Influence of ion interaction on lead removal by a polyamide nanofiltration membrane

Desalination 362 (2015) 84–92

Contents lists available at ScienceDirect

Desalination journal homepage: www.elsevier.com/locate/desal

Influence of ion interaction on lead removal by a polyamide nanofiltration membrane Saber Mehdipour a, Vahid Vatanpour b,c, Hamid-Reza Kariminia a,⁎ a b c

Department of Chemical & Petroleum Engineering, Sharif University of Technology, P.O. Box: 11155-9465, Tehran, Iran Faculty of Chemistry, Kharazmi University, P.O. Box 15719-14911, Tehran, Iran Novel Technology Research Group, National Petrochemical Company, Petrochemical Research and Technology Company, P.O. Box 1435884711, Tehran, Iran

H I G H L I G H T S • • • • •

Increasing lead ion concentration resulted in a higher retention of this ion at a constant pressure. Increasing anion valences of the interfering ions resulted in a higher lead ion rejection and a lower permeate flux. Decreasing or increasing the pH of the lead ion solutions caused a higher lead ion rejection. Altering the pH of the lead ion solutions from their natural pH, decreased the permeate flux. The presence of monovalent cations didn't reduce lead ion rejection, significantly.

a r t i c l e

i n f o

Article history: Received 14 December 2014 Received in revised form 20 January 2015 Accepted 22 January 2015 Available online xxxx Keywords: Nanofiltration Heavy metals Zeta potential Cation interaction Lead removal

a b s t r a c t Retention of lead(II) ions on a polyamide nanofiltration membrane was investigated. Effects of different factors including operating pressure, lead ion concentration, anion nature, pH and composition of feed on the lead ion rejection were studied. The solutions used consisted of Pb(NO3)2, PbCl2 and PbSO4 in the single-salt solution system and Pb(NO3)2, Cu(NO3)2, Zn(NO3)2, Cd(NO3)2, NaNO3 and NH4NO3 in the binary-salt solution system. The influence of divalent and monovalent cations including cadmium, copper, zinc, sodium and ammonium on the rejection of lead ion was examined. The transmembrane pressure and lead ion concentration varied between 10 and 40 bar and 20 and 400 mg Pb2+/L, respectively. It was observed that increasing the pressure and initial feed concentration 2− − resulted in a higher lead ion rejection. By replacing NO− 3 with Cl andSO 4 , the rejection of lead ion increased about 2% and 9%, respectively. Applying anions with higher valences resulted in a higher lead ion rejection and lower permeate flux. Maximum permeate flux and minimum lead ion rejection was observed at pH 5.6. In the binary salt solutions, the rejection of lead did not change significantly in the presence of monovalent cations. However, the presence of divalent cations caused a substantial reduction in the lead ion rejection. © 2015 Elsevier B.V. All rights reserved.

1. Introduction The presence of heavy metals in wastewaters causes many problems due to their persistence in the environment. There is also a high risk of heavy metal accumulation in the body tissues of living organisms due to their high solubility in the aquatic environments [1]. Industries that generate heavy metals such as Pb, Cd, Zn, Cu, Ni and Cr are considered as the most hazardous industries [2]. Among the heavy metals, lead is a very dangerous agent due to its side effects on the human body even at concentrations as low as 0.01–5 mg/L [3]. To date, numerous techniques have been considered for development of efficient and inexpensive treatment methods to improve the quality of effluents. Heavy metal removal can be achieved by conventional processes such as chemical ⁎ Corresponding author. E-mail address: [email protected] (H.-R. Kariminia).

http://dx.doi.org/10.1016/j.desal.2015.01.030 0011-9164/© 2015 Elsevier B.V. All rights reserved.

precipitation [4–7], solvent extraction [8], ion exchange [9–11], adsorption [12–14] and metal oxides [15]. Despite being efficient and cheap, these methods are suffering from substantial disadvantages such as incomplete removal of heavy metals, high energy consumption, slow removal rates and generation of toxic sludge [16]. On the path to tackle these difficulties and find less expensive and more efficient methods, membrane separation processes have attracted a high attention. Nanofiltration (NF) technology has been used to treat lead containing wastewaters. In this method, the rejection of lead ion can be high enough to achieve acceptable quality for the filtered water [17–20]. Nanofiltration is a membrane filtration process that stands between ultrafiltration and reverse osmosis in which the separation characteristics are based on the sieving effect and membrane surface charge, though the NF membranes have been modified in many commercial applications [21]. Nanofiltration membranes can remove divalent ions and low molecular weight organic materials [22]. Three phenomena contribute to the ion transport through

S. Mehdipour et al. / Desalination 362 (2015) 84–92

85

Pressure

NF membranes that include size exclusion, charge effect (based on Donnan exclusion and dielectric exclusion), and different diffusivity and solubility of existing ions [23]. It is known that various parameters such as feed solution concentration, composition and pH value are effective on the interactions between solutes of the ionic solution and solute– membrane interactions [24,25]. However, in the real industrial wastewaters, there is a mixture of several heavy metals in which the interfering cations can influence on the desired cation removal. Nevertheless, there are a few studies dealing with lead ion removal in the mixtures consisted of interfering cations [19,20,26]. To the best of our knowledge, there is only one study about the effect of anions nature on the lead ion rejection and permeate flux that is related to lead ion removal in the presence of a single salt [27]. In this work, the retention of lead ion of a single salt solution containing Pb(NO3)2, PbCl2 or PbSO4 was investigated to study the influence of anion nature (monovalent and divalent) on the lead ion removal. Besides, the effects of different conditions i.e. pressure of 10–40 bar, pH of 3–7 and lead ion concentration of 20–400 mg/L were examined on the rejection of lead ions. Then, the binary-salt solutions consisted of lead ion and another cations of Zn(NO3)2, Cu(NO3)2, Cd(NO3)2, NaNO3 or NH4NO3 were examined in order to study the effect of interfering monovalent and divalent cations on the removal efficiency of lead ion by nanofiltration process.

Pressure

Membrane module

P

P

Retentate valve

Permeate sampling balance

PC

Flowmeter Heat-exchanger Cooling water Feed tank

valve Drain

Feed

Pump Fig. 1. Schematic diagram of nanofiltration experimental setup.

2. Experimental 2.1. Membrane A NE 4040-90 membrane manufactured by CSM, South Korea was used in this study. The used membrane was a thin-film polyamide composite membrane in spiral wound element configuration. The module was opened, and the flat sheet membrane was used in our experiments. As reported by the manufacturer, NaCl and MgSO4 (2000 ppm) rejection of the membrane was 85–95% and 97%, respectively.

experimental work. Each membrane was washed by circulating demineralized water for at least one hour at the pressure of 30 bar followed by compacting for one hour at the maximum pressure used in this study (40 bar) to avoid any compression effect. Retentate and permeate were both returned to the feed tank. To measure the permeate flux, the amount of the collected solution passed the membrane was determined using an intelligent gravimetric analyzer connected to a computer. When the concentrations of permeate and feed reached a constant value, the operation reached the steady-state condition. To confirm this condition, the conductivity of both solutions was measured.

2.2. Materials 2.5. Electro-kinetic (zeta) potential measurement Chemicals used in this work were all of analytical grade. All the metal solutions were prepared by dissolving enough mass of each material in a high purity demineralized water (conductivity b 1 µS/cm and pH 6.3 ± 0.1). Lead nitrate Pb(NO3)2, lead chloride PbCl2, lead sulfate PbSO4, copper nitrate Cu(NO3)2·3H2O, zinc nitrate Zn(NO3)2·6H2O, cadmium nitrate Cd(NO3)2·4H2O, sodium nitrate NaNO3, ammonium nitrate NH4NO3, sodium hydroxide, nitric acid, hydrochloric acid and sulfuric acid were all obtained from Merck Chemicals, Germany. 2.3. Analytical methods Elemental analysis of lead was conducted for the feed and permeate solutions by measuring the conductivity of these solutions at 25 °C using a conductometer LF 330/340 (WTW, Germany), while the concentration of the cations in the feed and permeate of binary-salt solutions was determined using a Varian AA240FS Atomic Adsorption Spectrometer (US). 2.4. Membrane test unit All experiments were performed with a cross-flow filtration unit equipped with a flat-sheet cell as presented schematically in Fig. 1. The effective filtration area of the filtration cell was 36 cm2. The temperature of feed solution was maintained at the constant value of 25 ± 2 °C using a heat exchanger. The transmembrane pressure was varied between 10 and 40 bar by changing the feed pressure. The solution was pumped from the feed tank into the membrane cell by means of a volumetric pump. The membrane was immersed overnight in a mixture of ethanol/ water solution with equal volumetric amount before using in any

The separation mechanism in nanofiltration membranes is usually described in terms of charge and sieve effect [28]. Therefore, knowledge about the surface charge is very useful in explaining the test results. Information about the charge of membrane surface provided by membrane manufacturer is rare. Determination of electrokinetic potential or zeta potential is a very appropriate method to measure the charge of membrane surface and determine the electrical properties of the membrane. Zeta potential of the membranes was determined by streaming potential method along the membrane surface using an EKA Electro Kinetic Analyzer instrument (Anton Paar, Austria). In the streaming potential method, movement of the electrolyte solution through a capillary system creates a streaming potential where its relation with the zeta potential of the capillaries is given by Smoluchowski– Helmholtz approach [29–31]: ζ¼

dU η L dp ε  ε ∘ Q  R

ð1Þ

where ζ is the zeta potential, dU/dp is the slope of streaming potential versus pressure, η is the electrolyte viscosity, εo is the permittivity, ε is the dielectric constant of the electrolyte, L is the length of the capillary system, Q is the cross-sectional area of the capillary system and R is the AC resistance of the cell using an electrolyte solution. According to Fairbrother and Mastin approach [32], for electrolyte concentrations greater than 10−3 M, the ratio L/Q·R in Eq. (1) can be replaced by KB that is the specific electrical conductivity of the electrolyte solution outside the capillary system. Zeta potential measurements were conducted using a clamping cell. Schematic of the clamping cell and the geometric parameters defining a

86

S. Mehdipour et al. / Desalination 362 (2015) 84–92

single capillary was the same as presented by Walker et al. [33]. Prior to zeta potential measurements, the membrane was washed by circulating deionized water across the membrane surface for 20 min. Measurements were made with 0.001 M KCl electrolyte solution. During the measurements, pH of the solution varied in the range of 3 to 7 by adding 0.1 N solution of NaOH and HCl. Due to construction design of clamping cell, the measurements were performed by firmly pressing the test surface against a PMMA grooved spacer. Therefore, the zeta potential of the sample was calculated according the following relation: ζ s ¼ 2ζ tot −ζ PMMA

ð2Þ

where ζtot is the zeta potential of sample plus PMMA, ζPMMA is the zeta potential of PMMA, and ζs is the net zeta potential of the sample.

absolute temperature, respectively. ψ is the electrical potential within the film layer, and Jv is the permeate flux. Terms on the right-hand side of Eq. (3) represent the transport of an ion due to diffusion, electro-migration and convection. The ionic molar flux Ji is expressed as: J i ¼ J v C i;p :

ð4Þ

Due to constant electric potential gradient for each ion under the steady-state membrane filtration, there is a zero net electric current that can be expressed as: n

FΣi¼1 zi J i ¼ 0:

ð5Þ

Eq. (3) can be simplified to: 2.6. Operating procedure After preparing the synthetic lead solutions by adding enough amounts of various lead salts to demineralized water, experiments were carried out at several pressures at constant pH of 5.6 ± 0.1 and temperature of 25 ± 2 °C. Concentrations of lead in Pb(NO3)2 singlesalt solution was 20, 50, 100, 200 and 400 mg/L at the pressures of 10, 20, 30 and 40 bar, respectively. However, in PbCl2 and PbSO4 cases due to low solubility of these salts in water, the experiments were performed at only one concentration (20 mg/L of Pb2+) at the pressures of 10, 20, 30 and 40 bar. The influence of initial feed pH (in the range of 3 to 7) on the lead rejection was investigated by adding HNO3, HCl, H2SO4 and NaOH solutions to the solutions containing 20 mg Pb2 +/L from Pb(NO3)2, PbCl2 and PbSO4 salts. Experiments to compare the effect of interfering cations on lead rejection were carried out at the concentrations of 50, 100, 200 and 400 mg/L of each cation against a solution with the lead concentration of 100 mg/L. 3. Data assessment In the high-pressure operations, solvent passes through the membrane in which partial solute permeation occurs. The larger solutes and non-permeated ones accumulate adjacent to the membrane surface within the boundary layer (film layer) where a concentration profile is developed (Fig. 2). This phenomenon is called concentration polarization and can be described by film theory [34]. Under this condition, the mass balance of the electrolyte solution inside the film layer is J i ¼ −Di;∞

dci zi F dψ − cD þ ci J v dx RT i i;∞ dx

ð3Þ

where Ji is the molar flux of ion i, Di,∞ is the diffusion coefficient of ion i in the bulk solution, zi is the charge number of ion i, ci is the ion i concentration, F, R and T are the Faraday constant, ideal gas constant and

J i ¼ −Dsalt

dci þ ci J v dx

ð6Þ

where Dsalt is the effective diffusivity of the salt which is defined as: Dsalt ¼


Dþ D− zþ z− zþ Dþ −z− D−

ð7Þ

By solving Eq. (6), the concentrations of solutes on the membrane surface will be obtained [34] using the following boundary conditions: x ¼ 0; c ¼ cb and x ¼ δ; c ¼ w  c   J ci;w ¼ ci;p þ ci;b −ci;p exp v k

ð8Þ

where δ is the thickness of film layer, ci,w and ci,p are the solute concentration on the membrane surface and in the permeate solution, respectively. For single salt solution, k is the mass transfer coefficient defined as: k¼

Dsalt : δ

ð9Þ

The mass transfer coefficient is calculated by a Sherwood number (Nsh) correlation which is defined as a function of Reynolds number (Re) and Schmidt number (Sc) as follows: NSh ¼

k dh a a ¼ a1 Re 2 Sc 3 Dsalt

ð10Þ

where dh is the hydraulic diameter of the feed channel passing over the membrane, and a1, a2 and a3 are the empirical coefficients related to hydrodynamic conditions. Reynolds number and Schmidt number are expressed as follows: Re ¼

ρ u dh μ

ð11Þ

Sc ¼

μ ρ Dsalt

ð12Þ

where u is the fluid velocity in the channel with hydraulic diameter of dh, ρ and μ are the density and dynamic viscosity of the feed solution, respectively. By measuring the solute concentration in the feed solution (ci,b), the observed rejection is given by: R∘ ¼ 1−

Fig. 2. Schematic representation of mass balance for a system containing two charged ions in one salt.

ci;p : ci;b

ð13Þ

Because of concentration polarization phenomenon, the solute concentration in the vicinity of the membrane (ci,w) is higher than that in

S. Mehdipour et al. / Desalination 362 (2015) 84–92 100

the bulk solution and can be obtained from Eq. (8). The real (intrinsic) rejection (R) is defined as the following: ci;p : ci;w

90

ð14Þ

4. Results and discussion

(a)

95

Rejecon (%)

R ¼ 1−

87

4.1. Surface charge properties of NF membrane

85 80 75

20 ppm 50 ppm 100 ppm 200 ppm 400 ppm

70

4.2. Experiments using single salt solution

65 60 5

10

15

20

25 Pressure (bar)

30

35

25 KCL, 0.001M

20 15 10

ζ (mv)

5 0 0

-5

1

2

3

4

5

6

7

pH

-10 -15 -20 -25 Fig. 3. Zeta potential (ζ) of NE4040-90 membrane against pH using 1 mM KCl solution.

45

R² = 0.9928 R² = 0.9817

100 80 60

20 ppm 50 ppm 100 ppm 200 ppm 400 ppm

40

4.2.1. Influence of pressure and metal concentration on lead rejection To study the influence of lead concentration on the Pb2+ retention and permeate fluxes of the NF membrane, the initial concentration of lead ion in the feed solution was adjusted by adding definite amount of Pb(NO3)2 to demineralized water. The experiments were performed at the natural pH of Pb(NO3)2 solution equal to 5.6. The rejection of lead ion after 2 h experiment and water fluxes versus pressure are shown in Fig. 4a and b, respectively. Fig. 4a represents the trend of Pb2+ (Pb(NO3)2) rejection versus pressure at different concentrations of Pb2+ ion. The rejection of lead ion improved with an increase in the pressure, reaching a maximum value between 88.5 and 97.5% at various concentrations. Increasing the pressure at high feed concentrations had less effect on the lead ion rejection. When the transmembrane pressure increases, two phenomena eventuate, simultaneously: 1) by increasing the pressure, more solute is forced to the membrane surface that leads to concentration polarization which consequently solute rejection will be decreased; 2) by increasing the pressure, the solvent flux will be increased, but the solute transport across the membrane is hindered by steric and electrical effects [20]. As a result, an increase in solute rejection is obtained as water permeation becomes greater at higher pressures. However, the solute diffusion is not affected by the pressure since it is controlled by the solute concentration [36,37]. Fig. 4a also

40

R² = 0.9926 R² = 0.9904 R² = 0.992

(b)

120

Flux (kg/m2h)

The experimental data for calculating surface zeta potential of NE 4040-90 membrane in the pH range of 3.0 to 7.0 using 1 mM KCl electrolyte are indicated in Fig. 3. Membrane charge is a function of several phenomena where adsorption of differently charged solutes of the electrolyte and detachment of functional groups of the membrane [20] are the most important factors. As shown, the zeta potential (ζ) varies with the pH. The membrane is positively charged at pH 3.0 and zeta potential of the membrane surface decreases at the higher pH values until the isoelectric point (pH 5.3) and is negatively charged above this value. The net charge of the membrane around the isoelectric point is almost zero where the membrane behaves like a nonpolar surface [35].

20 0 5

10

15

20

25 30 Pressure (bar)

35

40

45

Fig. 4. Influence of pressure at five different concentrations of 20, 50, 100, 200 and 400 mg/L of Pb2+ (a) Pb2+ intrinsic rejection (b) permeated fluxes (25 °C, pH 5.6).

shows that the lead rejection is increased with concentration of lead. The analysis of retention data shows that at higher concentrations, lead rejection is increased significantly. Krieg et al. [38] found that sulfate and chloride rejection increases at higher concentrations and the rejection of calcium and sodium increases with their concentrations, too. They described these results as a verification of the membrane charge which was negative at very low concentrations of cations and positive at higher concentrations of cations. Also, this behavior can be explained by membrane fouling due to the adsorption of solute on the membrane surface that decreases the effective pore size of the NF membranes [39]. By increasing the lead ions concentration, adsorption of these ions on the membrane surface could be a reason for this behavior. As a result of adsorption of Pb2 + on the membrane surface, the membrane charge becomes positive that eventuates in more retention of lead ions [40]. The trend of fluxes in several lead ion concentrations against pressure is exhibited in Fig. 4b. The permeate fluxes increased almost linearly with the applied transmembrane pressure for each concentration. 4.2.2. Pressure effect on permeate flux and lead rejection in PbCl2 and PbSO4 solutions Experiments aiming to study the effect of monovalent or divalent anions were performed at a constant concentration of lead ion (20 mg Pb2+/L) and constant pH values equal to 5.6 for Pb(NO3)2 solution, 5.4 for PbSO4 and 6.5 for PbCl2 solutions at 10 to 40 bar. Fig. 5a shows the effect of pressure on the lead ion retention. The trend of the higher pressure the higher the lead ion retention was observed, between 10 and 40 bar. The retention of lead ion was enhanced when the charge of the associated anion was increased. The observed order of the lead ion retention was PbSO4 N PbCl2 N Pb(NO3)2 which shows that the surface forces are stronger when divalent anions are present. This behavior can be explained by Donnan exclusion phenomenon [25]. The abovementioned

88

S. Mehdipour et al. / Desalination 362 (2015) 84–92 100

difference between osmotic pressure of the boundary solution (film layer) and that of the bulk solution is insignificant, and concentration polarization is negligible [43].

(a)

95 90

Rejecon (%)

85 80 75 sulfate 70 chloride nitrate

65 60 5

10

15

20

25

30

35

40

45

Pressure (bar) 180

(b)

160

R² = 0.9957

140

R² = 0.9901

Flux(kg/m2.h)

120 R² = 0.995 100 80 60 nitrate

40

chloride 20

sulfate

4.2.3. Influence of solution pH Fig. 6a shows the rejection of Pb2+ as a function of feed pH for 20 mg 2+ Pb /L with different anion solutions. It can be seen that lead ion is highly rejected for all studied anions tested. The difference in lead ion rejection as a function of pH for divalent anion is smaller than that obtained for monovalent anions. This has also been observed in other studies with different NF membranes [44]. A minimum lead ion rejection was observed when the pH was equal to the natural pH of the solution (in case of PbCl2 and PbSO4 solution) or slightly higher than the natural pH (in case of Pb(NO3)2 solution). As shown in Fig. 6a, lead ion rejection has a maximum value at pH 3.0. It drops to a minimum value near the natural pH of solutions. For instance, the rejections of lead ion were 92.5, 89 and 88.6% for PbSO4, PbCl2 and Pb(NO3)2 solutions, respectively. When the pH of solutions reached to 4.0 and 3.0, − the concentration of proton and the associated anions (NO− 3 , Cl and 2− SO4 ) was increased and the system acted like a ternary solution (Pb2+, H+ and anion). By adding acids and increasing the concentration of common anion between salt and acid, a higher concentration of anion leads to more and easy passage of anion across the membrane. In order to maintain the electro-neutrality condition, because of its very smaller size, more protons are crossed across the membrane and the Pb2+ ions were rejected strongly due to their higher valence. Mehiguene et al. [45] expressed that at higher acidity, protons might neutralize the negative sites of the membrane surface, which can reduce the repulsion between

0 5

10

15

20

25

30

35

40

45

Pressure (bar) Fig. 5. Influence of pressure on (a) Pb2+ intrinsic rejection and (b) permeated fluxes using PbSO4, PbCl2 and Pb(NO3)2 solutions (Pb2+ ion concentration of 20 mg/L at 25 °C).

order is observed because the divalent SO 2− ion is strongly rejected by 4 the negatively charged functional groups of membrane compared to the − 2+ counter ion). Accordingly, monovalent NO − 3 or Cl (for the same Pb cations are also rejected to ensure the electro-neutrality at both sides of the membrane. The retention is increased by increasing the valence of a co-ion (in this case, anion) due to the growing repulsion between the membrane surface and higher valance ions [41]. It seems that in the cases of PbCl2 and Pb(NO3)2, hydration energy is an important factor among influencing factors on the selectivity of a nanofiltration membrane. The hydration energy of Cl¯ and NO¯3 are 378 and 317 kJ/mol (Table 1), respectively. The nitrate ions are less hydrated than the chloride ions that results in less rejection of cations associated to Pb2+ [42]. At higher pressures, the size effect of solutes does not seem to have an important role in the rejection of Pb2 + ion and the high pressure of the feed flow is dominant to the sieve effect. As can be seen in Fig. 5b, the membrane flux remains linear against pressure. The

Table 1 Diffusion coefficient [49,50], hydration radii and hydration energy of ions [51]. Ion 2+

Pb Cd2+ Cu2+ Zn2+ Na+ NH+ 4 NO− 3 − Cl SO2− 4 a

Di,∞ × 10−9 (m2/s)

RHa (nm)

Hydration energy (kj/mol)

0.945 0.719 0.714 0.703 1.334 1.957 1.902 2.032 1.065

0.260 0.337 0.337 0.348 0.183 0.124 0.128 0.120 0.229

−1485 −1809 −2099 −2047 −406 −307 −314 −378 −1047

The hydration radii for all ions are determined using the Stokes–Einstein equation.

Fig. 6. Influence of pH on (a) Pb2+ observed rejection (%) and (b) permeated fluxes with PbSO4, PbCl2 and Pb(NO3)2 solutions (Pb2+ concentration of 20 mg/L at 25 °C and 20 bar).

S. Mehdipour et al. / Desalination 362 (2015) 84–92

anions and membrane surface. In fact, at lower pH values, the charge of the membrane becomes more positive that influences the lead ion rejection which is highly governed by the charge effect rather than the sieve effect. At the pH values higher than the natural pH of the solutions, the rejection of lead ion increased while the concentration of Na+ and the negative charge of the membrane showed an increase as well (Fig. 3). Pore size of the membrane becomes smaller at higher pH values due to dissociation of carboxyl groups of the membrane and the repulsion between them that leads to a higher solute rejection [35,46]. At a higher pH values of the feed solution (about 7.0), metal ions are capable of forming complexes with OH− ion which can lead to the formation of the insoluble hydroxides [39]. Membranes reject these hydroxides easily and therefore, the rejections of lead ions will be greater. On the other hand, NE 4040-90 membrane posses both carboxyl groups (\COOH) and ammonium groups (\NH+ 3 ) where both at low and high pH values, ammonium groups and carboxyl groups are protonated and deprotonated, respectively. In both cases, electrostatic repulsion between the charged groups results in an increase in the rejection of the corresponding solutes [46]. Fig. 6b shows the permeate flux versus the pH of the feed solution. The permeate flux increased against pH until a maximum value which attained at the natural pH of each solution. This might be due to the occurrence of isoelectric point of the membrane at the natural pH of solutions. Similar results have been reported in previous studies [20,35,36]. Childress and Elimelech [46] explained that this observation is related to three factors that include membrane pore size, electro-viscous effect and osmotic pressure gradient at the membrane surface. According to these phenomena, when the membrane surface possess a charge (positive or negative), membrane pore size becomes smaller. This is due to the repulsion between the charged groups on the active layer of the membrane and the solutes which results in a less flux and a higher rejection, accordingly. At the pH 3.0, the solute rejection was at the maximum level and the osmotic pressure near the membrane surface was higher. Higher osmotic pressure causes a decrease in the net driving pressure which leads to a decrease in the permeate flux [35,46].

89

Fig. 7. Effect of the nature and the concentration of the added monovalent cations on the lead ion rejection: (a) Na+ and (b) NH+ 4 at 25 °C and 20 bar.

4.3. Lead rejection in binary-salt solutions Many industrial effluents contain more than one heavy metal. Having enough knowledge about the interactions between these cations and their co-anions is important to achieve a desirable removal condition. In this regard, influence of divalent and monovalent cations including cadmium, copper, zinc, sodium and ammonium on the rejection of lead ion (from Pb(NO3)2) was examined. In order to study the effect of the interfering cations concentrations on lead retention, the concentration of lead ion was maintained constant, but the amount of other cations was changed. The results showing lead ion removal in the presence of interfering cation in a binary solution containing monovalent cations and divalent cations are presented in Figs. 7 and 8, respectively. As shown in Fig. 7a and b, when concentration of the monovalent cations increases, the retention of lead almost remains constant. In the presence of sodium and ammonium with the concentrations of 50 mg/L, the rejection of lead ion was 92.2 and 91.2%, respectively. At the cation concentration of 400 mg/L, the rejection was 90 and 89.5%, respectively. When the lead concentration was 100 mg/L at the TMP equal to 20 bar (in the single salt solution), the rejection of lead ion was 92% which did not change significantly. Even in the presence of sodium ion, the rejection of lead ion increased slightly compared to the case that a single salt was present. It seems that in this case; presence of more permeable cation (sodium) resulted in more rejection of the lead ion. Su et al. reported a higher Ca2 + rejection in the presence of Na+ ion (more permeable than Ca2+) as compared with a single salt solution [47]. Hodgson expressed that the permeability of a given ion may become extremely high when a less permeable ion is present [48]. In this study, in the presence of Na+ ions, the permeability of Pb2+ ions

decreased which resulted in a higher rejection for this ion. The rejections of Na+ and NH+ 4 increased to 85.5 and 84.3%, respectively (at the concentration of 100 mg/L) and then remained almost at a constant level. In the binary salt mixture with a higher nitrate concentration (compared to the single salt solution), the lower valence of this ion leads to an easy passage of nitrate ions across the membrane. Lead ions were strongly rejected by the membrane because of their high valence. More sodium and ammonium ions passed through the membrane to keep the charge balance across the membrane. The rejection of other solutions containing lead ion together with divalent cations are represented in Fig. 8a, b and c. The rejection of lead ion decreases at higher concentrations of the interfering cations. A substantial decrease in the lead ion rejection was observed when the concentration of added divalent cations increases. The rejection of lead ion was 84, 83.7 and 85.4% in the presence of 100 mg/L of Cu2+, Zn2+ and Cd2+, respectively. A comparison between the effects of the presence of different cations (with the concentration of 100 or 400 mg/L) on the rejection of lead ion is given in Fig. 9. As shown, higher concentration of these cations has a negative effect on the lead ion rejection. Divalent cations had more influence on the reducing the lead ion rejection compared to monovalent cations. The sequence of interfering cations on reducing lead ion rejections was Cu2+ N Zn2+ N Cd2+. This difference for cations with equal valence is related to the hydration energies of cations, their hydrated radii [20,38,41,43] and the diffusivities of ions [19,36]. It seems that hydration energy of cations is an important factor here. As shown in Table 1, the order of cation's hydration energy is

90

S. Mehdipour et al. / Desalination 362 (2015) 84–92 100

(a) 95

Rejecon (%)

90 Cu(NO3)2 85

Pb(NO3)2

80 75 70 0

50

100

150

200

250

300

350

400

450

Cu2+ concentraon (mg/L)

Cu2+ N Zn2+ N Cd2+ N Pb2+ where Pb2+ possesses the smallest value among these ions. Therefore, presence of any other cation causes a significant reduction in Pb2 + rejection. Considering the order of hydrated radii (Zn2 + N Cu 2 + ≈ Cd 2 + N Pb2 + ) proves that Pb 2 + has the smallest value which is a sign of less hydration of lead cations. This fact, results in the formation of smaller size hydrated cation in the solution that can transport easier due to the steric effects. Another reason for the observed rejection order can be attributed to the diffusion coefficients of the existing cations in the solution. Assuming an approximation for the largeness of diffusion coefficients of cations equal to those observed in aqueous solution, they would be in the following order: Pb2 + N Cd2 + N Cu2 + N Zn2 +. Therefore, Pb 2 + could pass through the membrane more easily due to its higher diffusion which results in a lower rejection.

100

(b)

5. Conclusions

(c)

A polyamide nanofiltration membrane was examined for the removal of lead ions. It was observed that the lead ion rejection depends on the applied pressure, feed concentration, feed pH and nature of interfering cations. The maximum lead rejection (using lead nitrate solution) was 97.5% at 30 bar and 400 mg/L of Pb2+ concentration. At the pressure of 40 bar with different lead ion concentrations (20 mg/L to 400 mg/L), the removal of lead ion varied from 81% to 96.5%. Changing the associated anions had a significant effect on the lead ion rejection. was applied as an interfering anion, the rejection of lead When SO2− 4 − ion increased to 90% while in cases of NO− 3 and Cl , its rejection was 81 and 83%, respectively. pH played as a critical factor that influenced the solute rejection and the flux of permeate as well. The charge of the membrane was significantly influenced by pH change. The charge of the membrane was positive below the pH 5.3 and the membrane was negatively charged above this value. The presence of associated anions at the both high and low pH values of the feed resulted in a higher rejection of the lead ion and a lower flux of the permeate. The presence of monovalent cations (Na+, NH+ 4 ) did not cause a remarkable decrease in the lead ion rejection. Even in the presence of Na+, more rejection of the lead ion was observed in comparison with the single salt solution. However, increasing the interfering divalent cation concentration (Cu2+, Zn2+ and Cd2+) decreased the lead ion rejection significantly. The highest reduction in the lead ion rejection was observed in a solution containing Cu2 + ions. The order of rejection of the cations found to be Cu2 + N Zn2 + N Cd2 + N Pb2 +. This behavior is due to difference between the hydration energy of cations where Cu2+ had the highest value that caused the most reduction in the lead ion reduction.

Rejecon (%)

95 90 85 80 Zn(NO3)2

75

Pb(NO3)2 70 0

50

100

150

200

250

300

350

400

450

Zn2+ concentraon (mg/L) 100

Rejecon (%)

95 90 85 80 Cd(NO3)2

75

Pb(NO3)2 70 0

50

100

150

200

250

300

350

400

450

Cd2+ concentraon (mg/L) Fig. 8. Effect of the nature and the concentration of the added divalent cations on the lead ion rejection (100 mg/L) in the binary solution: (a) Cu2+, (b) Zn2+ and (c) Cd2+ at 25 °C and 20 bar.

100 100 mg/l

95

92.2 90

90

90.9

400 mg/l 89.5 84.6

Rejecon (%)

85

85.38

83.74 79.75

80.22

80 74.6

75 70 65 60 55 50 Na+

NH4+

Cu2+

Zn2+

Cd2+

Fig. 9. Influence of cation type and concentration on the Pb2+ ion rejection in the binary solution.

List of symbols

U P L Q R KB Ji Di,∞ zi ci x F R T Jv Ci,p Dsalt Ci,b

streaming potential pressure of electrolyte flow length of capillary system cross-sectional area of capillary system AC resistance of cell specific electrical conductivity molar flux of ion “i” diffusion coefficient of ion “i” in bulk solution charge number of ion “i” ion concentration distance Faraday constant ideal gas constant absolute temperature permeate flux concentration of ion “i” in permeate solution effective diffusivity of salt concentration of ion “i” in feed bulk solution

S. Mehdipour et al. / Desalination 362 (2015) 84–92

Ci,w concentration of ion “i” in membrane surface k mass transfer coefficient Sherwood number NSh hydraulic diameter of feed channel dh a1, a2 and a3 empirical coefficients Re Reynolds number Sc Schmidt number u fluid velocity in the channel observed rejection Ro R intrinsic rejection hydration radius RH

Greek symbols

ε εo ρ ζ ζs ζtot ζPMMA η ψ δ μ

dielectric constant of electrolyte permittivity density of feed solution zeta potential net zeta potential of the sample zeta potential of sample plus PMMA zeta potential of PMMA electrolyte viscosity electrical potential within the film layer thickness of film layer dynamic viscosity of feed solution

Acknowledgment The authors would like to thank the Petrochemical Research & Technology Co. (Tehran, Iran) for the financial support of this project (grant no. 0870289208) and Dr. Yousof Mohammadi for his extensive support during this research. References [1] M.J. González-Muñoz, M.A. Rodríguez, S. Luque, J.R. Álvarez, Recovery of heavy metals from metal industry wastewaters by chemical precipitation and nanofiltration, Desalination 200 (2006) 742–744. [2] S. Babel, T.A. Kurniawan, Cr(VI) removal from synthetic wastewater using coconut shell charcoal and commercial activated carbon modified with oxidizing agents and/or chitosan, Chemosphere 54 (2004) 951–967. [3] S. Prakash, M. Kumar, B.P. Tripathi, V.K. Shahi, Sol–gel derived poly(vinyl alcohol)-3(2-aminoethylamino) propyl trimethoxysilane: cross-linked organic–inorganic hybrid beads for the removal of Pb(II) from aqueous solution, Chem. Eng. J. 162 (2010) 28–36. [4] M.M. Matlock, B.S. Howerton, D.A. Atwood, Chemical precipitation of heavy metals from acid mine drainage, Water Res. 36 (2002) 4757–4764. [5] Y. Ku, I. Jung, Photocatalytic reduction of Cr (VI) in aqueous solutions by UV irradiation with the presence of titanium dioxide, Water Res. 35 (2001) 135–142. [6] J.L. Huisman, G. Schouten, C. Schultz, Biologically produced sulphide for purification of process streams, effluent treatment and recovery of metals in the metal and mining industry, Hydrometallurgy 83 (2006) 106–113. [7] A. Özverdi, M. Erdem, Cu2+, Cd2+ and Pb2+ adsorption from aqueous solutions by pyrite and synthetic iron sulphide, J. Hazard. Mater. 137 (2006) 626–632. [8] J.H. Lee, S.J. Nam, Equilibrium in the extraction of Pb(II) by D2EHPA and dithizone, Hwahak Konghak 28 (1990) 264–270. [9] E. Pehlivan, T. Altun, Ion exchange of Pb2+, Cu2+, Zn2+, Cd2+, and Ni2+ ions from aqueous solution by Lewatit CNP 80, J. Hazard. Mater. 140 (2007) 299–307. [10] S.Y. Kang, J.U. Lee, S.H. Moon, K.W. Kim, Competitive adsorption characteristics of Co2+, Ni2+, and Cr3+ by IRN-77 cation exchange resin in synthesized wastewater, Chemosphere 56 (2004) 141–147. [11] B. Alyüz, S. Veli, Kinetics and equilibrium studies for the removal of nickel and zinc from aqueous solutions by ion exchange resins, J. Hazard. Mater. 167 (2009) 482–488. [12] E.B. Susan, J.O. Trudy, B. Mark, A.D. Dean, A review of potentially low-cost sorbent for heavy metals, Water Res. 33 (1999) 2469–2479. [13] V.G. Serrano, A.M. Garcia, A.E. Mansilla, C.V. Calahorro, Adsorption of mercury, cadmium and lead from aqueous solution on heat-treated and sulphurized activated carbon, Water Res. 32 (1998) 1–4.

91

[14] M. Ghaedi, F. Ahmadi, A. Shokrollahi, Simultaneous preconcentration and determination of copper, nickel, cobalt and lead ions content by flame atomic absorption spectrometry, J. Hazard. Mater. 142 (2007) 272–278. [15] D. Dong, L. Liu, X. Hua, Y. Lu, Comparison of lead, cadmium, copper and cobalt adsorption onto metal oxides and organic materials in natural surface coatings, Microchem. J. 85 (2007) 270–275. [16] H. Eccles, Treatment of metal-contaminated wastes: why select a biological process? Trends Biotechnol. 17 (1999) 462–465. [17] M.J.G. Muñoz, M.A. Rodriguez, S. Luque, J.R. Alvarez, Recovery of heavy metals from metal industry wastewaters by chemical precipitation and nanofiltration, Desalination 200 (2006) 742–744. [18] S.M.C. Ritchie, D. Bhattacharyya, Membrane-based hybrid processes for high water recovery and selective inorganic pollutant separation, J. Hazard. Mater. 92 (2002) 21–32. [19] S. Bouranenea, P. Fievet, A. Szymczykb, M.E. Samarc, A. Vidonnea, Influence of operating conditions on the rejection of cobalt and lead ions in aqueous solutions by a nanofiltration polyamide membrane, J. Membr. Sci. 325 (2008) 150–157. [20] C. Gherasim, J. Cuhorka, P. Mikulasek, Analysis of lead(II) retention from single salt and binary aqueous solutions by a polyamide nanofiltration membrane: experimental results and modelling, J. Membr. Sci. 436 (2013) 132–144. [21] K. Linde, A.S. Jonsson, Nanofiltration of salt solutions and landfill leachate, Desalination 103 (1995) 223–232. [22] G.L. Amy, B.C. Alleman, C.B. Cluff, Removal of dissolved organic matter by nanofiltration, J. Environ. Eng. 116 (1990) 200–205. [23] H. Starthmann, Membrane separation processes, J. Membr. Sci. 9 (1981) 121–189. [24] A.E. Childress, M. Elimelech, Effect of solution chemistry on the surface charge of polymeric reverse osmosis and nanofiltration membranes, J. Membr. Sci. 119 (1996) 253–268. [25] Y.Z. Xu, R.E. Lebrun, Investigation of the solute separation by charged nanofiltration membrane: effect of pH, ionic strength and solute type, J. Membr. Sci. 158 (1999) 93–104. [26] C.V. Gherasim, P. Mikulášek, Influence of operating variables on the removal of heavy metal ions from aqueous solutions by nanofiltration, Desalination 343 (2014) 67–74. [27] W. Saikaew, S. Mattaraj, R. Jiraratananon, Nanofiltration performance of lead solutions: effects of solution pH and ionic strength, WSTWS 10 (2010) 193–200. [28] J. Schaep, B. Van der Bruggen, C. Vandecasteele, D. Wilms, Influence of ion size and charge in nanofiltration, Sep. Purif. Technol. 14 (1998) 155–162. [29] M. Smoluchowski, Handbook of Electricity and Magnetism, vol. 2, Barth, Leipzig, 1921. 366. [30] H.A. Abramson, Electrokinetic phenomena and their application to biology and medicine, J. Phys. Chem. 38 (8) (1934) 1128–1129. [31] H.J. Jacobasch, J. Schurz, Characterization of polymer surfaces by means of electrokinetic measurements, Progr. Colloid Polym. Sci. 77 (1988) 40–48. [32] F. Fairbrother, H. Mastin, Studies in electroendosmosis, J. Chem. Soc. 125 (1924) 2319–2330. [33] Sh.L. Walker, S. Bhattacharjee, E.M.V. Hoek, M. Elimelech, A novel asymmetric clamping cell for measuring streaming potential of flat surfaces, Langmuir 18 (2002) 2193–2198. [34] W.R. Bowen, A.W. Mohammad, Diafiltration by nanofiltration: prediction and optimization, J. AIChE 44 (1998) 1799–1812. [35] M.R. Teixeira, M.J. Rosa, M. Nystrom, The role of membrane charge on nanofiltration performance, J. Membr. Sci. 265 (2005) 160–166. [36] B.A.M. Al-Rashdi, D.J. Johnson, N. Hilal, Removal of heavy metal ions by nanofiltration, Desalination 315 (2013) 2–17. [37] S. Pérez-Sicairos, S.W. Lin, R.M. Félix-Navarro, H. Espinoza-Gómez, Rejection of As(III) and As(V) from arsenic contaminated water via electro-cross flow negatively charged nanofiltration membrane system, Desalination 249 (2009) 458–465. [38] H.M. Krieg, S.J. Modise, K. Keizer, H.W.J.P. Neomagus, Salt rejection in nanofiltration for single and binary salt mixtures in view of sulphate removal, Desalination 171 (2004) 205–215. [39] Z. Wang, G. Liu, Z. Fan, X. Yang, J. Wang, S. Wang, Experimental study on treatment of electroplating wastewater by nanofiltration, J. Membr. Sci. 305 (2007) 185–195. [40] C.V. Chung, N.Q. Buu, N.H. Chau, Influence of surface charge and solution pH on the performance characteristics of a nanofiltration membrane, Sci. Technol. Adv. Mater. 6 (2005) 246–250. [41] K. Mehiguene, Y. Garba, S. Taha, N. Gondrexon, G. Dorange, Influence of operating conditions on the retention of copper and cadmium in aqueous solutions by nanofiltration: experimental results and modelling, Sep. Purif. Technol. 15 (1999) 181–187. [42] A.P. Padilla, H. Saitua, Performance of simultaneous arsenic, fluoride and alkalinity (bicarbonate) rejection by pilot-scale nanofiltration, Desalination 257 (2010) 16–21. [43] L. Paugam, S. Taha, G. Dorange, P. Jaouen, F. Quéméneur, Mechanism of nitrate ions transfer in nanofiltration depending on pressure, pH, concentration and medium composition, J. Membr. Sci. 231 (2004) 37–46. [44] A. Santafé-Moros, J.M. Gozálvez-Zafrilla, J. Lora-García, Nitrate removal from ternary ionic solutions by a tight nanofiltration membrane, Desalination 204 (2007) 63–71. [45] K. Mehiguene, S. Taha, N. Gondrexon, J. Cabon, G. Dorange, Copper transfer modeling through a nanofiltration membrane in the case of ternary aqueous solution, Desalination 127 (2000) 135–143. [46] A. Childress, M. Elimelech, Relating nanofiltration membrane performance to membrane charge (electrokinetic) characteristics, Environ. Sci. Technol. 24 (2000) 3710–3716.

92

S. Mehdipour et al. / Desalination 362 (2015) 84–92

[47] M. Su, D. Wang, X. Wang, M. Ando, T. Shintani, Rejection of ions by NF membranes for binary electrolyte solutions of NaCl, NaNO3, CaCl2 and Ca(NO3)2, Desalination 191 (2006) 303–308. [48] T.D. Hodgson, Selective properties of cellulose acetate membranes towards ions in aqueous solutions, Desalination 8 (1970) 99–138.

[49] J.A. Dean, Lange's Handbook of Chemistry, 15th Edition McGraw-Hill, New York, 1999. [50] P. Vanysek, Ionic conductivity and diffusion at infinite dilution, in: D.R. Lide (Ed.), CRC Handbook of Chemistry and Physics, 83rd ednCRC Press, Boca Raton, 2002. [51] C.S.G. Philips, R.J.P. Williams, Inorganic Chemistry, Clarendon, Oxford, 1965.