Poly(Amidoamine) Dendrimer Modified Superparamagnetic Nanoparticles as an Efficient Adsorbent for Cr(VI) Removal: Effect of High-Generation Dendrimer on Adsorption Performance

In this work, three different generations of Poly(amidoamine) dendrimers (PAMAM) decorated on magnetic Fe3O4 composites (Fe3O4@SiO2–G1, Fe3O4@SiO2–G3, Fe3O4@SiO2–G5) were fabricated and characterized by FTIR, XRD, TEM, TGA, VSM and XPS. The obtained composites were used for Cr(VI) removal. Batch adsorption studies showed that the adsorption reached equilibrium within 60 min, and the optimal pH was 3.0. The result of adsorption kinetics was simulated by the pseudo-second-order model. The adsorption equilibrium isotherm was well fitted with the Langmuir adsorption model. Furthermore, thermodynamics calculations revealed that the adsorption process was endothermic and spontaneous. Importantly, adsorption capacity of Cr(VI) obeyed the sequence of Fe3O4@SiO2–G1 < Fe3O4@SiO2–G5 < Fe3O4@SiO2–G3, 3 generation of PAMAM (G3) was the optimal for adsorption capacity of Cr(VI). The maximum theoretical Cr(VI) adsorption capacity (qm) of Fe3O4@SiO2–G3 was 334.45 mg/g, and removal ration remained above 89.5% after five cycles of adsorption–desorption. Thus, Fe3O4@SiO2–G3 is predicted to be an efficient adsorbent for the adsorption of Cr(VI) from aqueous solution, and the obtained results can help in the generation optimization during fabrication of dendrimer modified adsorbents.


Introduction
Water contamination arises from heavy metal ions has attracted great attention due to its high toxicity, non-degradability, carcinogenicity, and ecological unbalance [1][2][3]. Of the heavy metals, chromium is widely used in the fields of electroplating, cement production, paper processing and textile industry [4]. Chromium ion is harmful to the environment and human health because of its biological build-up and tremendous toxicity. Cr(III) and Cr(VI) are the primary Cr ion states in aqueous solution. The latter is more toxic and carcinogenic [5]. Therefore, the development of low cost materials and effective techniques for the removal of Cr(VI) is imperative.
A various number of water treatment processes like chemical precipitation [6], oxidation-reduction [7], ion exchange [8,9], electrodialysis [10], membrane filtration, photocatalysis [11][12][13] and adsorption [14,15] have been employed to decontaminate heavy metal pollution. Adsorption has been found out to serve as an efficient and economical technology for dealing with Cr(VI) due to its easy operation, low cost and environmentally friendly. To date, Numerous adsorbents including activated carbon [16], biological adsorbent [17], resin [18] and magnetic nanoparticles [19] have been extensively applied for Cr(VI) removal. Among these materials, magnetic nanoparticle Fe 3 O 4 is the most attractive due to its advantages of biocompatibility, cheap price and especially superparamagnetic characteristics [20], that is widely applied in the fields of electrode, drug delivery, catalyst, thermal therapy, adsorption and so on [21][22][23][24][25]. Magnetic separation adsorbent based on Fe 3 O 4 nanoparticle is outstanding due to its convenience and high efficiency. However, naked Fe 3 O 4 nanoparticle is easy to be oxidized in air and aggregated in solution, that reduces its adsorption ability and selectivity for heavy metal ions. Surface-modified with functional groups is the most efficient method to overcome these disadvantages [26]. The surface of Fe 3 O 4 nanoparticle is easy to be functionalized by silica [27], amine groups [28], oxygen-containing groups (carboxyl, hydroxyl) [29][30][31] and sulfur-containing groups (sulfhydryl, sulfonic acid) [32]. The existence of these substances can improve the stability of Fe 3 O 4 nanoparticles and prevent aggregation. Sharma et al. [27] designed a visible light active photocatalyst Fe 3 O 4 /ZnO/Si 3 N 4 for the degradation of dyes from aqueous solution, that Si 3 N 4 -as an effective support for metal oxides to avoid the agglomeration. In addition, functional groups can selectively adsorb specific heavy metal ions through complexation effect, electrostatic attraction or ion exchange, thus improving the adsorption selectivity and capacity. For instance, Alhokbany et al. prepared Schiff base polymer decorated magnetic Fe 3 O 4 composite (MPOP) for removing of Cd(II) and Hg(II) from aqueous solution, the interaction between N atom and metal ion dominated the adsorption of Cd(II) and Hg(II) [33]. Fu  ] for aqueous Pb(II) and Hg(II) removal, adsorption mechanism is that Hg(II) preferentially binds to sulfur functional groups and Pb(II) tends to be adsorbed by nitrogen groups in poly(C3N3S3) matrix [34].
Poly(amidoamine) dendrimer (PAMAM) is regarded as a kind of biopolymer, which is widely studied in many fields, such as drug delivery, gene therapy, ultrasensitive sensor, catalyst carrier and adsorption in wastewater treatment, because of its excellent biocompatibility, highly branched three-dimensional structure, lots of reactive terminal groups, and monodispersed molecular weight [35][36][37]. Tremendous amine groups located within its interior and its surface enable PAMAM to bind with heavy metal ions by complexation effect, ion exchange and electrostatic attraction [38]. Nevertheless, PAMAM dendrimer is easily soluble in aqueous media, which makes its separation/recovery difficult to recycle. So PAMAM dendrimer is usually anchored on support substances such as silica [39,40], graphene [41], and magnetic substance [42]. Furthermore, adsorption capacity is related to the dendrimer generation of PAMAM. Niu et al. found that the increase of adsorption capacity of Hg(II) and Ag(I) with dendrimer generation (G2 > G1 > G0.5), which is attributed to high content of sulfur group that high generation dendrimer possess [43]. However, as we all known that with the increase of generation, the steric hindrance increases, and serious surface defects of PAMAM appears, thus the adsorption capacity must be affected. As far as we know, the influence of high-generation dendrimer of PAMAM on adsorption capacity for Cr(VI) has not been studied.
In this work, three different generations of PAMAM dendrimer were immobilized on magnetic Fe 3 O 4 composites (Fe 3 O 4 @SiO 2 -G1, Fe 3 O 4 @SiO 2 -G3, Fe 3 O 4 @SiO 2 -G5) to combine the merits of the two components for the efficient removal of Cr(VI), and compared the adsorption capacity of Cr(VI) between the three different adsorbents. The effective factors on the adsorption capability such as adsorbent dosage, adsorption time, temperature, initial concentration and pH of sample solution were examined and optimized. In addition, the adsorption isotherms, adsorption kinetics, thermodynamic and reusability were investigated to evaluate the adsorption properties and adsorption mechanism.

Characterization
The synthetic materials were characterized by Fourier-transform infrared spectra (FTIR), X-ray powder diffraction (XRD), transmission electron microscopy (TEM), Thermogravimetric analysis (TGA), Vibrating sample magnetometer (VSM) and X-ray photoelectron spectroscopy (XPS). FTIR spectra were obtained on a Perkin-Elmer Spectrometer in the range from 4000 to 400 cm −1 . The crystalline structure of Fe 3 O 4 , Fe 3 O 4 @SiO 2 -G1, Fe 3 O 4 @SiO 2 -G3 and Fe 3 O 4 @ SiO 2 -G5 were characterized by XRD (PANalytical Empyrean, Netherlands) with Cu Ka radiation (k = 1.54178 Å) at a scanning rate of 2°/min in the 2θ range from 4° to 70°. The particle size and morphology of nanocomposites were observed by TEM (JEM-2010 (HT)). TGA measurements were performed on Diamond TG/DTA 6300 (Perkin-Elmer USA) at a heating rate of 5 °C/min under nitrogen. Magnetic hysteresis loops were characterized by VSM (Lakeshore 7404) at room temperature with a maximum external field of Hm = 200,000 Oe. The surface element composition and element valence of the adsorbent were characterized by XPS (Thermo Scientific K-Alpha) with Al Kα ray source and C 1s (284.80 eV) for calibration.

Preparation Procedure
The synthetic route of Fe 3 O 4 @SiO 2 -G1, Fe 3 O 4 @SiO 2 -G3 and Fe 3 O 4 @SiO 2 -G5 is presented in Scheme 1. Briefly, APTMS coated Fe 3 O 4 was synthesized according to the reported method [44,45]. About 3.0 g of APTMS coated Fe 3 O 4 and 5.0 mL of methyl acrylate were dispersed in 80 mL methanol under N 2 . The suspension was reacted at room temperature for 72 h. The resulting particles were washed with methanol and then reacted with 10.0 mL 1,2-diaminoethane in 100 mL methanol at room temperature for another 72 h. After that, Fe 3 O 4 @SiO 2 -G1 was obtained after separation and the subsequent washed with ethanol. The above reaction cycles (Michael addition and amidation reaction) were repeated until 3 generation of PAMAM dendrimer modified Fe 3 O 4 (Fe 3 O 4 @SiO 2 -G3) and 5 generation of PAMAM dendrimer modified Fe 3 O 4 (Fe 3 O 4 @SiO 2 -G5) were obtained. It was noteworthy that the addition amount of methyl acrylate and 1,2-diaminoethane increased with the increase of the dendrimer generation, which was kept at 50% excess and 2 times, respectively. The obtained composites were washed with methanol and dried in vacuum at 50 °C for 12 h.

Batch Adsorption Experiments
K 2 Cr 2 O 7 was dissolved with deionized water as a stock Cr(VI) solution (1000 mg/L). The Cr(VI) solutions with different initial concentrations were prepared by dilution of the stock Cr(VI) standard solution with deionized water. The effects of five parameters on the adsorption of Cr(VI) by Fe 3 O 4 @SiO 2 -G1, Fe 3 O 4 @SiO 2 -G3 and Fe 3 O 4 @SiO 2 -G5 were studied. The range of the tested parameters was as follows: initial pH (3-10), adsorbent dosage (4-16 mg), adsorption time (10-150 min), temperature (293-308 K), and initial Cr(VI) concentration (10-120 mg/L). For a typical example, when measuring the effect of the adsorption time, 10 mg of adsorbent was added into a 150 ml stoppered conical flask containing 50 ml of 50 mg/L Cr(VI) solution, with an initial pH of 3.0, and shaking with 150 rpm at 303 K. After designated time periods (10,15,20,30,40, 60, 90, 120, 150 min), the Cr(VI) solution samples were collected by a 5 mL syringe and filtered immediately through a 0.45 μm microfiltration membrane. The concentration of Cr(VI) was determined by atomic absorption spectrophotometry (AAS, AA240FS-GTA120, USA) [46]. The pH was regulated by HCl (0.1 M) and NaOH (0.1 M). All experiments were carried out triplicate in order to effectively reduce the experimental error.
The difference between the initial concentration and the adsorption equilibrium concentration is allowed to evaluate the amount of metal ion adsorbed. The equilibrium adsorption capacity (q e , mg/g) of the absorbent towards Cr(VI) is calculated by Eq. (1) [43]: where C 0 and C e (mg/L) are the initial and equilibrium concentrations of Cr(VI) in the solution, respectively, V (L) is the volume of the solution, and m (g) represents the weight of adsorbent.  Comparing the change of Fe 3 O 4 @SiO 2 -G3 after adsorption, it can be seen that Fe 3 O 4 @SiO 2 -G3 is also existed as isolated cluster with a nearly spherical shape, but the cluster size increases, with the size of about 70 nm (Fig. 2c). This phenomenon may be due to the electrostatic attraction between Cr(VI) and the adsorbent.
The magnetic hysteresis loops of Fe 3 O 4 and the nanocomposites are shown in Fig. 3, the coercivity and magnetic remanence are almost zero for them, indicating superparamagnetic nature of the adsorbents. The saturation magnetization is 38.55, 37.88 and 31.86 emu/g for Fe 3 O 4 @SiO 2 -G1, Fe 3 O 4 @SiO 2 -G3, and Fe 3 O 4 @SiO 2 -G5, which decreases by 4.90, 5.57, and 11.59 emu/g respectively as compared with Fe 3 O 4 , indicating that amounts of nonmagnetic matter is covered on the surface of Fe 3 O 4 magnetic core, which further proving the successful preparation of nanocomposites. The remaining saturation magnetization in nanocomposites is still adequate for their efficient separation by using an external magnetic field [45]. As shown in Fig. 3(inset), the black homogeneous dispersion of Fe 3 O 4 @SiO 2 -G3 is easily attracted to the wall of vial in the presence of an external magnetic field. The results demonstrate that as-synthetic nanocomposites can be rapidly separated and recovered from solution by simple magnetic separation owing to the strong superparamagnetism, thus facilitating manipulation in the practical applications.

Effect of Solution pH for Cr(VI) Removal
The solution pH has considerable effect on metal ion adsorption as it influences not only on the existing forms of heavy metal ion, but also the surface charge of an adsorbent. To investigate the optimal pH value for Cr(VI) removal, a series of experiments with different initial pH (3-10) were carried out. As presented in Fig. 4, With the solution pH increases from 3 to 10, the adsorption for Cr(VI) is gradually decreased for the three adsorbents. This is attributed to the different forms of chromium    [48]. The maximum Cr(VI) adsorption capacity is obtained at pH 3, which is partly attributed to the strong electrostatic interactions between anions (Cr 2 O 7 2− and HCrO 4 − ) and protonated imidazole N + in the three adsorbents. When the solution pH is greater than 3, due to the deprotonation of N and an increase in the number of bivalent anions (CrO 4 2− ), both enhance the electrostatic repulsion of the adsorbents and chromium ions, which decline the adsorption capacity for Cr(VI).
Furthermore, the adsorption capacity decreases by the sequence of Fe 3 O 4 @SiO 2 -G3 > Fe 3 O 4 @SiO 2 -G5 > Fe 3 O 4 @ SiO 2 -G1, the adsorption capacity of Fe 3 O 4 @SiO 2 -G3 and Fe 3 O 4 @SiO 2 -G5 are stronger than that of Fe 3 O 4 @SiO 2 -G1, which is attributed to lower content of amine groups that Fe 3 O 4 @SiO 2 -G1 possess. But interestingly we found that the adsorption capacity of Fe 3 O 4 @SiO 2 -G5 was relatively poor than that of Fe 3 O 4 @SiO 2 -G3. This may be due to excessively high generation not only causes the large sterically hindrance, which hinders the contact of chromium ions with the amine groups located within the interior of PAMAM, but also causes serious surface defects of PAMAM, that decreasing amine groups on the surface of Fe 3 O 4 @SiO 2 -G5. The above result indicates that the affinity of the composites increases with dendrimer generation increases only in low generation range, but excessively high generation of PAMAM is unfavourable for the adsorption. Above all, the optimal pH is 3.0, which is selected for the subsequent experiments. In addition, the removal of Cr(VI) is affected by generation of PAMAM dendrimer, 3 generation of PAMAM (G3) is the optimal for adsorption capacity of Cr(VI).

Effect of Adsorbent Dosage for Cr(VI) Removal
Adsorbent dose is an important factor which determines the presence of number of active sites on the surface of adsorbent [49]. The adsorption capacity of Cr(VI) varies with the adsorbent dosage. Figure 5 shows the effect of the mass of adsorbent on the adsorption of Cr(VI). It is observed that the adsorption capacity decreases with the increase in the adsorbent dose, the adsorption capacity of Cr(VI) by Fe 3 O 4 @SiO 2 -G3 decreases from 438.75 to 155.16 mg/g with the dosage increases from 4 to 16 mg, this effect can be explained by the active sites of adsorbent are not be completely occupied by chromium ions at high dosages of adsorbent. To consider operation consumption, while also ensuring a high adsorption capacity, the requirement of active sites was fulfilled at 10 mg dose for concentration of 50 mg/L, and q e was 239.75 mg/g.

Effect of Temperature for Cr(VI) Removal and Thermodynamic Study
The effect of temperature of Cr(VI) adsorption on adsorbent was investigated in the temperature range 293-308 K. As shown in Fig. 6a, it is noticed that an increase in temperature promotes an increase in the amount of Cr(VI) adsorbed by all adsorbents, which indicates that the adsorption process of Cr(VI) is endothermic. This is attributed to the increment in the kinetic energy of the adsorbent molecules with the increment in temperature, the increased activated sites created on the adsorbents, and the increase in kinetic motion of chromium ions. The different thermodynamic parameters of adsorption can reflect its spontaneity, randomness, endothermicity or exothermicity. Thermodynamic parameters such as standard free enthalpy ΔG 0 is calculated with Eq. (3), standard enthalpy ΔH 0 and standard entropy ΔS 0 are calculated from the slope and intercept of the plot of K d versus 1/T (Fig. 6b)

using Eq. (4) [50]:
where K d is the distribution constant; a (g/L) is the adsorbent dose; q e (mg/g) is the adsorption capacity at equilibrium; C e (mg/L) is the concentration at equilibrium of adsorbate in solution; ΔG 0 (kJ/mol) is the standard Gibbs energy change;

Kinetic Studies on Cr(VI) Removal
To investigate the adsorption rate of Cr(VI) on adsorbents and measure the adsorption equilibrium, the adsorption kinetics were studied. The adsorption kinetic plots of the adsorbents for Cr(VI) are illustrated in Fig. 7a. The adsorption of the adsorbent proceeds speedily during 0-40 min. Then, it conducts relaxedly and approaches equilibrium at 60 min. The equilibrium adsorption capacity of Fe 3 O 4 @ SiO 2 -G1, Fe 3 O 4 @SiO 2 -G3 and Fe 3 O 4 @SiO 2 -G5 for Cr(VI) is 150.80, 240.55 and 227.55 mg/g, respectively. The sharp increase in the initial stage is mainly due to the presence of abundant active sites on adsorbents and high concentration of Cr(VI), which facilitates the contact of active adsorption sites with chromium ions, leading to the rapid capture of chromium ions by the adsorbents [48]. When the adsorption time continues to extend, Cr(VI) concentration reduces, and the active adsorption sites of the adsorbent are largely occupied by chromium ions, decreasing the contact opportunity and resulting in the decrease of adsorption rate until approach equilibrium. To ensure the system reached adsorption equilibrium, the contact time was set to 60 min in the following experiments.
To explore the adsorption mechanism, the experimental data were fitted by kinetic equations, and the correlation coefficients were obtained. The two equations represent the pseudo-first-order kinetic model and pseudo-second-order kinetic model, which can be described below in Eqs. (5) and (6) [51]: Pseudo-first-order equation:  Pseudo-second-order equation: where q e (mg/g) is the equilibrium adsorption amount of Cr(VI), q t (mg/g) is the adsorption amount of Cr(VI) at time t (min), k 1 and k 2 are the equilibrium rate constant of pseudofirst-order adsorption and pseudo-second-order adsorption models, respectively. The fitting curves and related parameters of the pseudofirst-order model and pseudo-second-order model are shown in Fig. 7b, c and Table 2. The adsorption kinetic for the three adsorbents fit Pseudo-second-order model well as indicated by the higher correlation coefficients (R 2 ). The calculated equilibrium adsorption capacity (q e,cal ) that derived from pseudo-second-order model is more in agreement with the experimental value (q e,exp ), further confirms the reasonable of describing the adsorption kinetic process of Cr(VI) by pseudo-second-order model, suggesting that the chemical interactions are possibly involved in the adsorption processes. In the current work, an occurrence of chemical adsorption for the adsorbent can be understood as the complexation effect of chromium ions with the N-and O-containing groups on adsorbents. Furthermore, the generation of PAMAM with no impact on the adsorption mechanism.

Effect of Initial Concentration for Cr(VI) Removal and Adsorption Isotherms
The effect of the initial Cr(VI) concentration on the adsorption capacity of adsorbents for Cr(VI) was investigated. As shown in Fig. 8, with the increase of initial Cr(VI) concentration, the adsorption capacity of the three adsorbents gradually increase. The adsorption capacity of   of Cr(VI) increases from 10 to 120 mg/L. The enhanced adsorption for Cr(VI) by the raising initial concentration is due to the high concentration gradient under high concentration, which promotes the diffusion of chromium ions to the active sites of the adsorbent. Therefore, the removal for Cr(VI) by the adsorbent is increased.
Two adsorption isotherms (i.e., Langmuir and Freundlich model) were used to describe the equilibrium adsorption according to Eqs. (7) and (8) [52]: Langmuir adsorption isotherm: Freundlich adsorption isotherm: where q e (mg/g) and C e (mg/L) are the amounts of Cr(VI) absorbed on adsorbent and the Cr(VI) concentration at adsorption equilibrium, respectively. k L (L/mg) is the Langmuir constant related to the energy of adsorption, q m (mg/g) refers to the maximum adsorption capacity, k F and n are constants for the Freundlich isotherm. The fitting curves and relative parameters of each model are shown in Fig. 9 and Table 3. Compared with Freundlich model, Langmuir model provides better correlation coefficients (R L 2 > R F 2 ). This result demonstrates that the adsorption of Cr(VI) onto the nanocomposites sites follows a monolayer/homogenous pattern [53]. It is suggested that chromium ions are highly likely to interact with the active sites of the nanocomposites through a complexation process [43]. The maximum adsorption capacity (q m ) calculated by the Langmuir model is 245.10, 334.45, and 326.80 mg/g for Fe 3 O 4 @SiO 2 -G1,  Table 4. The results show that Fe 3 O 4 @SiO 2 -G3 has an excellent adsorption capacity for Cr(VI) and is an effective adsorbent.

XPS Study on Adsorption of Cr(VI) on Fe 3 O 4 @ SiO 2 -G3
To further investigate the adsorption mechanism, The XPS analysis of Fe 3 O 4 @SiO 2 -G3 before and after adsorption was carried out. The XPS spectra is shown in Fig. 10, the peaks of Fe, O, N, C, Si are appeared in the XPS spectrum before adsorption in Fig. 10a (Fig. 10c). As well as the N 1s consists of two fitting peaks of CONH and C-N groups, which centers at 400.39 and 398.68 eV, respectively (Fig. 10e).
After adsorption of Cr(VI), the peak of Cr 2p is presented in the XPS spectra (Fig. 10a), indicating the successful uptake of Cr(VI) onto Fe 3 O 4 @SiO 2 -G3. From the high-resolution XPS spectrum (Fig. 10b), the optimum fitting is achieved by deconvoluting Cr 2p spectrum into four peaks. The peaks at 576.38 eV and 586.23 eV are ascribed to Cr (III) 2p 3/2 and 2p 1/2 , respectively, while the peaks at 581.25 eV and 590.09 eV are belonged to Cr(VI) 2p 3/2 and 2p 1/2 , respectively. The results indicated that the chromium adsorbed onto the surface of Fe 3 O 4 @SiO 2 -G3 contained both Cr(III) and Cr(VI), because Cr(VI) was reduced to Cr(III) during adsorption.
The O1s peak of CONH changes from 531.76 to 531.90 eV after capture of Cr(VI) in Fig. 10d, but the O1s peaks of Si-O-Si and Fe-O have no change before and after adsorption. The N 1s peak of CONH group moves from 400.39 to 400.57 eV, while that of C-N group moves from 398.68 to 398.90 eV after adsorption (Fig. 10f). The above increasement of binding energy demonstrated the involvement of N and O for Cr(VI) adsorption by Fe 3 O 4 @SiO 2 -G3. Because there are five electrons in the outer layer of the N atom, three of which have been paired to form bonds, the remaining pair of lone pairs are easy to provide a lone pair of electrons to form a complex. N shares electrons with Cr in this process, resulting in a decrease in electron density of N and an increase in binding energy. With the aid of electron donors, a large amount of Cr(VI) was reduced from Cr(VI) to Cr(III), which was then adsorbed. Meanwhile, O may also share electrons with Cr to form a coordination complex [43]. According to the above considerations, the number of Nand O-containing functional groups is the critical factor that determine the adsorption property of Fe 3 O 4 @SiO 2 -G3.

Reusability
Reusability is an important parameter property for evaluating the performance of the adsorbent. The adsorption/ desorption process of Cr(VI) on Fe 3 O 4 @SiO 2 -G3 was investigated using 0.5 mol/L NaOH solution as eluent. The reusability property was determined by considering five adsorption-desorption cycles. Figure 11 shows that the Cr(VI) adsorption capacity of Fe 3 O 4 @SiO 2 -G3 gradually decreases with the number of cycles, but not significantly. After five cycles of use, the removal ration of Fe 3 O 4 @ SiO 2 -G3 for Cr(VI) remains above 89.5%, indicating that Fe 3 O 4 @SiO 2 -G3 has good stability and reusability.

Conclusions
In this work, three different generations of PAMAM dendrimer immobilized on magnetic Fe 3 O 4 composites (Fe 3 O 4 @SiO 2 -G1, Fe 3 O 4 @SiO 2 -G3, Fe 3 O 4 @SiO 2 -G5) were successfully prepared and used to remove Cr(VI). The solution pH of 3.0 is favored for Cr(VI) removal. Adsorption approaches equilibrium at 60 min for all the adsorbents. The results of adsorption kinetics and isothermal studies indicate that the adsorption behavior conform to both the pseudo-second-order kinetic model (chemical interaction) and Langmuir isotherm model (homogeneous monolayer adsorption). Furthermore, the thermodynamics model demonstrates that adsorption is endothermic and spontaneous. More interestingly, we found that the uptake of Cr(VI) obeyed the sequence of Fe 3 O 4 @SiO 2 -G1 < Fe 3 O 4 @ SiO 2 -G5 < Fe 3 O 4 @SiO 2 -G3, indicating that the affinity of the nanocomposites increases with dendrimer generation increases only in low generation range, but excessively high generation of PAMAM is unfavourable for the adsorption due to the large steric hindrance and serious surface defects of PAMAM. As-synthesized Fe 3 O 4 @SiO 2 -G3 exhibits competitive adsorption capacity (334.45 mg/g) for Cr(VI) than most alternative adsorbents mainly because of its regular structure and the abundant active sites located within its interior and its surface. Regeneration suggests the removal Fig. 10 The XPS total survey spectra of Fe 3 O 4 @SiO 2 -G3 (a), high resolution Cr 2p spectrum (b), high resolution O 1s spectrum (c, d) and high resolution N 1s spectrum (e, f) before and after adsorption of Cr(VI) ration of Fe 3 O 4 @SiO 2 -G3 for Cr(VI) remains above 89.5% after five cycles of adsorption-desorption. This work shows that Fe 3 O 4 @SiO 2 -G3 is an efficient and reusable adsorbent for the adsorption of Cr(VI) from aqueous solution, and the obtained results could be useful for the generation optimization during fabrication of dendrimer modified adsorbents.