Adsorptive removal of arsenic from groundwater using chemically treated iron ore slime incorporated mixed matrix hollow fiber membrane

Accepted Manuscript Adsorptive removal of arsenic from groundwater using chemically treated iron ore slime incorporated mixed matrix hollow fiber membrane Somak Chatterjee, Sirshendu De PII: DOI: Reference:

S1383-5866(16)31640-9 http://dx.doi.org/10.1016/j.seppur.2017.02.019 SEPPUR 13548

To appear in:

Separation and Purification Technology

Received Date: Revised Date: Accepted Date:

7 September 2016 31 January 2017 6 February 2017

Please cite this article as: S. Chatterjee, S. De, Adsorptive removal of arsenic from groundwater using chemically treated iron ore slime incorporated mixed matrix hollow fiber membrane, Separation and Purification Technology (2017), doi: http://dx.doi.org/10.1016/j.seppur.2017.02.019

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.

Adsorptive removal of arsenic from groundwater using chemically treated iron ore slime incorporated mixed matrix hollow fiber membrane

Somak Chatterjee and Sirshendu De *

Department of Chemical Engineering, Indian Institute of Technology, Kharagpur Kharagpur – 721302, India.

* Corresponding author: Tel: +913222283926 Fax: +913222255303 Email – [email protected]

1

Abstract Iron ore slime was chemically treated and impregnated in polysulfone hollow fibers for treatment of arsenic contaminated water. Improvement of arsenic removal capacity of the treated slime was observed due to deposition of metallic hydroxide/oxyhydroxide on its surface. Successful incorporation of iron ore slime within the membrane matrix was confirmed by X-ray diffractograms. Scanning electron micrographs confirmed the blockage of pores within polysulfone membranes by the additive, resulting in decrease of porosity, permeability and molecular weight cut off. However, improved hydrophilicity, higher surface roughness and arsenic uptake capacity of prepared membranes were observed. Mechanism of arsenic removal by the mixed matrix membrane was governed by electrostatic attraction mediated adsorption. Membrane with the highest additive percentage (10 wt%) produced arsenic free water for 14 hours for real life feed solution at 11.5 L/m2 h, when operated in dynamic mode. Exhausted membrane was regenerated for three cycles using synthetic solution. Breakthrough time for arsenic removal was reduced from 28 hours to 22 hours and 14 hours after second and third cycle, respectively. Maximum interference effect on removal process is shown by dissolved sulfate ions. The membrane was also able to remove microorganisms and iron simultaneously from real life feed solution below their WHO approved permissible level. Keywords: mixed matrix membrane; polysulfone; iron ore slime; arsenic; adsorption

2

1. Introduction Arsenic is one of the pervasive contaminant and ranks 20th in terms of natural abundance. Concentration of arsenic in natural rocks varies from 0.5-20 mg/g [1]. Arsenic can be mobilized under both acidic and basic conditions amongst other heavy metals. Natural factors, such as geochemical weathering of soil and biological actions facilitate mobilization of arsenic within groundwater [2]. Anthropogenic activities like hydraulic fracturing, mining of fossil fuels, use of arsenic contained herbicides, pesticides and livestock result in leaching to the underground water table. Generally, arsenic exists in two major forms in groundwater: arsenite and arsenate, typically in the ratio of 70:30. Mobility and toxicity of arsenite is more profound than arsenate. Contamination and successive intake of different forms of arsenic in groundwater lead to severe health implications, like skin discoloration, lesions and cancers [2]. Due to its lethal action in the human body, world health organization (WHO) has set the maximum concentration limit of arsenic in drinking water to 10 µg/L [3]. Arsenic has been reported in groundwater of countries like, Argentina, Bangladesh, China, Chile, India and Mexico [4]. Alluvial plains of India, especially along the Ganges and Brahmaputra rivers have severe problems of arsenic contamination in groundwater [4]. Several remediation techniques have been used to remove arsenic from groundwater. For example, coagulation by alum [5], adsorption by iron oxides and activated alumina [6, 7], biological accumulation within fern species, like, Pteris vittata [8]. However, each of these processes has some inherent limitations. Alum coagulation requires pre-oxidation of the primary stream. Adsorption by iron oxides and activated alumina is highly pH sensitive and the adsorbents require frequent regeneration [9]. Biological removal techniques still need to be standardized and show poor efficiency [10]. Nanofiltration membranes are attractive alternative, but have high operating cost and dominant fouling problems [11]. However, mixed matrix membranes are efficient replacement in this context. Fabrication of these membranes requires use of an inorganic additive in the polymeric matrix. These inorganic additives enhance the selectivity of the membrane towards a targeted species, reducing fouling and improving hydrophilicity of the membrane [12]. However, expensive and complex additives increase the cost of these membranes [13, 14]. Therefore, inorganics must be selected based on their performance, size, complexity in manufacturing and most importantly cost. It is interesting to note that wastes from steel producing industries, like, slag and slime have been used previously to treat heavy metals, due to their high iron and alumina 3

content [15]. The uptake capacity of these metals can further be increased by appropriate treatment techniques [16]. Selection of base polymer also plays a significant role in the performance of a membrane. Polysulfone (PSF) belongs to the class of thermoplastic polymers having adequate toughness and stability at high temperature [17]. Polyvinylpyrrolidone (PVP) is a pore former and also increases the mechanical stability of the membrane [17]. The present study highlights the scope of impregnating chemically treated iron ore slime (IOS) in PSF-PVP matrix to produce hollow fiber membranes for arsenic filtration from water. Different properties of prepared hollow fibers were assessed. Adsorption capacities of inorganic additives and mixed matrix membrane (MMM) were evaluated. Dynamic filtration capability of the best membrane was evaluated to estimate the breakthrough point. Regeneration efficiency of the membrane for three cycles was also studied. Performance of MMM in presence of other co-existing anions and arsenic contaminated groundwater was also investigated.

Experimental procedure 2.1. Materials Polysulfone (average molecular weight: 22.4 kDa) was purchased from M/s, Solvay Chemicals, Mumbai, India. N, N-dimethyl formamide (DMF: density 944 kg/m3) and polyethylene glycol (PEG) (molecular weight: 100 kDa, 200 kDa, 35 kDa, 20 kDa, 10 kDa, 6 kDa, 4 kDa, 400 Da) were supplied by M/s, Merck (India) Ltd., Mumbai, India. Polyvinyl pyrrolidone (average molecular weight: 44 kDa) and dextran (molecular weight: 70 kDa) were obtained from M/s, Sigma Aldrich Chemicals, USA. For preparation of synthetic feed solution, sodium arsenate heptahydrate (Na2HAsO4.7 H2O) and sodium arsenite (NaAsO2) were used, supplied by M/s, Loba Chemie Pvt. Ltd., Mumbai, India. Sodium hydroxide, used for regenerating the membrane, was procured from local market. 2.2. Preparation of treated IOS IOS particles (obtained from Tata Steel plant, Jamshedpur, 22.47° N, 86.12° E) were chemically treated to increase its arsenic uptake capacity. Schematic of treatment technique is shown Fig. S1 (see supplementary section). Briefly, 50 g of IOS was stirred with 200 mL of hydrochloric acid (6 N), between a temperature of 50-60ºC for three hours. The solid-acid 4

quantity was kept at a ratio of 1:4 [18]. Next, the residual solid was hydrolyzed using sodium hydroxide (4 N), till pH 7. Finally, the hydrolyzed mass was washed repeatedly to remove any trace of sodium chloride [18]. Treated IOS was grounded to form fine particles and their size distribution was measured using Mastersizer (model: 2000) supplied by M/s, Malvern instruments, UK, at room temperature. 2.3. Spinning of hollow fiber MMM Nomenclatures of different membranes along with the viscosities of casting solution (measured by rheometer, model: Physica MCR 301, Anton Paar, USA) is given in Table 1. At first, pre-determined quantity of PVP was dissolved in DMF followed by PSF and treated IOS. The solution was stirred by a mechanical stirrer for two hours at 333 K. The solution was then homogenized at room temperature (303 ± 2 K) for six hours in an ultrasonic cleaner (Piezo-U-Sonic, India) [12]. Dry wet method was employed to spin the hollow fibers [19]. 2.4. Surface characterization of the MMM Surface characterization of the prepared membranes was carried out in terms of scanning electron microscope (model: ESM – 5800, JEOL, Japan and ZEISS EVO 60 Scanning electron microscope with oxford EDS detector), atomic force microscope (model: 5500 AFM, Agilent Technologies, USA) and transmission electron microscope (JEOL 2010 HRTEM) equipped with a Gatan multi-scan CCD camera). Zeta potential of the membranes was measured using zetacad (Zetacad-DC, manufacturer: - CAD instruments, Les Essarts-le-Roi, France). FTIR (Fourier Transform Infrared Spectroscopy, supplied by M/s, Perkin Elmer, CT; model: Spectrum 100) of the hollow fiber MMM was conducted before and after experiment with As(V) solution. X-ray diffraction (XRD) patterns of treated IOS, hollow fiber MMM and pure polymeric hollow fiber membrane was carried out by X-ray Diffractometer (M/s, PANalytical, model: Xpert Pro, The Netherlands). The pore volume distribution and surface area of the prepared membranes were determined by Quantachrome instruments, Florida, USA (model AUTOSORB-1), using nitrogen as the adsorption medium (degassing temperature: 343K, time: 24 h). Contact angle of different membranes (using sessile drop method at room temperature) was measured by a goniometer (M/s, Rame-Hart instrument co., USA, model number: 200-F4 series).

5

2.5. Inherent membrane property characterization Inherent membrane properties, like, permeability, porosity, MWCO and As(V) rejection capacity of the prepared MMM were determined using a cross flow hollow fiber setup [19]. Permeability of the membrane was obtained from the slope of the straight line between pure water flux and transmembrane pressure drop (TMP). Bulk porosity of the membrane was determined gravimetrically. MWCO is the molecular weight of the solute that is retained 90% by the membrane. Arsenic rejection capacity of the membrane is determined by analyzing the concentration of solute after filtration. Details of these measurements are already studied [12, 19]. 2.6. Equilibrium studies of additive and MMM Equilibrium study was conducted in an orbital shaker (speed: 150 rpm; time: 24 hours; pH: 7  0.2) to obtain the maximum adsorption capacity of treated, untreated IOS and MMM, both in arsenite and arsenate media [12]. Two types of isotherm equations, i.e., Langmuir and Freundlich were used to fit the adsorption data by non linear regression analysis. Sum of squared errors (SSE) between the fitted and experimental values were obtained to determine the best fitted isotherm [12]. 2.7. Dynamic cross flow ultrafiltration experiments Dynamic filtration of the selected membrane was performed in a cross flow hollow fiber setup in recycle mode (permeate is pumped back to the feed solution, to keep the feed concentration constant) [19]. The effect of operating conditions (cross flow rate and TMP) was analyzed, and optimum conditions were selected in terms of arsenic removal efficiency. Filtration experiments for long duration were carried out to evaluate the stability of the membrane (both in terms of permeate flux and concentration). Experiments were terminated when arsenic concentration in permeate exceeded the allowable limit (10 µg/l) and the membrane was regenerated using 1 (M) sodium hydroxide solution for 24 hours [20]. Regeneration was performed for two cycles. Effects of coexisting anions (phosphate, sulfate, carbonate and bicarbonate) and pH of feed on arsenic removal ability of the membrane were assessed. Total arsenic concentration in all streams was measured using Atomic absorption spectrophotometer (Model: Aanalyst

6

700 coupled with MHS-15, PerkinElmer Instruments, USA). Measurement error is two percent of reported data. 2.8. Filtration experiments with arsenic contaminated groundwater sample Arsenic removal efficiency of the selected membrane was evaluated for groundwater obtained from an affected area in West Bengal, India (Viz., Rajarhat, Kolkata, 22.35°N, 88.2°E). The concentration of different ions (both anion and cation) present in the sample was measured by ion chromatograph (Model: 883 Basic IC Plus, Metrohm, Switzerland). Microorganisms in feed and filtrate was determined by standard colony counting and MPN method [12].

3. Result and discussions 3.1. Properties modification of treated IOS Untreated IOS has 84% ferric oxide and 7.8% alumina (Refer Table S1). However, ferric oxide remains in a possible inert state, thereby the arsenic adsorption capacity of IOS is low (1.4 mg/g). It has been observed previously that chemical activation of materials enhances the arsenic adsorption capacity [18]. Treatment technique used in this study can be divided in two parts. In first step, acid treatment of the raw IOS is performed. Under this treatment step, iron and aluminium oxides are dissolved in the residual solution, in their ionic forms. Chemical reaction in this treatment step can be summarized as,

Fe3O4 (solid )  8H   2Fe3 (aqueous)  Fe2 (aqueous)  4H 2O

(1)

Fe2O3 (solid )  6H   2Fe3 (aqueous)  3H 2O

(2)

Al2O3 (solid )  6H   2 Al 3 (aqueous)  3H 2O

(3)

In second step, hydrolysis of the acid dissolved mass with alkali has been performed to deposit oxyhydroxide or hydroxide of the respective ions onto the surface of IOS. It is important to note here that the coating layer increases the adsorption affinity for arsenic [18]. The chemical reaction highlighting the alkali treatment step can be summarized as,

7

Fe3  3OH   FeOOH (iron oxyhydroxide precipitate)  H 2O

(4)

Fe2  2OH   Fe(OH )2

(5)

4Fe(OH )2  O2  2H 2O  4FeOOH (ironoxyhydroxide precipitate )  4H 2O

(6)

Al 3 3OH   AlOOH (alu min iumoxyhydroxide precipitate)  H 2O

(7)

Finally, the adsorbent is washed several times to ensure that the final pH of treated IOS is 7 to remove any trace of alkali, acid and sodium chloride. Sharp differences in different properties of treated and untreated IOS are clearly visible from Table S2. It can be observed that the surface area of the treated additive increases to 135 m2/g, along with the significant increase in pore volume from 0.05 to 0.22 ml/g. This leads to corresponding increase in arsenic adsorption capacity (16.2 mg/g for As(V) and 7.5 mg/g for As(III)), for treated IOS. Increase in adsorption capacity due to activation leads to generate less volume of waste sludge (spent adsorbent). The volume weighted mean diameter of treated IOS particle is 25 μm. Other notable changes in properties are pHZPC, conductivity and surface pH and these are mentioned in Table S2. 3.2. Surface property of treated IOS and MMM Surface morphology of IOS and mixed matrix hollow fibers is shown in Figs. 1(a) and (b). Top view of IOS along with that of M0 and M3 membranes is presented in Fig. 1(a). It is observed from Figs. 1(a) (i) and (ii), that there is a sharp difference in structure of IOS, after treatment. Comparing these figures, porosity of IOS increases after acid-alkali treatment. Deposition of metallic oxyhydroxide is clearly evident from the scanning electron microscopic (SEM) images. This can be corroborated by the enhancement in arsenic adsorption capacity due to coating layer on the additive [18]. Top view of the M0 and M3 membranes are shown in Fig. 1(a)-(iii) and (iv), respectively. Surface of M0 membranes shows the presence of pores. However, due to addition of treated IOS, top surface of M3 membrane does not show the porous structure at the same magnification. This is due to deposition of treated IOS particles in the polymeric matrix [12-14]. Blockage of pores decreases the permeability and porosity and is discussed in the subsequent sections. TEM images of treated IOS are presented in Fig. 1(a)-(v). Porous structure is evident along the surface of treated IOS. Gradual magnification of the TEM images shows the presence of

8

parallel lattice fringes along the surface of the particle. Interlayer lattice distance between the closest parallel fringes is 0.16 nm corresponding to goethite and kaolinite phases in IOS particles [21]. Cross sectional images of the cast membrane are presented in Fig. 1(b). It is visible that addition of treated IOS affects the membrane morphology significantly. All the membranes fabricated show an asymmetric structure, consisting of thin dense layer followed by finger like pores. These are formed due to rapid phase inversion during membrane formation [22]. On addition of the treated IOS in the membrane matrix, the macrovoids are elongated, as observed from Fig. 1(b)-(ii). Figs 1(b)-(iii) and (iv) show distribution of these macrovoids is more pronounced at higher IOS concentration. This is due to hydrophilicity of the IOS particles inducing water transport into the polymeric matrix. It may be observed that the viscosity of spinning solution (refer Table 1) increases with treated IOS concentration. Generally, viscous polymeric solution results in delayed demixing of solvent and non-solvent forming a denser microstructure [23]. However, in the present case, the hydrophilicity of IOS particles plays a dominant role to attract the non-solvent into polymeric matrix during phase inversion leading to larger macrovoids. As observed from Fig. 1(a) (iii) and (iv), the top surface of MMM loses its porosity due to impregnation of IOS, the permeability of hollow fibers decreases with IOS concentration and the results are presented in subsequent sections. However, arsenic removal capacity increases with increasing concentration of additives within the membrane. Atomic force microscopic (AFM) images of the prepared membranes are presented in Fig. 2. It can be observed that the surface pattern of the fabricated hollow fiber membranes shows distinct changes due to addition of treated IOS. The surface profile of M0 membrane shows that the average roughness is 40 nm that increases to 48 nm for M1 and 54 nm for M2 membrane. Presence of inorganic additive is clearly visible on the surface of MMM. For membrane M3, the average roughness is the highest (97 nm). Increase in roughness occurs with treated IOS concentration due to reaching of hydrophilic additives to the inner skin surface during phase inversion [24]. Increase in surface roughness improves adsorptive property of the membrane [12-14]. 3.3. XRD analysis XRD patterns of pure polymeric hollow fiber membranes (M0), MMM hollow fiber membrane (M3) and treated IOS are presented in Fig. 3. Treated IOS particles show high 9

intensity diffraction peaks at 2Ɵ=31.10 and 38.70. The first peak represents presence of mixture of phases, like goethite, quartz and kaolinite. The second one corresponds to goethite and hematite phases (PCPDF card no 290173) [21]. Lower intensity diffraction peaks are also present at 2Ɵ= 24.20, indicating mixture of goethite and kaolinite phases [25]. Presence of weakly crystalline ferrihydrite is also evident from peaks at 2Ɵ= 40.70 and 49.50. Silicon dioxide phase corresponds peaks like 2Ɵ = 53.90, 62.30 and 64.10, respectively [25]. Pure polymeric membrane shows a broad peak at 2Ɵ =170, signifying characteristic peak of polysulfone [26]. Also, a weak diffraction peak at 2Ɵ = 200 is observed, indicating the presence of PVP [27]. Existence of all the peaks within the diffractograms of M3 membrane proves the incorporation of treated IOS within the PSF-PVP polymeric matrix. 3.4. Membrane porosity and permeability Variation of porosity and permeability of the hollow fiber MMM with treated IOS concentration is presented in Fig. 4. Porosity of M0 membrane is 75% that decreases to 46% for M3 membrane. This is due to the blockage of pores by treated IOS particles as confirmed by the SEM images (refer Fig. 1). As a result, membrane permeability decreases correspondingly with IOS concentration. It is observed that the permeability for M-0 membrane is 6 x 10 -10 m/Pa.s and it reduces to 2.6 x 10 -10 m/Pa.s for M-3 membrane, due to subsequent narrowing of macrovoids and blockage of pores. 3.5. Pore size distribution, MWCO, contact angle and average pore radius Variation of cumulative pore volume distribution of different cast membranes is shown in Fig. 5(a). It is observed from this figure that the pore volume becomes invariant beyond a diameter of 100 Å. Hence, detection of pores beyond this limit is less probable. Also it can be noticed, that the pore volume gradually decreases with concentration of treated IOS within the cast membranes. For example, the cumulative pore volume is 0.12 cc/g, corresponding to a pore diameter of 100 Å for PSF membrane. This gradually decreases to 0.04 cc/g, corresponding to the same pore diameter, for M-1 membrane. Decrease in surface area (refer Table 1) corroborates the reduction in pore size. For example, surface area of M0 membrane is 136 m2/g and it decreases to 34 m2/g for M-3 membrane. This observation is in line with SEM images that pore blocking of the membrane intensifies with IOS concentration.

10

Variation of MWCO, contact angle and pore radius with treated IOS concentration in the hollow fiber membrane can be observed from Fig. 5(b). MWCO of M-1 membrane is 75 kDa and it reduces to 45 kDa for M-3 membrane. Similarly, the pore radius of fabricated membrane decreases from 85 Å to 63 Å. Presence of any inorganic particles within the membrane matrix restricts the molecular motion of the polymeric chain, thereby reducing its mean free volume [30] and decreasing the pore size and MWCO of the membrane. It can also be observed from contact angle data that the hydrophilicity of the membrane increases with treated IOS concentration in spinning solution. For example, the contact angle of the membrane decreases from 750 (M-0 membrane) to 630 (M-3 membrane). Treated IOS particles being hydrophilic in nature, come up to the top surface of the membrane during phase inversion, thereby increasing the hydrophilicity of the membrane [28]. 3.6. Membrane selection Profiles of As(V) concentration and permeate flux of different membranes, are shown in Fig. 6. It can be observed from Fig. 6(a) that As(V) concentration in permeate at any time decreases with treated IOS concentration in the membrane. For example, As(V) concentration for M0 membrane is 90 µg/L after 30 minutes of filtration. For M1 membrane, it drops to 20 µg/L. These observations suggest that pure polymeric membrane does not have any arsenic removal capacity. The concentration drops further down to 0.2 µg/L for M3 membrane after the same filtration time. Also, it can be observed from Fig. 6(b) that the permeate flux at any time instant decreases with treated IOS concentration in the membrane. For example, the permeate flux for M0 membrane after 30 minutes of filtration is 70 L/m2.h and it decreases to 18 L/m2.h for M3 membrane. This is due to reduction in membrane permeability with treated IOS concentration, as explained earlier. However, in terms of arsenic removal capability, membrane M3 is selected for all subsequent experiments, as it confirms the WHO limit. FTIR spectra of M-3 membrane are shown in Fig. 6(c). Presence of adsorbed water molecules is confirmed by small peaks corresponding to wavelengths at 4006-3704 cm-1. Stretching vibration of hydrogen bond formed between hydroxyl radical on the surface of iron oxide (Fe2O3) and water molecules corresponds to wavelength at 3240 cm-1. Similarly, hydroxyl stretch of goethite gives rise to peak at wavelengths 755 cm-1 [29]. Peak at 472 cm-1 corresponds to stretching vibrations of iron-oxygen bound complex compounds. The peak at 1114 cm-1 is due to vibration of silica group present in IOS. Aryl-sulfone vibration of polymeric chain corresponds to the wavelength at 1398 cm-1 [30]. Aluminium oxide 11

stretching is confirmed by peaks at 660-604 cm-1. Bands at 2041 cm-1, 1644 cm-1 and 1502 cm-1 refer to methoxy stretch, alkenyl stretch and skeletal C-C vibrations [30]. Adsorption of arsenate by the membrane is confirmed by suppression of peaks at each wavelength. 3.7. Equilibrium uptake capacity Batch adsorption experiments were conducted to obtain the uptake capacity of the untreated and treated IOS as well as M-3 membrane in different solutions, i.e., arsenate and arsenite. Corresponding figure is shown in Fig. 7. Isotherm constant values and error values [31] are reported in Table S3. It is observed from Fig. 7 that the uptake capacity of IOS is increased after chemical treatment, in both As(V) and As(III) solution. For example, the adsorption capacity in As(V) solution for untreated and treated IOS is 1.5 mg/g and 16.2 mg/g. Similarly, the uptake capacity in As(III) solution is 1.4 mg/g and 7.5 mg/g for untreated and treated IOS, respectively. Increase in arsenic uptake capacity is due to the deposition of iron and aluminium hydroxide/oxyhydroxide on the surface of IOS, as explained earlier [18]. However, it can be observed that the uptake capacity of treated IOS is less in case of As(III) compared to As(V). This happens due to uncharged nature of As(III) in solution (H3AsO30) in the drinkable pH range, i.e., from 6.5 to 8.5 [4]. Hence, charge mediated attraction of As(III) to the surface of adsorbent is negligible. For example, Langmuir uptake capacity of treated IOS in As(V) and As(III) is 16.2 mg/g and 7.5 mg/g, respectively (refer Table S3). Uptake of arsenic species by prepared MMM is observed on impregnation of treated IOS within the polymeric matrix. For example, the uptake capacity for M-3 membrane is 10.5 mg/g and 4.2 mg/g for As(V) and As(III), respectively. It is observed from the SSE values (given in Table S3) that the Freundlich isotherm explains the adsorption characteristics better than the Langmuir isotherm. 3.8. Removal mechanism Variation of zeta potential of membrane M-3 with pH and effect of pH in removal of arsenic from water is shown in Figs. 8(a) and 8(b), respectively. It is observed from Fig. 8(a) that the zeta potential of the membrane decreases with pH of the solution. Finally, the pH at the zero point charge (pHzpc) of M-3 membrane is 7.5. At this pH, the surface of the membrane is devoid of any charges [32]. The surface becomes negatively charged once the solution pH is above this. Hence, the membrane surface is unable to attract negatively charged arsenate species, resulting in weak removal at higher pH, as observed from Fig. 8(b). For example, at pH 12, the concentration of arsenic in permeate is 50 µg/L. But, the opposite 12

phenomenon is observed when the solution pH is lower than 7.5, when the membrane surface is positively charged [32]. Hence, electrostatic attraction mediated adsorption of arsenate species like, H2AsO4-, HAsO4- and AsO43- (dominant species of arsenate in groundwater) occurs, resulting in higher removal efficiency [4]. For example, the permeate concentration is 1 µg/L at pH of 2.5. The interaction of arsenate with iron and aluminium oxyhydroxide of treated IOS can be represented as,  M  OH  H 2 AsO4 ( As(V )in the pH rangeof 4  8)  MHAsO4  H 2O

(8)

 M  OH 2  H 2 AsO4 ( As(V )in the pH rangeof 4  8)  MHAsO4  H 3O

(9)

where, M stands for the metal (iron or aluminium). The density of the active sites (  M  OH ) increases with acid- base treatment and it increases adsorption of arsenate on the membrane surface [18]. In case of As(III), the dominant species are H3AsO3 and sparsely distributed H2AsO3 - in the pH ranging from 3-8 [4]. Therefore, electrostatic mediated attraction is quite low for As(III) as observed from Fig. 7. This is the main reason for oxidation of As(III) to As(V) to evoke maximum uptake capacity in a removal process. The ionic radius of arsenic species (0.46 Å) is much smaller than the pore radius of the membrane (63 Å). Therefore, contribution of sieving mechanism is insignificant [33]. This can be verified from Fig. 6(a), where it is observed that the removal percentage of As(V) by M0 membrane is only around 15%. Hence, it can be concluded, that adsorption is the dominant removal mechanism, facilitated by electrostatic attraction. 3.9. Cross flow filtration in total recycle mode 3.9.1. Effect of operating conditions Removal efficiency and solute flux of the membrane is affected by variation of operating conditions, i.e., TMP and cross flow rate [28]. The membrane was subjected to different sets of operating conditions for a period of 40 hours at feed arsenic concentration of 100 µg/L. Permeate arsenic concentrations and flux profile of the membrane at various operating conditions is presented in Fig. 9(a). It is observed from this figure, that the arsenic concentration in the filtrate increases with TMP (at a constant cross flow rate). At higher pressure, the contact time between the arsenic and treated IOS of the membrane reduces, 13

thereby decreasing the adsorption efficiency. As a result, concentration of arsenic in permeate is increased. For example, at 28 hours, arsenate concentration at 21 kPa and 34 kPa (10 L/h cross flow rate) is 10 µg/L and 13 µg/L, respectively. High permeate flux is observed at a higher TMP due to increased driving force [12, 28]. For example, at a time period of two hours, the flux for 21 kPa and 34 kPa is 17 L/m2 h and 22 L/m2 h, respectively (cross flow rate of 10 L/h). Similarly, on increasing the cross flow rates of filtration, arsenate concentration of the permeate increases. For example, it can be observed from Fig. 9(a) that the concentration is 13 µg/L at 10 L/h and it increases to 17 µg/L at 20 L/h (TMP being kept fixed at 34 kPa). This happens due to lower adsorption efficiency for reduced contact time at higher cross flow rates [12, 28]. Also, permeate flux is observed to increase from 22 L/m2 h to 27 L/m2 h, due to lower adsorptive resistances at 20 L/h cross flow rate. 3.9.2. Stability of membrane and regeneration studies The stability of the membrane is judged on the basis of permeate flux and arsenate concentration. The breakthrough time of the membrane for a specific run is defined as the filtration period till when the arsenate concentration in the permeate is below the safety limit, i.e., 10 µg/L. Operating conditions are selected as 21 kPa and 10 L/h, respectively (highest removal efficiency is observed at this operating condition) for feed concentration 100 µg/L. Permeate concentration and flux profile are shown in Fig. 9(b). Adsorption efficiency is rapid during the initial filtration period. However, after a certain time, i.e., 28 hours, the adsorption sites start getting saturated and the resultant arsenate concentration in the permeate increases. Also it is observed from this figure that the breakthrough time decreases to 22 hours and 14 hours, after first and second regeneration cycle, respectively. Decreased removal efficiency occurs after each cycle due to incomplete desorption, thereby leading to lower breakthrough time. In this study, adsorptive fouling plays a dominant role during initial filtration experiments (till 30 hours). For example, permeate flux decreases from 17 L/m2 h to 15.3 L/m2 h. However, after this, the adsorptive resistance becomes insignificant, thereby leading to almost steady flux (15 L/m2 h). Flux decrease occurs after each regeneration cycle. For example, the initial flux decreases to 14 L/m2 h and 12 L/m2 h after first and second regeneration cycle, respectively. The concentration of leached out iron from the membrane surface in permeate does not exceed 0.1 mg/L, for the entire operational time. Also, permeate

14

stream was filtered through whatman filter paper (pore size 2.5 µm) to check the leaching of any treated IOS particles (25 µm) and no particle was found to be retained. Permeate concentration and flux profiles for arsenite spiked water are shown in Fig. S2. The breakthrough time for arsenite filtration is 14 hours, which is considerably lower than arsenate spiked solution. This is due to attraction of sparsely distributed H2AsO3- species, in the pH ranging from 3-8, as discussed previously. This can be corroborated by the adsorbed amount of arsenite (4.2 mg/g) when compared to arsenate (10.5 mg/g) (refer Table S3). Permeate flux is 14 L/m2 h, and it decreases to 12.5 L/m2 h after 30 hours of filtration and remains steady thereafter. 3.9.3. Effect of coexisting anions Variation of arsenic removal capacity of M-3 membrane with concentration of different coexisting anions is shown in Fig. 9(c). The anions selected for this study are sulfate, nitrate, carbonate and bicarbonate (arsenate feed concentration of 100 µg/L and 21 kPa TMP and 10 L/h flow rate). Arsenic removal capacity of the membrane in absence of any anion is 95%. However, with increasing sulfate concentration, the removal percentage drops significantly. For example, the removal percentage is below 60% at sulfate concentration of 400 mg/L. This takes place due to competitive adsorption of sulfate on the active sites within the membrane [34]. But this is not exactly the same in case of bicarbonate. For example, the removal percentage remains 90% even at a higher concentration of the ion, i.e., 400 mg/L. According to the figure, the effect of anions on the removal of arsenic by M-3 membrane is of the order sulfate>carbonate>nitrate>bicarbonate. Bulkiness and charge of coexisting ions are the main reasons for this order [34]. For example, the charge of sulfate and carbonate is the same (-2). However, the molar mass of sulfate (96 g/mol) is higher than carbonate (60 g/mol). Hence, sulfate affects arsenic removal percentage to a greater extent than carbonate. Similarly, nitrate affects the removal more than bicarbonate, though both of them carry single negative charge. Also, the order of effect is higher for dual charged anion (sulfate and carbonate) than single charged anion (nitrate and bicarbonate). 3.10. Performance of MMM for treatment of arsenic contaminated groundwater The membrane is also subjected to a dynamic cross flow run in presence of a real-life arsenic contaminated feed. The total duration of the experiment was 40 hours. Feed and permeate samples were collected at regular intervals. Detailed characterization of the feed 15

was performed with respect to other coexisting ions and parameters, like, alkalinity, pH, total dissolved solids, etc (refer Table 2a). The arsenic concentration in the feed was 90 µg/L. Total arsenic concentration and flux profile of real life feed is presented in Fig. 10(a). Permeate flux at the onset of the experiment is 11.5 L/m2 h. For a time period of 25 hours, the flux decrement is rapid, due to fouling. For example, the permeate flux at 25 hours is 10 L/m2 h. After this time, steady state flux of 9.5 L/m2 h is attained due to minimization of adsorptive fouling. It may also be noted that adsorption sites are saturated and concentration rises till it surpasses the permissible limit. Permeate concentration was below 10 µg/L for 14 hours. However, for synthetic feed solution, the breakthrough time is 28 hours. Decrease in the breakthrough time is due to competition between arsenic and other dissolved ions within the feed [34]. Total bacterial colony of the feed water was also measured and variation is shown in Fig. 10(b) [19]. It is observed that the feed has bacterial concentration of 6 x 104 CFU/100 mL, which decreased to 4.6 x 104 CFU/100 mL, after one hour of operation. After 14 hours of operation, bacterial concentration in permeate was 0 CFU/100 mL, which was found to obey the permissible range of drinking water [3]. Even after 30 hours of operation, no bacterial colonies were observed in the permeate. Therefore, prepared MMM was also successful in removal of coliform bacteria from contaminated stream. 3.11. Comparison of performances with other membranes Performance of different mixed matrix membranes towards arsenic removal from water is shown in Table 2(b). There are few works reported on arsenic filtration from water i.e., four in flat sheet membranes and three in hollow fiber membranes. Cost of graphene oxide impregnated polysulfone membranes is high [35]. Amongst others, the highest adsorption capacity is for PVDF-zirconia blend flat sheet membranes (32.3 mg/g for arsenate) [14] and PSF-zirconia nanoparticles blend hollow fiber membranes (95.3 mg/g) [36]. PVDF-zirconia membranes [14] have the highest specific flux amongst others reported (8.4 L/ m2 h kPa). However, the hollow fibers find a variety of usage in practical scenarios compared to the flat sheet membranes due to higher surface area, low operating pressure and simplicity in operation [36]. Also, the power requirement is substantially low, 4.7 x 10-3 (kWh per m2 of membrane area per m3 of water processed) for PSF-treated IOS membrane. Iron-chitosan hollow fiber was operated in vacuum and its performance is not in line with pressure driven membrane systems [37]. Performance index (refer Table 2(b)) of PSF-treated IOS membrane is higher than PVDF-zirconia [14] and Fe-Mn binary oxide impregnated PES membranes [13]. However, it is much lower than PSF-zirconia nanoparticles blend hollow 16

fiber membranes. But, these nanoparticles are synthetic product and filtered stream needs further treatment to ensure no leaching [38]. It is also observed that manufacturing cost of PSF-treated IOS membrane is low as compared to other hollow fiber membranes (refer Table 2(b)).

4. Conclusions IOS was collected and treated using sodium hydroxide and hydrochloric acid to increase its arsenate and arsenite adsorption capacity by 16 times and 7 times, respectively. Treated IOS was doped in polysulfone hollow fiber membrane. Scanning electron micrographs showed that macrovoids were formed due to treated IOS addition and some pores were blocked by them. The permeability was decreased from 6 x 10 -10 m/s Pa to 2.6 x 10-10 m/s Pa with addition of 10 wt. % of treated IOS particles. Corresponding reduction in porosity was from 75% to 46%. However, hydrophilicity and arsenic removal capacity of the membrane were improved. The contact angle was decreased from 750 to 630 and arsenic removal percentage was increased from 15% to 90%. MWCO and pore radius of membrane M3 (treated IOS 10 w/w %) were 45 kDa and 63 Å. M-3 membrane showed best performance with a breakthrough time of 28 hours for arsenic filtration from synthetic feed, in a continuous cross flow mode (permeate flux 17 L/m2 h). The same time was decreased to 14 hours for a real-life feed due to presence of other larger sized pollutant.

Acknowledgements This work is partially supported by a grant from the SRIC, IIT Kharagpur under the scheme no. IIT/SRIC/CHE/SMU/2014-15/40, dated 17-04-2014 and funding from INAE chair professorship of Prof. S. De. The authors are also thankful to Tata Steel, Jamshedpur, for providing raw material. Any opinions, findings and conclusions expressed in this paper are those of the authors.

17

References [1] S. K. Acharyya, S. Lahiri, B. C. Raymahashay and A. Bhowmik, Arsenic toxicity of groundwater in parts of the Bengal basin in India and Bangladesh: the role of quaternary stratigraphy and Holocene sea-level fluctuation, Environ. Geol., 2000, 39, 1127–1137. [2] D. Das, A. Chatterjee, M. Samanta, G.B.K. Chowdhury, T.R. Chowdhury, B.A. Fowler, Arsenic metabolism and toxicity to freshwater and marine species, in: B.A. Fowler (Ed.), Biological and Environmental Effects of Arsenic, vol. 6, Elsevier Science Publisher, Amsterdam, 1983, pp. 155–170. [3] World Health Organization, Guidelines for Drinking Water Quality, 2006. [4] A. Mukherjee, P. Bhattacharya, K. Savage, A. Foster, J. Bundschuh, Distribution of geogenic arsenic in hydrologic systems: controls and challenges, J. Contam. Hydrol. 99 (2008) 1-7. [5] M.M.T. Khan, K. Yamamoto, M.F. Ahmed, A low cost technique of arsenic removal from drinking water by coagulation using ferric chloride salt and alum, Water Sci. Technol. 2 (2002) 281–288. [6] V. Lenoble, O. Bouras, V. Deluchat, B. Serpaud, J.-C. Bollinger, Arsenic adsorption onto pillared clays and iron oxides, J. Colloid Interf. Sci. 255 (2002) 52–58. [7] T.S. Singh, K.K. Pant, Equilibrium, kinetics and thermodynamic studies for adsorption of As(III) on activated alumina, Sep. Purif. Technol. 36 (2004) 139–147. [8] S. Tu, L.Q. Ma, Interactive effects of pH, arsenic and phosphorus on uptake of As and P and growth of the arsenic hyperaccumulator Pterisvittata L. under hydroponic conditions, Environ. Exp. Bot. 50 (2003) 243–251. [9] M.V. Norton, Y.J. Chang, T. Galeziewski, S. Kommineni, Z. Chowdhury, Throw-away iron and aluminium sorbents versus conventional activated alumina for arsenic removal—pilot testing results, in: Proceedings—Annual Conference, Washington, DC, Am. Water Works Assoc. (2001) 2073–2087. [10] Y. Sag, Biosorption of heavy metals by fungal biomass and modeling of fungal biosorption: a review, Sep. Purif. Methods 30 (2001) 1–48.

18

[11] B. Van der Bruggen, M. Manttari, M. Nystrom, Drawbacks of applying nanofiltration and how to avoid them: a review, Sep. Purif. Technol. 63 (2008) 251-263. [12] S. Chatterjee, S. De, Adsorptive removal of fluoride by activated alumina doped cellulose acetate phthalate (CAP) mixed matrix membrane, Sep. Purif. Technol. 125 (2014) 223-238. [13] R. J. Gohari, W. J. Lau, T. Matsuura and A. F. Ismail, Fabrication and characterization of novel PES/Fe–Mn binary oxide UF mixed matrix membrane for adsorptive removal of As(III) from contaminated water solution, Sep. Purif. Technol. 118 (2013) 64–72. [14] Y-M. Zheng, S- W. Zhou, K. G. N. Nanayakkara, T. Matsuura, J. Paul Chen, Adsorptive removal of arsenic from aqueous solution by a PVDF/zirconia blend flat sheet membrane, J. Membr. Sci. 374 (2011) 1–11. [15] C. Oh, S. Rhee, M. Oh, J. Park, Removal characteristics of As(III) and As(V) from acidic aqueous solution by steel making slag, J. Hazard. Mater. 214 (2012) 147-155. [16] V. K. Gupta, Equilibrium uptake, sorption dynamics, process development, and column operations for the removal of copper and nickel from aqueous solution and wastewater using activated slag, a low-cost adsorbent, Ind. Eng. Chem. Res. 37 (1998) 192-202. [17] M. J. Han, S. T. Nam, Thermodynamic and rheological variation in polysulfone solution by PVP and its effect in the preparation of phase inversion membrane, J. Membr. Sci. 202 (2002) 55-61. [18] S. De and A. Maiti, Arsenic Removal from Contaminated Groundwater, TERI Press, New Delhi, India, 2012. [19] B. K. Thakur, S. De, A novel method for spinning hollow fiber membrane and its application for treatment of turbid water, Sep. Purif. Technol. 93 (2012) 67-74. [20] X. Shi, G. Tal, N. P. Hankins, V. Gitis, Fouling and cleaning of ultrafiltration membranes: A review, J. Water. Process. Eng. 1 (2014) 121-138. [21] E. Yusiharni, R. Gilkes, Rehydration of heated gibbsite, kaolinite and goethite: An assessment of properties and environmental significance, Appl. Clay. Sci. 64 (2012) 6174.

19

[22] N. Abdullah, R. J. Gohari, N. Yusof, A. F. Ismail, J. Juhana, W. J. Lau, T. Matsuura, Polysulfone/ hydrous ferric oxide ultrafiltration mixed matrix membrane: Preparation, characterization and its adsorptive removal of lead (II) from aqueous solution, Chem. Eng. J. 289 (2016) 28-37. [23] P. S. T. Machado, A. C. Habert, C. P. Borges, Membrane formation mechanism based on precipitation kinetics and membrane morphology: flat and hollow fiber polysulfone membranes, J. Membr. Sci. 155 (1999) 171-183. [24] M. Khayet, K. C. Khulbe and T. Matsuura, Characterization of membranes for membrane distillation by atomic force microscopy and estimation of their water vapour transfer coefficients in vacuum membrane distillation process, J. Membr. Sci. 238 (2004) 199–211. [25] S. Roy, A. Das, Characterization and processing of low-grade iron ore slime from the jilling area of India, Miner. Process. Extr. Metall. Rev. 29 (2008) 213-231. [26] P. F. Andrade, A. F. de-Faria, S. R. Oliveira, M. A. Z. Arruda, M. C. Goncalves, Improved antibacterial activity of nanofiltration polysulfone membranes modified with silver nanoparticles, Water. Res. 81 (2015) 333-342. [27] A. Rawat, H. K. Mahavar, S. Chauhan, A. Tanwar and P. J. Singh, Optical band gap of polyvinylpyrrolidone/polyacrilamide blend thin films, Indian J. Pure Appl. Phys. 50 (2012) 100–104. [28] S. Chatterjee, S. De, Adsorptive removal of arsenic from groundwater using a novelhigh flux polyacrylonitrile (PAN) - laterite mixed matrix membrane, Environ. Sci.: Water Res. Technol. 1 (2015) 227-243. [29] L. Panda, B. Das, D. S. Rao, Studies on removal of lead ions from aqueous solution using iron ore slimes as adsorbent, Korean. J. Chem. Eng. 28 (2011) 2024-2032. [30] J. Coates, Interpretation of infrared spectra, a practical approach, in Encyclopedia of Analytical Chemistry, ed. R. A. Meyers, John Wiley & Sons Ltd., Chichester, 2000, pp. 10815–10837. [31] K. Foo, B. Hameed, Insights into the modeling of adsorption isotherm systems. Chem. Eng. J. 156 (2010) 2–10. 20

[32] S. S. Tripathy, A. M. Raichur, Abatement of fluoride from water using manganese dioxide coated activated alumina, J. Hazard. Mater. 153 (2008) 1043–1051. [33] D. R. Latulippe, A. M. Mika, R. F. Childs, R. Ghosh, C. D. M. Filipe, Flux performance and macrosolute sieving behaviour of environment responsive formed-in-place ultrafiltration membranes, J. Membr. Sci. 342 (2009) 227–235. [34] X. Meng, G. P. Korfiatis, S. Bang and K. W. Bang, Combined effects of anions on arsenic removal by iron hydroxides, Toxicol. Lett. 133 (2002) 103–111. [35] R. Rezaee, S. Nasseri, A. H. Mahvi, R. Nabizadeh, S. A. Mousavi, A. Rashidi, A. Jafari, S. Nazmara, Fabrication and characterization of a polysulfone-graphene oxide nanocomposite membrane for arsenate rejection from water, J. Environ. Health. Sci. Eng. 13 (2015) 1-11. [36] J. He, T. Matsuura and J. P. Chen, A novel Zr-based nanoparticles embedded PSF blend hollow fiber membrane for treatment of arsenate contaminated water: Material development, adsorption and filtration studies, and characterization, J. Membr. Sci. 452 (2014) 433–445. [37] M. S. S. Dorraji, A. Mirmohseni, F. Tasselli, A. Criscuoli, M. Carraro, S. Gross and A. Figoli, Preparation, characterization and application of iron (III)-loaded chitosan hollow fiber membranes as new bio-based As(V) adsorbent, J. Polym. Res. 21 (2014) 399–412. [38] D. Mohan, C. U. Pitmann Jr., Arsenic removal from water/wastewater using adsorbentsA critical review, J. Hazard. Mater. 142 (2007) 1-53.

21

List of Figures Fig. 1(a): Scanning electron micrographs representing (i) untreated IOS (1000 X magnification); (ii) treated IOS (1000 X magnification); (iii) top surface of M0 membrane (5000 X magnification); (iv) top surface of M3 membrane (5000 X magnification) & (v) TEM image of treated IOS. Fig. 1(b). Scanning electron micrographs representing cross sectional surface morphology of (i) M0 membrane; (ii) M1 membrane; (iii) M2 membrane; (iv) M3 membrane. Fig. 2. AFM images and surface roughness of hollow fiber membranes (a) M0; (b) M1; (c) M2 and (d) M3. Fig. 3. XRD analysis of M0, Treated IOS and M3 membrane. Fig. 4. Variation of porosity and permeability of membranes with increasing treated iron ore slime dosages. Fig. 5. Variation of (a) cumulative pore volume and (b) different characterization properties (MWCO, pore radius and contact angle). Fig. 6. Effect of increasing iron ore slime dosages in membrane on (a) As(V) concentration in the permeate ; (b) permeate flux at an operating condition of 21kPa and 10 l/h and (c) FTIR analysis of membrane M3 (before and after As(V) adsorption). Fig. 7: Adsorption isotherm studies for treated, untreated IOS and M3 membrane at 298 K, pH 7. Fig. 8: Variation of (a) zeta potential and (b) As(V) concentration of the membrane with pH of solution. Fig. 9: (a) Effect of operating conditions; (b) effect of regeneration and (c) effect of coexisting anions Fig. 10: (a) Permeate flux, total arsenic and (b) microbial concentration profile of a real-life feed solution.

List of Tables Table 1. Nomenclature and different physical parameters of the membranes. Table 2(a). Characteristics of permeate and feed of arsenic contaminated real life feed sample. Table 2(b). Comparison of performance of different mixed matrix membranes in arsenic contaminated water filtration. 22

Supplementary Information for Adsorptive removal of arsenic species from groundwater using polysulfonetreated iron ore slime mixed matrix hollow fibre membrane

Somak Chatterjee and Sirshendu De *

Department of Chemical Engineering, Indian Institute of Technology, Kharagpur, Kharagpur – 721302, India.

* Corresponding author: Tel: +913222283926 Fax: +913222255303 Email – [email protected]

23

50 gm of Iron ore slime

200-210 ml 6.0 (N) hydrochloric acid

Acid treatment of Iron ore slime (i) Temperature 5060 0C (ii) stirrer speed: 300-400 rpm (iii) duration of treatment 2.5 - 3h

Distillation of acid treated iron ore slime

Recovery of ~70-75 % hydrochloric acid

Hydrolysis of residual, using 4.0 (N) sodium hydroxide

Final pH 6.5-7.0; room temperature

Finally washed by water until pH of washed water reaches ~ 7.0 and the residual mass has no sodium chloride. Air dried, grounded and stored.

Used as additive in membranes

Fig. S1. Schematic of treatment procedure of iron ore slime.

24

16

70 Feed : 100 g/L

60

Membrane M3 21 kPa 10 l/h

2 Permeate flux (L/m h)

14

50

40

12 30

20

10 WHO Limit: 10 g/L

10

8

0

0

5

10

15

20 25 Time (Hours)

30

35

40

As (III) concentration in the permeate (g/L)

(c )

Operating conditions

45

Fig. S2. Variation of permeate flux and concentration at 21 kPa and 10 l/h for As(III) of M3 membrane.

25

Table S1. Composition of IOS. Compound

Weight% (± 2%)

Fe2O3

84

Al2O3

7.8

SiO2

6.7

Others*

1.5

(* Others: Na2O: 0.2; MgO: 0.21; P2O5: 0.27; SO3: 0.13; CaO: 0.22; TiO2: 0.2; MnO: 0.06; CuO: 0.18; ZnO: 0.02; ZrO2: 0.01)

26

Table S2. Comparison of different properties of untreated and treated IOS.

Parameters

Untreated IOS

Treated IOS

Particle size (µm)

19±2

25±2

Surface area (m2/g)

32.3±10

135.7±10

Pore volume (ml/g)

0.05±0.01

0.22±0.01

Bulk density (g/ml)

1.8±0.2

1.7±0.2

pHZPC

6.5±0.5

7.6±0.5

Conductivity (1:5 ratio of IOS and

68±10

165±10

6.5±0.5

7.3±0.5

1.5±0.5

16.2±0.5

1.4±0.6

7.5±0.6

distilled water) (µs/cm) pH (1:5 ratio of IOS and distilled water) Maximum adsorption capacity of As(V) (mg/g) at 298 K Maximum adsorption capacity of As(III) (mg/g) at 298 K

27

Table S3. Evaluation of different isotherm parameters at T = 303 K, pH 7. Adsorbent

Langmuir Model

Adsorbate (synthetic water solution) Vm (mg/g)

K

(l/mg)

Freundlich Model

SSE

Kf

n

SSE

mg((n-1)/n) l1/n/g)

Untreated IOS

As(V)

1.5±0.2

0.1±0.01

0.1±

0.4±0.04

3.6±0.3

0.02±0.002

Treated IOS

As(V)

16.2±1.6

5.4±0.5

2.5±

9.5±0.9

5.5±0.5

5.7±0.5

Untreated IOS

As(III)

1.4±0.1

0.05±0.005

0.01±

0.2±0.02

2.6±0.2

0.01±0.001

Treated IOS

As(III)

7.5±0.8

0.01±0.001

2.3±

0.4±0.04

1.5±0.1

2.4±0.2

M3 membrane

As(V)

10.5±1.1

0.17±0.02

5.2±

2.6±0.2

2.8±0.2

4.2±0.4

M3 membrane

As(III)

4.2±0.4

0.01±0.001

0.4±

0.2±0.02

1.4±0.1

0.4±0.04

28

(i)

(ii)

(iii)

(iv)

(v) Fig. 1(a): Scanning electron micrographs representing (i) untreated IOS (1000 X magnification); (ii) treated IOS (1000 X magnification); (iii) top surface of M0 membrane (5000 X magnification); (iv) top surface of M3 membrane (5000 X magnification) & (v) TEM image of treated IOS. 29

(i)

(ii)

(iii)

(iv)

30

Fig. 1(b). Scanning electron micrographs representing cross sectional surface morphology of (i) M0 membrane; (ii) M1 membrane; (iii) M2 membrane; (iv) M3 membrane.

31

(a)

(40 nm)

(b)

(48 nm)

(c)

(54 nm)

(d)

(97 nm)

Fig. 2. AFM images and surface roughness of hollow fiber membranes (a) M0; (b) M1; (c) M2 and (d) M3.

32

2500 M3 membrane

2000 1500

Intensity

1000 8000

Inclusion of succesive phases in the MMM Goethite+quartz+Kaolinite

6000

Treated IOS

Goethite+ Hematite Ferrihydrite Goethite+ Kaolinite

4000

Silicon dioxide

2000 3000

M0 Membrane

Polysulfone

Polyvinylpyrollidone

2000 1000 0

15

30

45 60 75 2 (Degrees)

90

105

Fig. 3. XRD analysis of M0, Treated IOS and M3 membrane.

33

120

100

10 Porosity Permeability

8

(0.02)

7 Porosity

(0.02)

60

6

(0.02) (0.02)

4

20

2

Error bar is 5 % of reported data Standard deviation of data is reported in brackets () 3 number of measurements done for each point

0

(m/Pa.s)

3

10

40

5

Permeability x 10

80

9

1 0

0

2 4 6 8 10 Concentration of treated IOS (wt. %)

Fig. 4. Variation of porosity and permeability of membranes with increasing treated iron ore slime dosages.

34

0.16

(a )

Cumulative pore volume (cc/g)

0.14 M0 membrane M1 membrane M2 membrane M3 membrane

0.12 0.10 0.08 0.06 0.04 0.02 0.00 0

100

Membrane properties

90

100

200

300

400

500 0 Pore width (A )

600

700

800

MWCO (kDa)

(b)

Pore radius (A) Contact angle (degree)

(0.08)

(0.1) (0.05)

80

(0.15)

70

(0.02)

(0.2) (0.09) (0.1)

(0.15)

60

(0.2)

(0.09)

50 (0.12)

40 30 Standard deviations are reported in brackets ()

20

3 number of measurements done for each point

10 0

1

2 3 4 5 6 7 8 9 10 11 Concentration of treated IOS (wt. %)

Fig. 5. Variation of (a) cumulative pore volume and (b) different characterization properties (MWCO, pore radius and contact angle). 35

As(V) concentration in the permeate (g/l)

110 100

(a )

Error is 3% of reported data

90 80 70

P= 21 kPa;

M0

Q= 10 Lph;

M1

C0= 100g/L.

M2 M3

60 50 40 30 20

WHO Limit= 10 g/l

10 0 0

50

100

150

200

250

Time (min)

80

(b)

M0

Error is 5% of reported data

M1

70

M2

60 P= 21 kPa;

2

Permeate flux (l/m .h)

M3

50

Q= 10 Lph; C0= 100g/L.

40 30 20 10 0

50

100 150 Time (min)

36

200

250

300

8.5

After As(V) adsorption

8.0

Before As (V) adsorption

7.5

755 1114

(c )

1502

% Transmittance

7.0

1644

6.5

1398

2041

6.0 5.5

472 660

4006-3704

3240

604

5.0 4.5 4.0 3.5 3.0 4000

3000

2000

1000

0

-1 Wavelength (cm )

Fig. 6. Effect of increasing iron ore slime dosages in membrane on (a) As(V) concentration in the permeate ; (b) permeate flux at an operating condition of 21kPa and 10 l/h and (c) FTIR analysis of membrane M3 (before and after As(V) adsorption).

37

As (III) untreated IOS

Adsorption capacity (mg/g)

As (III) treated IOS

24 22 20 18 16 14 12 10 8 6 4 2 0

As (V) untreated IOS Initial concentration range = 10-100 mg/L;

As (V) treated IOS

shaker speed= 150 rpm;

As (V) M3

shaking time= 24 hours; Temp= 303 K;

As (III) M3

volume= 90 mL;adsorbent weight= 0.45 g

Langmuir isotherm fit

Error is 3% of reported data

0

20

40 60 Ce (mg/L)

80

100

120

Fig. 7: Adsorption isotherm studies for treated, untreated IOS and M3 membrane at 303 K, pH 7.

38

Zeta potential (mV)

10

(a )

Error is 2 % of reported data

5 pHzpc=7.5

0

-5

-10

Concentration of As (V) in the permeate (g/L)

3

70

4

5

(b)

6

7

pH

8

9

10

11

12

Operating Conditions

Feed: 100 g/l;

60

Membrane M3 21 kPa 10 L/h

50

Error is 3 % of reported data

40 30 20 pHZPC

WHO LIMIT

10 0 2

4

6 8 pH of solution

10

12

14

Fig. 8: Variation of (a) zeta potential and (b) As(V) concentration of the membrane with pH of solution.

39

Feed= 100 g/l

55

Membrane M3 21 kPa 10 l/h

60 2 Permeate flux (l/m .h)

34 kPa 10 l/h

50

Error is 3% of reported data

45 40 35

40

30 25 20

20

15 WHO LIMIT

10 5

0 0

5

10

15

20

25

30

35

40

0 45

As (V) concentration in the permeate (g/l)

(a )

34 kPa 20 l/h

Time (h) Feed= 100 mg/l Membrane M3 First run

70

Error is 3 % of reported data

Second run Third run

2 Permeate flux (l/m .h)

60

60

(b)

50

50

40

40

30

30

20

20

10

10 WHO Limit

0 0

5

10

15

20 25 Time (h)

40

30

35

40

0 45

As(V) concentration in the permeate (g/l)

70

100 90

% Removal of As (V)

80 70

(c )

60 

50

Error is 3% of reported data

40 30 20 10

Operating Conditions

sulfate

Feed:- 100 g/l

carbonate

M3 Membrane

Nitrate

21 kPa 10L/h

bicarbonate

0 0

100 200 300 Concentration of Anions (mg/l)

400

Fig. 9: (a) Effect of operating conditions; (b) effect of regeneration and (c) effect of coexisting anions.

41

70

14 2 Permeate flux (L/m h)

Total arsenic in feed: 90 g/Ll

(a )

Membrane M3

60

21 kPa:: 10 L/h

Error is 3% of reported data

13

50

12

40

11

30

10

20

9

10 WHO LIMIT

8

4 Bacterial Concentration x 10 (CFU/100 ml)

0

5

10

(b)

15

20 25 Time (h)

30

35

0 45

40

Total arsenic concentration in the permeate (g/l)

15

Membrane M3

6

21 kPa:: 10 l/h

Feed

Error is 2% of reported data

4

2

0 0

5

10 15 20 Time of operation (h)

25

30

Fig. 10: (a) Permeate flux, total arsenic and (b) microbial concentration profile of a real-life feed solution.

42

Table 1. Nomenclature and different physical parameters of the membranes.

Nomenclature

PSF (w/w%)

PVP (w/w%)

Iron ore slime (w/w

Viscosity of

%)

casting solution

Surface area of prepared

(mPa s)* membranes (m2/g) M-0

18

2

0

830±20

136±10

M-1

4.5

1018±20

50±10

M-2

6.5

1631±20

44±10

M-3

10

3375±20

34±10

*Shear rate 100 1/s; T=90 sec.

43

Table 2(a). Characteristics of permeate and feed of arsenic contaminated real life feed sample. Parameter

Feed

Permeate

Permeate Drinking water

(1 hour)

(after 14

specifications (*)

hours) pH

8.1±0.2

7.5±0.2

7.2±0.2

6.5-8.5

Electrical

346±17

330±16

319±16

-

244±12

230±11

215±10

500

Alkalinity (mg/L)

440±22

400±20

380±19

200

Turbidity (NTU)

6.4±0.3

0.4±0.02

0.1±0.005 5

Iron (mg/L)

2±0.1

1.3±0.065

BDL

0.3

Sodium (mg/L)

49.2±2.5

47.7±2.3

42±2.1

-

Hardness (mg/L)

173±8.7

147±7.4

135±6.8

300

Potassium

5.3±0.3

5.0±0.3

4.3±0.2

-

Arsenic (µg/L)

90±4.5

7±0.3

9.5±0.5

50

Chloride (mg/L)

35±1.8

33±1.7

30±1.5

250

Nitrite (mg/L)

0.5±0.005

0.02±0.001

BDL

-

Nitrate (mg/L)

0.7±0.035

0.03±0.001

BDL

45

Microbial count

54±2.7

27±1.4

0

0

(7.2 ±0.4 ) x 105

(4.5 ±0.3 )x 104 0

0

Conductivity (µS/cm) Total dissolved solids (mg/L)

(Faecal Coli form) MPN/100 ml Microbial count (Entire colony forming, dead or alive) CFU/100 ml

‘*’ represents specifications according to Bureau of Indian standards (BIS); ‘-’ represents no limit. 44

Table 2(b). Comparison of performance of different mixed matrix membranes in arsenic contaminated water filtration. Type

Litres of water treated per gram of the membrane (L/g)

Specific flux of membrane (L/m2*h*kPa)

Energy Performance Index Consumption (Litres of water till (kWh per m2 breakthrough/ unit membrane area gram of inorganics per m3 of added) processed water) Flat Sheet Membranes 10 -

Cost of membrane/m2 (USD) (**)

Reference

PES-Fe Mn binary oxide PVDF Zirconia blend PAN-laterite-PVP Graphene oxidepolysulfone

16.17 (arsenite)

0.9

32.3 (arsenate)

8.4

0.015

4

667

[14]

0.4 (arsenate) -

0.9 0.1

0.025 -

20.3 -

20 6252

[28] [35]

Zirconia nanoparticles-PSF Iron (III)- Chitosan Treated IOS-PSFPVP

95.2 (arsenate)

1.25

Hollow Fiber Membranes -

141

354

[36]

12.5 (arsenate) 2.1 (arsenate) 0.05( arsenite)

2.2E-4 0.8

3.6 X 10-9 4.7 X 10-3

14.5

1233 73

[37] This study

45

[13]

Highlights 

Treated iron ore slime was impregnated in polysulfone hollow fiber membranes.



Mixed matrix membranes showed improvement in properties with impregnation.



Membranes successfully removed arsenic along with bacteria and iron from water.



Arsenic free water was filtered for 28 hours at 17 L/m2h by M-3 membrane.



Regeneration was performed for three cycles.

46