Assessment of the chemical stability of nanofiltration and reverse osmosis membranes employed in treatment of acid gold mining effluent

Accepted Manuscript Assessment of the chemical stability of nanofiltration and reverse osmosis membranes employed in treatment of acid gold mining effluent Bárbara C. Ricci, Carolina D. Ferreira, Larissa S. Marques, Sofia S. Martins, Beatriz G. Reis, Míriam C.S. Amaral PII: DOI: Reference:

S1383-5866(16)30739-0 http://dx.doi.org/10.1016/j.seppur.2016.11.007 SEPPUR 13340

To appear in:

Separation and Purification Technology

Received Date: Revised Date: Accepted Date:

14 June 2016 5 November 2016 7 November 2016

Please cite this article as: B.C. Ricci, C.D. Ferreira, L.S. Marques, S.S. Martins, B.G. Reis, M.C.S. Amaral, Assessment of the chemical stability of nanofiltration and reverse osmosis membranes employed in treatment of acid gold mining effluent, Separation and Purification Technology (2016), doi: http://dx.doi.org/10.1016/j.seppur. 2016.11.007

This is a PDF file of an unedited manuscript that has been accepted for publication. As a service to our customers we are providing this early version of the manuscript. The manuscript will undergo copyediting, typesetting, and review of the resulting proof before it is published in its final form. Please note that during the production process errors may be discovered which could affect the content, and all legal disclaimers that apply to the journal pertain.

ASSESSMENT OF THE CHEMICAL STABILITY OF NANOFILTRATION AND REVERSE OSMOSIS MEMBRANES EMPLOYED IN TREATMENT OF ACID GOLD MINING EFFLUENT Bárbara C. Ricci a*; Carolina D. Ferreira a; Larissa S. Marquesa; Sofia S. Martins; Beatriz G. Reis a; Míriam C. S. Amarala

a

Department of Sanitary and Environmental Engineering - Federal University of Minas

Gerais. P.O. Box 1294, ZIP 30.270-901, Belo Horizonte, MG, Brazil. * Corresponding author: Tel.: +55 31 3409-3669; fax: +55 31 97166-1087. E-mail address: [email protected] . ABSTRACT As reported in our previous work, the association of nanofiltration (NF) and reverse osmosis (RO) has proved to be a promising treatment for pressure oxidation process (POX) effluent, generated in gold ore processing. Despite the excellent performance achieved, it was essential to evaluate if the membrane performance could be impaired by continuous exposure to effluent. Accordingly, this study aimed to investigate the effect of continuous exposure to POX effluent on characteristics of nanofiltration (MPF34) and reverse osmosis (TFC-HR) membranes, used in POX effluent treatment. The membranes were immersed in the effluent and in acid solutions in order to simulate the continuous exposure during the treatment. The effect of exposure on NF membrane was evaluated based on the rejections of glucose, magnesium sulfate, sulfuric acid, and cobalt and nickel sulfates. The effect of exposure on the RO membrane was evaluated by assay of the rejections of sodium chloride and sulfuric acid. For both membranes, the effects of the exposure on the hydraulic permeability, hydrophobicity, and chemical and morphological characteristics were also evaluated. For MPF-34, the cobalt and nickel rejection decreased by 33% during exposure period, indicating a reduction of the NF ability in separation of metals from acid solution. TFC-HR membrane exhibited satisfactory stability under the acidic conditions employed in this investigation, thus exhibiting potential for practical implementation.

KEYWORDS: pressure oxidation effluent, nanofiltration, reverse osmosis, sulfuric acid recovery, metal recovery, water reuse. 1

INTRODUCTION

The pressure oxidation process (POX) is frequently employed as a pre-oxidation step for refractory gold ores. This technique consists of the hydrometallurgical oxidation of concentrated ore pulp performed under high pressure (~20 atm) and temperature (135200°C). During the pressure oxidation process, the sulfide in the host mineral matrix is oxidized by oxygen, producing soluble metal sulfates and sulfuric acid. As a result, the gold originally occluded in the sulfide mineral is completely released, allowing a high gold recovery by cyanide leaching[1, 2]. Despite the efficiency of POX, it is responsible for the generation of a large volume of liquid effluents. Those are characterized by its high acidity and substantial metal content, some of these of great economic value, like Cu, Co and Ni. In order to treat this effluents, industry usually carry out a neutralization using alkalizing agents such as quicklime or hydrated lime, a method with high simplicity and easy to operate. However, the production of large volume of precipitate (sludge) with significant heavy metal content is the main drawback of this technology. It also represents a risk for the environment, requiring appropriate disposition, which leads to high costs for its final disposal as well[3, 4]. In view of the high acid demand of the mining industry, the commercial value of metals that are present in POX effluent, and the high water consumption during gold ore processing, the recovery of dissolved metals, sulfuric acid, and water from POX effluent consists in a potential eco-efficient alternative to reduce neutralization costs and an opportunity to add value to this effluent. In this context, the membrane separation process, especially the combination of nanofiltration (NF) and reverse osmosis (RO), is a promising treatment technology. NF allows higher permeate fluxes in comparison to RO but it is still highly efficient on retaining organic compounds and multivalent salts [5]. NF is also known to be charged and present a membrane isoelectric point (IEP), that is usually between 3 and 6. Thus, at acid pH, the membrane surface shows a positive charge [6, 7], a phenomenon that directly influence NF retention mechanisms. In this way, when treating acid waters like POX effluent, NF positively charged membrane can strongly reject multivalent metal

cations while acidic anions are allowed to permeate almost freely through the membrane [8]. This characteristics make NF a very promising technology for the separation of metal species from acid streams. RO, in turn, is a highly efficient process on the retention of dissolved salts and organic molecules with low molar mass [5], achieving 99% of retention of those. Its mainly application is on seawater desalination, being responsible for the production of 20% of all desalinated water worldwide [9]. Its characteristics made RO to become increasingly attractive as an industrial effluents treatment technology especially when water recovery is desired [10, 11]. As RO also produces a very concentrated stream when producing a high quality permeate, it can be also apllied for altering the concentration of existing solutions, as reported by Ahrsan et al. [12]. Those authors obtained a concentrate with an acid content 400% greater than the feed acetic acid solution using RO. The integration of NF and RO for treatment of acidic streams contaminated with metals has been reported in recent years in several articles and patents. Some applications include phosphoric acid purification[13, 14]; purification of nitric acid used in the production of picture tubes [15]; metal and acid recovery from electroplating effluent [16];and sulfuric acid purification [17]. In addition to the cited applications, the association of NF and RO has proved to be a promising treatment for effluent from the gold mining industry, specifically, pressure oxidation process effluent. As reported in our previous paper [18], this association allowed for the recovery of a purified acid stream, which may be reused in the ore processing; the production of a metal enriched stream, which may be transferred to a subsequent metal recovery stage; and the generation of high quality reuse water. Despite the excellent performance achieved, it is essential to evaluate if the membrane performance could be impaired by continuous exposure to effluent. Several authors assessed the effect of continuous exposure to acid solutions on the performance of NF and RO membranes [14, 19-22]. However, effects of continuous exposure were not always evident. Some studies reported that the same type of membrane can be chemically stable or unstable depending on the characteristics of the solution to which is exposed [19, 22]. Improvement of the membrane performance with continued exposure to acid solution has even been reported[14]. Thus, it is necessary to evaluate the stability of the membrane towards the specific solution to which will be exposed.

In this context, the effect of continuous exposure to the pressure oxidation effluent on NF (MPF-34) and RO (TFC-HR) membranes characteristics was investigated. Accordingly, the membranes were immersed in the effluent and acid solutions in order to simulate continuous exposure to effluent during the treatment process proposed in our previous work. For the NF membrane, the effect of exposure on the rejections of glucose and magnesium sulfate was evaluated. In addition, the effect of exposure on the NF performance in the separation of sulfuric acid and metal sulfates was evaluated. For the RO membrane, the effect of exposure on the rejections of sodium chloride and sulfuric acid was evaluated. For both membranes, the effect of exposure on the hydraulic permeability, hydrophobicity, and chemical and morphological characteristics was also assessed. 2 2.1

EXPERIMENTAL Pressure-oxidation process effluent

The effluent from the pressure-oxidation process was supplied by a gold mining company located in the state of Minas Gerais, Brazil. The effluent is characterized by high acidity (median pH was 1.30) and high metal concentrations (Table 1). Additionally, some metals of high commercial value are present in this effluent, such as cobalt and nickel. Table 1: Physicochemical characterization of the microfiltrated pressure oxidation effluent used for immersion of NF membranes Parameter pH Conductivity TSS TDS Cu Co 2.2

Unit mS.cm-1 mg.L-1 mg.L-1 mg.L-1 mg.L-1

Value 1.46 26.56 17 23,782 134.80 39.60

Parameter Ni Ca Mg Fe Mn Al

Unit mg.L-1 mg.L-1 mg.L-1 mg.L-1 mg.L-1 mg.L-1

Value 214.90 423.00 2,429 430.00 89.58 340.00

Parameter As Total acidity Free acidity SO42-

Unit mg.L-1 gCaCO3.L-1 gCaCO3.L-1 mg.L-1

Value 34.05 10.28 6.89 20,230

Nanofiltration and reverse osmosis membranes

The flat sheet NF and RO membranes used in this study were supplied by Koch Membranes. Some characteristics of these membranes as provided by the supplier are shown in Table2.

Table2:Characteristics of nanofiltration and reverse osmosis membranes

Membrane Type Membrane Chemistry Molecular Weight Cut-Off (Da) Operating pH Maximum Operating Pressure (bar) Maximum Operating Temperature (°C) Glucose Rejection (%) NaCl Rejection (%)

Nanofiltration

Reverse Osmosis

MPF – 34

TFC – HR

Proprietary composite

Proprietary polyamide

200

2

0 – 14

4 – 11

35

41

70

45

95 1

2

35 1

99.553

thin

film

composite

1

The test conditions specified by Koch were: RO water at 440 psi (30 bar), 86°F (30°C). Feed solution for rejection tests with 3% glucose or 5% NaCl. 2 Not specified 3 The test conditions specified by Koch were: 2,000 mg/L NaCl solution at 225 psi (15.50 bar), 15% recovery, 77°F (25°C), and pH 7.5

2.3

Experimental setup

Microfiltration was used as effluent pre-treatment. The experimental microfiltration (MF) apparatus comprised two tanks (membrane and permeate storage tank), a diaphragm pump, rotameters for the permeate, a pressure indicator, and a submerged hollow-fiber module supplied by Pam Membranas Seletivas Ltda. (filtration area: 0.04 m2; average pore size: 0.4μm; polymer-base: polyetherimide). Microfiltration was performed at a constant pressure of 0.7 bar and up to a recovery rate of 80%. The NF and RO experiments were performed with a bench scale unit (Figure1). The experimental apparatus comprised: a feed tank; a centrifugal pump connected to a speed controller; a rotameter; a valve for pressure adjustment; a stainless steel membrane module; a manometer, and a thermometer.

P T Retentate

Permeate Valve

FI

Membrane Module

Feed Tank

Rotameter

Centrifugal Pump

Figure1: Schematic diagram of the experimental set-up. The stainless steel membrane module had a diameter (2R) of 9 cm, providing a filtration area of 63.60cm2. The radial inlet radius of the cell (r1) was 64mm and the internal channel height (2h) was 1 mm. The membranes tested were properly cut before being placed in the cell and a feed spacer was placed over the membrane to promote flow distribution. Tests were performed in batch concentration mode with continuous collection of the permeate and recycling the retentate to the feed tank. A fixed feed flow rate of 2.4 LPM and a 15% recovery rate were adopted, except when another value is specified. During the tests, the permeate flow rate was monitored periodically. Flux data were normalized to 20°C using a correction factor calculated from the ratio of the water viscosity at the permeation temperature and the water viscosity at 20°C as shown in Equation(1) .

(1) where J(T) is the measured permeate flux at a given operation temperature, water dynamic viscosity at a given operation temperature,

is the

is the water dynamic

viscosity at 20°C and J(20°C) is the normalized flux at 20°C. The retentate concentration of component i (Cri) was calculated from the mass balance as shown in Equation(2):

(2) where Cpi and Cfi respectively correspond to the concentrations of component i in the permeate and feed solutions and

corresponds to the effluent recovery rate defined as

the percentage of feed that is convert in permeate. Due to the concentration polarization phenomenon, the concentration of a particular solute i at the membrane surface is higher than that found in the bulk solution. Thus, the rejection data were interpreted in terms of real rejection, as shown in Equation(3).

( 3) Here, Rare and Chi are respectively the real rejection and concentration of solute i at the membrane surface. The concentration of solute i at the membrane surface can be predicted by using the film model, given by Equation(4)[23].

(4) Here, Cm,i, Cp,i, and Cb,i are respectively the concentrations of solute i at the membrane surface, in the permeate, and in the bulk; J is the permeate flux; and ki is the mass transfer coefficient of solute i in the test cell. The mass transfer coefficient for solute i in a radial tangential flow filtration cell in laminar regime (Re<1800) can be obtained from the Sherwood number (Sh), according to Equation(5)[24].

(5)

Where h is half of the cell internal height; Di is the diffusion coefficient of solute i; Re is the Reynolds number; Sc is the Schmidt number; R is the filtration cell radius. The Schmidt number is given by Equation(6). (6)

Here, υ is the kinematic viscosity of water at 20 C. The Reynolds number is given by Equation(7). (7)

Here, u0 is the average tangential velocity and dh is the cell hydraulic diameter. The average tangential velocity was determined by means of Equation(8)[25]. (8)

Here, Q is the feed flow rate, h is half of the cell internal height, and rlnrd is the logarithmic mean radius calculated from Equation(9)[25].

(9)

Here, R is the filtration cell radius and r1 is the radial inlet radius of the cell. The hydraulic diameter, dh, was estimated by means of Equation(10)[25]. (10) The Reynolds number calculated for the operational conditions used in this study was 839. This value is lower than 1800 and thus characterizes laminar flow. 2.4

Analytical Methods

Metal concentrations were determined by atomic absorption spectrometry (Atomic Absorption Spectrophotometer - GBC - AVANTA). The sulfuric acid concentration was determined by neutralization titration using a NaOH standard solution. The pH of the samples was measured using a Qualxtron QX 1500 pH meter. The sulfate concentrations were determined by ion chromatography (Dionex ICS-1000 ion chromatograph equipped with AS-22 and ICS 12a columns). The glucose concentrations were determined by analyzing the total organic carbon (Shimadzu TOCV CNP TOC analyzer). Calibration curves were constructed for measurement of the

NaCl and MgSO4 concentrations using ion conductivity. A Hanna HI 983 conductivity meter was used to determine the conductivities of the samples. 2.5

Cleaning procedure and water permeability determination

Prior to the filtration tests, the NF and RO membranes were subjected to a cleaning procedure. This procedure consisted of ultrasonication with citric acid solution at pH 2.5 followed by ultrasonication with 0.1% Noah solution, for 20 min each. 2.6

NF and RO membrane immersion procedure

The effects of continuous exposure to effluent on the characteristics of the NF and RO membranes were evaluated by immersing two pristine membrane fragments in each of the solutions presented in Table 3for a period of eight weeks. At two week intervals, fragments were removed from the solution and rinsed with deionized water. Subsequently, filtration tests were carried out for one of the fragments, as will be described in Section2.7. For the other fragment, a fraction was cut out for analysis of the morphological and chemical characteristics, as described in Section 2.8. After collecting a membrane fragment and proceeding with filtration tests, the immersion solutions were replaced and the fragments were re-immersed. This procedure was repeated until the eight week period was completed.

Table 3: Description of the solutions used in the evaluation of the effects of continuous exposure to effluent on NF and RO membrane characteristics Solution Membrane NFA MPF-34

Description Effluent from pressure oxidation process microfiltrated 0.15 mol.L-1 sulfuric acid solutiona

NFB

MPF-34

ROA

TFC-HR

0.08 mol.L-1 sulfuric acid solution

ROB

TFC-HR

0.35 mol.L-1 sulfuric acid solutiona

Details Physicochemical characterization of the microfiltrated pressure oxidation effluent used in immersion of NF membranes was presented in Table 1. The NFB solution concentration was calculated to simulate six months of exposure to real effluent from the pressure oxidation process during the eight weeks of the test. The ROA solution concentration was calculated to simulate exposure to the RO retentate produced from treatment of POX effluent proposed in our previous work [18]. The ROB solution concentration was calculated to simulate six months of exposure to RO retentive during the eight weeks of the test.

a

Concentrations were calculated considering the total exposure time . The membrane immersed in NFA would be exposed to 0.046 mol.L-1 H2SO4 for one day . This concentration was estimated from acidity of POX effluent. So, at the end of 6 months, the total exposure would be 8.34 mol/L. In order to reach the same total exposure at 8 weeks (56 days) NFB concentration was calculated as 0.15 mol/L . Calculation was performed in a similar way to ROB solution.

2.7

Filtration tests

Prior to the filtration tests, the membranes were stabilized with deionized water at a pressure of 10 bar. The temperature was maintained at 20 ± 2°C and a feed volume of 2000 mL was used in all experiments. The following tests were performed for the NF and RO membranes prior to the immersion procedure and were repeated every two weeks of exposure unless another frequency of analysis is indicated. For the NF membranes, the following tests were performed: •Hydraulic permeability determination •Evaluation of magnesium sulfate rejection •Evaluation of metal and sulfuric acid rejection •Evaluation of glucose rejection (performed before and at the end of eight weeks of immersion)

For RO membranes, the following tests were performed: • Hydraulic permeability determination • Evaluation of sulfuric acid rejection • Evaluation of sodium chloride rejection The methods for each test are described below. 2.7.1 Hydraulic permeability determination The hydraulic permeability of the membranes was determined by monitoring the permeate flow stabilized at pressures of 10, 8, 6, and 4 bar using deionized water as feed. The flow values obtained were corrected to 20°C by means of Equation(1). The hydraulic permeability (K) was obtained from the slope of the linear regression of the pressure versus normalized flux data. 2.7.2 Rejection measurements for NF For the NF membranes, the effect of continuous exposure to acid solutions (NFA and NFB) on selective separation of sulfuric acid from metal salts was assessed by filtering a synthetic solution containing cobalt and nickel sulfates and sulfuric acid. The synthetic solution consisted of 97 mg.L-1 of nickel and 17 mg.L-1 of cobalt. The solution pH was adjusted to 1.3by addition of sulfuric acid. Solutions were prepared with concentrations based on the median values found in pressure oxidation effluent. The effect of membrane exposure on the magnesium sulfate and glucose rejection was also evaluated. Magnesium sulfate retention tests were performed by using a 2,000 ppm solution as the feed. Glucose retention tests were performed by using a 500 ppm solution as the feed. Glucose rejection was determined as a function of permeate flux for pressures of 4 to 12 bar for pristine membranes and after 8 weeks of exposure. 2.7.3 Rejection measurements for RO The effect of continuous exposure of the membrane to acid solutions (ROA and ROB) on sulfuric acid rejection was evaluated by filtering a synthetic solution of sulfuric acid (pH = 1.3). The synthetic feed concentration was selected so as to simulate the RO feed conditions when treating pressure oxidation process effluent. The effect of membrane

exposure on the sodium chloride rejection was also evaluated by using a 2,000 ppm solution as the feed. 2.8

Evaluation of morphological and chemical characteristics of membranes

2.8.1 Scanning Electron Microscopy (SEM) The effect of continuous exposure to effluent on the membrane surface morphology was analyzed by using a FEI Quanta 200 scanning electron microscope (SEM). Prior to analysis, membrane samples were coated with a gold layer. 2.8.2 Atomic force microscopy (AFM) The membrane samples were also analyzed by atomic force microscopy to observe additional changes in the surface morphology. Tapping mode AFM was performed with an Asylum MFP-3D-SA/AFM microscope (Asylum Research) over an area of 5 × 5 μm2. A silicon probe (AC240TS-R3, Olympus, Tokyo, Japan) was used. After scanning, the images were flattened with order 2 to remove curvature and slope. After flattening, the root-mean-squared roughness (RRMS) was determined. 2.8.3 ATR - FTIR In order to evaluate changes in the membrane surface composition due to continuous exposure to effluent, attenuated total reflection Fourier transform infrared (ATR-FTIR) spectroscopy analysis was performed. ATR-FTIR experiments were carried out using a Shimadzu FTIR IR Prestige-21instrument equipped with an attenuated total reflectance (ATR) accessory. The spectrum was obtained in the range of 400–4000 cm−1 at 4 cm−1 resolution. 2.8.4 Contact angle measurements In order to verify changes in the membrane hydrophobicity, contact angle measurements were performed by using a DIGIDROP-DI goniometer (GBX Instruments). This system is equipped with a CCD camera and an automated liquid dispenser. The contact angle was determined by placing deionized water droplets (10 μL) on the surface of the samples using a syringe. All experiments were carried out at room temperature. An image was acquired and then analyzed to determine the average angle between the droplet and the surface. Three consecutive measurements were performed at room

temperature using the Surface Energy mode of the software, which allows direct measurement of the contact angle (in degrees). 3 3.1

RESULTS AND DISCUSSION Effect of exposure to effluent on the performance and characteristics of MPF-34 membrane

As show in Table4, the membrane permeability increased with increasing exposure time, and this effect was more pronounced for the membrane exposed to acidic solution (NFB solution). Furthermore, there was a reduction in the magnesium sulfate rejection coefficient, and the salt selectivity was considerably reduced for the membrane exposed to the acid solution. Table4: Results of water permeability (Lp) and rejections of magnesium sulfate (RMg), nickel sulfate (RNi), cobalt sulfate (RCo), and sulfuric acid (Racid), determined for MPF34 membrane as a function of the duration of exposure to NFA and NFB solutions a

Time Solution (Week)

NFA

NFB

0 2 4 6 8 0 2 4 6 8

Lp (L.h-1.bar-1.m-2) 2.9 4.3 5.6 7.1 8.5 3.1 5.2 9.2 11.1 13.3

Filtration Experiments RMgb RNic (%) (%) 98.6 98.0 92.8 90.6 89.7 80.9 66.1 75.2 66.2 65.2 97.1 98.1 94.0 80.2 69.8 70.7 35.0 22.8 18.7 33.0

RCoc (%) 98.6 92.8 89.7 66.1 66.2 97.1 94.0 84.3 35.0 18.7

Racidd (%) 7.7 10.3 14.5 2.6 13.8 11.8 -0.9 0.2 3.9 37.5

a

The indexes 0, 2, 4, 6, and 8 indicate the number of weeks of exposure. The MgSO4 real rejection was determined assuming a salt diffusion coefficient of 0.85 × 10 -9m2.s-1[26]. c The real rejection of nickel and cobalt were determined by considering the respective metal diffusion coefficients: 0.61 × 10-9 m2.s-1and 0.72 × 10-9 m2.s-1[26]. d The real rejection of sulfuric acid was determined by considering an acid diffusion coefficient corresponding to 2.60 × 10-9 m2.s-1[26]. b

Comparison of the results obtained for the membrane exposed to the effluent (NFA solution) and the synthetic acid solution (NFB solution) showed that at the end of 8 weeks (56 days) of exposure to the effluent (NFA solution), the magnesium sulfate rejection coefficient was reduced by 33%. For the membrane exposed to the synthetic acid solution (NFB solution), where the expected degradation at the end of four weeks

would be equivalent to 3 months of exposure to effluent (NFA solution), the rejection coefficient decreased by only 28%. Thus, the simulated solution failed to reproduce the effect of real effluent on the degradation of the membrane, which highlights the importance of the effect of the effluent matrix in degradation. It was also observed that the membrane showed low rejection coefficients for sulfuric acid (Table4), indicating that this species freely permeated the nanofiltration membrane during the exposure period. Regarding the metal selectivity for the membrane exposed to the effluent (NFA solution), there was a reduction of 33% in the values of the rejection coefficient. For the membrane exposed to acidic solution (NFB solution), a maximum reduction of 80% was observed. Thus, the performance of the membrane for separation of the cations of the acid solution was reduced after eight weeks of exposure, with the most pronounced effect for the membrane exposed to the acidic solution. The increase in hydraulic permeability and the observed reduction of selectivity may be related to degradation of the selective layer of the nanofiltration membrane. If the presence of pores is considered, the increased permeability and reduced selectivity may be related to an increase of the pore size. In order to understand the observed increase in the hydraulic permeability and the reduced selectivity, the rejection of a neutral solute (glucose in this case) was evaluated. The glucose rejection (Table 5) was examined by using the solution-diffusionimperfection model and Spigler-Kendem model associated with the Steric Hindrance Pore (SHP) model.

Table 5: Glucose rejection (R) and permeate flux (JV) as a function of applied pressure (P)a Solution P (bar) 4 6 8 10 12

NFA b

Rr (%) 0 8 84.76 33.74 87.79 36.83 88.87 42.01 90.70 46.12 91.41 48.67

NFB -1

-2

Jv (L.h m ) 0 8 16.98 35.72 25.46 54.70 35.96 76.39 44.42 97.08 53.56 114.61

b

Rr (%) 0 8 74.79 9.85 81.17 9.32 83.15 15.25 86.37 16.78 88.08 27.18

Jv (L.h-1m-2) 0 8 15.72 46.00 23.37 71.42 31.68 94.05 39.55 116.64 46.03 137.93

a

The indexes 0 and 8 indicate the number of weeks of exposure. The real rejection of glucose was determined by considering a diffusion coefficient corresponding to 0.67×10-9m2.s-1[27]. b

As shown in Table 5, there was a reduction of the glucose rejection and an increase in the permeate flux at the end of 8 weeks for both of the membranes (immersed in the effluent and in the acidic solution). As observed for the hydraulic permeability and magnesium sulfate rejection analyses, the most pronounced effect was observed for the membrane immersed in the NFB solution. The effect of exposure on the parameters of the solution-diffusion imperfection model was evaluated according to the approach proposed by Gonzalez et al. (2006). In this model, it is recognized that there may be small defects (pores) on the membrane surface, whereby transport can occur. Thus, the solution-diffusion-imperfection model consists of a modification of the solution-diffusion model that aims to include the contribution of the convective flux through the pores of the membrane to the total flux. The rejection of solute i in this model is given by Equation(11)[14]. (11) Where R is the real rejection of solute i; Δπ is the osmotic pressure differential; Lp is the hydraulic permeability; K1 is the solute permeability coefficient; K2 is the coupled coefficient that describes the flux through the pores of the membrane. For the glucose solutions evaluated, the osmotic pressure differential was considered to be close to the osmotic pressure of the feed, the value of which was determined using the Van't Hoff Equation and corresponded to 7 × 10-2 bar. Once the applied pressure

(

) was at least 4 bar,

thus, the

approximation is valid.

Equation(11) can therefore be simplified, resulting in Equation(12). (12) Thus, considering (1/R) - 1 as a dependent variable and 1/(

as the independent

variable, a linear equation of the type y = ax + b is obtained from Equation(12). Thus, by performing a linear regression of the (1 / R) -1 data versus 1 / (

), the

coefficients K1 and K2 may be determined (Table6). Table6: Parameters for solution-diffusion imperfection model equation for MPF-34 membrane before (week 0) and after 8 weeks of immersion Solution

Week

K1(10-6 m.s-1)

K2(10-8 m.s-1.bar-1)

NFA

0

0.40

4.35

NFA

8

13.03

153.94

NFB

0

1.01

3.57

NFB

8

104.06

771.35

As shown in Table6, an increase in the contribution of the flux through the membrane pores as a function of exposure time was evidenced by the increase in the observed value of the parameter K2. This result suggested that exposure of the membranes to the effluent may result in an increase of the pores sizes, which implies an increase in both the solvent and solute fluxes, resulting in a reduction of the rejection and an increase of the hydraulic permeability. In order to assess whether exposure of the membranes resulted in an increase in their effective pores, the approach proposed Wang et al. (1995). In this approach, the pore radius was estimated by using experimental data for the rejection of a neutral solution (glucose) as a function of the permeate flux. The results were analyzed using the Spiegler-Kedem and SHP (Steric Hindrance Pore) models. The Spiegler-Kedem model Equation is given by Equation(13)[7]: (13) Where R is the real rejection of solute i; J is the permeate flux; P is the permeability of solute i;

is the reflection coefficient.

As shown in Equation(13), the rejection of a given solute increases with the increase of the permeate flux, reaching a limit that corresponds to the reflection coefficient. Because the contribution of the diffusive flux of the solute can be neglected if the permeate flux becomes infinite, the reflection coefficient is a characteristic of the convective transport of the solute. A reflection coefficient equal to 1 indicates that the convective transport of the solute is completely prevented or that no convective transport occurs at all. The last scenario represents the ideal case of reverse osmosis membranes in which no pores are available for convective flux. A reflection coefficient smaller than 1 is obtained if the solute is sufficiently small to penetrate into the pores of the membrane[7]. For neutral solutes where the convective flux is not influenced by electrostatic effects, separation occurs according to the size exclusion mechanism. Thus, only the ratio of the radius of the solute (rs) to the pore radius (rp) of the membrane determines the reflection coefficient, as indicted in Equation(14) deduced from the SHP model [7]. (14) As can be seen from Equation(14), the SHP model can be used to estimate the pore radius for a given membrane. For this purpose, the reflection coefficient of a given solute must be determined by Equation(13) and then the pore radius must be calculated by using Equation(14). Thus, in this study, the

and P values were determined from the nonlinear fit of

Equation(13) to the real glucose rejection data as a function of the permeate flux. Figure shows the experimental data for real rejection of glucose as a function of the permeate flux as well as the curve obtained from the nonlinear fit (dashed lines). A good correlation was obtained. Once the

and P parameters were determined, the pore radius (rp) was determined from

Equation(14). In order to determine the pore radius, it was considered the radius of the solute (rs) as glucose Stokes radius, which corresponds to 0.365 nm [27]. The pore size was determined for the membranes prior to the immersion process and at the end of 8 weeks (Table 7).

NFA-W0

NFA-W8

100

80

R (%)

60

40

20

0 0

20

40

60

80

-1

100

120

-2

Jv (L.h .m )

(a) NFB-W0

NFB-W8

100

80

R (%)

60

40

20

0 0

20

40

60

80 -1

100

120

140

-2

Jv (L.h .m )

(b) Figure 2: Experimental rejection data for glucose as a function of permeate flux and the curve obtained from the nonlinear fit (solid line) for week 0 (W0) and week 8 (W8). (a) Results for the membrane immersed in the effluent (solution NFA) and (b) results for the membrane immersed in the acidic solution (NFB solution).

Table 7: Reflection coefficient values ( ), glucose permeability (P), and pore radius (rp) as a function of the duration of exposure to NFA and NFB solutions Solution Week

P

rp

(L.h-1.m-2.bar-1) (nm) NFA

0

0.92 2.13

0.42

NFA

8

0.52 30.40

0.61

NFB

0

0.93 4.43

0.41

NFB

8

0.36 167.80

0.74

A reduction in the values of the reflection coefficients was observed after 8 weeks of exposure. This decrease may be related to the increase of the solute convective flux contribution caused by the increase of pores sizes of the membrane, resulting from its degradation. The reduction of the selectivity to magnesium sulfate and nickel and cobalt metals may also be related to the increase of membrane pore size (Table 7). It is known that for charged solutes, rejection occurs by both charge and size exclusion mechanisms [28]. As in the case of neutral solutes, increasing the membrane pore size can cause a reduction of the rejection of charged species. As can be seen from Equation(14), the reflection coefficient of a given species depends on the ratio of its radius (rs) to the membrane pore radius (rp). Thus, when the rs/rp ratio approaches 1, a 100% reflection coefficient is expected, indicating a high solute rejection. In this context, at week 0 (Table8), the hydrated radius/membrane pore radius ratio for magnesium, nickel, and cobalt species was close to 1, which resulted in a high rejection of these solutes. However, due to the increase of the membrane pore size, the ratio was reduced for the membrane immersed in the effluent and in the acid solution after 8 weeks. This fact resulted in a reduction of the reflection coefficients for each species and therefore in a reduction of the rejection coefficients. Notably, the reduced rejection may also be related to changes in the membrane surface charge, but the effect of increasing the size of the pores seemed to be relevant to the reduction.

Table8: Effect of exposure to NFA and NFB solutions on the ration of hydrated metal radius of magnesium, cobalt, and nickel (rs) to membrane pore radius(rp) rs/rp Solution

Week Mg

Ni

Co

NFA

0

1.021 1.009 0.959

NFA

8

0.706 0.698 0.663

NFB

0

1.035 1.023 0.972

NFB

8

0.577 0.570 0.542

Thus, the results obtained by means of solution-diffusion imperfection model as well as those obtained by the Spiegler-Kendem/SHP model indicated that there was an increase in the convective flux due to the increase of the membrane pore size. This increase was caused by membrane degradation due to exposure to the effluent (NFA solution) and the acid solution (NFB solution). Degradation of the selective layer of MPF-34 was demonstrated by analysis of the membrane surface by SEM (Table 3). Prior to immersion in effluent (NFA solution), the membrane surface appeared homogenous, while after 8 weeks, the emergence of dark areas indicated changes in the selective layer and the increase in pore size.

(a)

(b)

Figure 3: SEM micrographs of the surface of the MPF-34 membrane: (a) prior to exposure and (b) after eight weeks of exposure to effluent (NFA solution).

In order to verify the effect of exposure to NFA and NFB acid solutions on the chemical composition of the MPF-34 membrane, the membranes were analyzed by infrared spectroscopy using attenuated total reflectance (ATR-FTIR) (Figure 4).

MPF NFA W0

MPF NFA W4

2,0 1485

1,8

1103

1578 1296

1,6

1319

1,4

Absorbance

1238

MPF NFA W8

1147

1,2 1,0 0,8 0,6 0,4 0,2 0,0 2000

1800

1600

1400

1200

1000

800

600

400

-1

Wavenumber ( cm )

(a) MPF 34 NFB W0

MPF 34 NFB W4

2,0 1485

1,8

1147 1103

1578 1293

1,6

1319

1,4

Absorbance

1238

MPF 34 NFB W8

1,2 1,0 0,8 0,6 0,4 0,2 0,0 2000

1800

1600

1400

1200

1000

800

600

400

-1

Wavenumber ( cm )

(b) Figure 4: Infrared spectra of MPF-34 membrane as a function of exposure time for (a) membrane exposed to the NFA solution and(b) membrane exposed to the NFB solution. In attenuated total reflection infrared spectrometry, the membrane intermediate and active layers are analyzed; thus, the spectra consist of a superposition of the spectra of each of the layers. Bands from the intermediate polymeric layer of the membrane were

identified; this layer possibly consists of polysulfone (PSU) or polyethersulfone (PES) (Table 9). Since the composition of the selective layer of the MPF-34 membrane is patented, identification of the polymer bands of this layer is beyond the scope of this study. Despite the observed degradation of the selective layer of the membrane (Figure 3), it was not possible to detect the appearance or disappearance of bands as a function of the exposure time of the membrane, suggesting no occurrence of chemical changes on the membrane surface. Table 9: Main IR bands of polysulfide(PSU) and polyethersulphone (PES) found in theMPF-34 membrane in the region between2000 and400cm-1 Assignment Wavenumber Vibration ( cm-1) 1583, 1485, C=C Phenyl group 1103 1323, 1294 O=S=O Anti-symmetric stretching PSU/PES 1238 C-O-C Stretching 1148

O=S=O stretching

Intensity

Ref.

Strong Middle Strong

[29, 30]

Symmetric Strong

The effect of exposure on the hydrophilicity of the MPF-34 membrane was evaluated by determining the contact angle before and after four and eight weeks of exposure (Table 10). Prior to exposure, the MPF-34 membrane presented hydrophilic character, evidenced by the water contact angle. The hydrophilic character of the MPF-34 membrane was also verified by Gautam and Menkhaus (2014)[31] who found an angle of 39° for a virgin membrane fragment. The discrepancy between these angles can be attributed to differences in the operating conditions employed in each of the tests, such as the drop volume, temperature, humidity, and time measurements[32]. After four weeks of immersion in the effluent (NFA solution), a small reduction in the contact angle of the membrane was observed, which indicated slight degradation of the membrane within this period. The results of the nickel, cobalt, and magnesium sulfate rejection analyses corroborate this hypothesis. However, after eight weeks of exposure, there was an increase in the contact angle, which can be attributed to degradation of the selective layer and exposure of the support layer having greater hydrophobic character.

An increase in the contact angle was observed for the membrane exposed to acidic solution (NFB solution) after four weeks of immersion, which indicated exposure of the support layer with hydrophobic character. However, comparison of the results obtained in week 4 with that of week 8 showed a reduction in the hydrophobic character of the membrane. For a hydrophobic material, reduction of the roughness leads to a reduction of the contact angle [33]; thus, one explanation for this observation may be that after exposure of the support layer, the layer underwent polishing, which resulted in reduction of the roughness and therefore the contact angle. Table 10: Contact angles determined for MPF-34 membrane prior to immersion (week 0) and after 4 and 8 weeks of immersion Solution NFA NFB

Contact angle (°) 0 4 8 46±2 42±1 56±5 46±2 55±2 44±1

AFM images of membrane MPF-34 are shown in Figure 5. Tapping mode AFM was used to obtain the roughness (RMS) of the MPF-34 prior and after eight weeks to exposure to NFA solution. It was observed an increase in RMS roughness of 3.195±2nm to 59.213nm, indicating degradation of the selective layer and exposing the support layer. This result corroborates with contact angle results presented previously.

(a)

(b) Figure 5: AFM images of MPF-34 prior (a) and after eight weeks (b) to exposure to NFA solution The results obtained for the MPF-34 membrane showed a reduction in the membrane selectivity at the end of 8 weeks of exposure for both the membrane exposed to the effluent and that exposed to the acid solution. For the membrane exposed to the effluent, a maximum reduction of 33% in the nickel and cobalt rejection coefficient was observed, resulting in recovery of acid with a lower degree of purity and a loss of noble metal in the permeate. 3.2

Effect of exposure to effluent on the performance and characteristics of the TFC-HR membrane

As shown in Table 11, the RO membrane permeability increased as a function of the exposure time, but this increase was significantly lower than that observed for the MPF34 membrane. By way of comparison, at the end of eight weeks, the permeability of the MPF-34 membrane exposed to the effluent (NFA solution) increased by 196%, while an increase of only 33% was obtained for the TFC-HR membrane exposed to the simulated concentrate from reverse osmosis (ROA solution).

Table 11: Hydraulic permeability, sodium chloride rejection, and sulfuric acid rejection determined for TFC-HR membrane as a function of duration of exposure to ROA and ROB solutions

Solution

Time Filtration Experiments (Week) Lp RNaClb Racidc

8

0.92 0.95 1.07 1.03 1.22

96.49 96.47 93.89 93.98 94.34

0

0.98

94.56 95.15

2

0.87

95.11 95.81

4

1.15

92.54 96.91

6

1.36

91.18 94.51

8

1.21

91.05 95.75

0 2 ROA

4 6

ROB

92.87 93.47 93.65 92.69 92.46

a

The indexes 0, 2, 4, 6, and 8 indicate the number of weeks of exposure. The real rejection of NaCl was determined by considering a diffusion coefficient of 1.61 × 10 -9m2.s-1 [26] c The real rejection of sulfuric acid was determined by considering an acid diffusion coefficient corresponding to 2.60 × 10-9m2.s-1[26] b

A slight decrease in the rejection to sodium chloride was observed after eight weeks of exposure. For the membrane exposed to the ROA solution, a reduction of 2.31% was observed, whereas for the membrane exposed to the ROB solution, a reduction of 3.71% was observed. The rejection of sulfuric acid remained above 92% throughout the exposure period (Table 11). These results suggest that during the evaluation period, the TFC-HR membrane exhibited good resistance to the acid conditions to which it was subjected, so that an only small reduction in its selectivity was observed, which does not invalidate the proposed use of the membrane.

Figure 6 presents the micrographs of the surface of TFC-HR prior to exposure of the membrane to ROB solution and after 8 weeks of exposure. The surface was not apparently modified after exposure to the solution; this observation corroborates the rejection results obtained in eight weeks of exposure.

(a)

(b)

Figure 6: SEM micrographs of the surface of the TFC-HR membrane: (a) prior to exposure and (b) after eight weeks of exposure to effluent (ROB solution). The effect of exposure to the acidic ROA and ROB solutions on the chemical composition of the TFC-HR membrane was evaluated by means of infrared spectroscopy using attenuated total reflectance (ATR-FTIR). The spectra obtained are shown in Figure 7.

TFC HR OIA W0

2,0 1,8 1,6

1238

1583 1485 1608 1664

1539

TFC HR OIA W4

1323

TFC HR OIA W8

1147 1103

1294

Absorbance

1,4 1,2 1,0 0,8 0,6 0,4 0,2 0,0 1800

1600

1400

1200

1000

800

-1

600

400

600

400

Wavenumber(cm )

(a) TFC HR OIB W0

2,0 1585

1,8 1,6

1238

1485

1606 1541 1323

TFC HR OIB W4

1148

TFC HR OIB W8

1103

1294

1662

Absorbance

1,4 1,2 1,0 0,8 0,6 0,4 0,2 0,0 1800

1600

1400

1200

1000

800 -1

Wavenumber (cm )

(b) Figure 7: Infrared spectra for TFC-HR membrane as a function of exposure time for (a) membrane exposed to the ROA solution and (b) membrane exposed to the ROB solution. As discussed previously, attenuated total reflection analyzes the intermediate and active membrane layers; thus, spectra consist of a superposition of the spectra of each of the layers. Thus, bands of the active layer (polyamide base polymer) and the support (polysulfone base polymer) were observed in the spectra of the TFC-HR membrane. The major bands detected and the characteristics of the polymers cited are presented Table 12.

Table 12: Main IR bands of polysulfone (PSU) and polyamide (PA) components of TFC-HR membrane in the region between 1800 and 400 cm-1 Assignment Wavenumber ( cm-1) 1664

PA

1608

1539

Vibration

Amide I band: (C=O stretching – dominant contributor, C–N stretching, and C–C–N deformation vibration in a secondary amide group) Aromatic amide (N–H deformation vibration or C=C ring stretching vibration) Amide II band (N–H in-plane bending and N–C stretching vibration of a –CO– NH– group) 1485, C=C Phenyl group

1583, 1103 1323, 1294

Intensity Weak

Weak

[29]

Weak

Strong

1238

O=S=O Anti-symmetric Middle stretching C-O-C Stretching Strong

1148

O=S=O Symmetric stretching

PSU

Ref.

[29, 30]

Strong

No appearance or disappearance of the bands presented in Table 12 was observed during the exposure time for the TFC for HR membrane based on comparison of the spectra shown in Figure 7. Thus, the results suggest that hydrolysis of the polymer polyamide did not occur, or occurred to a small extent, which confirms the stability of the membrane under the conditions used in this study. The effect of exposure to ROA and ROB solutions on the hydrophilicity of the TFC-HR membrane was verified by determining the contact angle of the virgin membrane and after eight weeks of exposure (Table 13). The TFC-HR membrane showed high hydrophilicity prior to the immersion process. The hydrophilic character of the TFC-HR membrane was also reported by Xu et al. (2006)[34], who determined a contact angle of 35° for a virgin membrane fragment. Both membranes exhibited increased hydrophilicity after 4 weeks of immersion, which may have been caused by the creation of more charged groups on the membrane surface due to the reaction of acid functional groups present in the polyamide polymer chain. However, after 8 weeks of exposure, the hydrophilicity declined. For a hydrophilic material, reduction of the roughness leads to an increase in the contact angle [33]; thus, the reduced hydrophilicity may be derived from selective layer polishing due to the continuous contact of the membrane with the acidic solutions, thereby resulting in an increase in the contact angle.

Table 13: Contact angles determined for TFC-HR membrane prior to immersion (week 0) and after 4 and 8 weeks of immersion

Solution ROA ROB

Contact angle (°) 0 4 8 36±4 22±4 51±2 36±4 19±8 52±2

AFM images of membrane TFC-HR are shown in Figure 8. Tapping mode AFM was used to obtain the roughness (RMS) of the TFC-HR prior and after eight weeks to exposure to ROB solution. It was observed a decrease in RMS roughness of 8.317nm to 7.950 nm, indicating the polishing of selective layer due to the continuous contact of the membrane with the acidic solutions. This result corroborates with contact angle results presented previously.

(a)

(b) Figure 8: AFM images of TFC-HR prior (a) and after eight weeks (b) to exposure to NFA solution The results for the TFC-HR membrane showed that this membrane presented good stability to the acidic conditions used in the study. Despite the good results, it is necessary to evaluate the effect of exposure of the membranes to the actual concentrate from the reverse osmosis stage and not just a simulated solution in order to determine whether, in addition to the acid, other concentrated constituents could result in a significant reduction in the membrane selectivity. 4

CONCLUSION

The results obtained indicated a decrease in the selectivity of the MPF-34 nanofiltration membrane at the end of 8 weeks of exposure to the effluent as well as to the acid solution. For the membrane exposed to the effluent, a maximum reduction of 33% in the nickel and cobalt metal rejection coefficient was observed. Thus, considering the treatment proposed in our previously work [18], acid would be obtained with a lower degree of purity and a portion of the noble metal in this permeate would be lost due to long term membrane exposure. Thus, to prolong the membrane performance, more stable membranes must be employed. The TFC-HR membrane exhibited satisfactory stability under the acidic conditions employed in this investigation, thus exhibiting potential for practical application.

5

ACKNOWLEDGEMENTS

The authors acknowledge the financial support and scholarships provided by the coordination of improvement of higher education personnel (capes), foundation for research support of minas gerais (Fapemig), and national council of technological and scientific development (Cnpq). 6

REFERENCES

[1] M. GRANATO, Metalurgia extrativa do ouro. Ministério de Minas e Energia. Centro de Tecnologia Mineral-CETEM. Brasília, Série Tecnologia Mineral, (1986). [2] F.P. Gudyanga, T. Mahlangu, R.J. Roman, J. Mungoshi, K. Mbeve, An acidic pressure oxidation pre-treatment of refractory gold concentrates from the Kwekwe Roasting Plant, Zimbabwe, Minerals Engineering, 12 (1999) 863-875. [3] P.S.M. Soares, L.D.S. Borma, Drenagem ácida e gestão de resíduos sólidos de mineração. Extração de Ouro-Princípios, Tecnologia e Meio Ambiente. Rio de Janeiro, RJ., (2002) 253276. [4] C.W. Corti, R.J. Holliday, Commercial aspects of gold applications: From materials science to chemical science, Gold Bulletin, 37 (2004) 20-26. [5] R.W. Baker, Membrane technology, Wiley Online Library, 2000. [6] G. Hagmeyer, R. Gimbel, Modelling the rejection of nanofiltration membranes using zeta potential measurements, Separation and Purification Technology, 15 (1999) 19-30. [7] J. Schaep, C. Vandecasteele, Evaluating the charge of nanofiltration membranes, Journal of Membrane Science, 188 (2001) 129-136. [8] A.I. Schäfer, A.G. Fane, T.D. Waite, Nanofiltration: principles and applications, Elsevier, 2005. [9] F. Fu, Q. Wang, Removal of heavy metal ions from wastewaters: A review, Journal of Environmental Management, 92 (2011) 407-418. [10] L. Qi, X. Wang, Q. Xu, Coupling of biological methods with membrane filtration using ozone as pre-treatment for water reuse, Desalination, 270 (2011) 264-268. [11] E. Kurt, D.Y. Koseoglu-Imer, N. Dizge, S. Chellam, I. Koyuncu, Pilot-scale evaluation of nanofiltration and reverse osmosis for process reuse of segregated textile dyewash wastewater, Desalination, 302 (2012) 24-32. [12] L. Ahsan, M.S. Jahan, Y. Ni, Recovering/concentrating of hemicellulosic sugars and acetic acid by nanofiltration and reverse osmosis from prehydrolysis liquor of kraft based hardwood dissolving pulp process, Bioresource Technology, 155 (2014) 111-115.

[13] M.P. Gonzalez, R. Navarro, I. Saucedo, M. Avila, J. Revilla, C. Bouchard, Purification of phosphoric acid solutions by reverse osmosis and nanofiltration, Desalination, 147 (2002) 315-320. [14] M. Gonzalez, I. Saucedo, R. Navarro, P. Pradanos, L. Palacio, F. Martinez, A. Martin, A. Hernandez, Effect of phosphoric and hydrofluoric acid on the structure and permeation of a nanofiltration membrane, Journal of Membrane Science, 281 (2006) 177-185. [15] D. Jakobs, G. Baumgarten, Nanofiltration of nitric acidic solutions from Picture Tube production, Desalination, 145 (2002) 65-68. [16] Z. Qinying, Y. Chun-e, F. Shiqi, Recovery device for metals in electroplating wastewater, in, Google Patents, 2013. [17] J. Tanninen, M. Manttari, M. Nystrom, Effect of electrolyte strength on acid separation with NF membranes, Journal of Membrane Science, 294 (2007) 207-212. [18] B.C. Ricci, C.D. Ferreira, A.O. Aguiar, M.C.S. Amaral, Integration of nanofiltration and reverse osmosis for metal separation and sulfuric acid recovery from gold mining effluent, Separation and Purification Technology, 154 (2015) 11-21. [19] A. Manis, K. Soldenhoff, E. Jusuf, F. Lucien, Separation of copper from sulfuric acid by nanofiltration, in: Fifth International Membrane Science & Technology Conference, 2003. [20] S. Platt, M. Nystrom, A. Bottino, G. Capannelli, Stability of NF membranes under extreme acidic conditions, Journal of Membrane Science, 239 (2004) 91-103. [21] J. Tanninen, S. Platt, A. Weis, M. Nystrom, Long-term acid resistance and selectivity of NF membranes in very acidic conditions, Journal of Membrane Science, 240 (2004) 11-18. [22] A.R. Guastalli, J. Labanda, J. Llorens, Separation of phosphoric acid from an industrial rinsing water by means of nanofiltration, Desalination, 243 (2009) 218-228. [23] M. Mulder, Basic principles of membrane technology, Springer Science & Business Media, 1996. [24] S. De, P. Bhattacharya, Prediction of mass-transfer coefficient with suction in the applications of reverse osmosis and ultrafiltration, Journal of Membrane Science, 128 (1997) 119-131. [25] V. Minnikanti, S. DasGupta, S. De, Prediction of mass transfer coefficient with suction for turbulent flow in cross flow ultrafiltration, Journal of Membrane Science, 157 (1999) 227239. [26] D.R. Lide, CRC handbook of chemistry and physics, CRC press, 2004. [27] X.-L. Wang, T. Tsuru, M. Togoh, S.-i. Nakao, S. Kimura, Evaluation of pore structure and electrical properties of nanofiltration membranes, Journal of chemical engineering of Japan, 28 (1995) 186-192. [28] J. Schaep, B. Van der Bruggen, C. Vandecasteele, D. Wilms, Influence of ion size and charge in nanofiltration, Separation and Purification Technology, 14 (1998) 155-162.

[29] C.Y. Tang, Y.-N. Kwon, J.O. Leckie, Effect of membrane chemistry and coating layer on physiochemical properties of thin film composite polyamide RO and NF membranes: I. FTIR and XPS characterization of polyamide and coating layer chemistry, Desalination, 242 (2009) 149-167. [30] H. Diallo, M. Rabiller-Baudry, K. Khaless, B. Chaufer, On the electrostatic interactions in the transfer mechanisms of iron during nanofiltration in high concentrated phosphoric acid, Journal of Membrane Science, 427 (2013) 37-47. [31] A. Gautam, T.J. Menkhaus, Performance evaluation and fouling analysis for reverse osmosis and nanofiltration membranes during processing of lignocellulosic biomass hydrolysate, Journal of Membrane Science, 451 (2014) 252-265. [32] Y. Baek, J. Kang, P. Theato, J. Yoon, Measuring hydrophilicity of RO membranes by contact angles via sessile drop and captive bubble method: A comparative study, Desalination, 303 (2012) 23-28. [33] D.J. Shaw, B. Costello, Introduction to colloid and surface chemistry, ButterworthHeinemann, Oxford, 1993. [34] P. Xu, J.E. Drewes, T.-U. Kim, C. Bellona, G. Amy, Effect of membrane fouling on transport of organic contaminants in NF/RO membrane applications, Journal of Membrane Science, 279 (2006) 165-175.

HIGHLIGHTS 

The chemical stability of nanofiltration and reverse osmosis membranes employed in treatment of pressure oxidation effluent was investigated in detail.



A decrease in the selectivity of the MPF-34 nanofiltration membrane at the end of 8 weeks of exposure to the effluent was verified.



The TFC-HR membrane exhibited satisfactory stability under the acidic conditions employed in this investigation, thus exhibiting potential for practical application.