Hierarchical multi-shell hollow micro–meso–macroporous silica for Cr(VI) adsorption

The development of easier, cheaper, and more effective synthetic strategies for hierarchical multimodal porous materials and multi-shell hollow spheres remains a challenging topic to utilize them as adsorbents in environmental applications. Here, the hierarchical architecture of multi-shell hollow micro–meso–macroporous silica with pollen-like morphology (MS-HMS-PL) has been successfully synthesized via a facile soft-templating approach and characterized for the first time. MS-HMS-PL sub-microspheres showed a trimodal hierarchical pore architecture with a high surface area of 414.5 m2 g−1, surpassing most of the previously reported multishelled hollow nanomaterials. Due to its facile preparation route and good physicochemical properties, MS-HMS-PL could be a potential candidate material in water purification, catalysis, and drug delivery. To investigate the applicability of MS-HMS-PL as an adsorbent, its adsorption performance for Cr(VI) in water was evaluated. Important adsorption factors affecting the adsorption capacity of adsorbent were systematically studied and Kinetics, isotherms, and thermodynamics parameters were computed via the non-linear fitting technique. The maximum capacity of adsorption computed from the Langmuir isotherm equation for Cr(VI) on MS-HMS-PL was 257.67 mg g−1 at 293 K and optimum conditions (pH 4.0, adsorbent dosage 5.0 mg, and contact time 90 min).


Results and Discussion
Synthesis of MS-HMS-PL. MS-HMS-PL was synthesized by a simple soft-templating hydrothermal method. The formation of multi-shell MS-HMS-PL implies the presence of multi-lamellar CTAB hollow spheres (Fig. 1, the formation of CTAB lamellar layers). The self-assembly of amphiphilic CTAB molecules in basic aqueous ammonia solution and the presence of 1-pentanol as co-surfactant leads to the formation of micelles and closed multi-layer hollow aggregates. The role of CTAB molecules is a structure-directing agent where in-situ polymerization of TEOS (in the basic condition) on the surface of preformed lamellar layers constructed by CTAB molecules leads to a hybrid organic-inorganic wall. After hydrothermal condition and complete polymerization of TEOS between CTAB layers (as depicted in Fig. 1), subsequent calcination was applied to remove the structure-directing agent and produce a stable MS-HMS-PL. In summary, we have developed a simple synthesis approach to the preparation of novel hierarchically architectured multi-shell hollow micro-meso-macroporous silica with pollen-like morphology. This material may have the potential for many uses in the fields of adsorption, analytical extraction, catalysis, carries, and drug release with prolonged-release time.

Structure characterization of MS-HMS-PL. XRD analysis.
To understand the crystalline structure of the MS-HMS-PL, L-XRD pattern of the sample is investigated. The L-XRD pattern of the MS-HMS-PL is shown in Fig. 2a. This silica material revealed only a broad diffraction peak at 2 θ 1° to 4°, implying the low ordering of silica framework in MS-HMS-PL.
FT-IR measurement. The FT-IR spectrum of MS-HMS-PL was recorded and is shown in Fig. S1. All characteristic absorption bands related to the nanoporous silica framework were observed for MS-HMS-PL (at 465 cm −1 : bending mode of Si-O-Si; at 802 cm −1 : symmetric mode of Si-O-Si vibration; at 964 cm −1 : stretching mode of Si-OH; at 1088 and 1235 cm −1 : asymmetric Si-O-Si vibration; 1626 cm −1 : bending mode of surface adsorbed H 2 O; 3247 cm −1 : O-H stretching modes of silanol groups) which were in close agreement with previously reported silica-based materials with nanoporous structures 9,19,20 . On a literature basis, a broad intense band between 3100-3750 cm −1 is attributed to the different silanol species (3650-3200 cm −1 : O-H stretching modes of strongly hydrogen-bonded sialons in the silica framework; above 3650 cm −1 : O-H stretching modes of isolated or free sialons) 21 .
Nitrogen adsorption-desorption isotherms. Nitrogen adsorption-desorption analysis was conducted with the aim of further investigating the structural feature of MS-HMS-PL. According to the IUPAC classification, MS-HMS-PL showed a combination of Type I(b) and IV isotherms with a Type H3 hysteresis loop at a relative pressure above 0.5 (Fig. 2b), suggesting the adsorption of N 2 molecules in the hollow voids (macropores, Fig. 1) [22][23][24] . Type I(b) isotherms are representative of materials having pore size distributions in the range of wider micropores (<2 nm) and possibly narrow mesopores (<∼2.5 nm). Type IV isotherms are given by mesoporous materials. The BET surface area and total pore volume measured to be 414.5 m 2 g −1 and 0.39 cm 3 g −1 for MS-HMS-PL (Table S1). Using the Barrett-Joyner-Halenda (BJH) method, the average pore sized of MS-HMS-PL is estimated to be 1.22 nm. Accordingly, the macropores are related to the hollow space in the middle of the MS-HMS-PL sub-microspheres, and the micro-/mesopores are located within the multi-lamellar layers of MS-HMS-PL. The mesopores with the pore size in the range of 2-4 nm are also observed in the BJH pore size distribution curve. All of the above results clearly indicate that MS-HMS-PL sub-microspheres possess a trimodal hierarchical pore architecture including micropores (the major constituent of lamellar shells), mesopores (the minor constituent of lamellar shells), and macropores (in the middle of sub-microspheres). This trimodal hierarchical pore structure is given in a schematic representation in Fig. 1.
Some typical hollow single, double, and multi-shell nanomaterials reported in the last decade are summarized in Table 1, and their properties including their composition, synthesis method, surface area, and pore size characteristics are also given. The BET surface area of MS-HMS-PL was 414.5 m 2 g −1 , surpassing most of the previously reported hollow nanomaterials. Furthermore, as shown in Table 1, unlike the reported materials in which mesopores are the major constituent of structure, the major constituent of lamellar shells in MS-HMS-PL sub-microspheres is micropores. With its trimodal hierarchical architecture and relatively high surface area, MS-HMS-PL could be a potential candidate for adsorption, extraction, catalysis, and drug delivery applications.

FE-SEM and TEM micrographs.
The FE-SEM micrographs of the MS-HMS-PL as shown in Fig. 2c-e revealed that this material is highly dispersed and possesses uniform pollen-like morphology. The magnified FE-SEM image (Fig. 2e) gave further information about the surface of the MS-HMS-PL with a wrinkled surface which can increase the accessible external surface area compared to the flat smooth surface. The TEM image of MS-HMS-PL (Fig. 2g) clearly shows that the particles possess a hollow spherical structure with a multi-lamellar shell. The schematic structure representation of MH-HMS-PL is shown in Fig. 2f.
Application of MS-HMS-PL as an adsorbent in the removal of Cr(VI) from aqueous solution. The important factors including pH, the dosage of adsorbent, contact time, initial Cr(VI) concentrations, and temperature, affecting the adsorption performance of MS-HMS-PL were systematically monitored. The adsorption kinetics, isotherms, and thermodynamics of Cr(VI) on MS-HMS-PL in aqueous solutions were investigated by conducting a series of kinetics, isotherms, and thermodynamic studies. The experimental adsorption data were fitted to the various kinetics and isotherms by the nonlinear method. According to the literature, the nonlinear method is a better way to obtain the kinetics and isotherm parameters because the linearization www.nature.com/scientificreports www.nature.com/scientificreports/ of nonlinear kinetic and isotherm expressions usually distort the fit by changing the error distributions 5-7,25-27 . The nonlinear forms of different kinetic (Eqs. 5-8) and isotherm models (Eqs. 9-11), as well as thermodynamic equations (Eqs. [12][13][14], and their parameters are tabulated in Table 2. pH and adsorbent dosage. The influence of pH on the removal percentage of Cr(VI) at three different dosages of adsorbent was tested and is shown in Fig. S2. From Fig. S2, the maximum removal percentage achieved at pH 4.0 and exceeded ~93% at all three dosages of the adsorbent. At this solution pH, the removal percentages of ~93%, ~97%, and ~99% were obtained for adsorbent dosages of 2.0, 5.0, and 8.0 mg, respectively. Accordingly, pH 4.0 and adsorbent dosage of 5.0 mg were used for next adsorption studies as optimum conditions. The removal percentage decreased continuously with the increase of pH values and ~72% Cr(VI) was removed at pH 8.0 and 5.0 mg adsorbent dosage. Decreasing the solution pH below 4.0 leads to a decrease in the removal percentage and the adsorbent removed ~67% Cr(VI) at pH 2.0 and 5.0 mg adsorbent dosage. According to the literature, silica-based adsorbents are deprotonated above pH 4.5-5.0 10,28,29 and this may be the reason for decreasing the  Fig. 3b. The experimental adsorption capacity after reaching equilibrium (Q e,exp /mg g −1 ) for certain initial concentrations are given in Table 3.
When the initial concentration of Cr(VI) was 5 mg L −1 , the adsorption capacity was 19.89 mg g −1 (%Removal = 99.46%). The adsorption capacity increased to 270.14 mg g −1 (%Removal = 33.77%) at an initial concentration of 200 mg L −1 . The adsorption of Cr(VI) occurs in three consecutive steps: (I) An initial fast step that lasts for 30 min (II) The slower adsorption process in the range from 30 to 90 min (III) An almost constant absorption step with very little variation in the adsorption rate where further increase in contact time does not reveal an increase in removal percentage of Cr(VI) or adsorption capacity of the adsorbent.
Consequently, 90 min was selected as optimum contact time to ascertain the equilibrium uptake of Cr(VI) to the MS-HMS-PL in the following experiments.
Kinetic and isotherm studies. In order to investigate the regulation of Cr(VI) uptake on MS-HMS-PL, four different kinetic models, viz., pseudo-first-order (PFO), pseudo-second-order (PSO), Elovich, and intra-particle diffusion (IPD), were used to fit the experimental data. Kinetics for the Cr(VI) adsorption onto MS-HMS-PL year Material Composition method S BET (m 2 g −1 ) D (nm) Ref. www.nature.com/scientificreports www.nature.com/scientificreports/ were systematically studied at eight different initial concentrations (5-200 mg L −1 ) and the results are given in Table 3. The corresponding curves in Fig. 3c are the PFO, PSO, and Elovich models after non-linear fitting at an initial concentration of 200 mg L −1 (figure caption for details). From Table 3, for non-linear kinetic models, the R 2 values of the PSO kinetic model were higher than those of PFO and Elovich models at all the initial Cr(VI) concentrations. For the initial concentration of 200 mg L −1 , Table 3, the R 2 values of the PFO, PSO, and Elovich models were respectively 0.8905, 0.9893, 0.9470, implying that PSO kinetic model gives the best fit to the data. Also, the experimental adsorption capacity at equilibrium, Q e,exp = 270.14 mg g −1 , was very close to the calculated adoption capacity at equilibrium, Q e,cal = 278.55 mg g −1 , obtained from the PSO model at the initial concentration of 200 mg L −1 . Accordingly, the adsorption process of Cr(V) by MS-HMS-PL could be well represented by the non-linear PSO kinetic model for all eight concentrations. Fitting of the nonlinear PFO, PSO, and Elovich kinetic models for eight different initial concentrations is depicted in Fig. S3.

MLV-CS
IPD kinetic model (Eq. 8) was used so as to gain a better understanding of the mechanism involved in the Cr(VI) adsorption on the MS-HMS-PL. Figure 3d shows the three-stage-adsorption process in the plots of Q t versus t 0.5 for the adsorption of Cr(VI) on the adsorbent, implying that more than one mechanism affected the adsorption process. Adsorption of Cr(VI) by hierarchically architectured MS-HMS-PL occurs in the three consecutive stages according to  Figure S4 is the plot of Qt vs. t 0.5 at different initial Cr (VI) concentrations and corresponding IPD kinetic parameters are given in Table 3.
Parameters of adsorption isotherms for a target adsorbate are of great significance for evaluating and predicting the adsorption behavior of an adsorbent. Two two-parameter isotherm models (namely the Langmuir and Freundlich models; Table 1 Eqs. 9 and 10) and a three-parameter isotherm model (namely Redlich-Peterson model; eq. 11) were used to evaluate the equilibrium data for adsorptive removal of Cr(VI) anions by MS-HMS-PL. The isotherm curves for adsorption of Cr(VI) at 293 K are shown in Fig. 3e (figure caption for details) and corresponding isotherm parameters are tabulated in Table 4. For Cr(VI) adsorption on MS-HMS-PL, the Langmuir model provide the better approximation of R 2 (R 2 = 0.9784) than those of Freundlich (R 2 = 0.8588) and R-P models (R 2 = 0.9752). Moreover, the maximum adsorption capacity computed from the Langmuir model (Q m,cal = 270.14 mg g −1 ) was close to the maximum adsorption capacity achieved in this experiment (Q m,exp = k IPD : intra-particle diffusion rate constant (mg g −1 min −0.5 )C: intra-particle diffusion constant related to the thickness of the boundary layer (mg g −1 ) Isotherm equations  www.nature.com/scientificreports www.nature.com/scientificreports/ 257.67 mg g −1 ) at 293 K, further demonstrating that the Langmuir isotherm model is better to express adsorption isotherm for Cr(VI) onto MS-HMS-PL. Figure S5 discloses non-linear isotherm plots of Cr(VI) adsorption on the MS-HMS-PL at four different temperatures and their corresponding isotherm parameters are given in Table 4. The values of Q m,cal calculated for the Langmuir model for adsorption of Cr(VI) onto MS-HMS-PL were estimated to be 257.67 (Q m,exp =270.14 mg g −1 ), 267.99 (Q m,exp = 282.89 mg g −1 ), 275.55 (Q e,exp = 291.51 mg g −1 ), and 281, 18 mg g −1 (Q m,exp = 302.02 mg g −1 ) at 293, 303, 313, and 323 K, respectively.
A useful dimensionless parameter called separation factor (R L , Table 2) is introduced in Langmuir's model to suggest the type of adsorption process where the values of R L = 0, R L = 1, R L > 1, and 0 <R L < 1 imply that the adsorption process is irreversible, linear, unfavorable, and favorable, receptivity 8,24 . The values of R L for adsorption of Cr(VI) onto MS-HMS-PL, as shown in Fig. 3f, lie between zero and unity, indicative of a favorable adsorption process.
Distribution coefficient values (K d , mL g −1 , Eq. 4) could also be used to gain more insight into the favorability of the adsorption process under set conditions: K d ∼10 2 (∼500 mL g −1 ): acceptable; K d ∼10 3 (∼5000 mL g −1 ): very good; K d ≥ 10 4 (∼50000 mL g −1 ): outstanding 7,15,30,31 . In a removal process, the larger K d values, the more affective the solid adsorbent is at holding and capturing the adsorbate. The K d values for adsorption of Cr(VI) on MS-HMS-PL (Fig. 3f) were in the range of 2.04×10 3 -7. Effect of temperature and thermodynamic studies. The influence of solution temperature on the adsorption capacity of adsorbent for Cr(VI) is shown in Fig. 4. The adsorption capacity increased as the solution temperature increased (Table 5, Fig. 4). To systematically characterize the physicochemical behavior of Cr(V) removal and to get a better insight into the mechanism of adsorption, the thermodynamic parameters of adsorption, viz., ∆ . G ads o , H ads o ∆ . , and S ads o ∆ . , were estimated. These crucial parameters could be estimated from the temperature-dependent isotherms using the Eqs. (12)(13)(14) given in Table 1  www.nature.com/scientificreports www.nature.com/scientificreports/ reported that the correct approach to calculating the aforementioned constants for an adsorption system is to perform non-linear fitting of the isotherms (usually Langmuir model) of the adsorption at various temperatures. For these calculations, it is considered that the aqueous solution of Cr(VI) is very diluted to consider that γ (the coefficient of activity, Eq. 12) is unitary. Thermodynamic equilibrium constants (K e o , Eq. 12) for each temperature could be calculated from the corresponding K L values for each isotherm after fitting the Langmuir equation to the adsorption isotherms (Eq. 9). The values of ∆H ads o and ∆S ads o can be estimated from the slope and intercept of van't Hoff equation (the plot of In K e o versus 1/T in Eq. 13), respectively. The calculated thermodynamic parameters are given in Table 5. According to the literature, the ∆ Comparative study. A comparison of the optimal conditions and adsorption performance of MS-HMS-PL for the adsorption of Cr(VI) with those of other adsorbent is shown in Table 6. It is clear that MS-HMS-PL showed high adsorption capacity for Cr(VI), calculated from Langmuir fitting, as compared to different synthetic, natural, and modified adsorbents. Such an outstanding removal capacity of MS-HMS-PL for Cr(VI) may be directly attributed to its high surface area and trimodal micro-meso-macroporous architecture which may provide fast mass transfer for adsorbate.

conclusion
In this study, a hierarchical multi-shell hollow micro-meso-macroporous silica with pollen like morphology designated as MS-HMS-PL was synthesized via a facile soft-templating method. BET and TEM analyses revealed that MS-HMS-PL sub-microspheres possess a trimodal hierarchical pore structure including micropores (the major constituent of lamellar shells), mesopores (the minor constituent of lamellar shells), and macropores (in the middle of sub-microspheres). With its high surface area (414.5 m 2 g −1 ), multi-shell structure, and facile synthesis route, MS-HMS-PL cloud be a good candidate for some applications such as adsorption, extraction, catalysis, and drug delivery. Accordingly, the adsorption performance of MS-HMS-PL for removal of Cr(VI) from aqueous solution was systematically studied. The effect of important adsorption factors-viz., pH. adsorbent dosage,   www.nature.com/scientificreports www.nature.com/scientificreports/ contact time, initial Cr(VI) concentration, and temperature-on the adsorption performance of MS-HMS-PL for Cr(VI) removal was evaluated. Kinetics, isotherms, and thermodynamic investigations were conducted to determine the adsorption mechanism and important parameters of adsorbent, and the non-linear fitting method was applied to calculated the kinetics and isotherms parameters. The estimated maximum adsorption capacity, according to the Langmuir model, was found to be 257.67 mg g −1 at pH 4.0, solution volume of 40 mL, the adsorbent dosage of 5.0 mg, the contact time of 90 min, stirring rate of 190 rpm, and solution temperature of 293 K.

Synthesis of MS-HMS-PL.
In a typical synthesis route, in a 250 mL bottle made from high-density polypropylene (HDPP) with a leakproof screw cap, 1.800 g CTAB was dissolved in 90 mL of pure water with magnetically stirring at 293 K for 10 min. Then, 7.7 mL of NH 4 + OH − was added to the bottle and the mixture was stirred for 20 min. Afterward, 2.7 mL of 1-pentanol was added to the mixture and stirred for 15 min followed by adding 7.5 mL of TEOS. The mixture was stirred at 293 K for 2 h and then maintained at an electric oven (343 K) for 24 h. The obtained white gel was washed with pure water and ethanol (96%) and dried at 333 K for 24 h. Finally, the white powder was calcined at 823 K for 6 h so as to remove the organic template CTAB. After specific time intervals (5-200 min), the Cr(VI)-loaded adsorbent particles were separated by centrifugation (3500 rpm, 5 min). The adsorption capacities of adsorbent at any time t (Q t , mg g −1 ), the adsorption capacity at equilibrium (Q e , mg g −1 ), and removal percentage were determined according to the following formula:  www.nature.com/scientificreports www.nature.com/scientificreports/ where C i is the initial adsorbate concentration (mg L −1 ), C t is the adsorbate concentration at a certain time t (mg L −1 ), C e is the adsorbate concentration at equilibrium (mg L −1 ), V is the volume of the liquid phase (L), and W is the mass of the adsorbent (mg).
The values of distribution coefficient (K d , mL g −1 ) at equilibrium can be calculated from the following formula: d i e i e e characterization. Low angle X-ray diffraction (L-XRD) analysis, Fourier transform infrared (FT-IR) spectroscopy, nitrogen adsorption-desorption measurement at 77 K, field emission scanning electron microscopy (FE-SEM), and transmission electron microscopy (TEM), were used for the characterization of the sample. Also, flame atomic absorption spectroscopy (FAAS) was used for the determination of the Cr(VI) concentration in the aqueous solutions. The details of the apparatus used in this work are given in Supplementary Information (S1. Experimental section, Apparatus).