Ion-exchange bonded H₂Ti₃O₇ nanosheets-based magnetic nanocomposite for dye removal via adsorption and its regeneration via synergistic activation of persulfate

Abstract

Magnetic nanocomposites (HTNSF) consisting of hydrogen titanate (H₂Ti₃O₇) nanosheets (HTNS) and maghemite (γ-Fe₂O₃) nanoparticles with varying weight-fraction (5–25 wt%) of the latter have been successfully synthesized by simple mechanical mixing of precursors in an aqueous solution having neutral solution-pH. A new model has been proposed to explain the typical attachment of γ-Fe₂O₃ nanoparticles to the edges of HTNS via an ion-exchange bond formation. The dye-adsorption properties of HTNSF magnetic nanocomposites have been investigated using the cationic methylene blue (MB) dye as a target pollutant. The new model satisfactorily explains a strong dependence of positive deviation (relative to the variation governed by the law-of-mixture) observed in the variation of dye-adsorption capacity on the similar variation observed in the pore volume of HTNSF magnetic nanocomposite as a function of weight-fraction of γ-Fe₂O₃ nanoparticles. The maximum MB adsorption capacity of 76 mg g⁻¹ is exhibited by HTNSF-10 sample which is higher than that (67 mg g⁻¹) of HTNS sample. The MB adsorption on the surface of HTNSF magnetic nanocomposite follows the pseudo-second-order kinetics and Langmuir and Dubinin–Kaganer–Radushkevich (DKR) isotherm models. The variation in the regression correlation coefficient (〈r²〉) values as a function of initial MB concentration strongly supports the Azizian analysis. The HTNSF-5 magnetic nanocomposite which contains the lowest weight-fraction (5 wt%) of γ-Fe₂O₃ nanoparticles shows the effective magnetic separation from the aqueous solution in 5 min. The reuse of HTNSF-5 magnetic nanocomposite has been successfully achieved through its regeneration via activation of persulfate (S₂O₈²⁻) in an aqueous solution involving the synergy effect in between thermal activation and that by the constituents of HTNSF-5 magnetic nanocomposite.

---

1. Introduction

Two dimensional hydrogen titanate (H₂Ti₃O₇) nanosheets (HTNS) are of major significance as a result of their novel physicochemical properties, such as ultrathin thickness, large specific surface-area, and quantum confinement effect resulting in their suitability in diverse applications including photonic structures, lithium ion batteries, photocatalysis, and catalysis.^1–13 Different research groups have synthesized HTNS with different chemical and crystal structures intended for different applications. Gateshki et al.¹ first reported the preparation of hydrogen titanate (H_0.67Ti_1.83O₄·H₂O) nanosheets by exfoliation of Cs_0.67Ti_1.83O₄·H₂O through intercalation of tetrabutylammonium (TBA). Wu and Xu et al.² synthesized hydrogen titanate nanosheets via hydrothermal treatment of anatase-titania (TiO₂) under highly alkaline conditions followed by washing of hydrothermal product using hydrochloric acid (HCl) typically under the supersonic irradiation condition. These nanosheets were subsequently used for deriving the nanosheets of monoclinic TiO₂. Takezawa and Imai³ described a new bottom-up route for the preparation of layered H₂Ti₃O₇ nanosheets using agar gel containing a titanium precursor and ammonium ions. Antonello et al.⁴ synthesized titanate nanosheets using the titanium(IV) iso-propoxide (Ti(OC₃H₇)₄) precursor under the reflux and vacuum conditions. The method was claimed as a novel strategy in the fabrication of optically active photonic structures. Zhao et al.,⁵ Lin et al.,⁶ Wu et al.,⁷ and Sutradhar et al.⁸ synthesized organic-stabilizer free layered protonic H₂Ti₂O₅·H₂O nanosheets via hydrothermal treatment and utilized them for the removal of Pb²⁺ cations and ciprofloxacin (CIP) from the aqueous solutions. Wang et al.⁹ reported the synthesis of titanate nanosheets on Ti foil via hydrothermal method and topotactically transformed them to a novel anatase/TiO₂(B) heterostructured nanosheet array which exhibited superior photocatalytic activity in the removal of Rhodamine B (RhB) dye from aqueous solution. Hareesh¹⁰ et al. processed flyash stabilized H₂Ti₃O₇ nanosheets via hydrothermal method for the removal of methylene blue (MB) dye from an aqueous solution via adsorption mechanism. Luan and Wang¹¹ synthesized H₂Ti₃O₇ nanosheets by exfoliation of a layered precursor via interacting TBA cations, followed by ion-exchange with Na⁺ ions and washing with water. 2D graphene/hydrogen titanate hybrid nanosheets were fabricated for application as anode materials in lithium-ion batteries. Zhou et al.¹² synthesized hydrogen titanate nanosheets through soft-chemical delamination of layered protonic titanates (LPT) using ethylamine and used them as non-light driven catalyst for the degradation of RhB dye via hydrogen peroxide (H₂O₂) activation. Kim et al.¹³ synthesized (3-aminopropyl)triethoxysilane (APTES) anchored H_0.67Ti_1.83O₄·H₂O nanosheets as a stable nanocontainer for DNA.

It appears that there are only few reports available in the literature which have attempted to utilize HTNS for the removal of synthetic-dyes from the aqueous solutions using either adsorption or photocatalysis mechanism.^9,10,12 Surprisingly, all of the layered titanate nanosheets utilized for various applications (including the dye-removal) are non-magnetic in nature. No attempt has been made to convert the non-magnetic titanate nanosheets to magnetic nanosheets for any of the aforementioned applications. It is realized that, in the dye-removal application, the magnetic nanocomposites provide ease of solid–liquid separation via the use of an external magnetic field which offer large benefits over the several issues associated with the separation of non-magnetic counterparts.¹⁴ In view this, the first major objective of this investigation has been set to synthesize, for the first time, the magnetic nanocomposites (HTNSF) based on the HTNS and maghemite (γ-Fe₂O₃) nanoparticles via a novel ion-exchange mechanism.¹⁵ The second major objective has been set to use this magnetic nanocomposite, for the first time, for the removal of cationic methylene blue (MB) dye from the aqueous solutions via adsorption mechanism and investigate the effect of weight-fraction of γ-Fe₂O₃ on the variation in the MB adsorption capacity of magnetic nanocomposite. MB dye has been utilized for different industrial applications such as the medical photodynamical therapy, chromoendoscopy, textile industries, analytical chemistry, and sensitizer in the solar energy conversion.¹⁶ All these applications, however, generate MB-charged effluents that may have carcinogenic and mutagenic properties towards the aquatic organisms. This is a serious cause of risk to the human life and eco-environment. Hence, the development of novel technology for the effective removal of MB dye from the aqueous solutions is of significance. The third major objective of present investigation has been set to demonstrate, for the first time, the magnetic separation of HTNSF nanocomposite using an external magnetic field followed by its regeneration and reuse via non-light driven synergistic activation of persulfate.¹⁶

It is to be noted that in comparison with the conventional dye-removal technologies such as the photocatalysis⁹ and electro-chemical oxidation,¹⁷ the proposed dye-removal technology can be conducted in the dark without the use of any external radiation (such as the UV, solar, visible, and infrared) and power sources (such as the voltage, sonicator, and microwave generator) which suggests considerable cost-saving in the dye-removal process. Moreover, further cost saving is achievable by the successful regeneration of catalyst via the activation of cheaper persulfate than that of costlier H₂O₂.

2. Experimental section

2.1 Synthesis of HTNS and HTNSF magnetic nanocomposites

The HTNS were first synthesized via combination of sol–gel and hydrothermal methods as described in detail elsewhere,¹⁰ (refer the section-A.1 of ESI†). In short, the nanocrystalline anatase-TiO₂ is coated on the surface of as-received flyash particles via sol–gel method and the obtained sol–gel product is subjected to hydrothermal treatment. The hydrothermal product is then subjected to two-step washing procedure for obtaining HTNS.

The HTNSF magnetic nanocomposites were synthesized in an aqueous solution using the process involving an ion-exchange mechanism operating in between HTNS and magnetic γ-Fe₂O₃ nanoparticles as described in detail elsewhere for the hydrogen titanate nanotubes (HTN),¹⁵ (refer the section-A.2 of ESI†). The as-synthesized magnetic nanocomposites are designated as HTNSF-X where X represents the weight-fraction of γ-Fe₂O₃ in the magnetic nanocomposite.

2.2 Characterization of HTNS and HTNSF magnetic nanocomposites

The samples were characterized using different analytic techniques for determining the morphology, average size, nanocrystallinity, chemical constituents, crystal structure, specific surface-area, pore-size distribution, band-gap energy (E_BG), magnetization curve, and zeta-potential. The details of characterization techniques utilized for this purpose are provided elsewhere (refer the section-A.3 of ESI†).

2.3 Dye-adsorption characteristics of HTNS and HTNSF magnetic nanocomposites

The dye-adsorption measurements were carried out using the cationic MB dye via procedure already described elsewhere¹⁵ (refer the section-A.4 of ESI†). The equilibrium MB adsorption data obtained is analyzed using the various adsorption isotherm (Langmuir, Freundlich, and Dubinin–Kaganer–Radushkevich (DKR)) and kinetics (Lagergren pseudo-first-order and pseudo-second-order) models.

2.4 Regeneration and reuse of HTNSF magnetic nanocomposite in the dye-removal application

The regeneration and reuse of catalyst in the dye-removal experiments were demonstrated using the HTNSF-5 magnetic nanocomposite. The regeneration of catalyst was typically attempted and compared using the aqueous solutions of H₂O₂ and potassium persulfate (K₂S₂O₈). The procedure used for the regeneration of HTNSF-5 magnetic nanocomposite having the pre-adsorbed (107 mg g⁻¹) MB on its surface, using the aqueous solution of H₂O₂ (ref. 15) and K₂S₂O₈,¹⁶ is described in detail elsewhere (refer the section-A.5 of ESI†).

2.5 Trapping of free hydroxyl radicals (˙OH)

The procedure followed for the detection of free ˙OH is described in detail in the section-A.6 of ESI.†

3. Results and discussions

3.1 Characteristics of HTNS and HTNSF magnetic nanocomposites

The TEM images of as-received magnetic nanoparticles at different magnifications are provided in Fig. 1. It is noted that the as-received magnetic nanoparticles are aggregated with the non-spherical aggregate size of 200–300 nm, Fig. 1a, and the nanocrystallite size in the range of 15–40 nm, Fig. 1b. The non-spherical aggregates with fine intra-aggregate porosity are further observed to form bigger agglomerates having relatively larger inter-aggregate pore size (100–300 nm). The continuous diffused rings observed in the SAED pattern, Fig. 1c, confirm the nanocrystalline nature of as-received magnetic particles. Moreover, the obtained SAED pattern is very similar to that reported by Morjan et al.¹⁸ and is indexed according to the face centered cubic (FCC) structure of maghemite (γ-Fe₂O₃) phase which is also in accordance with the XRD results obtained here.

Fig. 1 TEM images of as-received γ-Fe₂O₃ nanoparticles at lower (a) and higher (b) magnifications. The corresponding SAED pattern is shown in (c).

The TEM images of HTNS and HTNSF magnetic nanocomposites are presented at lower and higher magnifications in Fig. 2 and S1 (see the ESI†) respectively. The corresponding SAED patterns are shown as insets in the upper-left corners, Fig. 2. The HTNS sample, Fig. 2a and S1a (see the ESI†), clearly shows highly aggregated curled nanosheets having the average thickness of 5 nm. Surprisingly, two diffused concentric rings normally observed in the SAED pattern of hydrogen titanate which are indexed according to either monoclinic H₂Ti₃O₇ or orthorhombic protonic lepidocrocite titanate (H_xTi_2−x/4□_x/4O₄, where x ∼ 0.7 and □ is a vacancy),^19,20 are not observed in the SAED pattern presented in Fig. 2a. This is also consistent with the previous analysis reported by Hareesh et al.¹⁰ The SAED pattern instead consists of large number (>2) of closely spaced concentric rings which suggests the presence of unknown nanocrystalline phase in addition to that of hydrogen titanate phase. Interestingly, the SAED patterns presented in Fig. 2b–d (see the insets in upper-left corners) also contain number of closely spaced concentric rings; however, they resemble that of γ-Fe₂O₃, Fig. 1c, instead of those of hydrogen titanate. This is a result of the formation of HTNSF magnetic nanocomposites consisting of hydrogen titanate having lower crystallinity and varying weight-fraction of highly crystalline γ-Fe₂O₃. Due to difference in the crystallinity of two phases, the HTNSF magnetic nanocomposites exhibit SAED pattern corresponding to that of γ-Fe₂O₃.

Fig. 2 TEM images of HTNS (a), HTNSF-5 (b), HTNSF-10 (c), and HTNSF-25 (d) magnetic nanocomposites. The insets in the upper-left and lower-right corners show the corresponding SAED patterns and the attachment of γ-Fe₂O₃ nanoparticles at the edges of nanosheets respectively.

As demonstrated in Fig. S1 (see the ESI†), the average thickness of curled aggregated nanosheets is noted to increase from 5 nm to 23 nm with the increasing weight-fraction of γ-Fe₂O₃ from 0 to 25 wt%. This possibly suggests that spacing between the stacked nanosheets within the aggregates increases with the weight-fraction of γ-Fe₂O₃ which is reflected in an increase in the average thickness of nanosheets. As demonstrated in the lower-right corner insets of Fig. 2b–d (and also in Fig. S1b–d of ESI†), the magnetic γ-Fe₂O₃ nanoparticles are typically seen to be attached to the edges of nanosheets rather than on their surfaces. As shown in Fig. S2 (see the ESI†), the nanoparticles are also seen to be attached to the edges of HTNS which do not belong to those of γ-Fe₂O₃. However, in the latter case, these appear to be quartz (SiO₂) nanoparticles as per the EDX spectrum presented in Fig. S3a (see the ESI†). The latter confirms the presence of Si in the HTNS in exceptionally large amount; while, all HTNSF magnetic nanocomposites, Fig. S3b–d (see the ESI†) show the presence of larger amount of Fe relative to that of Si in agreement with the TEM and SAED analyses. This also suggests that the SAED pattern presented in Fig. 2a predominantly belongs to that of SiO₂ nanoparticles rather than that of γ-Fe₂O₃ or hydrogen titanate. Moreover, the intense Si peak as noted in Fig. S3a (see the ESI†) is also possibly contributed by the adsorption of silicate (SiO₄⁴⁻) ions on the surface of HTNS. The attachment of SiO₂ nanoparticles to the edges of nanosheets of hydrogen titanate and the presence of SiO₄⁴⁻ ions on their surfaces are ascribed to the mechanism of formation of nanosheets during hydrothermal treatment of TiO₂-coated flyash particles and the subsequent washing of hydrothermal product as proposed earlier by Hareesh et al.¹⁰

The powder XRD patterns of different samples are presented in Fig. S4 and S5 (see the ESI†); while, all identified phases, the diffracting angles and planes, and the corresponding JCPDS card numbers are tabulated in the Table S1 (see the ESI†). It is noted that the as-received flyash particles consists of SiO₂, mullite (Al₆Si₂O₁₃), and hematite (α-Fe₂O₃) having the weight-fraction of 69, 20, and 11 wt% respectively. The flyash–TiO₂ composite particles synthesized via sol–gel process show the successful deposition of anatase-TiO₂ on the surface of flyash particles. The product obtained after the hydrothermal treatment of flyash–TiO₂ core–shell composite particles exhibits two new broad peaks, which are identified as belonging to either monoclinic H₂Ti₃O₇ (with the interlayer spacing of 0.78 nm within the walls of nanotubes) or orthorhombic lepidocrocite-type or H₂Ti₂O₄(OH)₂-type hydrogen titanate structures, similar to the analysis presented earlier for the formation of nanotubes of hydrogen titanate via hydrothermal by Jose et al.²⁰ and others^15,21 (note: for further discussion on the mechanism of formation of HTNS, refer the section-H of ESI†). The XRD patterns of HTNSF magnetic nanocomposites show a gradual increase in the intensity of main peak, (311), of γ-Fe₂O₃ which is in accordance with its increasing weight-fraction in these samples. It is to be noted that the XRD pattern of magnetic nanoparticles is ascribed to that of γ-Fe₂O₃ as per the JCPDS card number of 39-1346 rather than to magnetite (Fe₃O₄) which has the XRD pattern almost similar to that of the former. Chen et al.,²² however, demonstrated that the XRD peaks of γ-Fe₂O₃ are slightly shifted to higher diffraction angles relative to those of Fe₃O₄. In consonance with this, in the present investigation it is noted that the obtained 2θ values for the magnetic nanoparticles, Table S1,† are slightly larger than those listed in the JCPDS card numbers 19-0629, 75-0033, and 85-1436 for Fe₃O₄. In contrast to this, these 2θ values are closer to those reported by Barakat et al.²³ and also comparable with those listed in the JCPDS card number 39-1346 belonging to γ-Fe₂O₃. Moreover, the average lattice parameter of 0.8365 nm is calculated which is closer to that (0.835 nm) of γ-Fe₂O₃ than that (0.840 nm) of Fe₃O₄ as reported by Chen et al.²² and Cai et al.²⁴ Hence, in this investigation, the structure of as-received magnetic nanoparticle is considered to be γ-Fe₂O₃.

The N₂ adsorption/desorption isotherms obtained using the different samples are presented in Fig. S6a (see the ESI†); while, the corresponding Barrett–Joyner–Halenda (BJH) pore size distribution graphs are shown in Fig. S6b (see the ESI†). The values of specific surface-area and pore volume as obtained for the different samples are tabulated in the Table S2 (see the ESI†). According to the IUPAC classification,²⁵ the as-received flyash particles exhibit close to type I reversible isotherm with the absence of any mesoporosity with the specific surface-area and pore volume as low as 2 m² g⁻¹ and 0.008 cm³ g⁻¹. The deposition of nanocrystalline anatase-TiO₂ via sol–gel method results in a change of the isotherm to type IV which is the characteristic feature of capillary condensation taking place in the mesopores. The flyash–TiO₂ composite particles exhibit the specific surface-area and pore volume of 21.5 m² g⁻¹ and 0.08 cm³ g⁻¹ which are higher almost by 10 times than those of as-received flyash. This is in contrast to the observation made by Shi et al.²⁶ who deposited TiO₂ on the surface of flyash via sol–gel method; however, did not observe any significant increase in the specific surface-area and pore volume possibly due to the relatively larger concentration (250 g l⁻¹) of flyash used in their experiments which significantly reduced the amount of TiO₂ deposited over each individual flyash particle. In the present investigation, the increased specific surface-area and pore volume after the anatase-TiO₂ deposition on the surface of flyash is further supported by the presence of small hysteresis loop which is of type H2 which in turn suggests the presence of ink-bottle type mesopores in the nanocrystalline anatase-TiO₂ coating. The HTNS and HTNSF magnetic nanocomposites exhibit the isotherms of type II and the associated hysteresis loop is of type H3 which strongly suggests the presence of aggregates of plate-like particles giving rise to slit-shaped pores. Comparison shows that the shape of isotherms and pore size distribution graphs as obtained for HTNS and HTNSF magnetic nanocomposites are similar to those reported recently for the magnetic nanocomposite consisting of molybdenum disulfide (MoS₂) nanosheets decorated with Fe₃O₄ nanoparticles by Song et al.²⁷ and for the layered protonated titanate nanosheets by Lin et al.⁶ (note: for further analysis of Fig. S6,† refer the discussion presented in the sections-I and J of ESI†).

The variation in the specific surface-area and pore volume as a function of weight-fraction of γ-Fe₂O₃ as obtained for HTNS and HTNSF magnetic nanocomposite samples is shown in Fig. 3. According to Lin et al.,⁶ the specific surface-area for the layered protonic titanate nanosheets generally lies within the range of 96–350 m² g⁻¹. The specific surface-area of HTNS (334 m² g⁻¹), Fig. 3a, is within this reported range (note: the dotted lines are obtained by joining the values of specific surface-area and pore volume of HTNS with those associated with pure γ-Fe₂O₃ nanoparticles which follow the law-of-mixture of these two phases without any synergy effect in between them). It is clearly seen that the obtained variations in the specific surface-area and pore volume do not follow the law-of-mixture of two phases. The values of both of these parameters exhibit positive deviations from the law-of-mixture as a function of increasing weight-fraction of γ-Fe₂O₃. This clearly suggests the synergy effect in between HTNS and γ-Fe₂O₃ nanoparticles which is described in the model presented in the Scheme 1. The latter also explains the ion-exchange mechanism responsible for the formation of HTNSF magnetic nanocomposite. As shown in the Scheme 1a, H⁺ ions are intercalated in between the H₂Ti₃O₇ nanosheets made up of TiO₆ octahedra; while, the as-received γ-Fe₂O₃ nanoparticles consist of Fe–O bonds on the surface with Fe having the oxidation state of +3. During the formation of HTNSF magnetic composite, the close approach of γ-Fe₂O₃ nanoparticles towards the aggregated H₂Ti₃O₇ nanosheets results in the ion-exchange reaction in between Fe³⁺ cations and H⁺ ions which replaces the intercalated H⁺ ions with Fe³⁺ cations. Since Fe³⁺ cations are already anchored to the surface of γ-Fe₂O₃ nanoparticles, the ion-exchange reaction results in the formation of ion-exchange bond at the interface of γ-Fe₂O₃ nanoparticles and aggregated H₂Ti₃O₇ nanosheets. Thus, the magnetic nanocomposite is formed in which the γ-Fe₂O₃ nanoparticles are attached to the aggregated H₂Ti₃O₇ nanosheets via ion-exchange bond typically at the edges of nanosheets where the ion-exchange reaction is highly likely to occur. This clearly explains the observation made in the TEM images, Fig. 2b–d (see lower-right corner insets) and S1b–d (see the ESI†), where γ-Fe₂O₃ nanoparticles are seen to be attached to the aggregated H₂Ti₃O₇ nanosheets typically at the edges of nanosheets. In the Scheme 1b, the nanosheets are shown to be highly aggregated for HTNS. However, in the case of HTNSF magnetic nanocomposite, due to the existence of strong repulsive forces between the adjacent γ-Fe₂O₃ nanoparticles, the nanosheets tend to move apart from each other. This may reduce the aggregation tendency of nanosheets by increasing the average stacking distance in between the nanosheets. It appears that the magnitude of repulsive force increases with the weight-fraction of γ-Fe₂O₃ nanoparticles within the HTNSF magnetic nanocomposites. As a result, the average stacking distance between the aggregated nanosheets may increase with the weight-fraction of γ-Fe₂O₃ nanoparticles. This is strongly supported by an increase in the average thickness of aggregated nanosheets from 5 nm to 23 nm with the weight-fraction of γ-Fe₂O₃ nanoparticles, Fig. S1 (see the ESI†). Moreover, it is noted that an increase in the average stacking distance between the aggregated nanosheets as per the mechanism proposed in the Scheme 1b should also result in an increase in the volume of slit-shaped pores, and hence, in the specific surface-area. This satisfactorily explains the positive deviations in the values of these parameters from the variation governed by the law-of-mixture with the increasing weight-fraction of γ-Fe₂O₃ nanoparticles as observed in Fig. 3.

Fig. 3 Variation in specific surface-area (a) and pore-volume (b) as a function of weight-fraction of γ-Fe₂O₃ nanoparticles.

Scheme 1 Model describing the mechanism of formation of HTNSF magnetic nanocomposite via creation of an ion-exchange bond (a) and the mechanism of increase in pore volume and specific surface-area of HTNSF magnetic nanocomposite with the attachment of γ-Fe₂O₃ nanoparticles at the edges of nanosheets (b).

3.2 MB adsorption characteristics of HTNS and HTNSF magnetic nanocomposites

The variation in the normalized MB concentration adsorbed at equilibrium on the surface of HTNS and HTNSF magnetic nanocomposites is presented in Fig. 4 for the initial MB concentration range of 7.5–250 μM at the initial solution-pH of 10. The adsorption equilibrium is seen to be attained within the first 10 min of contact time. As demonstrated in Fig. S7 (see the ESI†), the point-of-zero charge for HTNS and HTNSF-5 magnetic nanocomposite is determined to be at the initial solution-pH of 2.0 and 3.0 respectively. As a result, the surface of HTNS and HTNSF magnetic nanocomposite is highly negatively charged at the initial solution-pH of 10 due to both the adsorption of large amount of OH⁻ ions from the surrounding aqueous solution and the prior presence of SiO₄⁴⁻ ions on their surfaces. This results in a strong electrostatic attraction between the positively charged cations of MB and negatively charged surface of HTNS. In Fig. 4, the normalized equilibrium concentration of MB adsorbed on the catalyst surface is noted to decrease with the increasing initial MB concentration. The equilibrium amount of MB adsorbed (q_e, mg g⁻¹) on the surface per unit weight of catalyst can be calculated as,¹⁵where, C_o and C_e (mg l⁻¹) represent initial and equilibrium dye concentrations, m_ads (g) the mass of adsorbent, and V (l) the initial volume of solution. The obtained variation in q_e as a function of initial MB concentration is presented in Fig. S8 (see the ESI†); while, the equilibrium adsorption isotherm curves are presented in Fig. S9 (see the ESI†). The q_e values are noted to increase first with the initial MB concentration, then to reach a maximum value (q_m, adsorption capacity) at the initial MB concentration of 100–150 μM, and then to decrease with further increase in the initial MB concentration. The increase in q_e values is attributed here to the greater driving force of MB to occupy potential adsorption sites on the surface of catalyst; while, the decrease is ascribed to decrease in the negative surface-charge on the surface of catalyst due to decrease in the adsorption of OH⁻ ions as a result of their strong interaction with the positively charged MB cations in the surrounding aqueous solution. The obtained variation in q_m as a function of amount of γ-Fe₂O₃ is shown in Fig. 5. Relative to linear variation as governed by the law-of-mixture, a positive deviation in the variation of MB adsorption capacity as a function of amount of γ-Fe₂O₃ is observed. This clearly suggests a synergy effect in between the constituents of HTNSF magnetic nanocomposites in enhancing the MB adsorption capacity with the amount of γ-Fe₂O₃. The addition of magnetic component (γ-Fe₂O₃), having very low MB adsorption capacity (13 mg g⁻¹),¹⁵ to H₂Ti₃O₇ nanosheets-based adsorbents should result in a decrease in the dye-adsorption capacity of magnetic nanocomposite relative to that of non-magnetic HTNS. This has been demonstrated earlier for the Fe₃O₄–graphene magnetic nanocomposite by Yao et al.,²⁸ who reported the MB adsorption capacity of 70 mg g⁻¹ and 45 mg g⁻¹ for pure graphene and Fe₃O₄–graphene magnetic nanocomposite. Interestingly, in the present investigation, the nature of experimental variation in q_m follows the same trend as that of the pore volume and specific surface-area as noted in Fig. 3. Close inspection further reveals that the nature of experimental variation in q_m follows the same trend as that exhibited by pore volume rather than that shown by specific surface-area. This strongly suggests that the MB dye molecules preferentially occupy the potential adsorption sites available inside the volume of slit-type pores than those present on the surface of layered nanosheets. It is further noted that the HTNSF-10 magnetic nanocomposite possesses the maximum MB adsorption capacity of 76 mg g⁻¹ which is higher than that (67 mg g⁻¹) of HTNS. This trend is exactly opposite to the one reported in the literature by Yao et al.²⁸ Secondly, the maximum MB adsorption capacity value obtained here is higher than those (25, 44–45, and 42 mg g⁻¹) reported (or reviewed) recently for the other types magnetic nanocomposites such as MoS₂–Fe₃O₄,²⁷ graphene nanosheets (GNS)–Fe₃O₄,^28,29 and multi-wall carbon nanotubes (MWCNTs)–Fe₂O₃ ³⁰ respectively. Moreover, the maximum MB adsorption capacity value obtained here is also higher than those (30 and 60 mg g⁻¹) reported for the MWCNTs–Fe₃O₄ and activated carbon (AC)–Fe₃O₄ magnetic nanocomposites respectively as reported by Ai et al.²⁹ The MB adsorption capacity of CNTs and commercial adsorbent such as the activated carbon is reviewed to be 35–65 and 521 mg g⁻¹.^28,30 Thus, the MB adsorption capacity (67 mg g⁻¹) of HTNS as observed in this investigation is comparable with that of CNTs and graphene nanosheets²⁸ but lower than that of activated carbon. This is in agreement with the results obtained using the various carbon-based nanomaterials as sorbents as reported by Beless et al.³¹ (note: Li et al.,³² however, demonstrated that activated carbon may exhibit lower MB adsorption capacity normalized by BET surface-area compared with that shown by the graphene oxide and CNTs. By using the organic pollutants other than MB, Apul et al.³³ and Ersan et al.³⁴ also showed that under certain test conditions granular activated carbon may exhibit lower adsorption capacity relative to that exhibited by other carbon-based adsorbents).

Fig. 4 Variation in the normalized concentration of surface-adsorbed MB as a function of contact time as obtained for HTNS (a), HTNSF-5 (b), HTNSF-10 (c), and HTNSF-25 (d) magnetic nanocomposites. The initial MB concentration is varied as 7.5 (i), 30 (ii), 60 (iii), 90 (iv), 150 (v), 200 (vi), and 250 μM (vii). The initial solution-pH is 10.

Fig. 5 Variation in MB adsorption capacity as a function of weight-fraction of γ-Fe₂O₃ nanoparticles. The initial solution-pH is 10.

The time dependent MB adsorption on the surface of HTNS and HTNSF magnetic nanocomposites is analyzed using two different kinetics models – Lagergren pseudo-first-order^35,36 and pseudo-second-order^35,37 (refer the section-N of ESI†). The linear plots (see Fig. S10 of ESI†) with the regression correlation coefficient, 〈r₂²〉, approximately equal to one and q_e values approximately equal to that of experimentally observed values, q_e(Exp), (see the Table 1a) suggest that MB adsorption on the surface of HTNS and HTNSF magnetic nanocomposites follows the pseudo-second-order kinetics. In the Table 1b, the values of parameters of pseudo-first-order kinetics model have been tabulated typically for the lower and higher initial MB dye concentrations of 7.5 and 150 μM. It is noted that, for all samples, the lower the initial MB concentration, the lower the 〈r₁²〉 value and vice versa. As a result, the pseudo-first-order kinetics model which is not applicable for the lower initial MB concentration, becomes applicable at higher value of latter. The comparison with 〈r₂²〉 values at the initial MB concentrations of 7.5 and 150 μM for all samples, Table 1a, suggests that 〈r₂²〉 values slightly decrease at higher initial MB concentration. This suggests that the pseudo-second-order kinetics model which is applicable for the lower initial MB concentration, becomes inapplicable at higher value of latter. As demonstrated in Fig. 6, in the present investigation, this transition from the pseudo-second-order to Lagergren pseudo-first-order kinetics for HTNS typically occurs at the initial MB concentration of 150 μM which is highly consistent with the assumptions made in the theoretical derivations of these two kinetics models by Azizian.³⁵

Table 1 Values of parameters of pseudo-second-order (a) and Lagergren pseudo-first-order kinetics models (b) as obtained for different samples under the different test-conditions

(a)
Sample	[MB] μM	qe(Exp.) mg g^−1	qe mg g^−1	k2 g mg^−1 min^−1	〈r2^2〉
HTNS	7.5	5.8	5.81	2.12	0.999
	60	47	47.6	0.44	1
	90	67	76.9	0.003	0.998
	150	64	100	0.0003	0.959
HTNSF-5	7.5	5.9	5.95	0.81	1
	60	47	47.6	0.15	1
	90	69.8	76.9	0.005	0.999
	150	74.5	125	0.0002	0.973
HTNSF-10	7.5	5.9	5.95	—	1
	60	46.7	47.6	0.023	0.999
	90	66.7	71.4	0.004	0.999
	150	75.6	91	0.0008	0.989
HTNSF-25	7.5	5.7	5.78	1.5	1
	60	46.3	47.6	0.016	0.999
	90	64.4	76.9	0.0015	0.999
	150	68.3	111	0.0003	0.992

---

(b)
Sample	[MB] μM	qe(Exp.) mg g^−1	qe mg g^−1	k1 min^−1	〈r1^2〉
HTNS	7.5	5.8	1.47	0.083	0.597
HTNSF-5		5.9	2.5	0.19	0.822
HTNSF-10		5.9	—	—	—
HTNSF-25		5.7	4.86	0.193	0.978
HTNS	150	64	79.4	0.067	0.973
HTNSF-5		74.5	89.9	0.062	0.976
HTNSF-10		75.6	71	0.06	0.959
HTNSF-25		68.3	69.7	0.046	0.987

---

Fig. 6 Variation in the 〈r²〉 value as a function of initial MB concentration as obtained for HTNS using the Lagergren pseudo-first-order (i) and pseudo-second-order (ii) kinetics models.

The fitted curves for the equilibrium MB adsorption on the surface of HTNSF-5 magnetic nanocomposite obtained using the Langmuir,^15,38 Freundlich,^15,39 and Dubinin–Kaganer–Radushkevich (DKR),^15,40,41 isotherm models (refer the section-P of ESI† for the theoretical description of these models) are presented in Fig. S11 (refer the sections-Q and R of ESI†). The calculated values of different parameters of these equilibrium isotherm models are listed in the Table 2. The q_m values as obtained via Langmuir and DKR models match with the experimentally observed values (q_m(Exp)), along with 〈r²〉 values which are close to unity. This strongly suggests a monolayer adsorption of MB on the surface of HTNS and HTNSF magnetic nanocomposites. Moreover, R_L values (a dimensionless parameter which is described in the section-P of ESI†) for the MB adsorption on the surface of both HTNS and HTNSF magnetic nanocomposites lies in between 0.004 and 0.17, Table 2, which indicates that the adsorption process is favorable.¹⁵ In addition to this, both the negative values of ΔG⁰ and E values less than 8 kJ mol⁻¹, Table 2, strongly suggest the spontaneous MB adsorption on the surface of HTNS and HTNSF magnetic nanocomposites via electrostatic attraction mechanism.^15,41

Table 2 Values of parameters of different equilibrium adsorption isotherm models as obtained for different samples^a

Sample	qm(Exp.) mg g^−1	Langmuir					Fruendlich			DKR
		qm mg g^−1	KL l mg^−1	〈r〉^2	RL	ΔG kJ mol^−1	n	KF mg^1−1/n g^−1 l^1/n	〈r〉^2	qm mg g^−1	β mol^2 J^−2	〈r〉^2	E kJ mol^−1
a Note: qm: adsorption capacity; KL: Langmuir constant; RL: separation factor; ΔG: change in Gibb's free energy; n: a constant related to adsorption intensity; KF: Freundlich constant; β: a constant related to adsorption energy; E: adsorption energy.
HTNS	67.0	66.7	3.0	0.999	0.007–0.12	−37.0	1.35	57.4	0.851	69.7	6 × 10^−8	0.988	2.9
HTNSF-5	74.5	76.9	4.3	0.999	0.004–0.09	−38.0	1.23	94.6	0.974	73.2	4 × 10^−8	0.991	3.5
HTNSF-10	75.6	83.3	2.0	0.999	0.01–0.17	−36.0	1.36	56.9	0.856	73.3	6 × 10^−8	0.989	2.9
HTNSF-25	68.3	71.4	2.8	0.999	0.007–0.13	−37.0	1.40	55.9	0.842	70.3	6 × 10^−8	0.991	2.9

---

3.3 Magnetic separation, regeneration, and reuse of HTNSF magnetic nanocomposite

As reported earlier by Harsha et al.,¹⁵ the as-received γ-Fe₂O₃ nanoparticles used in this investigation are superparamagnetic in nature and exhibit a small hysteresis loop with the saturation magnetization, remanent magnetization, and coercivity values of 71 emu g⁻¹, 5.9 emu g⁻¹ and 59 Oe respectively. The saturation magnetization value of 71 emu g⁻¹ is comparable with that (76 emu g⁻¹) obtained for the bulk γ-Fe₂O₃ as reported by Millan et al.⁴² Hence, with the attachment of γ-Fe₂O₃ nanoparticles to HTNS via the formation of an ion-exchange bond, Scheme 1b, it is possible to obtain HTNSF magnetic nanocomposites with varying weight-fractions of γ-Fe₂O₃ nanoparticles. The magnetization curves obtained for different HTNSF magnetic nanocomposites are presented in Fig. S12 (see the ESI†); while, the obtained values of various magnetic properties are tabulated in the Table 3. Although the saturation magnetization is the smallest in the case of HTNSF-5 sample due to the lowest weight-fraction (5 wt%) of γ-Fe₂O₃ nanoparticles,⁴² it could be separated from an aqueous solution using an external magnetic field in 5 min (see the inset located at lower-right corner of Fig. S12 (see the ESI†)). Hence, the successful magnetic separation of HTNSF magnetic nanocomposite from the aqueous solution has been demonstrated in this investigation.

Table 3 Values of magnetic properties as obtained for different samples

Sample	Saturation magnetization (emu g^−1)	Remanent magnetization (emu g^−1)	Coercive field (Oe)
HTNS-5	4.3	0.4	−62.6
HTNSF-10	6.7	0.64	−58.2
HTNSF-25	29.0	2.96	−57.5
γ-Fe2O3	71.0	5.9	−58.7

---

After the magnetic separation of HTNSF magnetic nanocomposite from the treated aqueous solution, in order to reuse the catalyst for the next cycle of dye-adsorption, it is necessary to regenerate the catalyst either by desorption of previously adsorbed-dye or by completely decomposing the adsorbed-dye on the surface of catalyst. Since the first method merely transforms the adsorbed-dye from the solid phase to the liquid phase causing the secondary pollution,^43,44 in the present investigation, the second method of catalyst regeneration has been attempted and compared for H₂O₂ and K₂S₂O₈ activation techniques.^12,15,16

The obtained variation in the normalized concentration of MB adsorbed as a function of contact time, at the initial MB concentration of 90 μM, is presented in Fig. 7a for three consecutive cycles of dye-adsorption. It is noted that the amount of MB adsorbed at the equilibrium decreases from 97% to 13% after three consecutive cycles of dye-adsorption which results in the total MB adsorption of 107 mg g⁻¹ on the surface of HTNSF-5 magnetic nanocomposite. The regeneration of catalyst is then attempted using 30 wt% H₂O₂ solution to decompose MB dye which is adsorbed in the previous three cycles. However, as noted in Fig. 7a, only 8% of MB is adsorbed on the surface of HTNSF-5 magnetic nanocomposite in the fourth dye-adsorption cycle following its regeneration conducted using H₂O₂ solution. This strongly suggests that the regeneration treatment conducted using H₂O₂ solution could not decompose the previously adsorbed MB dye on the surface of HTNSF-5 magnetic nanocomposite. Surprisingly, Harsha et al.¹⁵ successfully demonstrated that H₂O₂ can be easily activated using the nanotubes of H₂Ti₃O₇ generating the superoxide radical ions (O₂˙⁻) and hydroxyl radicals (˙OH) which can attack and decompose the surface-adsorbed MB dye.

Fig. 7 (a) Variation in the normalized concentration of MB adsorbed as a function of contact time as obtained for HTNSF-5 magnetic nanocomposite under different test-conditions. The successive dye-adsorption cycles-1 to 3 (i–iii) are conducted before the regeneration treatment. The cycle-4 ((iv) and (v)) is carried out after the regeneration treatment conducted at room temperature (30 °C) in H₂O₂ (30 wt%) (iv) and K₂S₂O₈ (0.1 M (∼3 wt%)) (v) solutions respectively. For each dye-adsorption cycle, the initial MB concentration is 90 μM. (b) Variation in the normalized concentration of MB adsorbed after 1 h of contact time as a function of dye-adsorption cycle number as obtained for HTNSF-5 magnetic nanocomposite without (i) and with (ii) the involvement of regeneration treatment conduced in K₂S₂O₈ solution at 75 °C after each dye-adsorption cycle. The initial MB and K₂S₂O₈ concentrations are 90 μM and 2 mM (0.06 wt%) respectively.

Zhou et al.¹² also reported recently that the delaminated two-dimensional titanate nanosheets are superior non-light driven catalyst for the degradation of organic dyes (RhB) obtained through the generation of superoxide radical-ions via H₂O₂ activation. It is to be noted that, in the present investigation, the H₂Ti₃O₇ nanosheets are stabilized with SiO₄⁴⁻ ions and SiO₂ nanoparticles. Hence, there is a strong possibility that the potential sites available on the surface and edges of HTNSF-5 magnetic nanocomposite are already occupied by the foreign species reducing the concentration of surface-adsorbed O₂⁻ ions and H₂O₂. As per the reaction shown in the eqn (3), the adsorption of O₂⁻ ions and H₂O₂ in significant amount on the surface of HTNSF-5 magnetic nanocomposite appears to be an essential requirement in order to generate ˙OH in large amount. This requirement is seriously hampered due to the prior presence of SiO₄⁴⁻ ions and SiO₂ nanoparticles on the surface and at the edges of HTNS nanosheets. In addition to this, it also appears that the activation of H₂O₂ via the visible-light induced photo-Fenton reaction^45,46 and direct photolysis⁴⁷ is also not possible in this investigation (for the further discussion, refer the section-T of ESI†).

As an alternative, the regeneration treatment of HTNSF-5 magnetic nanocomposite after the third cycle of MB adsorption is conducted at room temperature (30 °C) by replacing H₂O₂ solution with K₂S₂O₈ solution. Surprisingly, it is noted that the fourth MB adsorption cycle conducted after the regeneration treatment in K₂S₂O₈ solution results in a drastic increase in the adsorption capacity of HTNSF-5 magnetic nanocomposite from 13% to 90%, Fig. 7a. This successfully demonstrates that the regeneration treatment conducted in the K₂S₂O₈ solution can decompose the previously adsorbed MB dye on the surface of HTNSF-5 magnetic nanocomposite. It is to be noted that persulfate (S₂O₈²⁻) is a strong oxidizing agent with the redox potential of 2.01 eV. Upon thermal, chemical, or photochemical activation, it is possible to generate sulfate radical-ion (SO₄˙⁻) which are relatively stronger oxidizing species with the redox potential of 2.6 eV.⁴⁸ Moreover, SO₄˙⁻ can further lead to the formation of ˙OH.

SO₄˙⁻ + H₂O → SO₄²⁻ + ˙OH + H⁺ (5)

These SO₄˙⁻ and ˙OH can then attack and degrade the surface-adsorbed cationic MB dye. In the present investigation, it appears that the thermal activation of S₂O₈²⁻ has a major contribution in the successful regeneration of HTNSF-5 magnetic nanocomposite.¹⁶

As demonstrated in Fig. 7b, the regeneration treatment conducted in K₂S₂O₈ solution at slightly higher temperature (75 °C), after each single cycle of MB adsorption, also maintains the original MB adsorption capacity for the consecutive five adsorption cycles with the reduced time (1 h) required for the regeneration treatment (note: the regeneration time at the room temperature is 3 h). The oxidation of MB (30–70 °C)¹⁷ and other organic pollutants in water such as sulfamonomethoxinum (80 °C),⁴⁸ 1,1,1-trichloroethane (20–50 °C),⁴⁹ and methyl tert-butyl ether (20–50 °C)⁵⁰ via thermal activation of S₂O₈²⁻ is reported in the literature by others. The findings in these earlier reports suggest that the degradation of organic pollutants via thermal activation of S₂O₈²⁻ is highly temperature dependent. In general, higher temperatures provide more energy to rupture O–O bonds of S₂O₈²⁻ and more readily produce reactive species such as SO₄˙⁻ and ˙OH, thereby leading to more rapid degradation of synthetic dyes in water. This is consistent with the results of ˙OH trapping experiments conducted at 30 °C without and with the presence of HTNS-5 magnetic nanocomposite as presented in Fig. 8a and b respectively. Without the presence of latter, Fig. 8a, the concentration of ˙OH produced is noted to increase continuously with the activation time (also see the inset in Fig. 8b). In view of the chemical reaction presented in the eqn (5), this strongly suggests corresponding increase in the concentration of SO₄˙⁻ with the activation time via chemical reaction presented in the eqn (4). The inset in Fig. 8b shows that the concentration of ˙OH produced within the first 20 min of activation time is larger in the presence of HTNS-5 magnetic nanocomposite than that produced in the absence of latter. This strongly suggests a positive synergy effect between the thermal activation and that by the constituents of HTNSF-5 magnetic nanocomposite. It appears that in the presence of HTNS-5 magnetic nanocomposite, having relatively smaller amount of γ-Fe₂O₃ nanoparticles with +3 oxidation state of Fe, the sulfate ions produced via chemical reaction presented in the eqn (5) may react with Fe³⁺ ions present on the surface of γ-Fe₂O₃ nanoparticles producing Fe²⁺ and SO₄˙⁻.⁴⁸

Fe³⁺ + SO₄²⁻ → Fe²⁺ + SO₄˙⁻ (6)

Fig. 8 Variation in PL intensity associated with the formation of 2-hydroxyterephthalic acid at different activation time of S₂O₈²⁻ in the absence (a) and presence (b) of HTNSF-5 magnetic nanocomposite. The activation time varies as – 0 (i), 5 (ii), 20 (iii), 40 (iv), and 60 min (v). The inset in (b) shows variation in the maximum PL intensity as a function of activation time in the absence (A) and presence (B) of HTNSF-5 magnetic nanocomposite.

Since the HTNSF-5 nanocomposite can absorb visible-light (see Fig. S13 of ESI†), the O₂˙⁻ generated on the surface may also produce SO₄˙⁻.

O₂ + e⁻ → O₂˙⁻ (7)

O₂˙⁻ + S₂O₈²⁻ → SO₄˙⁻ + SO₄²⁻ + O₂ (8)

The chemical reactions presented in the eqn (5), (6) and (8) satisfactorily explain the higher concentration of ˙OH observed within the first 20 min of activation time in the presence of HTNSF-5 magnetic nanocomposite. As observed in the inset of Fig. 8b, the concentration of ˙OH produced is noted to decrease and then to increase marginally with further increase in the activation time above 20 min. Such variation in the ˙OH concentration, and possibly that of SO₄˙⁻ concentration, is attributed to their consumption by Fe²⁺ ions⁵¹ and O₂˙⁻.

Fe²⁺ + ˙OH → Fe³⁺ + OH⁻ (9)

O₂˙⁻ + ˙OH → OH⁻ + O₂ (10)

Fe²⁺ + SO₄˙⁻ → Fe³⁺ + SO₄²⁻ (11)

O₂˙⁻ + SO₄˙⁻ → SO₄²⁻ + O₂ (12)

Nevertheless, above the activation time of 20 min, the combined concentration of SO₄˙⁻ and ˙OH appears to be sufficiently large enough to decompose MB adsorbed on the surface of HTNSF-5 magnetic nanocomposite, Fig. 7a. The comparison of Fig. 8a and b, thus, reflect a strong synergy effect in between the thermal activation and that by the constituents of HTNSF-5 magnetic nanocomposite in the generation of SO₄˙⁻ and ˙OH. Overall, the magnetic separation of HTNSF-5 magnetic nanocomposite followed by its regeneration and reuse via synergistic activation of S₂O₈²⁻ are successfully achieved in this investigation.

4. Conclusions

The HTNSF magnetic nanocomposites are successfully synthesized with varying weight-fraction of γ-Fe₂O₃ nanoparticles and utilized for the removal of cationic MB dye from the aqueous solutions via adsorption mechanism. A new model has been proposed to explain the formation mechanism of HTNSF magnetic nanocomposite involving the development of an ion-exchange bond via exchange of Fe³⁺ ions in place of H⁺ at the interface of constituents of magnetic nanocomposite. Moreover, the new model also explains the dependence of variation in the MB adsorption capacity of HTNSF magnetic nanocomposite on that of pore-volume and specific surface-area, as a function of increasing weight-fraction of γ-Fe₂O₃ nanoparticles, on the basis of repulsive forces created in between the adjacent γ-Fe₂O₃ nanoparticles. As a result, HTNSF-10 magnetic nanocomposite possesses the maximum MB adsorption capacity of 76 mg g⁻¹ which is higher than that (67 mg g⁻¹) of HTNS. The MB adsorption on the surface of HTNS and HTNSF magnetic nanocomposites follows the pseudo-second-order kinetics and Langmuir and DKR isotherm models. The variation in 〈r²〉 values as a function of initial MB concentration strongly supports the Azizian analysis which suggests the applicability of pseudo-second-order and Lagergren pseudo-first-order kinetics models at lower (<150 μm) and higher (>150 μm) initial MB concentrations. The successful recycling of HTNSF-5 magnetic nanocomposite, containing the lowest amount of γ-Fe₂O₃ nanoparticles, is demonstrated via its magnetic separation in 5 min followed by its regeneration by decomposing the previously adsorbed MB dye through the synergistic activation of S₂O₈²⁻ in an aqueous solution. A strong synergy effect is observed in between the thermal activation and that by the constituents of HTNSF-5 magnetic nanocomposite in the generation of SO₄˙⁻ and ˙OH. The regeneration of HTNSF-5 magnetic nanocomposite using H₂O₂ solution, however, could not be achieved due to the attachment of SiO₂ nanoparticles at the edges of HTNS and the presence of SiO₄⁴⁻ ions on their surfaces. Overall, the HTNSF magnetic nanocomposite provide an effective way for the removal of cationic MB dye involving the surface-adsorption of dye, its magnetic separation, followed by its regeneration and reuse via synergistic activation of S₂O₈²⁻ in an aqueous solution. The developed technology may be utilized in future for the removal of other types of synthetic-dyes from the aqueous solutions.

Acknowledgements

SS thanks CSIR, India for funding the projects # MLP0012, OLP216339, P81113 and providing the Senior Research Fellowship (SRF) to MJ. HP thanks BITS, Pilani, India for the financial support under the industry-university immersion program. Authors thank Mr Peermohammed, Mr Kiran, Mr Ajeesh P. Paulose, and Mr Arun Gopi (CSIR-NIIST, India) for conducting the BET, TEM/SAED/EDX, PPMS, and zeta-potential analyses respectively.

References

1. M. Gateshki, S. J. Hwang, D. H. Park, Y. Ren and V. Petkov, Chem. Mater., 2004, 16, 5153.
2. Y. Wu and B. Xu, Chin. Sci. Bull., 2005, 50, 2153.
3. Y. Takezawa and H. Imai, Small, 2006, 2, 390.
4. A. Antonello, M. Guglielmi, V. Bello, G. Mattei, A. Chiasera, M. Ferrari and A. Martucci, J. Phys. Chem. C, 2010, 114, 18423.
5. B. Zhao, F. Chen, X. Gu and J. Zhang, Chem.–Asian J., 2010, 5, 1546.
6. C. H. Lin, D. S. H. Wong and S. Y. Lu, ACS Appl. Mater. Interfaces, 2014, 6, 16669.
7. Q. Wu, X. Yang, J. Liu, X. Nie, Y. Huang, Y. Wen, J. Khan, W. U. Khan, M. Wu and T. An, ACS Appl. Mater. Interfaces, 2014, 6, 17730.
8. N. Sutradhar, A. Sinhamahapatra, S. K. Pahari, H. C. Bajaj and A. B. Panda, Chem. Commun., 2011, 47, 7731.
9. C. Wang, X. Zhang, Y. Zhang, Y. Jia, J. Yang, P. Sun and Y. Liu, J. Phys. Chem. C, 2011, 115, 22276.
10. P. Hareesh, K. B. Babitha and S. Shukla, J. Hazard. Mater., 2012, 229, 177.
11. X. Luan and Y. Wang, J. Mater. Sci. Technol., 2014, 30, 839.
12. C. Zhou, J. Luo, Q. Chen, Y. Jiang, X. Dong and F. Cui, Chem. Commun., 2015, 51, 10847.
13. T. W. Kim, I. Y. Kim, D.-H. Park, J.-H. Choy and S.-J. Hwang, Sci. Rep., 2016, 6, 21993 , Pages 1–12.
14. L. Thazhe, A. Shereef, S. Shukla, C. P. Reshmi, M. R. Varma, K. G. Suresh, K. Patil and K. G. K. Warrier, J. Am. Ceram. Soc., 2010, 93, 3642.
15. N. Harsha, K. V. Swathi Krishna, N. K. Renuka and S. Shukla, RSC Adv., 2015, 5, 30354.
16. A. Ghauch, A. M. Tuqan, N. Kibbi and S. Geryes, Chem. Eng. J., 2012, 213, 259.
17. S. Shukla and M. A. Oturan, Environ. Chem. Lett., 2015, 13, 157.
18. I. Morjan, F. Dumitrache, R. Alexandrescu, R. Birjega, C. Fleaca, I. Voicu, L. Gavrila, I. Soare, G. Filoti, V. Kuncser, G. Prodan, V. Ciupina and L. Vekas, Nanotech-2007: Nano Science and Technology Institute (NSTI) Conferenc and Expo, Sanata Clara, California, U.S.A., 2007, vol. 4, pp. 234–237.
19. M. Jose, M. P. Haridas and S. Shukla, J. Environ. Chem. Eng., 2014, 2, 1980.
20. M. Jose, N. Harsha, K. Suhailath, A. Peer Mohamed and S. Shukla, J. Environ. Chem. Eng., 2016, 4, 1295.
21. D. Wu, J. Liu, X. Zhao, A. Li, Y. Chen and N. Ming, Chem. Mater., 2006, 18, 547.
22. T. Chen, H. Xu, Q. Xie, J. Chen, J. Ji and H. Lu, Earth Planet. Sci. Lett., 2005, 240, 790.
23. N. A. M. Barakat, J. Mater. Sci., 2012, 47, 6237.
24. W. Cai and J. Wan, J. Colloid Interface Sci., 2007, 305, 366.
25. K. S. W. Sing, D. H. Everett, R. Haul, L. Moscou, R. A. Pierotti, J. Rouquerol and T. Siemieniewska, Pure Appl. Chem., 1982, 54, 2201.
26. Z. Shi, S. Yao and C. Sui, Catal. Sci. Technol., 2011, 1, 817.
27. H. J. Song, S. You, X. H. Jia and J. Yang, Ceram. Int., 2015, 41, 13896.
28. Y. Yao, S. Miao, S. Liu, L. P. Ma, H. Sun and S. Wang, Chem. Eng. J., 2012, 184, 326.
29. L. Ai, C. Zhang and Z. Chen, J. Hazard. Mater., 2011, 192, 1515.
30. M. Namvari and H. Namazi, Int. J. Environ. Sci. Technol., 2016, 13, 599.
31. B. Beless, H. S. Rifai and D. F. Rodrigues, Environ. Sci. Technol., 2014, 48, 10372.
32. Y. Li, Q. Du, T. Liu, X. Peng, J. Wang, J. Sun, Y. Wang, S. Wu, Z. Wang, Y. Xia and L. Xia, Chem. Eng. Res. Des., 2013, 91, 361.
33. O. G. Apul, Q. Wang, Y. Zhou and T. Karanfil, Water Res., 2013, 47, 1648.
34. G. Ersan, Y. Kaya, O. G. Apul and T. Karanfil, Sci. Total Environ., 2016, 565, 811.
35. S. Azizian, J. Colloid Interface Sci., 2004, 276, 47.
36. S. Lagergren, K. Sven. Vetenskapsakad. Handl., 1898, 24, 1.
37. Y. S. Ho, J. Hazard. Mater., 2006, 136, 681.
38. I. Langmuir, J. Am. Chem. Soc., 1916, 38, 2221.
39. H. M. F. Freundlich, Z. Phys. Chem., 1906, 57, 385.
40. M. M. Dubinin and L. V. Radushkevich, Proc. Acad. Sci. USSR, Phys. Chem. Sect., 1947, 55, 331.
41. C. K. Lee, K. S. Lin, C. F. Wu, M. D. Lyu and C. C. Lo, J. Hazard. Mater., 2008, 150, 494.
42. A. Millan, A. Urtizberea, N. J. O. Silva, F. Palacio, V. S. Amaral, E. Snoeck and V. Serin, J. Magn. Magn. Mater., 2007, 312, L5.
43. W. Konicki, D. Sibera, E. Mijowska, Z. Lendzion-Bielun and U. Narkiewicz, J. Colloid Interface Sci., 2013, 398, 152.
44. Y. R. Zhang, S. Q. Wang, S. L. Shen and B. X. Zhao, Chem. Eng. J., 2013, 233, 258.
45. W. Chu and C. C. Wong, Water Res., 2004, 38, 1037.
46. J. Y. T. Chan, S. Y. Ang, E. Y. Ye, M. Sullivan, J. Zhang and M. Lin, Phys. Chem. Chem. Phys., 2015, 17, 25333.
47. A. Lopez, A. Bozzi, G. Mascolo and J. Kiwi, J. Photochem. Photobiol., A, 2003, 156, 121.
48. Y. Ding, L. Zhu, J. Yan, Q. Xiang and H. Tang, J. Environ. Monit., 2011, 13, 3057.
49. X. Gu, S. Lu, L. Li, Z. Qiu, Q. Sui, K. Lin and Q. Luo, Ind. Eng. Chem. Res., 2011, 50, 11029.
50. K. C. Huang, R. A. Couttenye and G. E. Hoag, Chemosphere, 2002, 49, 413.
51. H. Narayani, H. Arayapurath and S. Shukla, Catal. Lett., 2013, 143, 807.

---

Footnote

† Electronic supplementary information (ESI) available. See DOI: 10.1039/c6ra14902b

---