Arsenate removal from aqueous solutions using magnetic mesoporous iron manganese bimetal oxides

Abstract

Magnetic mesoporous iron manganese bimetal oxide (MMIM) with high specific surface area, pore volume and well interconnected mesopores was synthesized via the nanocasting strategy using KIT-6 as a hard template. The morphology and physicochemical properties of the samples were characterized using SEM-EDS, TEM, XRD, VSM, FT-IR and XPS, etc. The obtained MMIM was used as an adsorbent to remove arsenate from aqueous solutions, and presented excellent performances for As(V) removal. The adsorption equilibrium data were well described by the Freundlich model, and solution pH values affected the removal efficiency of arsenate significantly, which was due to the isoelectric point (IEP) of the MMIM. The adsorption kinetics fitted a pseudo-second order model, and intra-particle diffusion was not the only rate-limiting step. In addition, MMIM exhibited a sensitive magnetic response and could be easily separated and recovered from aqueous solutions with an external magnetic field. Based on analysis results, possible mechanisms were discussed based on a combination of the results to different theoretical adsorption models.

---

1 Introduction

Arsenic is a toxic and carcinogenic chemical element widely distributed in the atmosphere, soils and natural waters, and arsenic contamination is considered to be one of the most serious environmental problems today.¹ Typically, widespread arsenic contamination of groundwater is caused by natural processes and a range of anthropogenic activities.^2,3 In natural water, soluble arsenic is primarily present in inorganic forms and exists in two predominant oxyanions: arsenite and arsenate,⁴ which depend significantly on redox and pH conditions.⁵ Ingesting arsenic-contaminated water is extremely detrimental to human health, such as cardiovascular, neurological and endocrine disorders, bladder and kidney cancers.^6,7 Consequently, most countries such as United States, the European Union and China,^8,9 are implementing maximum contaminant level (MCL) of 10 ppb recommended by the World Health Organization (WHO). However, in many parts of the world, such as Bangladesh, the west Bengal in India and the Inner Mongolia region in China, the arsenic concentration has far exceeded the standard values.^10,11 Therefore, effective removal of arsenic has becoming more and more urgent in water environment.

Compared with other conventional methods such as coagulation, reverse osmosis and biological treatment,^12–14 adsorption has been proven to be one of the most promising technologies for arsenic remediation, with advantages of having high removal efficiency, simplicity for operation, low cost and high recycle rate without harmful by-products. Among potential adsorbents, iron (hydr)oxides and iron-containing substances have been widely focused on for arsenic species removal due to their strong affinity and high selectivity for inorganic arsenic species in the sorption processes.² However, the serious drawback for those iron (hydr)oxides and iron-containing adsorbents is low adsorption capacity of arsenic, which limited the usage of these adsorbents. Recently, incorporating some other metal elements such as Zr, Ce, Mn and Co into iron oxides to prepare bimetal oxides is gaining considerable attentions,^15–17 since these bimetal oxides not only can inherit the advantages of parent oxides but also show obviously synergistic effect. It is found that iron oxides incorporated with Ce or Mn can significantly increase the arsenic adsorption capacity than the individual metal oxides from aqueous solutions.^2,18,19 Magnetic bimetal oxide nanoparticles, possessing high specific surface area, excellent adsorption ability and could be easily separated from solution by using an external magnetic field, have appealing properties by combining the virtues of the each component, which have enhanced properties in comparison with their monometallic nanoparticles.^17,20 However, the agglomeration of nanoparticles via the inter-particle dipolar force,²¹ leading to the loss of size effect and the decrease of surface area,²² which limits their practical application.

Since the discovery of ordered mesoporous silica, mesoporous materials have attracted more attention due to their unique properties such as uniform and adjustable mesopores, tunable and open pore architectures, high specific surface area and large pore volume.^23,24 Moreover, the stable and interconnected frameworks of their materials could be easily modified and functionalized without any change in mesostructure,^25,26 which could further tailor their adsorption performance. Therefore, these ordered mesoporous silicas are considered to be ideal hard template materials to prepare the ordered non-silica mesoporous metal/bimetal oxides. Compared with common bimetal oxides, bimetal oxides with mesoporous structures have unique physicochemical properties, such as intrinsic high specific surface area and large pore volume, regular and tunable pore size distribution, as well as stable and interconnected frameworks. Such features meet the requirement as excellent adsrobents, not only providing large space and interface capable of accommodating capacious guest species, but also enabling the possibility of specific binding, enrichment and separation.

On the basis of the above considerations, in the present study, a novel magnetic mesoporous iron manganese bimetal oxides (MMIM) with well-defined uniform mesopores was synthesized through the hard template approach, which combined the superiority of bimetal oxides and mesoporous materials. The obtained MMIM sample was used as an adsorbent and presented excellent performance for arsenate removal from aqueous solutions due to the high specific surface area and pore volume, and could be easily separated and recovered from the aqueous solutions using an external magnetic field after the arsenic adsorption. Characterization and analysis were performed by SEM-EDS, TEM, VSM, XRD and XPS before and after adsorption. Based on the results of analysis, possible adsorption mechanism of As(V) removal by MMIM from aqueous solutions was proposed.

2 Experimental section

2.1 Synthesis of KIT-6 template and MMIM

Three-dimensional mesoporous silica (KIT-6) with Ia3d symmetry was used as a hard template and was synthesized according to a previous report.²⁷ Briefly, 6.0 g Pluronic 123 (EO₂₀PO₇₀EO₂₀) dissolved in 217.0 mL of ultrapure water (18.2 MΩ cm⁻¹, Millipore) and 10.3 g of HCl (37%), then 6.0 g of n-butanol was added to the homogeneous solution at 35 °C. After vigorous stirring for 1 h, 12.9 g of TEOS (tetraethoxysilane) was added to the solution, which was then stirred for 24 h at this temperature. Subsequently, the white milky suspension was aged at 80 °C for another 24 h under static conditions. The white solid product was filtered, dried at 100 °C and finally calcined in air at 550 °C for 6 h.

The magnetic mesoporous iron manganese bimetal oxides was synthesized as follows: 1.0 g Fe(NO₃)₃·9H₂O, 0.246 g FeCl₂·4H₂O and 0.311 g Mn(NO₃)₂·4H₂O, which were used as iron and manganese precursors, were dissolved in 20 mL ethanol, and 1.0 g KIT-6 was added into this homogenous solution under vigorous stirring. The mixture was stirred for 2 h at room temperature, and then the ethanol was evaporated at 50 °C. These two synthetic steps were carried out in an anaerobic glove box to prevent Fe(II) oxidized to Fe(III) in mixture. The obtained dry powder was heated to 300 °C and calcined at this temperature for 6 h in high purity nitrogen atmosphere. The impregnation procedure was then repeated with half the amount of the metal precursors, and the precursors@silica composites were calcined 450 °C for 6 h in high purity nitrogen atmosphere; the resulting sample was treated with 2 M NaOH solution to remove the silica template, centrifuged, washed several times with ultrapure water and dried at 40 °C in air, and the final obtained powder was referred to as MMIM. For comparison, KIT-6 templated Fe₃O₄ was also was synthesized through the KIT-6, and the procedures were similar as the MMIM exception of the manganese precursors. The final obtained powder was denoted as templated Fe₃O₄.

2.2 Characterization of materials

Scanning electron microscopy (SEM) images and energy dispersion spectrometer (EDS) patterns of the samples were obtained on a Hitachi S-4800 apparatus. Transmission electron microscopy (TEM) images of the samples were obtained with a JEOL2010F instrument. Nitrogen adsorption/desorption isotherms were conducted with a Quantachrome Autosorb-iQ at 77 K, with degassing at 393 K prior to the measurements. The specific surface area, pore size distribution and pore volume of the samples were calculated to the BET (Brunauer–Emmett–Teller) method and Barrett–Joyner–Halenda (BJH) model, respectively. A Bruker D8 Advance Powder X-ray Diffractometer (XRD, Germany) using a Cu Kα (λ = 1.5406 Å) radiation source scanned from 10° to 90°, with a scanning speed of 1° min⁻¹ and a step size of 0.02°. The operation voltage and current were kept at 40 kV and 40 mA, respectively. The magnetic property was characterized using a vibrating sample magnetometer (VSM, lakeshore 7407, USA) at room temperature. The zero potential of the adsorbent materials before and after the adsorption were measured in the pH range from 3 to 10 using a Zetasizer apparatus (Nano Z, Malvern, U.K.). Fourier transform infrared (FT-IR) spectra of samples were measured on a Nicolet 5700 spectrometer with a range of 4000–400 cm⁻¹. The functional groups and the related oxidation states on the surface of the materials were analyzed by X-ray photoelectron spectroscopy (XPS) in a PHI 5000 Versaprobe spectrometer equipped with a rotating Al anode generating Al Kα X-ray radiation at 1486.6 eV.

2.3 Batch experiments

Solutions with different concentrations of As(V) prepared with Na₂HAsO₄·7H₂O (Analytical Grade, Sigma-Aldrich) and ultrapure water (18.2 MΩ cm⁻¹, Millipore). Batch experiments were performed in glass bottles (120 mL) containing 100 mL As(V) solution with required concentrations, and the dose of MMIM was 0.2 g L⁻¹. To test the effect of pH, the mixed solutions were adjusted to desired values ranging from 3.0 to 12.0 by adding HCl or NaOH solutions. For equilibrium experiments, a series of arsenate solutions was prepared ranging from 1 to 50 mg L⁻¹; the bottles were sealed with screw caps and placed in a thermostatic shaker (200 rpm) at 25 °C for 24 h to ensure equilibrium. Samples were collected at certain time intervals in 5 mg L⁻¹ and 10 mg L⁻¹ arsenic solutions for the adsorption kinetics study. To evaluate the regeneration and reusable properties, the As-adsorbed MMIM was eluted using 1 M NaOH solution and then washed with ultrapure water until a neutral pH was reached, dried in a vacuum, and reused in the next cycle of adsorption experiments. The adsorption/desorption cycles were repeated five times with 5 mg L⁻¹ of As(V) solution at pH 3.0.

Aqueous solution samples were immediately centrifuged, and supernatant were filtered with a 0.22 μm membrane, the residual arsenic concentrations were measured by Inductively Coupled Plasma-Optical Emission Spectrometry (ICP-OES, Agilent 720 ES, USA).

3 Results and discussion

3.1 Characterization of materials

The structure and morphology of the KIT-6 template were investigated with TEM and HRTEM (Fig. S1a–c†), the results confirmed that this KIT-6 template was a high-quality cubic Ia3d silica product, and the material consisted uniquely of large ordered bicontinuous mesoporous domains.²⁷ In addition, the small-angle power X-ray diffraction (SXRD) pattern (Fig. S1d†) of KIT-6 template shows two well-resolved scattering peaks which can be assigned to the (211) and (220) reflections of space group Ia3d, also suggesting a highly ordered 3D bicontinuous cubic mesostructure. Nitrogen adsorption/desorption isotherm of the KIT-6 template (Fig. S2†) was IV type with a sharp capillary condensation step at relatively high pressure, suggesting the existence of highly uniform channel-like mesopores,²⁸ and Fig. S2† inset illustrates a narrow pore size distribution. The BET surface area, pore size distribution and pore volume of template mesoporous silica (KIT-6) were 579.60 m² g⁻¹, 6.25 nm and 0.777 cm³ g⁻¹, respectively, which were shown in Table 1.

Table 1 Texture parameters of KIT-6 template and MMIM

Samples	BET (m^2 g^−1)	Pore size (nm)	Pore volume (cm^3 g^−1)
KIT-6 template	579.60	6.25	0.777
MMIM	143.88	5.88/11.52	0.647

---

SEM and EDS images of MMIM are shown in Fig. 1a and b, it was clearly observed that this obtained MMIM possesses large clusters nanoparticles. It was reported that with increasing aging temperature of KIT-6, the interconnectivity between these two channel systems increased.²⁹ In the present study, KIT-6 template obtained at a high aging temperature of 80 °C had higher interconnectivity between these two channel systems, and the nanocasted MMIM resulted in a rather dense structure. In addition, the pore system of the parent KIT-6 template was never completely filled in a nanocasting process, the nanocast material had to aggregate during the curing in parts of the parent template pore system.²⁹ These results made the obtained MMIM in the present study seemed to be hundreds of nanometers. The EDS analysis corresponding to the SEM image clearly revealed the existence of iron and manganese elements in the MMIM, and both the weight ratio and atomic ratio of Fe/Mn were approximately 3/1, which agreed well with the initial Fe/Mn molar ratio. This result suggested that iron and manganese were completely loaded on the channels of the KIT-6 template. The structure and morphology was further investigated with TEM. Highly well-ordered mesoporous structures of MMIM can be clearly seen in Fig. 1c. The lattice fringes were clearly visible with a spacing of 0.254 nm from HRTEM (Fig. 1d), which was in good agreement with the spacing of the (311) planes of Fe₃O₄.³⁰ No distinct lattices of manganese oxides were clearly observed, suggesting manganese oxides was amorphous in MMIM, which was consistent with WXRD analysis that was discussed later.

Fig. 1 SEM (a), EDS (b), TEM (c) and HRTEM (d) images of MMIM.

The specific surface area (SSA) and mesoporosity parameters of MMIM were investigated by nitrogen adsorption/desorption measurement. The specific surface area of MMIM was obtained by the Brunauer–Emmett–Teller (BET) method, the total pore volume was determined using the adsorbed volume at a relative pressure of 0.99, while pore size distributions of MMIM calculated from the adsorption branch of isotherm because the adsorption branch is highly preferred for pore size calculations and is hardly affected by any tensile strength effect phenomenon.³¹ Fig. 2a shows the isotherms of MMIM exhibited a type-IV with H1 hysteresis loop, typical characteristic of uniform mesoporous metal oxide prepared by the hard template method. This result revealed that the synthesized MMIM had well-defined uniform mesopores, which was in agreement with the results of TEM images. The detailed texture parameters of MMIM are also shown in Table 1. The obtained MMIM possesses a higher BET surface area and pore volume compared with other reported mesoporous oxides,^32–35 and this high surface area and pore volume may be the major reasons for its excellent capacity for arsenate removal from aqueous solutions. The Fig. 2a inset shows the bimodal pore size distribution of the MMIM: the small pore size was approximately 5.9 nm, while the large one was approximately 11.5 nm. Earlier studies have also indicated that KIT-6 template is composed of two channel systems that are connected to each other through micropores channels in the silica walls,^29,32,36,37 and the prevalence of these channels varies with the hydrothermal synthesis conditions. Nanocasted replica may grow within the pores of hard template in two ways, which may result in the formation of coupled replicas or uncoupled replicas, when metal oxides grow only in one of the channel systems of the KIT-6 template, the nanocast metal oxides show a bimodal pore size distribution.

Fig. 2 Nitrogen adsorption/desorption isotherm, pore size distribution (in inset) (a) and WXRD pattern (b) of MMIM.

Fig. 2b shows the WXRD pattern of MMIM. The diffraction peaks for the sample at 2θ values of 18.2°, 30.0°, 35.4°, 43.0°, 53.2°, 57.0° and 62.6°, corresponding to the reflection planes of (111), (220), (311), (400), (422), (511) and (440), respectively, could be indexed to magnetite Fe₃O₄ (JCPDS card no. 19-0629). The average crystallite size of the MMIM was calculated from the WXRD by using the Scherrer equation, and the value was approximately 10.7 nm. No principal peaks of manganese oxides appeared in the pattern of the MMIM nanoparticles in Fig. 2b, suggesting that the generated manganese oxides in MMIM may be amorphous, as no obvious diffraction peaks appeared in the WXRD pattern. Pure manganese oxide was also synthesized via the KIT-6 template in order to investigate the phase of the manganese species. Namely, the synthetic method was the same as that for MMIM, but the metal precursors was only manganese nitrate. Fig. S3a† shows the WXRD pattern of the pure manganese oxides, the broad peaks of the manganese oxides pattern indicated a poorly crystalline form, which was the major reason that no obvious diffraction peaks appeared in the WXRD pattern of MMIM. The major phase of manganese oxide were Mn₃O₄ (JCPDS no. 18-0803) and MnO₂ (JCPDS no. 44-0141), suggesting that Mn(II) was oxidized to Mn(III) and Mn(IV) during the high purity nitrogen atmosphere calcinations process, the same result was also deduced from XPS, which would be discussed later. However, compared with the WXRD pattern of the pure manganese oxides, the WXRD pattern of the pure iron oxides (templated Fe₃O₄), which was shown in Fig. S3b,† was not very different from that of MMIM, and the major phase of iron oxides was magnetite Fe₃O₄ (JCPDS card no. 19-0629).

The magnetic hysteresis curves measured at room temperature of the MMIM is shown in Fig. 3a. Saturation magnetization (M_s) is a key factor for the successful magnetic separation and the saturation magnetization of MMIM is 33.2 emu g⁻¹, which is lower than the reported for bulk magnetite (92 emu g⁻¹), the smaller particle size and the surface related effects such as surface disorder may be the main reasons for this result,^38,39 which is consistent with the obtained results of WXRD and TEM. Due to the nonlinear hysteresis loop with nonzero remnant magnetization (M_r) and coercivity (H_c), MMIM shows well pronounced ferromagnetic properties.⁴⁰ These results show that the MMIM could possess a sensitive response with an external magnetic field, such as a magnet (Fig. 3a inset), thus providing a potential advantage for the separation and recovery of adsorbent from aqueous solutions after the arsenic adsorption.

Fig. 3 Magnetization curves measured at room temperature for the MMIM. Inset: the dispersion and magnetic separation of the MMIM attracted by a magnet (a) and the FT-IR spectra (b) of MMIM and As(V)-loaded MMIM.

FT-IR analysis was conducted to reveal the surface nature of MMIM and As-loaded MMIM, which are shown in Fig. 3b. All the two samples exhibited broad and strong bands at 3426 cm⁻¹ and 1635 cm⁻¹, attributable to H–O–H stretching vibration and twisting vibration of water molecules, respectively, indicating the presence of physically-sorbed interstitial water molecules on the MMIM.⁴¹ The peak at 1015 cm⁻¹ corresponds to the bending vibration of the hydroxyl group (metal–OH),⁴² which was responsible for the formation of inner-sphere surface complexes.⁴³ Two adsorption peaks at 670 cm⁻¹ and 585 cm⁻¹ can be attributed to Fe–O–Fe and Fe–O stretching modes, respectively.^18,44 A new adsorption peak at 820 cm⁻¹ appeared on the spectra of As(V)-loaded MMIM, which was attributed to the stretching vibrations of Fe–As–O bond,^2,16 while the peaks at 1015 cm⁻¹ significantly weakened, indicating that some hydroxyl groups on the surface of MMIM were replaced by arsenate species. So it could be speculated that the substitution of metal–OH groups by arsenic ions played a key role in the adsorption process.

X-ray photoelectron spectroscopy (XPS) is a versatile analysis technique that was used to investigate the composition and chemical state of MMIM before and after arsenic adsorption. Fig. 4a shows the full-range survey spectra of MMIM and As(V)-loaded MMIM, which indicated that the major elements of MMIM were iron (Fe), manganese (Mn), oxygen (O) and carbon (C). The C 1s spectra before and after arsenic adsorption are presented in Fig. S4,† the peaks at 284.8 eV and 288.6 eV can be assigned to carbon dioxide and carbonate, respectively.^45,46 The carbon appearing on the spectra was due to exposure to air and water during sample preparation and reaction. The As 3d spectra on the full-range survey spectra of As-loaded MMIM revealed that the arsenate ions present in aqueous solutions had been adsorbed on the surface of MMIM. The O 1s narrow scans could be deconvoluted into three overlapped peaks corresponding to oxide oxygen (O²⁻), hydroxyl groups (OH⁻) and adsorbed water (H₂O). From Fig. 4b and c and Table 2, it was found that the O 1s spectra were quite different before and after adsorption. M–O (where M represents metal oxide substrate) increased from 44.504% to 48.535%, this slight increase may be attributed to the new oxygen obtained from As–O after arsenate was adsorbed on the surface of MMIM. However, the percentage of OH⁻ decreased from 36.789% to 32.821%. Arsenic was mainly removed through the hydroxyl groups (OH⁻) of the M–OH,¹⁸ and the high concentration of M–OH was proposed as the main reason for the high arsenic adsorption capacity of MMIM. However, by increasing the arsenic adsorbed on the surface of MMIM, some OH⁻ would be replaced by As(V), so the percentage of OH⁻ decreased after As-loaded MMIM. In was known that the binding energies of the different chemical states of the As 3d core level for As(III) and As(V) are 44.3–44.5 and 45.2–45.6 eV, respectively.^47,48 On the spectra of As(V)-loaded MMIM based on As 3d high-resolution deconvoluted spectra (Fig. 4d), only one As(V) 3d peak was detected at 45.58 eV, attributable to As(V)–O bonding, indicating that As(V) has been adsorbed on the surface of MMIM.

Fig. 4 Full-range XPS spectra of MMIM (a), O 1s spectra with three deconvolutions of MMIM (b), As(V)-loaded MMIM (c) and As 3d core levels of As(V)-loaded MMIM (d).

Table 2 Relative contents of O 1s in various chemical states

Samples	Chemical states	Binding energy (eV)	Percent (%)
MMIM	O^2−	529.25	44.504
	OH^−	530.77	36.789
	H2O	531.57	18.707
As(V)-loaded MMIM	O^2−	529.84	48.535
	OH^−	531.41	32.821
	H2O	532.24	18.644

---

The Fe 2p and Mn 2p XPS spectra are shown in Fig. S5;† the binding energies at 711.29 eV and 724.85 eV were assigned to Fe 2p_3/2 and Fe 2p_1/2 in Fig. S5a,† and the separation of the 2p_3/2 and 2p_1/2 spin–orbit levels was approximately 13.6 eV. It was in agreement with the literature that the peaks shifted to the higher binding energy and broaden for Fe₃O₄.⁴⁹ In addition, the absence of the satellite peak situated at approximately 719 eV, which was identified as a major characteristic of γ-Fe₂O₃, also indicated that the iron oxides in MMIM was Fe₃O₄, because the satellite peak in Fe₃O₄ would become less resolved or absent.⁵⁰ These results were in line with the result of WXRD pattern in Fig. 2b. The binding energies of MMIM at 642.2 eV and 653.9 eV were assigned to Mn 2p_3/2 and Mn 2p_1/2 in Fig. S5b,† indicating the presence of Mn(II), Mn(III) and Mn(IV) states of manganese,⁵¹ because the binding energies of Mn(II), Mn(III) and Mn(IV) were very closed to each other.⁵² This result indicated that although in high purity nitrogen atmosphere, part of Mn(II) was still oxidized into Mn(III) and Mn(IV) during the process of calcinations.

3.2 Arsenate adsorption isotherms and adsorption kinetics

MMIM, KIT-6 templated Fe₃O₄ and commercial Fe₃O₄ nanopowders were used to evaluate the adsorption capacity, and the As(V) adsorption isotherm data were fitted by both the Langmuir model (1) and the Freundlich model (2):where C_e is the equilibrium concentration of arsenic (mg L⁻¹); q_e is the equilibrium adsorption capacity (mg g⁻¹); q_m is the maximum adsorption capacity (mg g⁻¹); K_L is the adsorption constant (L mg⁻¹); and K_F and 1/n are Freundlich isotherm constants related to the adsorption capacity and the intensity of adsorption, respectively.

Fig. 5 shows the Langmuir and Freundlich isotherms for arsenate adsorption, and the calculated parameters are summarized in Table 3. It was found that the adsorption data of As(V) were better fitted by the Freundlich model from the correlation coefficients (R²). The maximum adsorption capacities of As(V) were 35.35, 18.85 and 10.86 mg g⁻¹, respectively, indicating that the obtained MMIM was higher uptake of As(V) than that of the template Fe₃O₄ and the commercial Fe₃O₄ nanopowders. Compared with commercial Fe₃O₄ (The BET surface area and pore volume of this commercial Fe₃O₄ nanopowders were 45.63 m² g⁻¹ and 0.213 cm³ g⁻¹, respectively), this enhancement in As(V) adsorption capacity was possibly due to the high specific surface area, large pore volume (143.88 m² g⁻¹ and 0.647 cm³ g⁻¹, respectively) and internal uniform mesopores of MMIM, which can enhance the accessibility of arsenic species to the active sites. However, the BET surface area and pore volume of templated Fe₃O₄ were 154.36 m² g⁻¹ and 0.593 cm³ g⁻¹, respectively. But the calculated maximum adsorption capacities of MMIM for As(V) was still much higher than the values of template Fe₃O₄. This result indicated that this magnetic mesoporous iron manganese bimetal oxides (MMIM) contained iron and manganese oxides presents the synergistic effect than monometallic iron oxides. Therefore, compared with commercial Fe₃O₄ and template Fe₃O₄, this obtained MMIM possesses the superiority of bimetal oxides and mesoporous materials, which not only can enhance arsenic adsorption capacity but also presence of the obviously synergistic effect than monometallic iron oxides (Fe₃O₄).

Fig. 5 Adsorption isotherms of As(V) on MMIM and at 25 °C. The initial arsenate concentration ranged from 1 to 50 mg L⁻¹, the dosage of MMIM and the initial pH values for the solutions were 0.2 g L⁻¹ and 3.0, respectively. Inset: adsorption isotherms of As(V) for commercial Fe₃O₄ and templated Fe₃O₄, the same conditions as MMIM except to the initial arsenate concentration.

Table 3 Estimated isotherm parameters for As(V) adsorption by MMIM, templated Fe₃O₄ and commercial Fe₃O₄

Samples	Langmuir isotherm model			Freundlich isotherm model
	KL (L mg^−1)	qm (mg g^−1)	R^2	n	KF	R^2
MMIM	0.643	35.35	0.9307	4.102	15.659	0.9836
Templated Fe3O4	1.565	18.85	0.9137	4.029	9.986	0.9927
Commercial Fe3O4	6.263	10.86	0.9128	5.489	7.320	0.9836

---

The results are also significantly higher than that of other reported related adsorbents, such as iron nanomaterials,⁵³ Magnetite–graphene hybrids,¹ Fe–BTC polymer,⁵⁴ γ-Fe₂O₃ flowers,³⁰ Iron oxide@carbon²¹ and cellulose@Fe₂O₃ composites.²² (Table S1†) The constant K_F values, which is defined as the arsenic adsorbed on the adsorbent at unit equilibrium concentration. The K_F values of MMIM, templated Fe₃O₄ and commercial Fe₃O₄ were 15.659, 9.986 and 7.320, respectively, indicating that MMIM exhibited higher adsorption capacity of As(V) than that of templated Fe₃O₄ and commercial Fe₃O₄ nanopowders, which agrees well with the conclusions drawn from the Langmuir model. The value of n from MMIM in Freundlich model for As(V) was greater than 1, suggesting this adsorption equilibrium isotherm is nonlinear, which can be attributed to adsorption site heterogeneity, electrostatic attraction and other sorbent–sorbate interactions.⁵⁵

The adsorption kinetics of As(V) are shown in Fig. 6a. The process is time dependent and the adsorption process can be divided into two steps: (i) the adsorption rate was considerably fast within the first 6 h, which may be due to the large number of available active sites on the surface of MMIM; (ii) the adsorption rate was slow in the subsequent step because the concentration gradients gradually reduce due to the accumulation of As(V) anions on the surface of MMIM, decreasing the adsorption rate of the second stage, and equilibrium capacity was reached within 24 h. This longer equilibrium time may indicate that specific adsorption occurred between the arsenic species and the surface of the adsorbent.⁵⁶ The pseudo-first order eqn (3) and pseudo-second order eqn (4) were applied to fit the experimental data:

where K₁ (h⁻¹) and K₂ (g mg⁻¹ h⁻¹) are the pseudo-first and pseudo-second order adsorption rate constants, respectively; q_e (mg g⁻¹) and q_t (mg g⁻¹) are the adsorption capacities at equilibrium and at any time t, respectively. The kinetic parameters estimated by nonlinear regression are represented in Table 4. Based on the estimated correlation coefficients (R²), the obtained equilibrium data were evidently described better by the pseudo-second order model for As(V). Moreover, the q values (q_e,cal) calculated from the pseudo-second order model were more consistent with the experimental q_e,exp values than those calculated from the pseudo-first order model, also demonstrating that the adsorption process of As(V) can be better fitted with the pseudo-second order model.

Fig. 6 Adsorption of As(V) on MMIM as a function of time (a) and intra-particle diffusion model for As(V) adsorption on MMIM (b) at 25 °C. The initial arsenic concentrations were 5 and 10 mg L⁻¹, the dosage of MMIM and the initial pH values for the solutions were 0.2 g L⁻¹ and 3.0, respectively.

Table 4 Kinetic parameters for As(V) adsorption on the MMIM

Anion species	Initial concentration (mg L^−1)	qe,exp (mg g^−1)	Pseudo-first order			Pseudo-second order
			K1 (h^−1)	qe,cal (mg g^−1)	R^2	K2 (g mg^−1 h^−1)	qe,cal (mg g^−1)	R^2
As(V)	5	17.255	0.1765	10.646	0.9786	0.0495	17.699	0.9991
	10	26.110	0.2283	18.015	0.9878	0.0338	26.774	0.9989

---

Since the above general kinetic models could not identify the rate-limiting step of As(V) diffusion on MMIM, an intra-particle diffusion model based on the theory proposed by Weber and Morris was used to analyze the rate-limiting step of adsorption,⁵⁷ as follows:

q_t = K_it^1/2 + C_i (5)

where K_i (mg g⁻¹ min^−1/2), the intra-particle diffusion rate constant of stage i, is estimated from the slope of the straight line of q_t versus t^1/2. C_i (mg g⁻¹) is the intercept of stage i. According to the model, if the intra-particle diffusion is the rate-limiting step of the entire adsorption process, the q_t versus t^1/2 relationship would be linear and the plot would pass through the origin. Otherwise, some other mechanism along with intra-particle diffusion must also be involved. For the solid–liquid adsorption process, the solute transfer is characterized by external mass transfer and/or intra-particle diffusion, and the relative mechanisms of adsorption included three steps: (i) transport of the solute from bulk solution through liquid film to the adsorbent's exterior surface; (ii) transport of the adsorbate within the pores of the adsorbent; (iii) adsorption of the adsorbate onto the exterior surface of the adsorbent.^58,59 Generally, the last step is the equilibrium reaction and is assumed to always be very rapid, and the resistance is hence assumed to be negligible. It was found that the plots for As(V) presented a multi-linear type from Fig. 6b, and there were three portions with different gradients, suggesting that intra-particle diffusion was not the only rate-limiting step and chemical complex reaction might be involved in the adsorption process. Therefore, the first two steps were limiting throughout the entire adsorption process on MMIM, including the fast instantaneous adsorption of arsenate from aqueous solutions attributed to the external diffusion and gradual adsorption stages, which correspond to the intra-particle diffusion.⁴⁸

3.3 Effect of pH on arsenate removal by MMIM

Solution pH value is an important parameter in the adsorption of cations and anions at the solid–liquid interface,⁶⁰ because hydrogen and hydroxyl ions concentration could strongly modify the redox potential and chemical species of the adsorbate as well as the surface charge of adsorbent.⁶¹ It was found that the solution pH had a pronounced effect on the arsenate removal in aqueous solutions from Fig. 7. The optimum pH was 3.0 for arsenate, and the removal efficiency decreased sharply with increasing solution pH values. This may be caused by zeta potential of the MMIM, which was also played an important role in the arsenate removal. As shown in Fig. 7 inset, the isoelectric point (IEP) of the MMIM was 5.1, and the surface of the adsorbent carried a positive charge when the pH was below 5.0 and a negative charge when the pH was above 5.0. Since the dissociation constants of H₃AsO₄ (pK_a1, pK_a2 and pK_a3) were 2.1, 6.7 and 11.2, respectively, and the As(V) ions existed as different ionic species depending on the aqueous solution pH values. Therefore, acidic solution (pH range from 3.0 to 5.0) favored the removal of As(V) due to electrostatic attraction between the negative As(V) species and the positively charged MMIM. While the pH was above 5.0, the increased electric repulsion between arsenate and MMIM negatively affected As(V) adsorption. Moreover, the competition of arsenate with hydroxyl groups in aqueous solutions also resulted in lower As(V) uptake.⁴⁸ As shown in Fig. 7 inset, the IEP of As-loaded MMIM was shifted to a lower pH, suggesting that arsenate species were adsorbed on the surface of the MMIM. In addition, the concentrations variation of total dissolved iron and manganese ions in MMIM system were also investigated, which is shown in Fig. S6.† The concentrations of total iron and manganese ions were increased with increasing the reaction time, however, both Fe and Mn were only approximately 3.0 mg L⁻¹ even under the acidic concentration of pH 3.0 after 24 h, which was negligible compared with the total adsorbent used (0.2 g L⁻¹). This result suggested that MMIM was stable in acidic aqueous solutions.

Fig. 7 Effect of pH on arsenate removal by MMIM at 25 °C. The initial As(V) concentration was 5 mg L⁻¹ and the dosage was 0.2 g L⁻¹. Inset: zeta potential of MMIM and As-loaded MMIM.

3.4 Regeneration and reusable property of MMIM

The regeneration and reuse of the adsorbent is important in considering the practical applicability. After adsorption of arsenate on the surface of MMIM, the MMIM was regenerated using 1 M NaOH solution, and this adsorption–regeneration cycles were carried out up to five times. Fig. 8 shows the adsorption capacity of the MMIM for As(V) in five consecutive adsorption–regeneration cycles. It was noted that after the fifth regeneration, the reduction in adsorption capacities of MMIM for As(V) was only approximately 9.5%. This obtained result indicated that the MMIM can be regenerated efficiently during the regeneration process, suggesting that the MMIM was desirable for potential in actual application.

Fig. 8 Regeneration reuse of the MMIM. The initial As(V) concentration was 5 mg L⁻¹ and the dosage was 0.2 g L⁻¹.

Based on the above analysis, the adsorption of As(V) on the MMIM in the present study followed a complex mechanism. As(V) removal efficiency by MMIM decreased with increasing solution pH values, indicating that As(V) adsorption by MMIM was not only through ligand exchange under this acidic solution, but also through Coulomb forces. These results suggested that both electrostatic attraction and surface complexation were the major mechanism for the As(V) removal, in addition, the hydroxyl groups on the surface of MMIM also directly exchanged with As(V) to coordinate to structural Fe(III).⁶²

4 Conclusions

A novel magnetic mesoporous iron manganese bimetal oxides (MMIM) was successfully synthesized via the nanocasting route from highly ordered mesoporous silica as a hard template. The obtained MMIM with high specific surface area and pore volume showed high performance for the removal of arsenate from aqueous solutions, and exhibited a sensitive magnetic response with an external magnetic field. The adsorption isotherms were well-described by Freundlich models, and the calculated adsorption capacity of As(V) was 35.35 mg g⁻¹. The adsorption kinetics fitted a pseudo-second order model, and intra-particle diffusion was not the only rate-limiting step in the adsorption process. The removal mechanisms was through a complex mechanism, including electrostatic attraction, surface complexation and As(V) ions directly replacement of hydroxyl groups (M–OH) on the surface of MMIM to form inner-sphere complexes. The regenerated MMIM could be reused for at least five cycles with a slight reduction in the adsorption capacity. It was concluded that this novel MMIM material in the present study is a potential adsorbent for the decontamination of arsenic-polluted water.

Acknowledgements

This work was financially supported, in part, by the National Natural Science Foundation of China (no. 51278356, 41372240), and the National Key Technologies R&D Program of China (no. 2012BAJ25B02).

References

1. V. Chandra, J. Park, Y. Chun, J. W. Lee, I. C. Hwang and K. S. Kim, ACS Nano, 2010, 4, 3979.
2. G. S. Zhang, J. H. Qu, H. J. Liu, R. P. Liu and G. T. Li, Environ. Sci. Technol., 2007, 41, 4613.
3. K. Müller, V. S. T. Ciminelli, M. S. S. Dantas and S. Willscher, Water Res., 2010, 44, 5660.
4. C. Shan and M. Tong, Water Res., 2013, 47, 3411.
5. S. R. Kanel, B. Manning, L. Charlet and H. Choi, Environ. Sci. Technol., 2005, 39, 1291.
6. L. Zeng, Water Res., 2003, 37, 4351.
7. X. Guo and F. Chen, Environ. Sci. Technol., 2005, 39, 6808.
8. A. H. Smith, P. A. Lopipero, M. N. Bates and C. M. Steinmaus, Science, 2002, 296, 2145.
9. H. Guo, Y. Ren, Q. Liu, K. Zhao and Y. Li, Environ. Sci. Technol., 2012, 47, 1009.
10. C. Jing, J. Cui, Y. Huang and A. Li, ACS Appl. Mater. Interfaces, 2012, 4, 714.
11. A. Mudhoo, S. K. Sharma, V. K. Garg and C. H. Tseng, Crit. Rev. Environ. Sci. Technol., 2011, 41, 435.
12. S. Song, A. Lopez-Valdivieso, D. J. Hernandez-Campos, C. Peng, M. G. Monroy-Fernandez and I. Razo-Soto, Water Res., 2006, 40, 364.
13. R. Y. Ning, Desalination, 2002, 143, 237.
14. L. Ma and W. X. Zhang, Environ. Sci. Technol., 2008, 42, 5384.
15. K. Gupta, T. Basu and U. C. Ghosh, J. Chem. Eng. Data, 2009, 54, 2222.
16. Y. Zhang, M. Yang and X. Huang, Chemosphere, 2003, 51, 945.
17. S. X. Zhang, H. Y. Niu, Y. Q. Cai, X. L. Zhao and Y. L. Shi, Chem. Eng. J., 2010, 158, 599.
18. Y. Zhang, M. Yang, X. M. Dou, H. He and D. S. Wang, Environ. Sci. Technol., 2005, 39, 7246.
19. G. S. Zhang, J. H. Qu, H. J. Liu, R. P. Liu and R. C. Wu, Water Res., 2007, 41, 1921.
20. L. Yu, X. J. Peng, F. Ni, J. Li, D. S. Wang and Z. K. Luan, J. Hazard. Mater., 2013, 246, 10.
21. Z. Wu, W. Li, P. A. Webley and D. Zhao, Adv. Mater., 2012, 24, 485.
22. X. Yu, S. Tong, M. Ge, J. Zuo, C. Cao and W. Song, J. Mater. Chem. A, 2013, 1, 959.
23. Y. Wan and D. Zhao, Chem. Rev., 2007, 107, 2821.
24. Z. Wu and D. Zhao, Chem. Commun., 2011, 47, 3332.
25. D. Bruhwiler, Nanoscale, 2010, 2, 887.
26. Y. Zhang, Z. A. Qiao, Y. Li, Y. Liu and Q. Huo, J. Mater. Chem., 2011, 21, 17283.
27. F. Kleitz, S. Hei Choi and R. Ryoo, Chem. Commun., 2003, 2136.
28. W. Teng, Z. Wu, D. Feng, J. Fan, J. Wang, H. Wei, M. Song and D. Zhao, Environ. Sci. Technol., 2013, 47, 8633.
29. H. Tüysüz, C. W. Lehmann, H. Bongard, B. Tesche, R. Schmidt and F. Schüth, J. Am. Chem. Soc., 2008, 130, 11510.
30. L. S. Zhong, J. S. Hu, H. P. Liang, A. M. Cao, W. G. Song and L. J. Wan, Adv. Mater., 2006, 18, 2426.
31. J. C. Groen, L. A. A. Peffer and J. Perez-Ramirez, Microporous Mesoporous Mater., 2003, 60, 1.
32. H. Tüysüz, E. L. Salabaş, E. Bill, H. Bongard, B. Spliethoff, C. W. Lehmann and F. Schüth, Chem. Mater., 2012, 24, 2493.
33. F. Jiao, A. Harrison, J. C. Jumas, A. V. Chadwick, W. Kockelmann and P. G. Bruce, J. Am. Chem. Soc., 2006, 128, 5468.
34. Y. Wang, J. Ren, Y. Wang, F. Zhang, X. Liu, Y. Guo and G. Lu, J. Phys. Chem. C, 2008, 112, 15293.
35. M. Jin, J. N. Park, J. K. Shon, J. H. Kim, Z. Li, Y. K. Park and J. M. Kim, Catal. Today, 2012, 185, 183.
36. F. Jiao, A. H. Hill, A. Harrison, A. Berko, A. V. Chadwick and P. G. Bruce, J. Am. Chem. Soc., 2008, 130, 5262.
37. C. Dickinson, W. Z. Zhou, R. P. Hodgkins, Y. F. Shi, D. Y. Zhao and H. Y. He, Chem. Mater., 2006, 18, 3088.
38. Y. Zhu, W. Zhao, H. Chen and J. Shi, J. Phys. Chem. C, 2007, 111, 5281.
39. J. Ge, Y. Hu, M. Biasini, W. P. Beyermann and Y. Yin, Angew. Chem., Int. Ed., 2007, 46, 4342.
40. F. Mou, J. Guan, H. Ma, L. Xu and W. Shi, ACS Appl. Mater. Interfaces, 2012, 4, 3987.
41. J. Zhou, S. Yang, J. Yu and Z. Shu, J. Hazard. Mater., 2011, 192, 1114.
42. M. V. Ananth, S. Pethkar and K. Dakshinamurthi, J. Power Sources, 1998, 75, 278.
43. S. Wang and C. N. Mulligan, Environ. Int., 2008, 34, 867.
44. C. Vichery, I. Maurin, P. Bonville, J. P. Boilot and T. Gacoin, J. Phys. Chem. C, 2012, 116, 16311.
45. X. Q. Li and W. X. Zhang, J. Phys. Chem. C, 2007, 111, 6939.
46. C. Y. Cao, P. Li, J. Qu, Z. F. Dou, W. S. Yan, J. F. Zhu, Z. Y. Wu and W. G. Song, J. Mater. Chem., 2012, 22, 19898.
47. S. Ouvrard, P. de Donato, M. O. Simonnot, S. Begin, J. Ghanbaja, M. Alnot, Y. B. Duval, F. Lhote, O. Barres and M. Sardin, Geochim. Cosmochim. Acta, 2005, 69, 2715.
48. B. Chen, Z. Zhu, J. Ma, Y. Qiu and J. Chen, J. Mater. Chem. A, 2013, 1, 11355.
49. X. Teng, D. Black, N. J. Watkins, Y. Gao and H. Yang, Nano Lett., 2003, 3, 261.
50. D. Zhang, Z. Liu, S. Han, C. Li, B. Lei, M. P. Stewart, J. M. Tour and C. Zhou, Nano Lett., 2004, 4, 2151.
51. R. Guillet-Nicolas, M. Laprise-Pelletier, M. M. Nair, P. Chevallier, J. Lagueux, Y. Gossuin, S. Laurent, F. Kleitz and M. A. Fortin, Nanoscale, 2013, 5, 11499.
52. D. Banerjee and H. W. Nesbitt, Geochim. Cosmochim. Acta, 1999, 63, 1671.
53. S. Luther, N. Borgfeld, J. Kim and J. G. Parsons, Microchem. J., 2012, 101, 30.
54. B. J. Zhu, X. Y. Yu, Y. Jia, F. M. Peng, B. Sun, M. Y. Zhang, T. Luo, J. H. Liu and X. J. Huang, J. Phys. Chem. C, 2012, 116, 8601.
55. Z. Wen, Y. Zhang and C. Dai, Colloids Surf., A, 2014, 457, 433.
56. G. Zhang, H. Liu, R. Liu and J. Qu, J. Hazard. Mater., 2009, 168, 820.
57. W. Weber and J. Morris, J. Sanit. Eng. Div., Am. Soc. Civ. Eng., 1963, 89, 31.
58. S. Rengaraj, Y. Kim, C. K. Joo and J. Yi, J. Colloid Interface Sci., 2004, 273, 14.
59. M. H. Kalavathy, T. Karthikeyan, S. Rajgopal and L. R. Miranda, J. Colloid Interface Sci., 2005, 292, 354.
60. L. Zeng, X. Li and J. Liu, Water Res., 2004, 38, 1318.
61. C. Escudero, N. Fiol, I. Villaescusa and J. C. Bollinger, J. Hazard. Mater., 2009, 164, 533.
62. A. U. Rajapaksha, M. Vithanage, L. Jayarathna and C. K. Kumara, Appl. Geochem., 2011, 26, 648.

---

Footnote

† Electronic supplementary information (ESI) available. See DOI: 10.1039/c4ra09746g

---