Yolk–shell structured magnetic mesoporous silica: a novel and highly efficient adsorbent for removal of methylene blue

In this study, a novel magnetic mesoporous silica with yolk–shell structure (Fe3O4@Void@m.SiO2) was successfully synthesized via a polymer-template assisted method. The Fe3O4@Void@m.SiO2 was characterized by using FT-IR, EDS, SEM, TEM, VSM, PXRD and nitrogen adsorption–desorption analyses. The Fe3O4@Void@m.SiO2 nanocomposite showed high efficiency in adsorption of an organic dye and water pollutant called methylene blue (MB) with 98.2% removal capability. Furthermore, the effect of different parameters in the adsorption of MB was investigated. Different models of kinetic were examined and compared with each other. The recoverability and reusability of designed Fe3O4@Void@m.SiO2 material were also studied under applied conditions.

After the first report about interesting materials with yolk-shell (YS) structure 1 , many researches have developed the synthesis methods and properties of these materials with different names such as nanorattle 2 , movable core/ shell 1 , core/shell with hollow interiors 3 and yolk/shell 4 . These nanomaterials are in the center of attention due to interesting properties such as high surface area, interstitial hollow space and low density 5,6 . These special properties make yolk-shell nanocomposites suitable to use in the fields of biomedical 7 , lithium batteries 8,9 , sensors 10,11 , catalysis [12][13][14][15][16] and adsorption 17,18 . Some of recently reported yolk-shell structured materials are Sn 4 P 3 @C 19 , Sn@ SnO/SnO 2 20 , Al@TiO 2 NPs 21 and Au-CeO 2 @ZrO 2 22 . Among different shells, silica-based ones is more important in catalysis and adsorption processes owing to its high loading capacity compared to other shells 23,24 . The preparation methods effect on features of the yolk, space, thickness, porosity and shape of YS materials. The more common methods used for the preparation of YSs are hard temple-assisted, soft template-assisted and templatefree 25 . Among these, the soft template method has attracted more attention due to easy removal of template and also economically friendly 26,27 .
Moreover, owing to superparamagnetic properties, easy separation and low toxicity, magnetic nanoparticles (MNPs) have been so interested in different fields such as biomedical 28,29 , magnetic resonance imaging 30 , drug delivery 31 and separation 32 . However, the most of MNPs suffer from disadvantages of aggregation, biodegradation and low capacity. In order to increase the stability and capacity, these NPs are composited with different species. Among different species, mesoporous silica is more attracted because of its high surface area, high pore volume and high capacity [33][34][35] .
As regards to water purification importance and lots of different pollutants that made by pharmaceutical, paper making, textile, leather, etc., different methods such as photocatalytic degradation, adsorption and oxidation have been used to eliminate water pollutants 36,37 . To date many adsorbents have been applied for the elimination of organic and inorganic pollutants from water. Among these, magnetic NPs have attracted more attention due to the advantages of easy magnetically separation, economically friendly and high efficiency. Some of recently developed systems are magnetic zeolites 38 , magnetic carbon nanotubes composites 39 , Zn/ferrite/ graphene oxide 40 and activated carbon/NiFe 2 O 4 41 . Although these adsorbents gave good to high efficiency in the removal of water pollutants, however, in spite of high adsorption capacity of the yolk-shell based magnetic nanocomposites, according to our knowledge, there is no report in the study of efficiency of these type nanomaterials in the removal of water pollutants. In view of the above, herein, a novel magnetic Fe 3 O 4 @mesoporous silica nanocomposite (Fe 3 O 4 @Void@m.SiO 2 ) with yolk-shell structure and high magnetic properties is prepared for adsorption of methylene blue (MB) from water. The kinetics, isotherm and equilibrium data for adsorption of MB have been analyzed and different models have been employed to understand the mechanism.
Adsorption process. For this, 5 mg of the Fe 3 O 4 @Void@m.SiO 2 was added in 15 mL of an aqueous solution of MB. This mixture was shaken continuously and the Fe 3 O 4 @Void@m.SiO 2 was magnetically separated immediately. The MB concentration was measured by UV-vis at 665 nm. The adsorption efficiency for MB in water solution was calculated by using Eq. (1).
On the other hand, the amount of adsorbed MB on the adsorbent (q e ) was calculated by using Eq. (2).
Herein, C 0 and C e are, respectively, the concentration of the initial dye solution and the residual dye solution that quantitatively estimated by linear regression equations resulted at different dye concentrations, V is solution volume in liters and m is adsorbent amount (mg).

Results and discussion
The magnetic mesoporous silica material (Fe 3 O 4 @Void@m.SiO 2 ) with yolk-shell structure was prepared from an intermediate structure (Fe 3 O 4 @RF) which was made by polymerization of HCHO and resorcinol (Fig. 2) on the Fe 3 O 4 NPs surface. The mesoporous silica shell was coated on the Fe 3 O 4 @RF by using CTAB and TMOS via a sol-gel approach. Finally, RF and CTAB were eliminated to deliver Fe 3 O 4 @Void@m.SiO 2 (Fig. 1).   Fig. 3.
As shown, the designed material has particles with spherical morphology and average size of 84 nm. These types of NPs are excellent candidates for adsorption processes. The TEM image of the Fe 3 O 4 @Void@m.SiO 2 nanomaterial also showed a yolk-shell structure with black cores (Fe 3 O 4 NPs) and mesoporous silica shell (Fig. 4).
The low-angle PXRD (LA-PXRD) pattern of Fe 3 O 4 @Void@m.SiO 2 showed an intense peak centered at 2 theta of 1 degree that is characteristic of nanomaterials with an ordered 2D hexagonal mesostructure (Fig. 5). These types of materials have high surface area and are very effective in the adsorption processes.
The wide-angle PXRD of Fe 3 O 4 @Void@m.SiO 2 showed six reflection peaks at 2θ of 63°, 57°, 54°, 43°, 35° and 30° (Fig. 6). This pattern is in good agreement with the PXRD pattern of Fe 3 O 4 NPs confirming high stability of the magnetite cores during material preparation. This analysis also showed a broad peak at 2 theta about 20 degree that is attributed to mesoporous silica shell 42 . Figure 7 shows the vibrating sample magnetometer (VSM) analysis of the designed nanomaterial. As shown, this material has good magnetic property. As expected, due to the nonmagnetic mesoporous silica shell and the large void space, the magnetization is reduced from 60 for Fe 3 8B). Interestingly, the latter peaks are disappeared in Fig. 8C confirming successful removal of CTAB and RF during calcination. It is important to note that, to prepare a material with mesoporous yolk-shell structure, both CTAB (containing aliphatic C-H) and RF (containing aromatic C-H) should be romed. After removal of CTAB a mesoporous silica shell is resulted, while after removal of RF the void space between shell and core is resulted.   www.nature.com/scientificreports/ The EDS was performed for elemental analysis of the material before adsorption process (Fig. 9). This analysis clearly showed the existence of oxygen, iron and silicon confirming high stability and well incorporation of expected magnetite cores and silica shells in the material network.
The nitrogen adsorption-desorption isotherm of the Fe 3 O 4 @Void@m.SiO 2 nanomaterial before adsorption process shows a type IV isotherm with an H2 hysteresis loop, which is characteristic of ordered mesostructures with high regularity (Fig. 10). The sharp capillary condensation steps occurred at a relative pressure of 0.45-0.97, indicate the large void space of Fe 3 O 4 @Void@m.SiO 2 and the porous silica shell 43 . The Brunauer-Emmett-Teller (BET) surface area and total pore volume of the material were also found to be 666.16 m 2 /g and 1.28 cm 3 /g, respectively. The Barrett-Joyner-Halenda (BJH) pore size distribution isotherm showed a bimodal size distribution at 7.1 and 12.2 nm related to the mesoporous shell and void space between yolk and shell, respectively (Fig. 11). These data are in good agreement with LA-PXRD and TEM results confirming the presence of a mesoporous shell and yolk-shell structure for the designed nanomaterial.
Adsorption Studies. The elimination of MB, as organic dye, was examined as a model to study adsorption ability of synthesized Fe 3 O 4 @Void@m.SiO 2 material. The time, dye concentration, amount of adsorbent and adsorption pH were optimized as well as different models were checked in kinetic and isotherm studies.
Effect of pH. The effect of pH was investigated due to the structure and ionization degree of the MB dye can be influenced by pH value 44 . The pKa of MB is 3.8 45 , therefore, at pH values above this, the preponderant MBs are cationic. The effect of pH values was studied during 5 min using 0.005 g of the adsorbent (Fig. 12). At acidic pH, the -OHs of mesoporous silica shell and magnetic core interact with MBs via H-bonding. At basic pH, Si-   (Fig. 13). Similar results have been reported for MB elimination using clay 46,47 and activated coal 48 . Another important factor for this adsorption process is H-bonding interactions between N-sites of dye and OH sites of adsorbent. It is important to note that the aforementioned interactions can occur on both inner and outer surfaces of mesoporous silica shell and also on outer surface of magnetite cores 49,50 . According these findings, a suitable mechanism for this adsorption process is proposed in Fig. 13. According to this experiment, at pH 9.0 the best result was obtained. Therefore, this was used as optimum pH in the subsequent tests.
Effect of adsorbent dosage and dye concentration. In next step, the amounts of adsorbent and dye concentration were optimized (respectively, Figs. 14 and 15). It is clear that the adsorbent amount and the numbers of active positions on its surface affect the adsorption rate. By increasing the amount of adsorbent, the vacant and unoccupied positions are increased and thus the percentage of removal is increased. While, after this, there is a little change in the adsorption process. As shown in Fig. 14, the optimum amount of adsorbent is 0.005 g for 15 mL of MB aqueous solution (5 ppm) with 98.2% removal. This confirms high performance of the designed material in the MB removal. In the next, the effect of dye concentration using a constant amount (0.005 g) of adsorbent was studied. As shown in Fig. 15, by increasing the amount of MB concentration, the removal performance of a certain amount of Fe 3 O 4 @Void@m.SiO 2 is reduced which it refers to disproportionate in amount of active sites  Effect of time. Figure 16 illustrates the time effect under optimum conditions. As seen, the rate of adsorption is high at first 3 min that can refer to adsorption by the outer surface of the shell. After that, this process is slowly increased with time. The latter adsorption can be happened on internal surface of the shell and also surface of the core. As shown, after 15 min the maximum adsorption is resulted.
Kinetics evaluation. To study the kinetics evaluation, the kinetic models of pseudo-first-order, second-order and Elovich were employed ( Table 1). The pseudo-first-order model is as follows: where q t is the adsorbed dye at t time (mg/g); q e is adsorbed dye at equilibrium (mg/g) and k 1 is the rate constant of adsorption (min −1 ). The pseudo-second-order model is also as follows: (3) log(q e − q t ) = log q e − k 1 t 2.303 Figure 11. BJH pore size distribution isotherm of the Fe 3 O 4 @Void@m.SiO 2 nanomaterial before adsorption process. www.nature.com/scientificreports/ the q e and k 2 were obtained via the plot of t q t versus t (Fig. 17, Table 1). The Elovich equation is also as follows:    Table 1 and Fig. 17, the pseudo-second-order model with a linear regression correlation coefficient (R 2 ) value of 0.997 is completely applicable for the process.
Isotherm study. To determine the maximum capacity of sorption, the Langmuir, Freundlich and Temkin isotherm models were employed. The Langmuir model is as follows: where q e is the equilibrium concentration of dye, q m is the maximum dye uptake and K L is Langmuir constant. These are determined by linearizing of Eq. (6) as shown in Eq. (7), (6) q e = q m K L C e 1 + K L C e  The Freundlich model is also as follows: The linear form of above equation is Eq. (9) in which K f and n are Freundlich constants The other isotherm model is Temkin with a linear form as follows: As shown in Table 2 and Fig. 18, the Langmuir model with an R 2 > 0.99 is the best isotherm for the process.
Thermodynamic study. Thermodynamic parameters (ΔG 0 , ΔH 0 and ΔS 0 ) were calculated by using following equations 51 : These parameters are listed in Table 3. The plot of ln K c versus 1/T is illustrated in Fig. 19. As shown, a satisfactory adsorption is resulted at RT and by increasing the temperature from 25 to 55 °C, a slight decrease in www.nature.com/scientificreports/ MB adsorption is observed indicating the process is exothermic. This was also confirmed by the negative ΔH°. The negative amount of ΔG 0 confirms that the MB adsorption on Fe 3 O 4 @Void@m.SiO 2 is achievable and spontaneous. The negative ΔS° also suggests a decrease in randomness at solid/solution interface 52 . Moreover, the amounts of ΔH 0 and ΔG 0 successfully indicate that MB molecules are both physically and chemically adsorbed into/onto adsorbent material.
EDS spectrum. To investigate the successful adsorption of MB molecules into/onto Fe 3 O 4 @Void@m.SiO 2 , the EDS analysis after adsorption process was performed (Fig. 20). As shown, the existence of new peaks of carbon, nitrogen, sulfur and chlorine in this spectrum confirms successful adsorption of MB molecules into/onto material.
Nitrogen adsorption-desorption and BJH pore size distribution isotherms. The nitrogen adsorption-desorption analysis of the Fe 3 O 4 @Void@m.SiO 2 nanomaterial after adsorption of MB was performed (Fig. 21). According to this analysis, the BET surface area and total pore volume of the Fe 3 O 4 @Void@m.SiO 2 nanomaterial after adsorption process were reduced to 142.32 m 2 /g and 0.35 cm 3 /g, respectively. The BJH pore size distribution isotherm after adsorption process also showed that the sizes of shell pores and void space between yolk and shell are 5.1 and 7.9, respectively (Fig. 22). Moreover, as shown in the Figs. 21 and 22, after the adsorption process the inten-   Recoverability and reusability studies. In next step, the recoverability and reusability of the Fe 3 O 4 @Void@m. SiO 2 were investigated. To do this, the adsorbed MB molecules on Fe 3 O 4 @Void@m.SiO 2 were desorbed by acidic ethanol (pH 2). Then, the adsorbent was recovered and reused under the same conditions as the first run. As  Comparison study. Next, the adsorption efficiency of our designed Fe 3 O 4 @Void@m.SiO 2 nanomaterial was compared with some reported adsorbents in MB adsorption ( Table 4). As shown, in the most of previous studies the adsorbent is not recovered and also in some cases higher temperature than RT is needed for absorption. Moreover, the time required for absorption in most of former reports is high. These findings successfully confirm higher tendency and excellent capacity of the Fe 3 O 4 @Void@m.SiO 2 for adsorption of MB molecules in comparison to previous adsorbents.

Conclusion
In summary, a magnetic yolk-shell structured nanomaterial with mesoporous shell (Fe 3 O 4 @Void@m.SiO 2 ) was prepared, characterized and used as an effective adsorbent for the removal of MB dye from aqueous solution. The adsorption study showed that the designed material is so efficient with high removal capacity. Optimum conditions to achieve maximum removal of 98.2% for adsorption process were 0.005 g of adsorbent, pH 9 and 15 min for 15 mL of 5 mg/mL MB solution. The kinetic and isotherm studies of adsorption process were investigated and different models were evaluated for the equilibrium data. The results showed that the Langmuir isotherm model and pseudo-second-order kinetic are successfully fitted. The maximum adsorption capacity of the material was