved with almost 100% phosphate recovery. Due to a very high affinity toward copper(II) with polyethylenimine on the CR20, no copper leakage was detected.

Acknowledgments This work was supported by the KIST Institutional Program (Project No. 2E24280), Korea Ministry of Environment as ‘Global Top Project’ (Project No.: GT-11-B-01-011-1), and National Research Foundation of Korea Grant funded by the Korean Government (Ministry of Science, ICT & Future Planning) (2013, University-Institute cooperation program). References



2þ 2 þ R-Cu2þ HPO2 4 þ 2Cl þ xH ! R  Cu Cl2 þ ð1  xÞHPO4

20

15

10

5

0 0

4. Conclusion

1

2

3

4

5

6

7

8

Regneration times Fig. 6. Comparing the equilibrium phosphate sorption, qe, after regeneration with 4% NaCl.

[1] D.J. Conley, H.W. Paerl, R.W. Howarth, D.F. Boesch, S.P. Seitzinger, K.E. Havens, C. Lancelot, G.E. Likens, Science 323 (2009) 1014. [2] CEEP, SCOPE NEWSLETTER; Institute for European Environmental Policy, 1999. [3] USEPA, EPA440/5-86-001, Washionton, DC, 1986. [4] M. Oliveira, A.V. Machado, R. Nogueira, Water Air Soil Pollut. 223 (2012) 4831. [5] V.L. Snoeyink, D. Jenkins, Water Chemistry, Wiley, 1980. [6] USEPA, EPA 625/1-87-001, Cincinnati, OH, 1987. [7] D. Jenkins, J.F. Ferguson, A.B. Menar, Water Res. 5 (1971) 369. [8] P. Cooper, T. Dee, G. Yang, 4th Annual Conference on Industrial Wastewater Treatment, Esher, UK, 1993. [9] R.I. Sedlak, Phosphorus and Nitrogen Removal from Municipal Wastewater: Principles and Practice, Lewis Publishers, 1991. [10] AWWA, Water Quality & Treatment Handbook, McGraw-Hill, New York, 1999. [11] USEPA, EPA 600/R-00/088, Washington, DC, 2000. [12] D.Y. Zhao, A.K. Sengupta, Water Res. 32 (1998) 1613. [13] F. Helfferich, J. Am. Chem. Soc. 84 (1962) 3237. [14] M. Chanda, K.F. Odriscoll, G.L. Rempel, React. Polym. 8 (1988) 85. [15] A. Ramana, A.K. Sengupta, J. Environ. Eng.-ASCE 118 (1992) 755. [16] B. An, T.R. Steinwinder, D.Y. Zhao, Water Res. 39 (2005) 4993. [17] A. Warshawsky, Modern Research in Ion Exchange, Martinus Nijhoff Boston, MA, 1986. [18] R.J.P.W.H.M.N.H. Irving, J. Chem. Soc. (1953) 3192. [19] D. Maity, D.C. Agrawal, J. Magn. Magn. Mater. 308 (2007) 46. [20] S.C.B. Myneni, S.J. Traina, G.A. Waychunas, T.J. Logan, Geochim. Cosmochim. Acta 62 (1998) 3499. [21] M. Ristic, E. De Grave, S. Music, S. Popovic, Z. Orehovec, J. Mol. Struct. 834 (2007) 454. [22] E.W.C. Bauman, J. Am. Chem. Soc 69 (1948) 2830. [23] D.Y. Zhao, A.K. SenGupta, Ind. Eng. Chem. Res. 39 (2000) 455. [24] L.M. He, L.W. Zelazny, V.C. Baligar, K.D. Ritchey, D.C. Martens, Soil Sci. Soc. Am. J. 61 (1997) 784. [25] J. Kim, M.M. Benjamin, P. Kwan, Y.J. Chang, J. Am. Water Works Assoc. 95 (2003) 77. [26] S.K. Zheng, J.J. Chen, X.M. Jiang, X.F. Li, Chem. Eng. J. 169 (2011) 194. [27] S. Sengupta, A. Pandit, Water Res. 45 (2011) 3318. [28] P.F. Lito, J.P.S. Aniceto, C.M. Silva, Water Air Soil Pollut. 223 (2012) 6133. [29] Y.F. Zhou, R.J. Haynes, Water Air Soil Pollut. 218 (2011) 457. [30] L.M. Blaney, S. Cinar, A.K. SenGupta, Water Res. 41 (2007) 1603. [31] C.Y. Chen, C.L. Chiang, P.C. Huang, Sep. Purif. Technol. 50 (2006) 15.