Well-dispersed magnetic iron oxide nanocrystals on sepiolite nanofibers for arsenic removal

Abstract

A novel nanostructure composed of magnetic iron oxide nanocrystals (MI) anchored on a sepiolite nanofiber backbone with excellent arsenic adsorption performance has been successfully developed. Sepiolites (SEPs) as typical nano-geomaterials with low cost, large specific surface (ca. 300 m² g⁻¹) and tunable surface chemistry are chosen as the host matrix. Transmission electron microscopy confirms that uniform Fe₂O₃ nanocrystals with an average particle size of ∼9 nm are spatially well-dispersed and anchored on the sepiolite backbone at a high Fe₂O₃ content of 33.2 wt%, rather than forming aggregates on the external surface. MI/SEPs have a high specific surface area, high loading amount, the non-aggregated nature of Fe₂O₃ nanocrystals, good dispersion and magnetic properties, making them promising for use as a separable adsorbent for As(III) removal with high adsorption capacity and magnetic separation properties. The maximum adsorption capacity has a wonderful value of 50.35 mg g⁻¹ for As(III) on the MI/SEPs, which is higher than those of previously reported adsorbents. Moreover, MI/SEPs can reduce the concentration of As(III) from 140 to 1.5 μg L⁻¹. MI/SEPs also show high removal ratios of 96.4% without any pre-treatment in real groundwater with an arsenic concentration of 456.5 μg L⁻¹.

---

Introduction

Sepiolites as fibrous clay minerals with interpenetrated and regular mesopore systems have recently triggered enormous research activities because of their fascinating features such as large specific surface area, different pore sizes (micro- and mesopores), and surface free silanol groups (SiOH).^1,2 While they possess great potential in adsorption and separation, and energy storage and conversion, it is more intriguing to explore the possibility of enhancing and/or extending their properties by the formation of nanocomposites so that targeted applications can be designed based on synergic and cooperative effects between sepiolite nanofibers and the well-dispersed active nanocrystals.^3,4 Nanocrystals with a wide range of compositions can be readily introduced into sepiolite nanofibers through either a direct synthesis approach or a post loading method.^5,6 Normally, the nanocrystals can be well dispersed and/or confined in the case of a low metal content. However, at a high metal loading level, the metal species normally aggregate severely into large particles. It is extremely challenging to fix highly concentrated (e.g., >20 wt%) and uniformly dispersed nanocrystals on the sepiolite backbone without aggregating. For example, magnetic iron oxide/clays hybrid materials have been described in the literature, including both magnetite (Fe₃O₄) and maghemite (γ-Fe₂O₃).^7–9 However, most reports show randomly dispersed or even aggregated large nanoparticles.^10–13 Such disadvantages are largely a result of a lack of an efficient control to avoid substantial diffusion/aggregation of the metal precursors during their conversion into oxides, which is especially challenging in the case of high loading levels. Moreover, little attention has been paid to simultaneously exploring and utilizing the sepiolite nanofibers and the confined populated nanoparticles as active components for adsorption and separation. Therefore, it is fairly demanding to synthesize fibrous sepiolite-based nanocomposites that encapsulate high-content but uniformly dispersed and spatially separated nanoparticles while retaining an open mesopore system, because such features are highly desirable for adsorption which depends on molecular diffusion and transportation for their effectiveness.

Arsenic, which is toxic to man and other living organisms, presents potentially serious environmental problems throughout the world.^14–17 The US Environmental Protection Agency (EPA) has set the arsenic standard for drinking water at 10 parts per billion to protect consumers served by public water systems from the effects of long-term chronic exposure to arsenic.¹⁸ In natural water, arsenic is primarily present in inorganic forms and exists in two predominant species: arsenate (As(V)) and arsenite (As(III)). As(V) is the predominant species form in well-oxidized waters and As(III) predominates in reduced environments such as groundwater.^14,19 Among these, As(III) is much more toxic, soluble, and mobile than As(V). As(III) is generally accepted that it is more difficult to be removed due to its low affinity to various adsorbents.^20,21 To achieve effective As(III) removal, it is required to pre-treat As(III) by oxidizing it to As(V) and/or adjust the pH value before coagulation–precipitation/adsorption processes.^20,22,23 Apparently, the pre-treatment increases the operation cost and causes secondary pollution problems, thus it is disadvantageous for practical applications. It is highly desirable to develop an adsorbent for efficient and cost-effective removal of both As(V) and As(III) without any pre-treatment. Previous studies show that iron oxides (magnetite and maghemite) with sizes of about 12 and 3.8 nm exhibit significantly increased arsenic adsorption capacities compared with iron oxides with large particle sizes, which may result from more adsorption sites being exposed to arsenic species.^24,25 However, nanoparticles tend to aggregate into large particles easily, leading to deteriorated adsorption performance. Besides, it is difficult to practically apply nanoparticles in the wastewater treatment because small particles may cause difficulties in separation and/or diffusion. To overcome this difficulty, several researchers have combined iron oxides with carbon, carbon nanotubes (CNTs), macroporous siliceous foams (MOSF) and graphene-based materials.^26–29 However, the existing synthesis methods are expensive, complex, time-consuming, and environmentally unfriendly, leading to a low yield and limited practical applications. In addition, there are serious concerns about the health and environmental risks of iron oxide/CNTs/or MOSF/or graphene, once they have been released into the environment.

Herein, we report a pre-carbonization post-synthetic route for the construction of natural fibrous sepiolite loading magnetic iron oxide nanocrystals with high content (ca. 33 wt%) homogeneously dispersed and anchored on sepiolite nanofiber surface. This route is demonstrated to be a convenient, eco-friendly, and efficient approach for the high loading of well-dispersed and uniform nanocrystals. The synergic effects that combine sepiolite nanofiber with the uniformly dispersed iron oxides of high content make the materials ideal for arsenic removal with excellent adsorption and magnetic separation properties. The uptake capacities of As(III) are 50.35 mg g⁻¹. Moreover, the MI/SEPs can reduce the concentration of As(III) from 140 to 1.5 μg L⁻¹, far below the standard of 10 μg L⁻¹ suggested by WHO. The MI/SEPs were also applied in groundwater treatment, showing promising potential in practical applications for arsenic removal.

Experimental section

Materials

Sepiolites (SEPs) were purchased from Sigma. Other chemicals were purchased from Sinopharm Chemical Reagent Co., Ltd of China. All the chemicals were used as received without further purification. Deionized water was used throughout the experiments. The real groundwater samples were collected in the Jianghan Plain, Hubei.

Characterization

The morphology of as-prepared MI/SEPs composites was observed using a transmission electron microscope (Philips CM12 TEM/STEM) with the voltage of 120 kV. The composition and phase purity of the as-synthesized samples were analyzed by powder X-ray diffraction (XRD) with monochromatized Cu Kα (λ = 1.5406 Å) incident radiation by a Bruker AXS D8-Focus diffractometer operated at 40 kV voltage and 40 mA current with a scanning step of 0.05 second. Scanning electron microscopy (SEM) images of the product were taken on a Quanta200 scanning electron microscope operated at 20 kV equipped with an attached Oxford Link ISIS energy-dispersive X-ray spectroscopy (EDS). To test nitrogen adsorption–desorption curve and pore distribution, the samples must be in a 120 °C vacuum dried more than 12 h, and then measured on the Micromeritics ASAP 2020 type apparatus at 77 K. The FT-IR spectra were collected with an EQUNOX 55 instrument by diffuse reflectance scanning disc technique from 4000 to 400 cm⁻¹ at room temperature. Magnetic measurements were measured using a superconducting quantum interference device magnetometer (MPMS-XL5, Quantum Design). X-ray photoelectron spectroscopy (XPS) was obtained with a MULTILAB2000 electron spectrometer from VG Scientific using 300 W Al Kα radiation. The UV-vis spectra were measured on a Perkin Elmer Lambda 35 spectrophotometer. All samples were dried before characterization by an oven to remove the adsorbed water.

Synthesis of the SEPs@C nanofibers

In a typical procedure, 0.75 g SEPs was dispersed in 0.35 M glucose solution (57 mL) under vigorous magnetic stirring to form a homogeneous dispersion system at room temperature. Then, the mixed solution was put into a Teflon-lined stainless steel autoclave (80 mL). Finally, the container was closed and maintained at 160 °C for 48 h. After that, the autoclave was cooled to room temperature naturally. Then, the as-prepared nanofibers were obtained by filtering and rinsing with distilled water and anhydrous ethanol. Finally, the powders were dried in vacuum drying oven. The sample was denoted as SEPs@C.

Synthesis of the MI/SEPs

The synthetic process of magnetic MI/SEPs nanofibers follows a post-synthetic method by thermal decomposition of organic iron precursors shown in Scheme 1(I). In a typical synthesis, 0.4 g iron(III) acetylacetonate [Fe(acac)₃] was dissolved in 60 mL triethylene glycol (TEG) by ultrasonication and magnetic stirring. Then, 0.2 g SEPs@C was dispersed in the above solution by vigorous magnetic stirring. The suspension was transferred into a Teflon-lined stainless steel autoclave (80 mL in total volume), sealed, and maintained at 220 °C for 12 h. After being cooled to room temperature naturally, the products were diluted by absolute ethanol and washed several times with absolute ethanol (denoted MI/SEPs-N). For the preparation of the MI/SEPs, the as-prepared MI/SEPs-N was annealed at 250 or 500 °C for 4 h in air. The samples were denoted as MI/SEPs-250 and MI/SEPs-500, respectively.

Scheme 1 Schematic illustrations of the synthesis of the MI/SEPs (I) and arsenic removal by MI/SEPs (II).

Arsenic adsorption

Arsenic trioxide (As₂O₃, FW 197.84) was used as the source of As(III). The pH values were adjusted with HCl or NaOH solution using a pH meter (PXSJ-226, Shanghai Leici Instrument Factory, China). The concentration of the arsenic in the solution was determined by hydride generation atom fluorescence spectrometry (AFS-230E, China) immediately after the separation of the formed solid precipitate from the solution. The limit of detection is 10 μg L⁻¹. The quantity of toxic ions adsorbed by per unit mass of the sample, q_e (mg g⁻¹), is calculated as follow:

q_e = (C_i − C_e)M/V

where C_i is the initial concentration of ions (mg L⁻¹), C_e is the equilibrium concentration of the ions (mg L⁻¹), M is the mass of the sample (g), and V is the volume of the solution (L).

Batch adsorption of As(III) was carried out at 298 K with an adsorbent loading of 0.2 g L⁻¹, and all experiments were run at pH 7.0. Typically, the adsorption isotherms were studied at variable initial As(III) concentrations (0.1–50 mg L⁻¹), followed by shaking for 24 h to achieve equilibrium. The kinetic studies were conducted at different intervals in 1 mg L⁻¹ of arsenic solution. The effect of anions on arsenic removal was tested in a solution of 1.0 mg L⁻¹ arsenic and 50 mg L⁻¹ anions. After adsorption and magnetic separation, the residual arsenic concentrations were measured so that adsorption curves could be obtained.

Results and discussion

Structural and morphological study of the MI/SEPs

The synthesis process of MI/SEPs is displayed in Scheme 1. Sepiolite [Si₁₂O₃₀Mg₈(OH)₄(H₂O)₄·8H₂O] is a nanofibrous magnesium silicate, with an ability to adsorb large quantities of water and polar organic molecules. Representative TEM images in Fig. 1a showed the pristine sepiolite fibers having needle morphology of 30–50 nm in diameter and 1–3 μm in length. These fibers formed bundles-like aggregates by surface interaction between individual needle-type particles. Large fibers were observed but they were formed by connected fibers as confirmed by detailed observation of dispersed single sepiolite fibers. A simply pre-carbonization treatment of sepiolites was performed, so as to introduce functional groups (e.g., hydroxyl, carbonyl, and carboxyl), which not only facilitate the dispersion of sepiolites but also serve as nucleation centers for the deposition of iron oxide nanocrystals (Fig. 1b). Next, through a facile high-temperature decomposition of the precursor iron(III) acetylacetonate and SEPs@C in liquid polyols, nearly monodisperse magnetite nanocrystals with high density were in situ anchored on the surface of sepiolite nanofibers (Fig. 1c). Subsequently, the nanocomposites were annealed as shown in Fig. 1d.

Fig. 1 TEM images of the pristine SEPs (a), SEPs@C (b), MI/SEPs-N (c), and MI/SEPs (d).

The corresponding transmission electron microscopy (TEM) images (Fig. 2a and b) of the MI/SEPs show that almost all the nanocrystals are selectively deposited on sepiolite nanofiber surface, but still spatially separated. The corresponding size-distribution histograms of iron oxide nanocrystals counted in Fig. 2c and d shows a narrow distribution of the nanocrystals from 5–15 nm with an average size of ∼9 nm. Elemental maps of MI/SEPs elucidate the uniform distribution of Fe₂O₃ nanocrystals along the sepiolite nanofibers at micron scale (Fig. S1 in ESI†). Fe and O elements are uniformly distributed and well correlated with the shape of the sample area.

Fig. 2 TEM images and the corresponding size-distribution histograms of MI/SEPs-250 (a and b) and MI/SEPs-500 (c and d).

The crystal structure of the SEPs and MI/SEPs were performed by X-ray diffraction (XRD) measurements and are shown in Fig. 3. For the pristine sepiolites, the characteristic diffraction peak at 7.288° corresponding to a d-spacing of 1.22 nm is assigned to (110) plane of SEPs. After the surface modified, the characteristic peak of SEPs remained in MI/SEPs and a number of new peaks appeared. It can be shown that the predominant phase of iron oxide is the maghemite (γ-Fe₂O₃) in the as-prepared composite, and no peaks assigned to impurities are observed. As increasing the annealing temperature, the more stable hematite (α-Fe₂O₃) begins to appear in the MI/SEPs-500. The different phases of the composites greatly impact the arsenic adsorption performance.

Fig. 3 XRD spectra of the pristine SEPs, MI/SEPs-250, and MI/SEPs-500.

The successful anchoring of well-dispersed iron oxide nanocrystals on the SEPs was further confirmed by X-ray photoelectron spectroscopy (XPS) analysis. The survey spectra of MI/SEPs (Fig. 4a) show the presence of O and Fe, in accordance with the results of elemental maps. Two photoelectron peaks located at 710.4 and 724.0 eV are found in the Fe 2p spectra (Fig. 4b), which can be assigned to the Fe 2p_3/2 and Fe 2p_1/2 of γ-Fe₂O₃, respectively. Besides, a satellite peak located at 719.0 eV, which is the characteristic peak of γ-Fe₂O₃, can also be seen, indicating the well-dispersed iron oxide nanoparticles in the MI/SEPs-250 are γ-Fe₂O₃ rather than Fe₃O₄.³⁰ For the MI/SEPs-500, the satellite peak moved to 719.2 eV, due to the presence of α-Fe₂O₃.³¹ The XPS patterns are well in agreement with the XRD data and reveal that phase transformation could be achieved. The O 1s spectrum of MI/SEPs-500 (Fig. 4c) can be well-fitted to tow peaks at 530.2 and 532.6 eV, which are attributed to the binding energies of oxygen atoms bonded to Fe and Si, respectively. The binding energy of Si 2p is found to be 98.9 and 102.6 eV (Fig. 4d), in agreement with the binding energy values of Si–OH and Si–O.

Fig. 4 XPS survey scans of the pristine SEPs and MI/SEPs (a), the corresponding deconvoluted Fe 2p spectra of MI/SEPs-250 and MI/SEPs-500 (b), O 1s (c) and Si 2p (d) spectra of MI/SEPs-500.

FT-IR spectra of the pristine sepiolite and MI/SEPs are shown in Fig. S2.† The major band positions don't change after modification, suggesting that the basic crystal structures of sepiolites remain constant. However, for the MI/SEPs-250, a new absorption peak has appeared at 560 cm⁻¹ assigned to the Fe–O stretching vibrations of Fe₂O₃. The enhanced intensity for Fe–O is indicative of high iron loading in MI/SEPs-500. Besides, the FT-IR spectra indicate a strong decrease in intensity of the stretching O–H bands of Si–OH groups (at ca. 3720 cm⁻¹) covering the sepiolite surface. This behavior suggests the involvement of these surface groups in the interaction with the iron oxide nanocrystals, in a similar way to that observed in magnetite or anatase nanoparticles anchored on sepiolite surface.^5,8,32

The specific surface area (SSA) and pore-size distribution of MI/SEPs were performed by nitrogen adsorption–desorption isotherms with the density functional theory (DFT) methods, as shown in Fig. 5. The aspect of the N₂ isotherms of MI/SEPs samples is quite similar to that of pristine sepiolite. The SSA of SEPs, MI/SEPs-250, and MI/SEPs-500 are 297.19, 125.75, and 104.96 m² g⁻¹, respectively. The SSA of MI/SEPs is drastically decreased by 2–3 times than that of SEPs, such decreases correspond to an increase in the mean pore diameter from 6.19 to 10.65 nm. The coverage of the sepiolite surface by Fe₂O₃ nanocrystals produces a reduction of the specific surface area due to the micropores blockage, avoiding the accessibility of nitrogen molecules during the corresponding measurements. Typical IV isotherms with H3-type hysteresis loops (P/P₀ > 0.5) are observed for the MI/SEPs, indicating the presence of abundant mesoporous structure. The detailed features of the pore volume analysis are presented in Table S1 (ESI†).

Fig. 5 N₂ adsorption–desorption isotherms and corresponding pore size distribution curves (inset) of SEPs, MI/SEPs-250, and MI/SEPs-500.

Magnetization properties of MI/SEPs were investigated by superconducting quantum interference device (SQUID). Fig. 6a shows the temperature dependence of magnetization for the MI/SEPs. The curves are acquired between 5 and 300 K using zero-field-cooling (ZFC) and field-cooling (FC) procedures under an applied magnetic field of 100 Oe. The blocking temperature of the MI/SEP-250 and MI/SEP-500 is found to be about 200 K and 175 K, respectively, which is similar with the blocking temperature of the γ-Fe₂O₃ in previous study.³³ The saturation magnetization (M_s) of MI/SEPs-250 and MI/SEPs-500 are 31.95, and 29.53 emu g⁻¹, respectively, indicating that MI/SEPs have high magnetism which can be beneficial for reuse by magnetic separation. Fig. 6b displays the loops of MI/SEPs which exhibit very low remanence and coercivity.

Fig. 6 ZFC and FC magnetization curves versus temperature plot of MI/SEPs under an applied magnetic field of 100 Oe (a). Magnetic hysteresis curves of MI/SEPs at 300 K (b), and MI/SEPs dispersed water solution and magnetic separation (inset).

The parameters are summarized in Table S2.† The magnetic intensities are lower than bulk γ-Fe₂O₃ due to the presence of SEPs and the small size of γ-Fe₂O₃ nanocrystals. The MI/SEPs composites exhibit a superparamagnetic state with small remnant magnetization and coercivity at room temperature which is desirable for many practical applications, so that strong magnetic signals at small magnetic fields are obtained. Moreover, it can be found from inset of Fig. 6b that MI/SEPs could be well dispersed in water, which is beneficial for adsorption of water-soluble pollutants. The arsenic-loaded MI/SEPs can be easily separated by using a magnet, suggesting that MI/SEPs can be potentially used as a separable adsorbent. After magnetic separation, the concentration of residual MI/SEPs in the treated samples has been estimated using the UV-vis absorption spectra.¹⁸ It is noteworthy to observe that there is almost none of residual MI/SEPs, as shown in Fig. S3.†

Arsenic adsorption properties

The MI/SEPs can be adopted ideal materials in the removal of arsenic from groundwater. To investigate the adsorption property for As(III), adsorption isotherms have been obtained with different initial concentrations ranging from 0.1 to 50 mg L⁻¹. The data of As(III) adsorption were fitted with Freundlich and Langmuir isotherm models as shown in Fig. 7, and the calculated isotherm parameters are summarized in Table S3.† Based on the determination coefficient (R²), the Langmuir model fits the experimental data better than the Freundlich model. The maximum adsorption capacities of As(III) are 35.15 mg g⁻¹ for MI/SEPs-250 and 50.35 mg g⁻¹ for MI/SEPs-500, respectively. These values are much higher than those of previously reported adsorbents which are showed in Table S4.†

Fig. 7 Adsorption isotherms of arsenic on MI/SEPs at 298 K. The initial concentration ranged from 0.1 to 50 mg L⁻¹, the dosage of adsorbents was 0.2 g L⁻¹, and the initial pH values for the solutions were 7.0.

In order to further research the applicability of the MI/SEPs and exhibit these excellent performances in groundwater treatment, the adsorption rate of As(III) ions with an initial concentration of 1 mg L⁻¹ on the MI/SEPs have been specially studied at 298 K (Fig. 8). The adsorption process is rapid during the initial 60 min, which may be due to the large number of available sites in the initial stage. Thereafter the adsorption efficiency of As(III) increases very slowly with increasing adsorptive time and finally the adsorption reached equilibrium after 1 and 3 h for MI/SEPs-250 and MI/SEPs-500, respectively. The pseudo-first order and pseudo-second order kinetic models were applied to further investigate the adsorption performances of As(III) on MI/SEPs, as shown in Fig. S4.† The calculated kinetic parameters are presented in Table S5.† According to the correlation coefficient (R²), the adsorption processes of As(III) on MI/SEPs were better described by the pseudo-second order kinetic model. These results indicate that the as-obtained MI/SEPs have large adsorption capacities for As(III) and can be potentially used as an excellent adsorbent for As(III) in groundwater treatment.

Fig. 8 Adsorption of arsenic on MI/SEPs as a function of time and pseudo-second order curves. The initial concentration was 1 mg L⁻¹, the dosage of adsorbents was 0.2 g L⁻¹, and the initial pH values for the solutions were 7.0.

In nature, more than one anion will normally be competing for the adsorption sites. Thus it makes sense to investigate the competitive adsorption between different anions as well as single anion adsorption. In addition, investigation of the competition between the anions can provide insight into the reactions occurring on the surface. The effects of fluoride, phosphate, bicarbonate, nitrate, sulfate, carbonate and chloride on the sorption of arsenic on the adsorbent have been studied (Fig. 9). The obtained results show that the tested anions have no adverse influence on the As(III) removal excepting fluoride and phosphate. The adsorbents have certain adsorption for fluoride in the coexistence of arsenic and fluoride, which is in agreement with the adsorption results (Fig. S5†). The maximum adsorption capacities of the MI/SEPs-250 composites are 35.15 mg g⁻¹ for As(III) and 45.48 mg g⁻¹ for fluoride, respectively. Because of the similar chemical structure and dissolution constants, the competitive adsorption occurs on the iron oxide surface.³⁴ There are sites on the surface that were specific for each ion as well as some nonspecific sites on which both ions could be adsorbed.

Fig. 9 The effect of coexisting anions on arsenic adsorption of MI/SEPs. The initial concentration of arsenic and anions was 1.0 and 50 mg L⁻¹, respectively. The dosage of adsorbents was 0.2 g L⁻¹, and the initial pH values for the solutions were 7.0.

MI/SEPs exhibit an excellent adsorption performance not only in the relatively high arsenic concentration range but also in the low concentration range. As shown in Fig. 10a, with the initial arsenic concentration of 140 or 60 μg L⁻¹, the equilibrium concentration of As(III) can be both reduced to lower than 5 μg L⁻¹ even by MI/SEPs-250 with the lower adsorption capacity, which could reach the drinking water standard value or even less. It is suggested that the good arsenic adsorption ability of MI/SEPs in the low concentration range. Considering the practical applicability of the adsorbent for removal of As(III) under actual groundwater, the adsorption experiments were performed in the real groundwater sample from Jianghan Plain with the initial arsenic concentration of 456.5 μg L⁻¹. Although large amounts of other species exist in the groundwater (Table S6†), MI/SEPs still shows high removal ratios of 96.4% which could reach the drinking water standard value or less (Fig. 10b). The MI/SEPs after the adsorption test can be regenerated by stirring with aqueous sodium hydroxide solution at 25 °C for 4 h. In the second cycle, the removal efficiency of regenerated MI/SEPs is 93.8% (Fig. S6†). These results further confirm applicability of MI/SEPs composites in arsenic removal from real groundwater without any pre-treatment.

Fig. 10 Arsenic(III) adsorption of MI/SEPs-500 with the initial concentration of 140 and 60 μg L⁻¹ (a), and arsenic uptake of MI/SEPs in the real groundwater (Jianghan Plain) (b).

XPS was used to confirm the adsorption of arsenic by MI/SEPs-500. The survey scan of core level binding energy of MI/SEPs-500 before and after arsenic adsorption is shown in Fig. 11a. Four prominent peaks corresponding to As 3d, As 3p, As 3s, and As (A) can be clearly observed at low binding energy, indicating the presence of As–O.^9,18,35 A more detailed electronic structure of these arsenic complexes can be found in high resolution As 3d deconvoluted spectra (Fig. 11b), which show three prominent peaks at 43.3, 45.1, and 46.4 eV associated with As–OH bonds, As₂O₃, and As₂O₅, respectively. The As(V) peak was detectable, indicating that partially oxidation of As(III) occurred on the surface of the MI/SEPs-500. The O 1s spectrum of arsenic adsorbed MI/SEPs-500 plotted in Fig. 11c, can be well-fitted to three peaks at 530.0, 530.4, and 532.6 eV, which are attributes to the binding energies of oxygen atoms bonded to As, Fe, and Si, respectively. It is clear that the intensity of Fe–OH became decline and it plays a key role in arsenic adsorption. According to the reported studies, the As(III) species have a high affinity for the iron oxide surface, which forms inner-sphere bidentate, binuclear As(III)–Fe(III) complexes.^36–38 On the basis of all above results, the As(III) species were directly adsorbed on surface of the MI/SEPs by the formation of surface complexes such as monodentate, bidentate-binuclear, and bidentate-mononuclear complexes. And a little portion of As(III) was subsequently oxidized to As(V) possibly by air. The proposed mechanism for arsenic removal by MI/SEPs was developed in Fig. 12.

Fig. 11 XPS spectra at low binding energy of MI/SEPs-500 before and after arsenic adsorption (a), and the deconvoluted O 1s (b) and As 3d (c) spectra of As(III)-loaded MI/SEPs-500.

Fig. 12 Schematic diagram for the proposed mechanism of arsenic adsorption on MI/SEPs.

This high performance of arsenic absorption of MI/SEPs nanostructures originates directly from the fascinating features such as high porosity and open pore network of the sepiolites which facilitate fast arsenic diffusion and transportation as well as promote the accessibility of iron oxides. Additionally, the high-content, uniformly-dispersed, and spatially separated iron oxide nanocrystals that are decorated on SEPs fibers will provide increased active sites for arsenic capture.

Conclusions

In summary, magnetic Fe₂O₃/sepiolite composites with novel nanoarchitectures and excellent performance for arsenic capture have been rationally designed and fabricated by using a pre-carbonization post-synthetic route. This pre-carbonization process is crucial to avoid the problem of aggregation. In this way, highly concentrated crystalline iron oxides (up to 33.2 wt%) can be selectively loaded on sepiolite nanofiber surface. SEM elemental mapping technique demonstrates that Fe₂O₃ nanocrystals are spatially well-dispersed on sepiolite nanofiber surface without blocking the open pore network of sepiolite, which gives rise to increased adsorption sites and subsequent high arsenic adsorption capacities. In addition to high uptake performance, arsenite(III) can be removed without any pre-treatment. Moreover, the residual arsenic concentration can be successfully reduced to 1.5 μg L⁻¹. The composite also demonstrates excellent performance in groundwater treatment. Finally, the synthesis method is quite facile controllable and versatile for selectively and exclusively loading a series of other nanoparticulate metal oxides on the sepiolite nanofiber surface. These materials will be found many applications in sorption, catalysis, and energy storage and conversion.

Acknowledgements

This work was supported by the National Basic Research Program of China (973 Program, Grant no. 2011CB933700) of the Ministry of Science and Technology of China, the National Natural Science Foundation of China (Grant no. 51371162 & 51344007) and the Fundamental Research Funds for the Central Universities.

Notes and references

1. A. Gómez-Avilés, M. Darder, P. Aranda and E. Ruiz-Hitzky, Angew. Chem., Int. Ed., 2007, 46, 923.
2. E. Ruiz-Hitzky, M. Darder and P. Aranda, J. Mater. Chem., 2005, 15, 3650.
3. E. Ruiz-Hitzky, M. Darder, F. M. Fernandes, E. Zatile, P. Aranda and F. J. Palomares, Adv. Mater., 2011, 23, 5250.
4. R. Fernández-Saavedra, P. Aranda and E. Ruiz-Hitzky, Adv. Funct. Mater., 2004, 14, 77.
5. P. Aranda, R. Kun, M. A. Martín-Luengo, S. Letaïef, I. Dékány and E. Ruiz-Hitzky, Chem. Mater., 2008, 20, 84.
6. A. Esteban-Cubillo, J. F. Marco, J. S. Moya and C. Pecharromán, J. Phys. Chem. C, 2005, 15, 3650.
7. G. H. Zhao, J. Z. Wang, Y. F. Li, X. Chen and Y. P. Liu, J. Phys. Chem. C, 2011, 115, 6350.
8. Y. G. Alfaro, P. Aranda, F. M. Fernandes, B. Wicklein, M. Darder and E. Ruiz-Hitzky, Adv. Mater., 2011, 23, 5224.
9. J. Yang, H. W. Zhang, M. H. Yu, I. Emmanuelawati, J. Zou, Z. G. Yuan and C. Z. Yu, Adv. Funct. Mater., 2014, 24, 1354.
10. J. Liu, S. Z. Qiao, Q. H. Hu and G. Q. Lu, Small, 2011, 7, 425.
11. A. H. Lu, J. J. Nitz, M. Comotti, C. Weidenthaler, K. Schlichte, C. W. Lehmann, O. Terasaki and F. Schuth, J. Am. Chem. Soc., 2010, 132, 14152.
12. Y. H. Deng, Y. Cai, Z. K. Sun, D. Gu, J. Wei, W. Li, X. H. Guo, J. P. Yang and D. Y. Zhao, Adv. Funct. Mater., 2010, 20, 3658.
13. X. P. Dong, H. R. Chen, W. R. Zhao, X. Li and J. L. Shi, Chem. Mater., 2007, 19, 3484.
14. W. R. Cullen and K. J. Reimer, Chem. Rev., 1989, 89, 713.
15. I. Ali, Chem. Rev., 2012, 112, 5073.
16. D. K. Nordstrom, Science, 2002, 296, 2143.
17. S. Fendorf, H. A. Michael and A. van Geen, Science, 2010, 328, 1123.
18. J. Ma, Z. L. Zhu, B. Chen, M. X. Yang, H. M. Zhou, C. Li, F. Yu and J. H. Chen, J. Mater. Chem. A, 2013, 1, 4662.
19. D. Mohan and C. U. Pittman Jr, J. Hazard. Mater., 2007, 142, 1.
20. B. A. Manning and S. Goldberg, Environ. Sci. Technol., 1997, 31, 2005.
21. J. F. Ferguson and J. Gavis, Water Res., 1972, 6, 1259.
22. M. J. Scott and J. J. Morgan, Environ. Sci. Technol., 1995, 29, 1898.
23. G. S. Zhang, J. H. Qu, H. J. Liu, R. P. Liu and G. T. Li, Environ. Sci. Technol., 2007, 41, 4613.
24. C. T. Yavuz, J. T. Mayo, W. W. Yu, A. Prakash, J. C. Falkner, S. Yean, L. L. Cong, H. J. Shipley, A. Kan, M. Tomson, D. Natelson and V. L. Colvin, Science, 2006, 314, 964.
25. T. Tuutijarvi, J. Lu, M. Sillanpaa and G. Chen, J. Hazard. Mater., 2009, 166, 1415.
26. Z. X. Wu, W. Li, P. A. Webley and D. Y. Zhao, Adv. Mater., 2012, 24, 485.
27. X. L. Yu, S. R. Tong, M. F. Ge, J. C. Zuo, C. Y. Cao and W. G. Song, J. Mater. Chem. A, 2013, 1, 959.
28. V. Chandra, J. Park, Y. Chun, J. W. Lee, I. C. Hwang and K. S. Kim, ACS Nano, 2010, 4, 3979.
29. S. J. Zhang, X. Y. Li and J. P. Chen, Carbon, 2010, 48, 60.
30. J. Lu, X. L. Jiao, D. R. Chen and W. Li, J. Phys. Chem. C, 2009, 113, 4012.
31. G. B. Sun, B. X. Dong, M. H. Cao, B. Q. Wei and C. W. Hu, Chem. Mater., 2011, 23, 1587.
32. E. R. Hitzky, J. Mater. Chem., 2001, 11, 86.
33. Q. Han, Z. H. Liu, Y. Y. Xu, Z. Y. Chen, T. M. Wang and H. Zhang, J. Phys. Chem. C, 2007, 111, 5034.
34. H. S. Zhao and R. Stanforth, Environ. Sci. Technol., 2001, 35, 4753.
35. S. Vadahanambi, S. H. Lee, W. J. Kim and I. K. Oh, Environ. Sci. Technol., 2013, 47, 10510.
36. S. Fendorf, M. J. Eick, P. Grossl and D. L. Sparks, Environ. Sci. Technol., 1997, 31, 315.
37. B. A. Manning, S. Fendorf and S. Goldberg, Environ. Sci. Technol., 1998, 32, 2383.
38. Y. H. Wang, G. Morin, G. O. Nguema, F. Juillot, F. Guyot, G. Calas and G. E. Brown, Environ. Sci. Technol., 2010, 44, 109.

---

Footnote

† Electronic supplementary information (ESI) available: The further characterizations of MI/SEPs samples such as SEM, Mapping, FT-IR, and adsorption data. See DOI: 10.1039/c5ra01592h

---