Fabrication of Fe₃O₄ based cellulose acetate mixed matrix membranes for As(III) removal from wastewater

Abstract

Water poisoning due to arsenic is getting worse worldwide because of its serious health hazards and carcinogenic nature. A productive method is required to remove it from water to protect the environment and human life. In this direction, iron oxide (Fe₃O₄)/cellulose acetate (CA)-based mixed matrix membranes (MMMs) were fabricated by varying the concentration of Fe₃O₄ nanoparticles from 0–2 wt% using the phase inversion method for efficient As(III) removal. The impact of Fe₃O₄ on the membranes' surface morphology and mechanical properties was analyzed through scanning electron microscopy (SEM) and ultimate tensile strength (UTS). Fourier transform infrared spectroscopy (FTIR) and X-ray diffraction (XRD) were performed for chemical functionalities and phase structure analysis. Atomic Adsorption Spectrophotometry (AAS) is used to detect the As(III) concentration in water samples. The As(III) adsorption experiments were performed at different concentrations with varying time intervals, and the coefficient of determination and sum of square error function were used to conduct the analysis. The results were best fitted into the Langmuir isotherm model (R² > 0.99) with a maximum adsorption capacity of 90.3 mg g⁻¹. The pseudo-second-order and Weber-Morris models were used to examine intra-particle diffusion as a rate-limiting step. According to membrane performance tests, the nanoparticles' addition increased the hydrophilicity and water flux, improving the membranes' permeability, wettability, and porosity. It was found that a 2 wt% loading of Fe₃O₄ nanoparticles in the MMM achieved a maximum percentage As(III) removal efficiency of 93%. This study shows that these membranes can efficiently remove As(III) from contaminated water because of their adsorption and filtration properties.

---

Water impact

Due to the growing risk that arsenic poses to the environment, this research focused heavily on the most economical and practical approach for removing arsenic from wastewater ensuring availability and sustainable management of water. The fabricated membranes are generally good candidates for removing arsenic from wastewater and have potential for other heavy metal removal and water purification applications.

1. Introduction

Every human needs water, and recent reports indicate that water usage is under significant stress, putting more pressure on the water treatment sector to keep up with the rising water demand. According to the 2023 Report on Sustainable Development Goals (SDGs), billions of people still suffer from recurring water shortages and accessibility issues in safe water for drinking purposes.¹ According to the United Nations, 80% of wastewater enters rivers without being treated first.² Water is a limited resource, and wastewater is a crucial resource for clean water supply. Industrial wastewater contains pollutants and heavy metals, posing threats to human health and aquatic life.^3,4

Arsenic (As) has been identified as the most dangerous heavy metal contaminating water bodies nowadays. Based on the oxidation potential and water pH, As is present in both organic and inorganic forms, which exist in natural and urban wastewater in various oxyanions, such as trivalent arsenite, As(III), and pentavalent arsenate, As(V). As(III) is considered much more poisonous and harder to remove in comparison to arsenate. In particular, long-term exposure to it significantly threatens human health, leading to chronic disorders, high blood pressure, skin problems, and cancer, even in small amounts.⁵ Groundwater comprises As(III) concentrations of 0.50 to 2.50 mg L⁻¹, and industrial wastewater is reported to contain more than 100 mg L⁻¹. Therefore, there is a need to meet the maximum contaminant level (MCL) of acceptable As(III) developed by the World Health Organization (WHO) and the US Environmental Agency Protection (US-EAP) for drinking water in the range of 50 to 10 μg L⁻¹.⁶ Due to this strict standard of arsenic contaminant concentration, several conventional terminologies like adsorption, ion exchange, electrocoagulation, flocculation, chemical precipitation, electro-kinetic technique, phytoremediation, and photocatalysis processes are employed for its removal from water. Significant issues with these previous techniques included generating massive amounts of harmful sludge, using huge volumes of chemicals, a high operational area and cost, and long time durations.⁷

Membrane separation processes have gained significance over other classical methods over the years to overcome the shortcomings detailed above due to their distinct selectivity, simplicity, reusability, high removal efficiency, ease of scaling up, small footprints, and lower energy consumption. Polymeric and ceramic membranes are commonly utilized in industry, but researchers are currently considering the new hybrid or mixed matrix membranes (MMMs), which combine the features of both membranes.^8,9 However, with the advent of MMMs, inserting the adsorbents inside the membrane bulk results in improved mechanical, thermal, magnetic, and electrostatic properties to remove the target contaminant.^10–12 Mohan and Pittman identified different kinds of adsorbents for arsenic removal, such as metal oxides like iron oxides, activated alumina, magnesia, titania, hydrotalcites, and phosphates, which have been mainly used because of their cheapness, smaller size, green nature, recyclability, and specific adsorption capacities for arsenic species.^13,14

Earlier literature explicitly declared that iron oxide-based nanoparticles were preferred because of their low cost, smaller size, and high reactivity. Various forms of iron oxide applied for As(III) removal are magnetite, maghemite, and hematite, but magnetite (Fe₃O₄) tested so far provides higher removal efficiencies, improved strength and stiffness, and the capability to build strong interactions with As(III).^15–17 Feng et al. successfully utilized Fe₃O₄ magnetic nanoparticles for removing As(III) on account of their capability to build inner sphere complexes with As(III) because of their high adsorption capacity, excellent thermal stability, specific surface area and chemical resistance, antifouling property, suitable anti-bacterial property, and biodegradability.^18,19 However, nanoparticles are unsuitable for adsorbents because their separation after treatment is a complicated process. Secondly, in flow through systems, more pressure drops and inter-particle forces will aggregate particles, causing a reduction in surface area.^20,21 Thus, to overcome these deficiencies, nanoparticles are added to the polymer membranes for the practical usage of nanoparticles in wastewater remediation.

Many polymers have been explored and used further for membrane synthesis in water purification.²² Cellulose acetate (CA) is the most favoured polymer in membranes. It is preferred because of the formation of a hydrogen bond as it consists of C=O and –OH groups, as well as for its ease of accessibility, low cost, higher hydrophilicity, adsorption property, non-toxic nature, and biocompatibility. Saranya et al. and Rajeswari et al. developed different adsorbents and nanoparticle-based CA MMMs for environmental amendments and heavy metal removals from textile industry effluents.^23,24 Durthi and Rajulapati studied As(III) removal using CA-incorporated zinc oxide (ZnO) nanoparticle-based MMMs with 58.77% removal efficiency.²⁵ Further research was made by Mithun Kumar using CA and its derivatives as additives in membranes with zirconium oxide (ZrO₂) and binary zinc–magnesium oxide, with 70–87% and 78–81% As(III) removal from the contaminated water, respectively.^26–28 For the polyethersulfone (PES) membrane, using graphene oxide-polyvinylpyrrolidone (GO-PVP) and nano zerovalent iron-kaolin (nZVI-Kaol) as nanocomposites for removing As(III) from synthetic wastewater results in an efficiency of 88.6 and 50%.^29,30 Another study observed 50% removal of As(III) using Fe₂O₃ in a polyvinylidene fluoride (PVDF) polymer matrix.³¹ There is a need to innovate a new membrane for high As(III) removal efficiency. This research deals with the influence of Fe₃O₄ in the CA membrane and its affinity towards As(III) removal from wastewater.

The impregnation of Fe₃O₄ as a potential As(III) nanoadsorbent in the CA polymer matrix improves the adsorptive properties and the membrane's selectivity, permeability, hydrophilicity, porosity, and removal efficiency. According to our study, no prior research has been reported regarding the fabrication of Fe₃O₄/CA MMMs, batch adsorption experiments for non-linear adsorption kinetic models and isotherms utilizing a solver add-in of Microsoft Excel, and filtration experiments for removal efficiency of As(III) in the previous literature. These membranes offer potential for both As(III) removal and high water flux, simultaneously paving the way for new practical applications. The objectives of this work are (i) fabrication of Fe₃O₄/CA MMMs using the phase inversion method; (ii) characterization of Fe₃O₄ and MMMs by several techniques like SEM, FTIR, XRD, and UTS; (iii) testing the membrane's performance in terms of water flux, permeability, wettability, and porosity; (iv) non-linear analysis of adsorption kinetics and isotherms to find out the adsorption mechanism and best fit model; (v) filtration experiments performed to calculate As(III) removal efficiency. As far as we know, Fe₃O₄-incorporated CA membranes were never fabricated before and used for removing As(III) from contaminated water.

2. Materials and methods

2.1. Materials

Cellulose acetate (CA, M_w = 50 000 g mol⁻¹), tetrahydrofuran (THF, purity ≥ 99.9%), iron oxide nanoparticles (Fe₃O₄, size = 50–100 nm), arsenic trioxide (As₂O₃, M_w = 197.84 g.mol⁻¹) and deionized water (DI water) were bought from Sigma Aldrich, USA. Paradise Gases Private Ltd., Pakistan, provided lab scale nitrogen (N₂ gas, purity ≥ 99.95%). There is no need for any additional purification technique, and all chemical reagents purchased are utilized directly.

2.2. Membrane preparation

The Fe₃O₄ nanoparticles embedded in CA MMMs were synthesized using the phase inversion method.³² The concentration of Fe₃O₄ nanoparticles varied by 0–2 wt% in the membrane's casting solution. The names and compositions of fabricated membranes are provided in Table 1.

Table 1 Name and composition of fabricated membranes

Membrane name	CA (wt%)	THF (wt%)	Fe3O4 (wt%)
M0	10	90	0
M0.5	10	89.5	0.5
M1	10	89	1
M1.5	10	88.5	1.5
M2	10	88	2

---

1 g CA was added into 7 mL THF under constant stirring conditions at 350 rpm overnight at an ambient temperature. Due to the magnetic nature of Fe₃O₄, it was impossible to use a magnetic stirrer. Therefore, the nanoparticles were dispersed in the remaining amount of solvent with the help of ultra-sonication for 30 min and then mixed with the already prepared solution. This solution is again placed on a magnetic stirrer until a considerable diffusion of nanoparticles occurs. After that, the mixture was ultra-sonicated for 45 min and degassed. The resultant homogenous mixture is then cast on a glass plate and, after a 30 s delay, immersed for 24 h into a non-solvent DI water coagulation bath. Membrane preparations were made at room temperature and kept in DI water before use. The synthesized membranes were cut to the required size for filtration experiments.

3. Characterization

3.1. Particle characterization

The shape, size, structure and uniform distribution of metal elements of Fe₃O₄ nanoparticles were observed by high resolution SEM. The XRD pattern was obtained on a diffraction system with Cu Kα radiation at a 2θ angle varying from 0 to 80° for the phase characterization. In order to verify the existence of different kinds of functional groups, the FTIR spectrum was collected.

3.2. Membrane characterization

3.2.1. Scanning electron microscopy (SEM). The technique was implemented for synthesized membranes' surface and cross sectional analysis. For SEM analysis, all the membranes were dried between different filter paper layers at 25 °C and soaked in liquid N₂ for several seconds until frozen and later fractured into smaller pieces. A thin layer of gold was applied to these pieces to increase conductivity, enabling electron interaction with atoms in the samples and preventing electrical charging. After that, it is laterally fixed on a brass plate using an adhesive tape. The top surface view and cross section images of these membranes were obtained by SEM (JSM-6490LA, JOEL, Tokyo, Japan) at a 10 mm distance and an applied current of 90 mA. The presence of Fe₃O₄ in the membranes was verified by Energy Dispersive X-ray (EDX) spectroscopy analysis with SEM.

3.2.2. X-ray diffraction (XRD). To check and evaluate the incorporation of Fe₃O₄ nanoparticles in the pure polymer and homogenous blending of membranes XRD analysis was performed by applying Cu Kα monochromatic radiation with an X-ray diffractometer (AG-XEUS Shimadzu). For the structural features of all the samples, the XRD spectrum was taken at a 0.04° step size and 1 s step size in the 2θ range 10–80°. The tube current and voltage were 5 mA and 20 kW, respectively.

3.2.3. Fourier transform infrared spectroscopy (FTIR). An FTIR (Perkin Elmer Spectrum One, USA) spectrometer was used to examine the membrane's composition and the interactions of the corresponding chemical bonds in the polymer chains with Fe₃O₄. The IR spectrum was collected at the wavenumber region from 4000 to 400 cm⁻¹.

3.2.4. Mechanical testing. The ultimate tensile strength testing and elongation until the membranes break is done to check the mechanical strength of the membranes. A universal testing machine (AG-X plus Shimadzu) was employed for tensile stress–strain testing in which membranes made into length and width of 20 mm and 10 mm subsequently hung in between the machine's jaw and ran at a speed of 0.5 mm min⁻¹.

3.3. Membrane performance tests

3.3.1. Filtration assembly. Filtration experiments were conducted in a 300 mL volume capacity, high pressure, chemically resistant, dead end stirred type filtration cell (Sterlitech HP4750). The assembly system is schematically illustrated in Fig. 1. The membrane size was 49 mm in diameter and the active surface area was 14.6 cm² with the highest pressure value of 69 bar. The required pressure was achieved with the aid of an attached inert nitrogen gas source.

Fig. 1 Dead end stirred cell schematic diagram.

3.3.2. Pure Water Flux (PWF). The parameter that determines the water volume passing through membranes per unit area in a unit of time is water flux. The fabricated membranes were placed in a dead end stirred cell for water flux measurements at different pressures at room temperature. A fixed quantity of water samples was obtained in steady state flow, and the time it took for these samples to reach the end of filtration was noted. The PWF of each membrane was obtained from eqn (1):Here, Q represents the water amount collected (L), t the time taken to collect water samples (h) and A the membrane's surface area (m²).³³

3.3.3. Pure Water Permeability (PWP). In this setup, pure DI water was transferred to the feed container and the flux of the pure water was measured by varying the pressures from 2, 3, 4, and 5 bar in a batch mode using inert nitrogen gas from the cylinder. PWP was obtained from eqn (2):where Q is the permeate's volume (L), t considers the time taken for collection of permeate (h), A is the membrane's active area (m²), and ΔP is the difference of pressure (bar) along both sides of the membrane.³⁴

3.3.4. Water content. Water content is the water absorption capacity of membranes, revealing the membrane's hydrophilic nature. For this purpose, the membranes soaked for 24 h in DI water were dried in between different layers of filter paper and immediately weighed for dry weight. After that, the membranes were put in a vacuum oven for 12 h at 50 °C until a consistency in weight was achieved for wet membrane weight. The weight difference in both wet and dry membranes accounts for the water content of each membrane, as shown in eqn (3):where W_w is the wet weight and W_d is the dry weight of the membrane (g). Values were taken thrice to avoid errors.³⁵

3.3.5. Membrane porosity. For the overall porosity measurements, the dry–wet weight method, in which the membranes' dry and wet weights were observed, was used for the membrane porosity calculations as well from eqn (4):Here W_w and W_d correspond to the wet and dry membrane weight (g), A the membrane active surface area (cm²), δ the thickness of the membrane (cm), and ρ pure water density at STP (g cm⁻³). In order to eliminate mistakes, the measuring process was done three times and average values are now stated.³⁶

3.3.6. Contact angle. Analysis of the water contact angle is necessary to examine whether the membrane surface's behavior is hydrophilic or hydrophobic. The sessile drop method is employed under standard conditions where deionized water is used for static contact angle measurements. Different water droplets of 1 μL were placed on the top surface of membranes using micro-syringes at arbitrary positions. The contact angle is measured by a contact angle goniometer using an optical subsystem to capture the profile of a pure liquid on a solid substrate. This system employs high-resolution cameras and software to capture and analyze the contact angle. For more accuracy, more than one measurement was taken at alternative points on the outer membrane surface to derive a mean contact angle value.³⁷

3.3.7. As(III) filtration experiments. The As(III) filtration experiments were performed in the same setup used before for water flux calculations. The As(III) stock solution was prepared from As₂O₃ using the method stated in the previously published literature.³⁸ Experiments were carried out at 3 bar adjusted pressure and permeate samples were collected at fixed time durations at room temperature. The quantitative amount of the As(III) metal ions was determined by AAS (Vario 6, Analytik Jena, Germany) in the permeate. N₂ as a purge gas was utilized due to its inert nature. The percentage removal of As(III) was deduced using eqn (5):where C_i and C_f are the initial and final concentrations of As(III) in the solution, respectively (mg L⁻¹).³⁹

3.3.8. Batch adsorption experiments. The prepared membranes were used for the kinetic and isotherm studies of As(III) adsorption. Adsorption experiments were performed in a batch system at room temperature with the pH adjusted to 7.0 ± 0.1 (ref. 40) and the initial concentration of As(III) set at 100 mg L⁻¹. In strong acidic or basic media, the As(III) adsorption on Fe₃O₄ nanoparticles is insignificant. As(III) shows stronger adsorption onto Fe₃O₄ nanoparticles due to its higher affinity for forming bonds with the iron oxide surface at a favorable pH in the slightly acidic to neutral range (pH 5–7).^41,42 As(III) adsorption was most effective at the pH level of 7.0.⁴³ At pH levels below 7, neutral As(III) species showed no substantial electrostatic interaction. At pH 8.0 and 9.0, As(III) adsorption reduced due to competition with hydroxyls (OH⁻) and deprotonation of Fe₃O₄ surfaces. The findings align with prior studies on the adsorption of multiprotonated oxyanion species on metal oxides.^44,45
3.3.8.1. Adsorption kinetics. In the batch adsorption experiment, a volume of 100 mL solution containing As(III) was used to dissolve the membranes with a weight quantity of 0.10 g. After that, the mixture was shaken for 560 min at 300 rpm in a rotary shaker because the membrane required more time to adsorb before reaching equilibrium. Using AAS, an aliquot (5 mL) was examined at various time intervals to determine the arsenic content. The experimental data was applied to the pseudo-first-order (PFO) and pseudo-second-order (PSO) kinetic models to determine the adsorption kinetics. The adsorption study using the non-linear type model of PFO is shown in eqn (6):

Q_t = Q_e[1 − exp(−K₁t)] (6)

Here Q_e and Q_t = adsorption capacity at equilibrium and relevant contact time, respectively (mg g⁻¹), K₁ = rate constant of PFO (1 min⁻¹).⁴⁶

The nonlinear form of the PSO model is described as eqn (7):

K₂ = PSO rate constant (g (mg⁻¹ min⁻¹)).⁴⁶

The kinetic model is also depicted by Web–Morris which is used to determine the rate-controlling step. According to this model, if a straight line passes through the origin after plotting the graph, intraparticle diffusion is the rate-limiting step; otherwise film diffusion is considered a rate-rate-limited step of adsorption kinetics.⁴⁷ The intraparticle diffusion is given by eqn (8):

q_t = k_pt^1/2 + C (8)

K_p represents the intra-particle diffusion rate constant (mg g⁻¹ min^1/2), and the constant C (mg g⁻¹) gives an idea of boundary layer thickness.⁴⁸

3.3.8.2. Adsorption isotherms. To analyze the adsorption isotherms, 100 mg/100 mL of all fabricated membranes was used to adsorb the initial concentration of As(III) solution (0–100 mg L⁻¹) at room temperature. With the help of 0.1 M HCL/NaOH, the pH level of the solutions was brought to 7.0 ± 0.1. The membrane samples were submerged in that solution and continuously agitated with a rotatory shaker at 300 rpm. The membranes were removed from the solution after 560 min, and AAS was used to determine As(III)'s equilibrium concentration. The experimental adsorption data were fitted into Freundlich and Langmuir isotherm models to determine each membrane's ability to bind As(III).

The relationship between the initial concentration C_o and the equilibrium concentration C_e of As(III) can be visualized using a mathematical isotherm model. Eqn (9)⁴⁹ was used to calculate Q_e.

Here m = mass of the membrane (mg), and V = volume of the adsorbent (mL). C_e was attained at a point where As(III) concentration remained unchanged after a specific amount of adsorbent's contact time, resulting in an overall zero transfer of As(III) adsorption from the membrane's surface (mg g⁻¹). The Langmuir isotherm in non-linear from is given in eqn (10):Here Q_o = constant associated with the membrane's maximum adsorption capacity (mg g⁻¹) and K_L = adsorption constant (L mg⁻¹).⁴⁶

The Freundlich isotherm in non-linear form is given by eqn (11):

Q_e = K_F·C^1/n_e (11)

Here K_F = adsorption capability and n = adsorbent degree of saturation constant. This isotherm represents heterogeneous (multilayer) adsorption on adsorbent surfaces.⁵⁰

4. Results and discussion

4.1. Analysis of Fe₃O₄ nanoparticles

Fig. 2a depicts the characterization results of Fe₃O₄ nanoparticles in which the SEM image is at ×10000 and 20 kV voltage magnification. The highly agglomerative form may be due to nanoparticles' highly reactive nature and their particle size varying from 50–100 nm. Elemental analysis shows atomic % of Fe as 40.2 and O as 59.8, respectively (Fig. 2b). In the FTIR spectrum (Fig. 2c), the characteristic peak at 570 cm⁻¹ is because of the metal oxide Fe–O functional group and the broad bands aside this correlate with the stretching vibrations of Fe–O bonds, specifically those of the crystalline lattice of Fe₃O₄ nanoparticles. The present peaks at 3423 cm⁻¹ and 1620 cm⁻¹ are because of the water molecule lattice.⁵¹

Fig. 2 a SEM image, b EDX mapping, c FTIR spectra, and d XRD pattern of Fe₃O₄ nanoparticles.

The XRD spectrum of Fe₃O₄ (Fig. 2d) shows that the reflection peaks at 2θ equal to 18.29, 30.24, 35.64, 43.38, 53.84, 57.52, 63.02, and 74.53 are attributed to diffraction from the (111), (220), (311), (400), (422), (511), (440), and (533) planes of Fe₃O₄ cubic inverse spinel, respectively. These characteristic peaks best correspond with the standard pattern of Fe₃O₄ (JCPDS 19-629). The absence of any additional peak represents the purity of nanoparticles.^51,52

4.2. Membrane characterization

4.2.1. SEM analysis of MMMs. The top and cross-sectional surface views of all the membranes at a 10 kV voltage and magnification of ×5000 ×1000 are presented in Fig. 3. From these micrographs, it can be inferred that the pure membrane had a dense and smooth surface compared to the other membranes. The concentration of Fe₃O₄ increased and the quantity of THF decreased after incorporating Fe₃O₄ nanoparticles into the membrane's casting solution. Hence, the affinity between the solvent and non-solvent becomes less, resulting in an enhanced porous morphology of MMMs. It could be seen that M_0.5 has more surface pores than M₀. Similarly, the number of pores and their sizes increase with an increasing quantity of the nanoparticles because the polymer and solvent interaction is diminished by the hindrance of the nanoparticles, and the solvent molecules' diffusion rate increases into the coagulation bath from the polymer matrix. Due to the hydrophilic nature of nanoparticles, the water penetration rate during the phase inversion process is faster, producing membranes with larger pore sizes giving higher water flux values than the pure CA membrane.⁵³

Fig. 3 Top surface and cross-section view of Fe₃O₄/CA membranes.

The viscosity of the casting dope solution increased with an increase in the composition of Fe₃O₄ nanoparticles. They tend to agglomerate on the membrane surface because of the delayed demixing rate. This phenomenon is observed on the top surface of M₂ (2 wt% Fe₃O₄ nanoparticles), where nanoparticles aggregate because of poor interactions and improper distribution in the polymer solution, causing the breakage of polymer chains.⁵⁴ The cross-sectional images consist of the top skin layer, which is in accordance with the spinodal decomposition mechanism that ensures selectivity and rejection. The middle layer is responsible for productivity, and the bottom layer provides mechanical support to the MMMs.⁵⁵ The creation of a large number of pores in the middle layer can be associated with the increase in the contents of hydrophilic nanoparticles that accelerate the de-mixing rate of the polymer solution with the non-solvent during phase separation.⁵⁶

The presence of macro-voids indicates faster solvent and non-solvent exchange during the phase inversion process, thus improving the overall membrane porosity.⁵⁷ With excessive loading of nanoparticles, the casting solution becomes more viscous and thermodynamically unstable, forming thicker membranes from M₀ to M₂. The bottom layer is smooth and dense, with no pores.⁵⁸

4.2.2. XRD patterns of MMMs. Fig. 4 presents detailed XRD patterns of the Fe₃O₄ nanoparticles, pure CA membrane, and 2 wt% CA/Fe₃O₄ membrane. The XRD spectrum of pure Fe₃O₄, as described above, has sharp characteristics peaks at 2θ = 30.24°, 35.64°, 43.38°, 53.84°, 57.52°, and 63.02° which are attributed to diffraction from the (220), (311), (400), (422), (511) and (440) planes with inverse spinel face-centered cubic structure.^51,52

Fig. 4 XRD patterns of Fe₃O₄, pure CA, and CA/Fe₃O₄ membranes.

The XRD spectrum of the pure CA membrane shows a sharp peak at 2θ = 12° that corresponds to the (001) plane, showing the semi-crystalline nature of the polymer. CA has low crystallinity because of its significant intermolecular attraction to hydrogen bonding between acetyl and hydroxyl groups.³³ However, the Fe₃O₄/CA mixed matrix membrane shows the signature peaks at 2θ = 12°, 30.24°, 35.64°, and 43.38°, which is in accordance with the CA polymer and the Fe₃O₄ nanoparticle characteristic peaks. It ensures that Fe₃O₄ is inserted into the CA polymer matrix.⁵⁹

4.2.3. FTIR spectrum of MMMs. To clarify the existence of Fe₃O₄ in membranes and the joining bond of Fe₃O₄ with CA, the FTIR spectrum of the pure CA membrane and all Fe₃O₄/CA MMMs, along with pure Fe₃O₄ nanoparticles, is demonstrated in Fig. 5.

Fig. 5 FTIR spectrum of Fe₃O₄, pure CA, and Fe₃O₄/CA membranes.

In the case of Fe₃O₄ nanoparticles, the decisive peak was at 570 cm⁻¹ assigned to the Fe–O group. In pure membrane M₀, the broadband at 3600–3200 cm⁻¹ is the characteristic of O–H stretching. The 1753 cm⁻¹ band was attributed to the functional group C=O of CA. Furthermore, peaks at 1251, 1373, 2128, and 2940 cm⁻¹ were assigned to the C–O (stretching), CH₃ (symmetric deformation), carbonyl C=O stretching vibrations, and CH₃ (asymmetric stretching). In all other membranes M_0.5, M₁, M_1.5, and M₂, there is no additional peak present in membranes.^60,61

4.2.4. Mechanical testing. Tensile strength and elongation break are the two main vital elements responsible for the mechanical testing of membranes. Fig. 6 exhibits clearly that there is a rising trend in tensile strength that interprets the strong bonding between the polymer matrix and nanoparticle filler. The addition of nanoparticles fills the channels in the membrane matrix, leading to continuous phase formation and enhanced mechanical strength. Incorporating Fe₃O₄ in the CA matrix demonstrates the arranged packing of polymer chains, which makes the CA crystalline structure more rigid in the Fe₃O₄/CA membrane. The pure CA membrane shows tensile strength at 23.9 ± 0.10 MPa with an increasing trend till 34.8 ± 0.20 MPa because of the proper distribution of nanoparticles within the polymer phase. Then, there is a gradual decrease to 32 ± 0.30 MPa, mainly marked by the increased porosity of the membrane with an increased percentage of Fe₃O₄ nanoparticles, as depicted by SEM images earlier.

Fig. 6 Mechanical properties of Fe₃O₄/CA membranes.

4.3. Membrane performance

4.3.1. Pure Water Flux (PWF). Fig. 7 displays the PWF variations for the membranes at different applied pressures. The PWF for the CA membrane was relatively low because of its dense top layer, and a rising trend of water flux was noted with an increase in nanoparticle concentration up to 1.5 wt% owing to the combined effect of its hydrophilicity and porosity. However, with the excessive loading of nanoparticles, the 2 wt% water flux decreases. It could be because of the viscosity increase and blocking of pores by particle agglomeration; also, the membrane's pore size decreased in view of the polymer's slow solidification during the process of phase inversion.⁶² The highest water flux value reported at 5 bar is 129 L m⁻² h⁻¹ for M_1.5, and the lowest value is 17 L m⁻² h⁻¹ for the pure CA membrane at 2 bar. This trend for all values is in accordance with the SEM interpretations.

Fig. 7 Pure water flux of Fe₃O₄/CA membranes.

4.3.2. Pure Water Permeability (PWP). The PWP values of the membranes obtained are demonstrated in Fig. 8. With an increase in pressure, there was an increase in the trend in the permeability of membranes. By contrast, the M₀ membrane had the lowest permeability of 8.5 L m⁻² h⁻¹ bar⁻¹ among all the MMMs. As described earlier, the addition of hydrophilic Fe₃O₄ nanoparticles enhanced both the porosity and hydrophilicity that contributed to better flux values in MMMs;⁶³ therefore, M_1.5 exhibited a higher water permeability value of 30 L m⁻² h⁻¹ bar⁻¹.

Fig. 8 Pure water permeability of Fe₃O₄/CA membranes.

4.3.3. Water content. The membrane's capacity for water content is indirectly related to membrane hydrophilicity and porosity. The water content values of all the membranes were computed using eqn (3), stated in Table 2. The water content percentage in the pure membrane was 44% and after loading Fe₃O₄ up to 2 wt%, it was 68%. Because of the hydrophilic nature of nanoparticles and their potential to create spaces, porosity, and large pore volumes in the polymer matrix, the water uptake in Fe₃O₄/CA MMMs will increase rather than in the pure CA membrane.⁵³

Table 2 Performance parameters of Fe₃O₄/CA membranes

Membranes	Water content (%)	Porosity (%)	Contact angle (°)
M0	44	66	68.9 ± 0
M0.5	49	68	65.5 ± 2
M1	55	71	60.7 ± 4
M1.5	61	73	52 ± 3
M2	68	74	47 ± 4

---

4.3.4. Membrane porosity. According to the results of porosity determined by eqn (4), the existence of nanoparticles had a dominating effect on membrane porosity (Table 2). The membrane with 2 wt% of Fe₃O₄ nanoparticles had the highest value of 74% of all other fabricated membranes. The reason was that Fe₃O₄ nanoparticles allow more water into the casting suspension during solvent and non-solvent exchange in the phase inversion process, thereby promoting the formation of more porous membranes with more and bigger pores.^24,34

4.3.5. Contact angle. Contact angle measurement is a direct evaluation of the membrane's hydrophilic or hydrophobic characteristics. In Table 2, the contact angle of the pure CA membrane is 68.9°, but the values decreased from 68.9° to 42.7° as the percentage of Fe₃O₄ nanoparticles increased from 0.5–2 wt%, indicating higher hydrophilicity. The membrane's surface becomes more hydrophilic as the contact angle decreases³⁷ (Fig. 9).

Fig. 9 Contact angle images for CA/Fe₃O₄ membranes.

4.3.6. Batch adsorption study of As(III) on Fe₃O₄/CA membranes. Fe₃O₄ nanoparticles have been extensively researched for their adsorption properties, particularly their ability to remove As(III) from wastewater. Studies show that under optimal conditions, Fe₃O₄ nanoparticles can achieve a maximum adsorption capacity of above 100 mg g⁻¹ for As(III) at pH 7.¹⁹ In order to determine the Fe₃O₄/CA membrane's performance and interactions with As(III)-contaminated water, a series of adsorption experiments were performed at different concentrations and time intervals.
4.3.6.1. Adsorption kinetics. The kinetic investigation is an important parameter to find out the impact of As(III) adsorption on the membranes. The adsorption kinetic curves for each Fe₃O₄/CA membrane are drawn in Fig. 10. The results demonstrated that in the first 120 min, there was a quick uptake of As(III), which was then followed by a slower adsorption phase and equilibrium was reached within 560 min. The kinetic study parameters and the experimentally calculated adsorption capacity are tabulated in Table 3.

Fig. 10 Adsorption kinetic models: (a) pseudo-first-order (b), pseudo-second-order and (c) Web–Morris intraparticle diffusion plots of As(III) removal from Fe₃O₄/CA membranes.

Table 3 Kinetic parameters of As(III) adsorption by Fe₃O₄/CA membranes

Membrane	Adsorption capacity at equilibrium	First-order model			Second-order model			Intraparticle diffusion model
	Q e,exp	Q e1	K 1	R ^2	Q e2	K 2	R ^2	K 1	C	R ^2
	mg g^−1	mg g^−1	min^−1		mg g^−1	g mg^−1 min^−1		mg g^−1 min^1/2
M0	37	36.4	0.006	0.98	46.0	0.0001	0.99	6.77	9.98	0.85
M0.5	42	39.8	0.012	0.97	46.4	0.0003	0.99	8.40	9.98	0.93
M1	63	60.65	0.011	0.97	71.5	0.0001	0.99	12.6	9.98	0.93
M1.5	83	77.7	0.015	0.96	88.5	0.0002	0.99	17.0	1.80	0.96
M2	90.3	86.7	0.016	0.97	97.5	0.0002	0.99	19.2	1.55	0.96

---

The result depicts that the experimental value is a little bit closer to the PFO, but the kinetic model is best fit to the PSO kinetic model, as supported by the high regression coefficient (R² > 0.99) value because R² is one of the significant factors in determining the kinetic model. The Fe₃O₄/CA membrane was thought to chemically adsorb As(III) species, according to the PSO kinetic model. It is recorded that as the amount of Fe₃O₄ increases from 0, 0.5, 1, and 1.5 to 2 wt%, the adsorption capacity values increase from 37, 42, 63, 83, and 90.3 mg g⁻¹.

Fig. 10 demonstrates the deviation of the curve from the origin, which indicates that intraparticle diffusion is not only the rate-limiting step; but in this film, diffusion is also involved.⁶⁴ The high value of the intercept is directly proportional to the boundary layer effect and suggests that surface adsorption is a rate-determining step. In the present study, initially, the intercept value is very high, which indicates the presence of external mass transfer resistance in the thick boundary layer, and when the dosage of Fe₃O₄ nanoparticles increases (M_1.5 and M₂), the intercept value is low, which signifies a minimum resistance to mass transfer and enhances the adsorption rate.^65,66 The result also depicts that the increase in the dosage value of Fe₃O₄ nanoparticles and the rate constant of intra-diffusion increase indicates a faster rate of diffusion, and a high value (19.22) is obtained at membrane M₂, which shows that adsorbate molecules readily move through the pores, which causes a faster adsorption process. The present study also shows that the rate constant value decreases when the boundary layer effect increases.⁶⁷

4.3.6.2. Adsorption isotherms. As(III) adsorption isotherm experiments on Fe₃O₄/CA membranes were done at various concentrations of As(III) (0, 20, 40, 60, 80, 100 mg L⁻¹) at different intervals of time (0, 15, 30, 60, 120, 240, 360, 480, 560 min). As indicated in Fig. 11, the results of the experiment were both fitted into the Freundlich and Langmuir isotherm models, and Table 4 lists the adsorption isotherm model parameters. The findings revealed that all the membranes' adsorption capacities increased along with the As(III) initial concentration. In comparison to the Freundlich isotherm model, the values of the experiments were successfully fitted into the Langmuir isotherm model based on the regression coefficient.

Fig. 11 Adsorption isotherm (Langmuir and Freundlich) model applied on Fe₃O₄/CA membranes.

Table 4 Isotherm parameters of As(III) adsorption by Fe₃O₄/CA membranes

Isotherm model	Parameters	Membranes
		M0	M0.5	M1	M1.5	M2
Langmuir	Q o	97.65	81.49	108.01	210.48	279.44
	K L	0.010	0.01952	0.0399	0.038	0.0494
	R ^2	0.994	0.992	0.992	0.999	0.999
Freundlich	N	1.7308	3.26	0.596	0.736	0.7907
	K F	0.749	0.6419	7.68	10.42	15.19
	R ^2	0.975	0.95381	0.946	0.997	0.996

---

Theoretically, As(III) adsorption was assumed to be on the single layer of the Fe₃O₄/CA membranes rather than the multi-layer when the results of the experiments of adsorption isotherms fitted the Langmuir model. M₂ performed better than the other membranes in an adsorption capacity comparison using an adsorption kinetic and isotherm analysis. The pure water flux and permeability values reduced as the Fe₃O₄ contents increased from 1.5 to 2.0 wt%, which could be caused by the agglomeration of excess Fe₃O₄ nanoparticle loading within the membrane, but this contributed to increased adsorption capacity because of the availability of more adsorption sites, as seen from the SEM image.

4.3.7. As(III) filtration study. In order to evaluate the removal efficiency of As(III) from water, filtration experiments were done using dead-end stirred cells by passing As(III) solution through all fabricated membranes. The permeate samples were collected and tested using AAS for the detection of arsenic concentrations. The removal efficiency was calculated using eqn (5), and the results are drawn in Fig. 12.

Fig. 12 Removal efficiency of Fe₃O₄/CA membranes for As(III).

As(III) removal through membranes involves two main mechanisms: adsorption and size exclusion. Whenever the As(III)-polluted water diffuses into the membrane's cross section, Fe₃O₄ nanoparticles can effectively adsorb arsenic via inner sphere complexation and produce an arsenic-free permeate.^68,69Fig. 13 revealed SEM–EDX mapping of the membrane done before and after the filtration experiment.

Fig. 13 SEM-EDX mapping of the M₂ membrane before and after As(III) removal.

Thus, As(III) is adsorbed on the surface of Fe₃O₄ by forming As–Fe complexes through a ligand–exchange mechanism. These interactions occur due to the electropositive charge on the As metal ion that acts as an electron acceptor and the electronegative charge on the oxygen atoms attached to Fe₃O₄ having a lone pair of electrons acting as an electron donor.⁷⁰ The As(III) removal percentage increased from 37% to 93%, with the percentage increase in Fe₃O₄ nanoparticles from 0–2 wt%. These values are associated with the more available active sites for the adsorbent Fe₃O₄ nanoparticles distributed on the membrane bulk, which provides a large surface area for incoming As(III) ions. In the case of Fe₃O₄, they have hydroxyl groups that when exposed to water bind or release H⁺ ions, which causes adsorptive removal behaviour because of the OH²⁺, OH⁻, and O⁻ functional groups. As(III) exhibits both metallic (As) and ligand (O) characteristics for binding to Fe₃O₄. This is the reason why As(III) and Fe₃O₄ will form a tight internal spherical complex.⁷¹Table 5 shows all the fabricated adsorptive membranes that have been used for As(III) adsorption. These membranes have varying concentration of nanoparticles and showed As(III) removal efficiencies from 50 to 96.9%.^{29–31,36,49,72–75} In the case of adsorption capacity, it can be concluded that the M₂ membrane, among all the other fabricated membranes, shows the highest As(III) adsorption capacity in wastewater.

Table 5 Adsorption capacity of As(III) by several adsorptive mixed matrix membranes

Membranes	Nanoparticles (wt%)	As(III) feed concentration (mg L^−1)	Adsorption isotherm model	As(III) adsorption capacity (mg g^−1)	As(III) removal efficiency (%)	References
PVDF/zirconia	20.0	100	Langmuir	21.50	96.9	72
PES/Fe-Mn binary oxide	22.5	100	Langmuir	73.50	87.5	73
PSF/iron ore slimes	10.0	100	Freundlich	7.50	90	74
PSF/HINM	1.5	100	Freundlich	41.09	90.8	36
CA/MnO2	1.5	100	Langmuir	50.34	—	49
PVDF/TiO2-HNT	0.5	0.25	Langmuir	0.18	94	75
PES/GO-PVP	0.5	100	Langmuir	88.6	88.6	29
PES/nZVI-Kaol	0.75	75	—	50	50	30
PVDF/Fe2O3	15	0.25	—	23	50	31
CA/Fe3O4	2.0	100	Langmuir	90.3	93	This study

---

5. Conclusion

In this research work, Fe₃O₄/CA MMMs were fabricated by varying the amount of Fe₃O₄ nanoparticles from 0 to 2 wt% using a phase inversion process and investigated for As(III) removal from contaminated water. The fabricated membranes were characterized by several techniques, like SEM, FTIR, XRD, and UTS, for membrane surface morphology, functional group detection, phase structure, and mechanical strength. The incorporation of Fe₃O₄ nanoparticles endows the polymeric membranes with enhanced hydrophilicity, optimized porosity, improved water flux and permeation in comparison to the pure CA membrane. The adsorption experiments showed that at 100 mg L⁻¹ As(III) concentration with 0.1 g of M₂ at 560 min, a maximum adsorption capacity of 90.3 mg g⁻¹ was attained. This system confirms the monolayer adsorption of As(III) on the Fe₃O₄/CA membrane, as depicted by non-linear analysis. The data was best fitted into the Langmuir adsorption isotherm and the Weber–Morris intraparticle diffusion model following pseudo-second-order kinetics. These membranes efficiently removed As(III) from water with better removal efficiency, comparable with the literature. It was observed that among all the membranes, the membrane with 2 wt% Fe₃O₄ nanoparticles exhibited the highest As(III) removal efficiency of 93%. In short, the addition of Fe₃O₄ nanoparticles can be a simple modification strategy that enhances the membrane's performance parameters with promising potential and can successfully treat As(III) contaminated water as defined by the WHO and USEPA to fulfill the MCL, i.e. 0.01 mg L⁻¹.

Abbreviations

CA	Cellulose acetate
THF	Tetrahydrofuran
MMMs	Mixed Matrix Membranes
As	Arsenic
As(III)	Arsenite
As(V)	Arsenate
As2O3	Arsenic trioxide
N2	Nitrogen
HCL	Hydrochloric acid
NaOH	Sodium hydroxide
DI Water	De-ionized water
Fe	Iron
C	Carbon
O	Oxygen
Cu	Copper
Fe3O4	Iron oxide
ZnO	Zinc oxide
ZrO2	Zirconium oxide
TiO2	Titanium oxide
MnO2	Manganese oxide
GO	Graphene oxide
nZVI	nano Zero-Valent Iron
PSF	Polysulfone
PES	Polyethersulfone
PVP	Polyvinylpyrrolidone
PVDF	Polyvinylidene fluoride
pH	Potential of hydrogen
OH	Hydroxyl group
C=O	Carbonyl group
CH3	Methyl group
WHO	World Health Organization
US-EPA	Environmental Agency Protection
SDGs	Sustainable Development Goals
SEM	Scanning Electron Microscopy
EDX	Energy Dispersive X-ray
FTIR	Fourier Transform Infrared Spectroscopy
XRD	X-ray Diffraction
UTS	Ultimate Tensile Strength
AAS	Atomic Absorption Spectroscopy
PWF	Pure Water Flux
PWP	Pure Water Permeability
PFO	Pseudo-First-Order
PSO	Pseudo-Second-Order
MCL	Maximum contaminant level
%	Percentage
Fig	Figure
Eq.	Equation
g	Gram
μg	Microgram
mg	Milligram
mA	Milliampere
mm	Millimeter
cm	Centimeter
L	Liter
m	Meter
nm	Nano Meter
kV	Kilovolt
°	Degree
kW	Kilowatt
s	Second
min	Minute
h	Hour
kN	Kilonewton
wt	Weight
°C	Degree Celsius
Δ	Differential
Q e	Adsorption capacity at equilibrium (mg g^−1)
Q t	Relevant contact time (mg g^−1)
K 1	Rate constant of Pseudo-First-Order (PFO) (min^−1)
Q o	Constant associated with membrane's maximum adsorption capacity (mg g^−1)
K L	Adsorption constant (L mg^−1)
K F	Adsorption capability (L mg^−1)
n	Adsorbent degree of saturation constant
K 2	Pseudo-Second-Order rate constant (g (mg^−1 min^−1))
W w	Wet weight (g)
W d	Dry weight of the membrane (g)
Q	Permeate's volume (L)
T	Time taken for collection of permeate (h)
A	Membrane's active area (m^2)
ΔP	Difference of pressure (bar)
K p	Intra-particle diffusion rate constant (mg g^−1 min^1/2)
C	Constant

Author contributions

Sidra Liaquat (investigation, acquisition of data, writing original draft), Sarrah Farrukh (project administration, supervision), Nasir Ahmad (resources analysis tool), Syed Shujaat Karim (review and editing), Erum Pervaiz (formal analysis), Ayesha Sultan and Subhan Ali (visualization).

Conflicts of interest

There are no conflicts to declare.

Acknowledgements

In order to perform this research, we gratefully acknowledge the fastest-growing Membranes for Applied Research (MEMAR) lab at the School of Chemical and Materials Engineering (SCME), National University of Sciences and Technology (NUST), Islamabad, Pakistan.

References

1. U. Water, Sustainable Development Goal 6 synthesis report on water and sanitation, Published by the United Nations New York, New York, 2018, vol. 10017.
2. W. H. Organization, 1 in 3 people globally do not have access to safe drinking water, UNICEF, WHO, Available from: https://www.who.int/news/item/18-06-2019-1-in-3-people-globally-do-not-have-access-to-safe-drinking-water-unicef-who.
3. A. Azimi, A. Azari, M. Rezakazemi and M. Ansarpour, Removal of heavy metals from industrial wastewaters: a review, ChemBioEng Rev., 2017, 4, 37–59.
4. L. Joseph, B.-M. Jun, J. R. Flora, C. M. Park and Y. Yoon, Removal of heavy metals from water sources in the developing world using low-cost materials: A review, Chemosphere, 2019, 229, 142–159.
5. W. Altowayti, N. Othman, S. Shahir, A. Alshalif, A. Al-Gheethi, F. Al-Towayti, Z. Saleh and S. Haris, Removal of arsenic from wastewater by using different technologies and adsorbents: a review, Int. J. Environ. Sci. Technol., 2021, 1–24.
6. W. H. Organization, A global overview of national regulations and standards for drinking-water quality, 2021.
7. S. Alka, S. Shahir, N. Ibrahim, M. J. Ndejiko, D.-V. N. Vo and F. Abd Manan, Arsenic removal technologies and future trends: a mini review, J. Cleaner Prod., 2021, 278, 123805.
8. D. Qadir, H. Mukhtar and L. K. Keong, Mixed matrix membranes for water purification applications, Sep. Purif. Rev., 2017, 46, 62–80.
9. I. Salahshoori, A. Seyfaee and A. Babapoor, Recent advances in synthesis and applications of mixed matrix membranes, Synthesis and Sintering, 2021, 1, 1–27.
10. T. Siddique, N. K. Dutta and N. R. Choudhury, Mixed-matrix membrane fabrication for water treatment, Membranes, 2021, 11, 557.
11. M.-C. Shih, An overview of arsenic removal by pressure-drivenmembrane processes, Desalination, 2005, 172, 85–97.
12. C. M. Nguyen, S. Bang, J. Cho and K.-W. Kim, Performance and mechanism of arsenic removal from water by a nanofiltration membrane, Desalination, 2009, 245, 82–94.
13. Y. J. Lim, S. M. Lee, R. Wang and J. Lee, Emerging materials to prepare mixed matrix membranes for pollutant removal in water, Membranes, 2021, 11, 508.
14. D. Mohan and C. U. Pittman Jr, Arsenic removal from water/wastewater using adsorbents—a critical review, J. Hazard. Mater., 2007, 142, 1–53.
15. A. Chiavola, E. Amato, M. Stoller, A. Chianese and M. Boni, Application of iron based nanoparticles as adsorbents for arsenic removal from water, Chem. Eng. Trans., 2016, 47, 325–330.
16. P. Sabbatini, F. Rossi, G. Thern, A. Marajofsky and M. F. de Cortalezzi, Iron oxide adsorbers for arsenic removal: a low cost treatment for rural areas and mobile applications, Desalination, 2009, 248, 184–192.
17. S. I. Siddiqui and S. A. Chaudhry, Iron oxide and its modified forms as an adsorbent for arsenic removal: a comprehensive recent advancement, Process Saf. Environ. Prot., 2017, 111, 592–626.
18. S. Lunge, S. Singh and A. Sinha, Magnetic iron oxide (Fe3O4) nanoparticles from tea waste for arsenic removal, J. Magn. Magn. Mater., 2014, 356, 21–31.
19. L. Feng, M. Cao, X. Ma, Y. Zhu and C. Hu, Superparamagnetic high-surface-area Fe3O4 nanoparticles as adsorbents for arsenic removal, J. Hazard. Mater., 2012, 217, 439–446.
20. X. Yu, S. Tong, M. Ge, J. Zuo, C. Cao and W. Song, One-step synthesis of magnetic composites of cellulose@ iron oxide nanoparticles for arsenic removal, J. Mater. Chem. A, 2013, 1, 959–965.
21. X. Zhang, Y. Wang, X. Chang, P. Wang and B. Pan, Iron oxide nanoparticles confined in mesoporous silicates for arsenic sequestration: effect of the host pore structure, Environ. Sci.: Nano, 2017, 4, 679–688.
22. A. Rabajczyk, M. Zielecka, K. Cygańczuk, Ł. Pastuszka and L. Jurecki, Nanometals-containing polymeric membranes for purification processes, Materials, 2021, 14, 513.
23. A. Rajeswari, E. J. S. Christy, G. I. C. Mary, K. Jayaraj and A. Pius, Cellulose acetate based biopolymeric mixed matrix membranes with various nanoparticles for environmental remediation-A comparative study, J. Environ. Chem. Eng., 2019, 7, 103278.
24. R. Saranya, Y. L. Thuyavan and G. Arthanareeswaran, Development of Adsorbents-based Cellulose Acetate Mixed Matrix Membranes for Removal of Pollutants from Textile Industry Effluent, J. Teknol. Lab., 2014, 70, 1–5.
25. C. P. Durthi, S. B. Rajulapati, A. A. Palliparambi, K. K. Anand and S. H. Sonawane, Studies on removal of arsenic using cellulose acetate–zinc oxide nanoparticle mixed matrix membrane, Int. Nano Lett., 2018, 8, 201–211.
26. M. Kumar, S. RaoT, A. M. Isloor, G. S. Ibrahim, N. Ismail, A. F. Ismail and A. M. Asiri, Use of cellulose acetate/polyphenylsulfone derivatives to fabricate ultrafiltration hollow fiber membranes for the removal of arsenic from drinking water, Int. J. Biol. Macromol., 2019, 129, 715–727.
27. M. Kumar, A. M. Isloor, T. S. Rao, A. F. Ismail, R. Farnood and P. Nambissan, Removal of toxic arsenic from aqueous media using polyphenylsulfone/cellulose acetate hollow fiber membranes containing zirconium oxide, Chem. Eng. J., 2020, 393, 124367.
28. M. Kumar, A. M. Isloor, S. R. Todeti, H. Nagaraja, A. F. Ismail and R. Susanti, Effect of binary zinc-magnesium oxides on polyphenylsulfone/cellulose acetate derivatives hollow fiber membranes for the decontamination of arsenic from drinking water, Chem. Eng. J., 2021, 405, 126809.
29. N. R. N. Abdul Ghani, S. S. Sulaiman, A. Tahreen and M. S. Jami, Polyether Sulfone-Graphene Oxide-Polyvinyl Pyrrolidone Nanocomposite Adsorptive Membrane for Arsenic Removal from Wastewater, J. Water Environ. Nanotechnol., 2021, 6, 121–137.
30. B. K. Selvan, K. Thiyagarajan, S. Das, N. Jaya, S. A. Jabasingh, P. Saravanan, M. Rajasimman and Y. Vasseghian, Synthesis and characterization of nano zerovalent iron-kaolin clay (nZVI-Kaol) composite polyethersulfone (PES) membrane for the efficacious As2O3 removal from potable water samples, Chemosphere, 2022, 288, 132405.
31. Y.-A. Boussouga, F. Tantish and A. I. Schäfer, Microporous Hematite-Loaded Composite Membrane for Arsenic(III) and Arsenic(V) Removal, ACS Appl. Eng. Mater., 2023, 1, 1164–1175.
32. L. Rozelle, J. Cadotte, R. Corneliussen, E. Erickson, K. Cobian and C. Kopp Jr, Phase inversion membranes, Encyclopedia of Separation Science, ed. M. Mulder, Academic Press, Cambridge, MA, USA, 2000, pp. 3331–3346.
33. R. Abedini, S. M. Mousavi and R. Aminzadeh, A novel cellulose acetate (CA) membrane using TiO2 nanoparticles: preparation, characterization and permeation study, Desalination, 2011, 277, 40–45.
34. T. Marino, F. Russo, L. Rezzouk, A. Bouzid and A. Figoli, PES-kaolin mixed matrix membranes for arsenic removal from water, Membranes, 2017, 7, 57.
35. A. Gholami, A. Moghadassi, S. Hosseini, S. Shabani and F. Gholami, Preparation and characterization of polyvinyl chloride based nanocomposite nanofiltration-membrane modified by iron oxide nanoparticles for lead removal from water, J. Ind. Eng. Chem., 2014, 20, 1517–1522.
36. A. M. Nasir, P. S. Goh and A. F. Ismail, Highly adsorptive polysulfone/hydrous iron-nickel-manganese (PSF/HINM) nanocomposite hollow fiber membrane for synergistic arsenic removal, Sep. Purif. Technol., 2019, 213, 162–175.
37. M. Batool, A. Shafeeq, B. Haider and N. M. Ahmad, TiO2 nanoparticle filler-based mixed-matrix PES/CA Nanofiltration membranes for enhanced desalination, Membranes, 2021, 11, 433.
38. A. Chand, P. Chand, G. Khatri and D. Paudel, Enhanced Removal Efficiency of Arsenic and Copper from Aqueous Solution Using Activated Acorus calamus Based Adsorbent, Chem. Biochem. Eng. Q., 2021, 35, 279–293.
39. R. Rezaee, S. Nasseri, A. H. Mahvi, R. Nabizadeh, S. A. Mousavi, A. Rashidi, A. Jafari and S. Nazmara, Fabrication and characterization of a polysulfone-graphene oxide nanocomposite membrane for arsenate rejection from water, J. Environ. Health Sci. Eng., 2015, 13, 1–11.
40. J. C. Hsu, C. J. Lin, C. H. Liao and S. T. Chen, Removal of As(V) and As(III) by reclaimed iron-oxide coated sands, J. Hazard. Mater., 2008, 153, 817–826.
41. S. Dixit and J. G. Hering, Comparison of arsenic(V) and arsenic(III) sorption onto iron oxide minerals: implications for arsenic mobility, Environ. Sci. Technol., 2003, 37, 4182–4189.
42. X. Huang, Y. Liu, X. Wang, L. Zeng, T. Xiao, D. Luo, J. Jiang, H. Zhang, Y. Huang and M. Ye, Removal of arsenic from wastewater by using nano Fe₃O₄/zinc organic frameworks, Int. J. Environ. Res. Public Health, 2022, 19, 10897.
43. C. H. Liu, Y. H. Chuang, T. Y. Chen, Y. Tian, H. Li, M. K. Wang and W. Zhang, Mechanism of Arsenic Adsorption on Magnetite Nanoparticles from Water: Thermodynamic and Spectroscopic Studies, Environ. Sci. Technol., 2015, 49, 7726–7734.
44. J. Perendija, Z. S. Veličković, I. Cvijetić, J. D. Rusmirović, V. Ugrinović, A. D. Marinkovic and A. Onjia, Batch and column adsorption of cations, oxyanions and dyes on a magnetite modified cellulose-based membrane, Cellulose, 2020, 27, 8215–8235.
45. S. Mandal, M. K. Sahu and R. K. Patel, Adsorption studies of arsenic(III) removal from water by zirconium polyacrylamide hybrid material (ZrPACM-43), Water Resour. Ind., 2013, 4, 51–67.
46. A. A. Khan, S. R. Naqvi, I. Ali, M. Arshad, H. AlMohamadi and U. Sikandar, Algal-derived biochar as an efficient adsorbent for removal of Cr (VI) in textile industry wastewater: Non-linear isotherm, kinetics and ANN studies, Chemosphere, 2023, 316, 137826.
47. Ş. Yilmaz, A. Zengin, T. Şahan and Ö. Selçuk Zorer, Utilization of a novel polymer-clay material for high elimination of hazardous radioactive contamination uranium(VI) from aqueous environments, Environ. Technol. Innovation, 2021, 23, 101631.
48. U. Ecer, Ş. Yilmaz and T. Şahan, Investigation of Mercury(II) and Arsenic(V) adsorption onto sulphur functionalised pumice: a response surface approach for optimisation and modelling, Int. J. Environ. Anal. Chem., 2020, 102, 1–21.
49. Z. Qiu, H. Chen, Z. Wang, T. Zhang, D. Yang and F. Qiu, Efficient removal of As (III) via the synergistic effect of oxidation and absorption by FeOOH@ MnO2@ CAM nano-hybrid adsorption membrane, Chemosphere, 2020, 258, 127329.
50. J. He, Y. Song and J. P. Chen, Development of a novel biochar/PSF mixed matrix membrane and study of key parameters in treatment of copper and lead contaminated water, Chemosphere, 2017, 186, 1033–1045.
51. V. Srivastava, P. Singh, C. Weng and Y. Sharma, Economically viable synthesis of FeO nanoparticles and their characterization, Pol. J. Chem. Technol., 2011, 13, 1–5.
52. P. Daraei, S. S. Madaeni, N. Ghaemi, E. Salehi, M. A. Khadivi, R. Moradian and B. Astinchap, Novel polyethersulfone nanocomposite membrane prepared by PANI/Fe3O4 nanoparticles with enhanced performance for Cu (II) removal from water, J. Membr. Sci., 2012, 415, 250–259.
53. C. Evangeline, V. Pragasam, K. Rambabu, S. Velu, P. Monash, G. Arthanareeswaran and F. Banat, Iron oxide modified polyethersulfone/cellulose acetate blend membrane for enhanced defluoridation application, Desalin. Water Treat., 2019, 156, 177–188.
54. R. Saranya, G. Arthanareeswaran, A. Ismail, D. D. Dionysiou and D. Paul, Zero-valent iron impregnated cellulose acetate mixed matrix membranes for the treatment of textile industry effluent, RSC Adv., 2015, 5, 62486–62497.
55. P. Sherugar, N. S. Naik, M. Padaki, V. Nayak, A. Gangadharan, A. R. Nadig and S. Déon, Fabrication of zinc doped aluminium oxide/polysulfone mixed matrix membranes for enhanced antifouling property and heavy metal removal, Chemosphere, 2021, 275, 130024.
56. X. Zhang, X. Fang, J. Li, S. Pan, X. Sun, J. Shen, W. Han, L. Wang and S. Zhao, Developing new adsorptive membrane by modification of support layer with iron oxide microspheres for arsenic removal, J. Colloid Interface Sci., 2018, 514, 760–768.
57. M. B. M. Y. Ang, K. P. O. Devanadera, A. N. R. Duena, Z.-Y. Luo, Y.-H. Chiao, J. C. Millare, R. R. Aquino, S.-H. Huang and K.-R. Lee, Modifying cellulose acetate mixed-matrix membranes for improved oil–water separation: Comparison between sodium and organo-montmorillonite as particle additives, Membranes, 2021, 11, 80.
58. A. S. M. Ali, M. M. Soliman, S. H. Kandil and M. Khalil, Emerging mixed matrix membranes based on zeolite nanoparticles and cellulose acetate for water desalination, Cellulose, 2021, 28, 6417–6426.
59. I. Shakeel, A. Hussain and S. Farrukh, Effect analysis of nickel ferrite (NiFe2O4) and titanium dioxide (TiO2) nanoparticles on CH4/CO2 gas permeation properties of cellulose acetate based mixed matrix membranes, J. Polym. Environ., 2019, 27, 1449–1464.
60. M. M. Ibrahim, T. Y. Fahmy, E. I. Salaheldin, F. Mobarak, M. A. Youssef and M. R. Mabrook, Role of tosyl cellulose acetate as potential carrier for controlled drug release, Life Sci. J., 2015, 10, 127–133.
61. S. H. Sonawane, A. Terrien, A. S. Figueiredo, M. Clara Goncalves and M. N. De Pinho, The role of silver nanoparticles on mixed matrix Ag/cellulose acetate asymmetric membranes, Polym. Compos., 2017, 38, 32–39.
62. R. Sahraei, T. Shahalizade, M. Ghaemy and H. Mahdavi, Fabrication of cellulose acetate/Fe3O4@ GO-APTS-poly (AMPS-co-MA) mixed matrix membrane and its evaluation on anionic dyes removal, Cellulose, 2018, 25, 3519–3532.
63. N. Tanne, R. Xu, M. Zhou, P. Zhang, X. Wang and X. Wen, Influence of pore size and membrane surface properties on arsenic removal by nanofiltration membranes, Front. Environ. Sci. Eng., 2019, 13, 19.
64. A. V. Vitela-Rodriguez and J. R. Rangel-Mendez, Arsenic removal by modified activated carbons with iron hydro(oxide) nanoparticles, J. Environ. Manage., 2013, 114, 225–231.
65. J. Wang and X. Guo, Rethinking of the intraparticle diffusion adsorption kinetics model: Interpretation, solving methods and applications, Chemosphere, 2022, 309, 136732.
66. F.-C. Wu, R.-L. Tseng and R.-S. Juang, Initial behavior of intraparticle diffusion model used in the description of adsorption kinetics, Chem. Eng. J., 2009, 153, 1–8.
67. A. E. Ofomaja, E. B. Naidoo and A. Pholosi, Intraparticle diffusion of Cr (VI) through biomass and magnetite coated biomass: A comparative kinetic and diffusion study, S. Afr. J. Chem. Eng., 2020, 32, 39–55.
68. R. Jain, Recent advances of magnetite nanomaterials to remove arsenic from water, RSC Adv., 2022, 12, 32197–32209.
69. W. Cheng, W. Zhang, L. Hu, W. Ding, F. Wu and J. Li, Etching synthesis of iron oxide nanoparticles for adsorption of arsenic from water, RSC Adv., 2016, 6, 15900–15910.
70. M. Ahmaruzzaman, Magnetic nanocomposite adsorbents for abatement of arsenic species from water and wastewater, Environ. Sci. Pollut. Res., 2022, 29, 82681–82708.
71. S. Yean, L. Cong, C. T. Yavuz, J. Mayo, W. Yu, A. Kan, V. Colvin and M. Tomson, Effect of magnetite particle size on adsorption and desorption of arsenite and arsenate, J. Mater. Res., 2005, 20, 3255–3264.
72. Y.-M. Zheng, S.-W. Zou, K. N. Nanayakkara, T. Matsuura and J. P. Chen, Adsorptive removal of arsenic from aqueous solution by a PVDF/zirconia blend flat sheet membrane, J. Membr. Sci., 2011, 374, 1–11.
73. R. J. Gohari, W. Lau, T. Matsuura and A. 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., 2013, 118, 64–72.
74. S. Chatterjee and S. De, Adsorptive removal of arsenic from groundwater using chemically treated iron ore slime incorporated mixed matrix hollow fiber membrane, Sep. Purif. Technol., 2017, 179, 357–368.
75. A. Moslehyani, R. Farnood, S. Tabe, T. Matsuura and A. F. Ismail, Novel Nanocomposite HNT-TiO2/PVDF Adsorptive Nanofiber Membranes for Arsenic (III) Removal, J. Membr. Sci. Res., 2020, 6, 416–423.

---