Magnetic hydroxyethyl cellulose spheres with efficient congo red removal

Ecofriendly adsorbent materials for the rapid and efficient removal of pollutant dyes are highly desired on account of concerns about environmental pollution and human health. Herein, novel magnetic HC/Fe3O4 spherical materials have been constructed via crosslinking hydroxyethyl cellulose (HC) by poly(ethylene glycol) diglycidyl ether (PGDE) followed by the introduction of magnetic Fe3O4 by a facile and effective strategy developed in this work. The morphology, structure and magnetic behavior, point of zero-charge (pHzpc) and Brunauer-Emmet-Teller (BET) of the spherical materials have been systematically investigated. Further, the spherical materials were utilized to remove congo red (CR-SO3Na) from aqueous solution under varying adsorption conditions. Meanwhile, the adsorption kinetics, thermodynamics and isothermics have been achieved to explore the adsorption process and possible adsorption mechanism of CR-SO3Na by the spherical materials. The materials show not only an efficient capacity of CR-SO3Na removal from aqueous solution, but also a sufficient magnetic property of the recovery of the materials from aqueous solution after adsorption. The spherical materials have great potential to be used as efficient adsorbents for the removal of dye-containing effluent.


Introduction
Congo red is a water-soluble anionic benzidine-based diazo dye and extensively applied in textile, paper, paint, pigments, rubber and plastics industries [1][2][3]. Nevertheless, this dye is non-biodegradable and carcinogenic, resulting in environmental pollution and human health problems [4][5][6]. Hence, it is vital to remove the dye pollutant from waste water. Among various technologies (adsorption, membrane separation, ion exchange, reverse osmosis, oxidation/ reduction and chemical precipitation) to remove dye pollutants from wastewater, adsorption is the most frequently used technology due to its flexibility for design, ease of operation, low cost, high efficiency for dye removal and wide applicability [7][8][9]. Over the past years, cellulose, as the most abundant biorenewable resource on the earth with outstanding advantages such as availability, low-price, nontoxicity and biodegradability [10][11][12], has been applied in fabricating adsorbent materials for the removal of congo red (CR-SO 3 Na) from water/wastewater. However, pure cellulose material displays a low adsorption efficiency for pollutants [13]. To overcome the issue, cellulose is often composited or modified with additional components which are able to attract pollutant dye molecules. Modified cellulose adsorbents include polyacrylamide grafted quaternized cellulose [14], surface quaternized cellulose nanofibrils [15], CaCO 3 -decorated cellulose [16], polyethyleneimine modified magnetic microcrystalline cellulose [17], zwitterionic celluloses [18] and cetyltrimethylammonium bromide modified cellulose nanocrystals [19]. Cellulose composite adsorbents include Fe(OH) 3 @cellulose hybrid fibers [20], hyperbranched polyethyleneimine/cellulose nanofiber [21], Chitosan/cellulose [22], graphene oxide modified cellulose Yao Hui and Rukuan Liu have contributed equally to this work.
* Airong Xu airongxu@haust.edu.cn * Sisi Liu liusisi274@126.com nanocrystal/poly(N-isopropyl acrylamide) [23], cellulosechitosan [24], sodium alginate/cellulose nanofibers/polyethyleneimine [25] and cellulose/methylene bisacrylamide [26]. These cellulose based materials composited or modified with additional components can efficiently adsorb congo red pollutant from waste water. In spite of their strong adsorption capacity for congo red, their practical applications are often limited. For example, cellulose nanofiber and nanocrystals involve chemicals and complexity in the extraction process [27,28]. The modifications of cellulose frequently involve quite time-consuming preparation process. Additionally, the insolubility of natural cellulose in water and most of common solvents often makes the preparation of cellulose composites complex [29,30]. Hydroxyethyl cellulose, a cellulose derivative, has attracted considerable attention because it not only possesses excellent biocompatibility and biodegradability similar to natural cellulose, but also is water-soluble [31,32]. This solubility of hydroxyethyl cellulose facilitates it to be functionalized via replacing the hydroxyl hydrogen atoms in hydroxyethyl cellulose [33][34][35], and thus making its practical application further possible. Some attempts have been made in the application of hydroxyethyl cellulose in the fabrication of dye adsorbents to remove methylene blue dye, rhodamine 6G, crystal violet and methyl violet [36][37][38][39]. Recently, Jana et al. developed hydroxyethyl cellulose composites by crosslink copolymerization of a mixture of acrylamide and N,Ndimethylacrylamide using N,N′-methylene-bis-acrylamide as a cross-linking agent onto hydroxyethyl cellulose backbone, showing a high monolayer adsorption capacity (102.4 mg g −1 ) for the removal of congo red [40]. However, these hydroxyethyl cellulose-based adsorbents commonly suffer from complex preparation procedures, low adsorption capacities of dye, or inconvenient recovery after dyes adsorption. Therefore, it is very necessary to improve the usability of the hydroxyethyl cellulose-based adsorbents by developing a facile but effective strategy.
Therefore, in this present work, we designed novel and magnetic HC/Fe 3 O 4 spheres using hydroxyethyl cellulose (HC), poly(ethylene glycol) diglycidyl ether (PGDE) and magnetic Fe 3 O 4 , which was not reported previously to the best of our knowledge. The spherical materials could be readily fabricated at ambient conditions. PGDE was employed to mainly crosslink HC, and magnetic Fe 3 O 4 was introduced into the spherical materials for the convenience of efficient separation of the HC/Fe 3 O 4 spherical adsorbents from aqueous solutions after adsorptions [41]. Further, the spherical materials were used for the adsorption performances of CR-SO 3 Na, and batch adsorption experiments were carried out to systematically investigate the effects of adsorption factors (pH, temperature, dosage, contact time, and initial dye concentration) on the adsorption of the HC/ Fe 3 O 4 spheres towards CR-SO 3 Na. The adsorption process was investigated using isotherm and thermodynamic and kinetic models to explore possible adsorption mechanism of CR-SO 3 Na by HC/Fe 3 O 4 . The recovery and the reusability of the HC/Fe 3 O 4 adsorbents were also estimated.

Fabrication of the magnetic HC/Fe 3 O 4 spheres
The magnetic HC/Fe 3 O 4 spheres were fabricated using the following procedure. A HC aqueous solution was obtained by dissolving 0.48 g of HC in 15.2 g of deionized water. To this HC aqueous solution, NaOH was added under stirring until NaOH was completely dissolved to gain a HC/NaOH solution. At this time, 0.96 g of PGDE was added to the HC/ NaOH solution, stirring for 10 min at ambient temperature to gain an HC/PGDE/NaOH solution.
87.6 g of paraffin was mixed with 17.8 g of Tween 80 under stirring to gain a liquid paraffin/Tween 80 mixture. Then, the HC/PGDE/NaOH solution was dropwise added to this mixture under fiercely stirring. After the addition of the HC/PGDE/NaOH solution was completed, continuously stirring for 4 h to obtain spherical material (denoted as HC sphere). Finally, the HC spheres were washed by water and ethanol, and then immersed in ethanol to remove water in the spheres. FeCl 3 ·6H 2 O and FeCl 2 ·4H 2 O were added to deionized water, stirred for 10 min, heated to 70 ℃ and then cooled to ambient temperature to a Fe 3+ /Fe 2+ solution. The whole process was protected by nitrogen to prevent Fe 2+ from being oxidized. In this solution, the molar ratio of Fe 3+ to Fe 2+ is 2, and the total concentration of Fe 3+ and Fe 2+ is 0.3 mol L −1 . The HC spheres fabricated above were immersed in the Fe 3+ /Fe 2+ solution for 3 h. The unabsorbed Fe 3+ /Fe 2+ solution was poured out. Then, the HC spheres which adsorbed the Fe 3+ /Fe 2+ solution were immersed in NaOH solution of 0.5 mol L −1 for 1 h to obtain wet magnetic HC/Fe 3 O 4 hydrogel spheres. After being washed with deionized water, the wet spheres were freeze-dried to obtain dried magnetic HC/Fe 3 O 4 spherical samples.
Neat Fe 3 O 4 was prepared using a similar approach as described above. 50 mL deionized water was added to a conical flask, followed by the addition of 2.7030 g FeCl 3 ·6H 2 O and 0.9941 g FeCl 2 ·4H 2 O (n(Fe 3+ ):n(Fe 2+ ) = 2:1, C(Fe 3+ + Fe 2+ ) = 0.3 mol L −1 ). After stirring for 10 min, the Fe 3+ /Fe 2+ solution was heated to 70 ℃, and then cooled to ambient temperature to gain a Fe 3+ /Fe 2+ solution. Then, 160 mL NaOH solution with a concentration of 0.5 mol L −1 was add to the Fe 3+ /Fe 2+ solution under stirring. At this time, solid Fe 3 O 4 was precipitated. After the addition of NaOH solution, the mixture was left for 1 h. Then, Fe 3 O 4 was collected, washed with deionized water and freeze-dried to obtain neat Fe 3 O 4 . The whole preparation process was protected with nitrogen to prevent Fe 2+ from being oxidized.

Characterization of the magnetic HC/Fe 3 O 4
The chemical groups of the magnetic HC/PGDE were analyzed by IR measurement (4 cm −1 resolution and 64 scans) in the range of 400−4000 cm −1 ; each sample was prepared into thin slice using KBr before determination.
The morphology of the fracture surface of the magnetic HC/Fe 3 O 4 was observed on a scanning electron microscope; the dried magnetic HC/Fe 3 O 4 spheres were frozen by liquid nitrogen and snapped immediately; prior to the observation, samples were sputtered with gold. Meanwhile, the energydispersive X-ray spectroscopy (EDX) equipped on it was conducted to determine the elemental compositions of the HC/Fe 3 O 4 .
The measurement of the Brunauer-Emmet-Teller (BET) surface area and adsorption average pore diameter of the HC/Fe 3 O 4 sample was performed using N 2 adsorption-desorption experiment using an accelerated surface area and porosimetry system (Quantachrome Autosorb IQ3, USA). All of the samples were outgassed at 120 °C for 6 h before measurement.

Adsorption investigation of CR-SO 3 Na
Batch adsorption investigations were completed using the HC/Fe 3 O 4 as a adsorbent to remove CR-SO 3 Na from an aqueous CR-SO 3 Na solution. The following adsorption factors were investigated including contact time, solution pH, temperature, adsorbent dose and initial CR-SO 3 Na concentration. The contact time investigations were completed at time = 0−120 min and a fixed initial CR-SO 3 Na concentration of 100 mg L −1 , adsorbent dosage of 0.5 g L −1 and 30 °C. The pH investigations were completed at pH = 5−10 and a fixed initial CR-SO 3 Na concentration of 100 mg L −1 , adsorbent dosage of 0.5 g L −1 and 30 °C for 20 h. The temperature investigations were completed at temperature = 25−45 °C and a fixed initial CR-SO 3 Na concentration of 100 mg L −1 and adsorbent dosage of 0.5 g L −1 for 20 h. The adsorbent dose investigations were completed at adsorbent dose = 0.1−1.2 g L −1 and a fixed initial CR-SO 3 Na concentration of 100 mg L −1 and 30 °C for 20 h. The initial CR-SO 3 Na concentration investigations were completed at concentration = 25−300 mg L −1 and a fixed adsorbent dosage of 0.5 g L −1 and 30 °C for 20 h. For the contact time, temperature, adsorbent dose and initial CR-SO 3 Na concentration investigations, the pH refer to the pH of aqueous CR-SO 3 Na solution.
Each experiment was repeated three times and the averages values were calculated. The adsorption capacity (q) and removal efficiency (R e ) of each dye were calculated using the following Eqs. (1-3): where, q e (mg g −1 ) and q t (mg g −1 ) are the equilibrium adsorption capacity and adsorption capacity at time t, respectively, C 0 , C e and C t were the initial, equilibrium and time t dye concentrations (mg L −1 ), respectively, and m (g) and V (L) were the mass of HC/Fe 3 O 4 and the volume of the dye solutions, respectively.

Desorption experiment
After CR-SO 3 Na adsorption process (CR-SO 3 Na concentration 100 mg·L −1 , 0.5 g L −1 of adsorbent dosage, pH = 5, 30 °C, 200 r min −1 , 12 h), HC/Fe 3 O 4 was separated by an external magnet, and then immersed in a desorbing agent in which ethanol and NaOH concentrations were 50 vol% and 0.1 mg L −1 . This mixture was oscillated at 200 r min −1 and 30 °C for 2 h. Then, the desorbed HC/Fe 3 O 4 was washed with deionized water till pH became neutral, then dried and reused for the next adsorption process of CR-SO 3 Na. The adsorption-desorption cycle was performed 5 times.

Determination the points of zero charge
The point of zero-charge (pH zpc ) of the HC/Fe 3 O 4 was determined at the pH range from 2.0 to 10.0 using a similar pHdrift method reported in literatures [42,43]. 50 mL of 0.01 M NaCl solution was added to a conical flask of 100 mL. The pH of the NaCl solution was adjusted to 2.0-10.0 using 0.1−0.5 M HCl or 0.1−0.5 M NaOH solution. The dissolved CO 2 was removed from the solution using N 2 until the initial pH becomes stable. Then, 10.0 mg of HC/Fe 3 O 4 was added to each NaCl solution with a preset initial pH (pH 0 ) to obtain a mixture. Each mixture was shaken at a rate of 200 rpm for 4 h at 25 °C and left overnight. Measurement of the final pH (pH f ) was then recorded. The difference between pH 0 and pH f values (ΔpH = pH f −pH 0 ) was calculated and plotted against pH 0 to give pH zpc at ΔpH = 0.

Preparation strategy of the magnetic HC/Fe 3 O 4
The recovery and reuse of the adsorbents after adsorption towards dye pollutants are of significance in terms of their practical application. Moreover, the magnetic HC/ Fe 3 O 4 spheres prepared in this work are primarily utilized in removing CR-SO 3 Na from aqueous solution. Therefore, Fe 3 O 4 was introduced into the HC/Fe 3 O 4 spheres for the convenience of its separation from aqueous solution after adsorption towards CR-SO 3 Na. Meanwhile, our previous investigations found that the adsorbents prepared from higher biomass concentrations could result in the decreased removal efficiency for pollutant dyes [44]. Thus, the concentration of HC was selected as low as possible, similarly for NaOH solution concentration and PGDE/HC mass ratio to reduce the amounts of NaOH and PGDE needed to prepare the HC/Fe 3 O 4 . Based on this conception, the concentration of HC was selected at 3 wt%, 5 wt% for NaOH solution concentration and 2 for PGDE/HC mass ratio. In the light of the above preparation strategy, the magnetic HC/Fe 3 O 4 spheres were readily prepared at ambient temperature: The HC/PGDE/NaOH solution was dropwise added to a paraffin/Tween 80 mixture under fiercely stirring, and simultaneously HC was crosslinked by PGDE in this process to gain gel spheres (see Fig. 1a); the spheres were successively washed by water and ethanol, and then immersed in ethanol to remove water in the spheres; the water − free spheres were immersed in a Fe 3+ /Fe 2+ solution for 3 h to introduce magnetic Fe 3 O 4 followed by being washed by deionized water and subsequently freeze-dried to obtain dried magnetic HC/ PGDE spheres (see Fig. 1b).    Fig. 3. Although the saturated magnetization (26.4 emu g −1 ) of the HC/PGDE is less than that of Fe 3 O 4 (55.8 emu g −1 ), its magnetic property is sufficient for the magnetic separation of HC/Fe 3 O 4 from aqueous solution (see the inset of the bottom right of Fig. 3). In addition, almost zero residual magnetic susceptibility is observed, indicating that the HC/ Fe 3 O 4 sample is superparamagnetic.  [45], suggesting that Fe 3 O 4 has been successfully incorporated during fabrication process. It is also found that no diffraction peak corresponded to the HC appears for the HC/Fe 3 O 4 . This mainly results from the crosslinking reaction of PGDE with the hydroxyl of HC, leading to the disappearance of the diffraction peak ascribed to the HC. Figure 4b shows

Effects of solution pH and adsorbent dose on CR-SO 3 Na Adsorption
The effect of solution pH on CR-SO 3 Na adsorption is shown in Fig. 5a. It is evident that the adsorption of CR-SO 3 Na is significantly impacted by solution pH. The adsorption capacity of the HC/Fe 3 O 4 towards CR-SO 3 Na reduces with the rise of pH. This primarily results from the following facts: CR-SO 3 Na is an anionic dye which can be dissociated into CR-SO 3 − anion and Na + cation in water; the protonated hydroxy groups in the HC/Fe 3 O 4 under acidic condition generate a electrostatic attraction towards the anion of CR-SO 3 Na; with increasing pH, the deprotonated hydroxy groups gradually disable the electrostatic attraction of the HC/Fe 3 O 4 towards the anion. Figure 5b shows dosage effect on CR-SO 3 Na adsorption. The removal efficiency of the HC/Fe 3 O 4 for CR-SO 3 Na considerably increases with increasing dosage. This is mainly because the more the amount of the dosage is, the more the available adsorption sites for CR-SO 3 Na adsorption also is. However, the equilibrium adsorption capacities of the HC/ Fe 3 O 4 for CR-SO 3 Na decrease with adsorbent dosage. This is mainly due to the fact that with the increase of dosage, the non-adsorbed sites in the HC/Fe 3 O 4 become more and more, thus leading to the decreased adsorption capacity.

Adsorption time effect and kinetics
The dependence of the adsorption capacity of on time is presented in Fig. 6a. At the initial stage of adsorption, the adsorption capacity of the HC/Fe 3 O 4 for CR-SO 3 Na enhances rapidly with time. This is mainly due to a large amount of available vacant adsorptive sites on the HC/ Fe 3 O 4 , being to the benefit of CR-SO 3 Na adsorption. With an elongation in time, vacant sites on the HC/Fe 3 O 4 are gradually occupied by CR-SO 3 Na molecules, and it takes a long time for CR-SO 3 Na molecules to be adsorbed on the vacant sites of the HC/Fe 3 O 4 , thus resulting in a decreased adsorption rate. At the last adsorption stage, the adsorption equilibrium achieves, and thus keep hardly variable adsorption rate.
The pseudo-first-order Eq. (4) and pseudo-second-order Eq. (5) models are used to investigate the adsorption process of CR-SO 3 Na by the HC/Fe 3 O 4 [46,47].
The adsorption kinetic curves of CR-SO 3 Na on the HC/ Fe 3 O 4 based on the pseudo-first-order and pseudo-secondorder model equations are shown in Fig. 6b and c, respectively, and the fitted parameters using the two equations are summarized in Table 1. The pseudo-first-order kinetic correlation coefficients (R 2 = 0.88826) for the adsorptions of CR-SO 3 Na by the HC/Fe 3 O 4 is much less than the pseudosecond-order kinetic correlation coefficients (R 2 = 0.99997). Moreover, the adsorption capacity ((q e ) cal = 182.48 mg g −1 ) calculated using the pseudo-second-order kinetic equation is extremely closer to the experimentally determined value ((q e ) exp = 180.16 mg g −1 ). However, the adsorption capacity ((q e ) cal = 49.62 mg g −1 ) calculated by the pseudo-first-order kinetic equation is much less than the experimental value ((q e ) exp = 180.16 mg g −1 ). Thus, the pseudo-second-order kinetic model is more proper to describe the adsorption of CR-SO 3 Na by the HC/Fe 3 O 4 than the pseudo-first-order kinetic model, suggesting that the adsorption of CR-SO 3 Na by the HC/Fe 3 O 4 are primarily chemisorption [48].
The adsorption kinetic data were further analyzed with the intraparticle diffusion kinetic Eq. (6) model to identify the diffusion mechanism during an adsorption process [47]. The fitted curve and parameters using this equation are shown in Fig. 6d and Table 1, respectively.
As seen in Fig. 6d, the curve of intra particle diffusion includes three linear sections. The first linear section of the curve with a large slope associates with the

Initial concentration effect and adsorption isotherm
As shown in Fig. 7, initial concentration (C 0 ) remarkedly effect on adsorption capacity (q e ) and removal efficiency. The adsorption capacity significantly enhances with initial   [50]. However, the removal efficiency displays a contrary trend with CR-SO 3 Na initial concentration. This results from the limited vacant adsorption sites on the HC/Fe 3 O 4 which are not enough to absorb more CR-SO 3 Na molecules at high initial concentration [51]. The interaction between the adsorbent (HC/Fe 3 O 4 ) and adsorbate molecule (CR-SO 3 Na) was investigated using the following adsorption isothermal models [49,[52][53][54].
Langmuir isothermal model: The Langmuir isothermal model assumes a monolayer adsorption process on a homogeneous adsorbent surface. In this equation, q max (mg g −1 ) and b (L mg −1 ) represent the maximal adsorption capacity and the adsorption intensity or Langmuir coefficient related to the affinity of the binding site, respectively. R L represents the separation factor, which is used to judge the degree of adsorption: R L = 0, adsorption does not occur; R L > 1, adsorption is difficult to occur; 0 < R L < 1, adsorption is easy to occur.
Freundlich isothermal model: The Freundlich isothermal model presumes a multilayer adsorption on a heterogeneous surface. In this equation, K F (7) C e q e = 1 bq max + C e q max (8) ln q e = ln K F + 1 n ln C e stands for the Freundlich isotherm constant, and n is a constant related to the adsorption strength. The larger n is, the stronger the heterogeneity of the material surface is; when n is about 1, meaning that the adsorbent has relatively homogeneous binding sites. Temkin isothermal model: The Temkin isothermal model assumes that the heat of adsorption of molecules exponentially decreases linearly with coverage. In this equation, RT/b is Temkin isotherm constant, and A is related to the heat of adsorption (J mol −1 ).
Dubinin-Radushkevich (D-R) model: The D-R isothermal model assumes a temperature dependent model used for both physisorption and chemisorption. In this equation, ε represents Dubinin-Radushkevich isotherm constant (Polanyi potential), q 0 is initial dye adsorption (mg g −1 ), and k ad is adsorption equilibrium constant.
The fitting curves from Langmuir, Freundlich, Temkin and D-R Eqs. (7)(8)(9)(10)(11)(12) are shown in Fig. 8, and the fitting parameters are summarized in Table 2. The Langmuir, Freundlich, Temkin and D-R isothermal correlation coefficients (R 2 ) are 0.996, 0.971, 0.996 and 0.752, respectively, indicating that the adsorption of CR-SO 3 Na by the HC/Fe 3 O 4 can be well described by the Langmuir, Freundlich and Temkin isothermal models instead of D-R isothermal model. R L is 0.029-0.24, indicating that the adsorption process is easy to occur. n is 3.075, indicating a multilayer adsorption process on a heterogeneous surface. The above findings indicate that CR-SO 3 Na is easily adsorbed onto the HC/Fe 3 O 4 ; the adsorption of CR-SO 3 Na by the HC/Fe 3 O 4 occurs simultaneously by a monolayer homogeneous adsorption process on a homogeneous adsorbent surface and a multilayer adsorption process on a heterogeneous surface; the heat of adsorption of molecules exponentially decreases linearly with coverage.

Adsorption temperature effect and thermodynamics
The equilibrium adsorption capacity of the HC/Fe 3 O 4 towards CR-SO 3 Na decreases with temperature ( Fig. 9), suggesting that the adsorption of CR-SO 3 Na on the HC/Fe 3 O 4 is an exothermic process, and low temperature is more conducive to the adsorption of CR-SO 3 Na on the HC/Fe 3 O 4 [4]. The thermodynamic parameters Gibbs free energy (ΔG 0 ), enthalpy change (ΔH 0 ) and entropy change (ΔS 0 ) allow to know the spontaneity, occurrence with absorption or release of energy and increase or decrease in entropy of the adsorption process from point of the thermodynamic [52,55]. The thermodynamic behavior of the adsorption of CR-SO 3 Na by the HC/Fe 3 O 4 were investigated using Eqs. (13)(14)(15) [52,55]: where, q e (mg g −1 ) and C e (mg g −1 ) stand for the adsorption amount and CR-SO 3 Na concentration at equilibrium, respectively. R (8.314 J mol −1 K −1 ) is the ideal gas constant. T (K) represents the absolute temperature. The thermodynamic parameters calculated using Eqs. (13)(14)(15) are shown in Table 3. Negative ΔG 0 values identify the spontaneity of the adsorption of CR-SO 3 Na by (14) ln

Reusability
The reusability of an adsorbent is an important indicator to evaluate its potential application in industrial wastewater treatment. Thus, it is essential to study the reusability of an adsorbent. Figure 10 shows the adsorption-desorption behavior of the HC/Fe 3 O 4 . After five desorption-adsorption cycles, the adsorption capacity of HC/Fe 3 O 4 for CR-SO 3 Na reduces by 13%, indicating that the HC/Fe 3 O 4 has a good reusability as an adsorbent material.

Possible adsorption mechanism
In order to investigate the possible adsorption mechanism of CR-SO 3 Na by the HC/Fe 3 O 4 , FT-IR spectra of the HC/Fe 3 O 4 before and after CR-SO 3 Na adsorption were determined and shown in Fig. 11a. For the convenience of understanding, the adsorbent after CR-SO 3 Na adsorption is named as HC/ Fe 3 O 4 -CR-SO 3 Na. After the adsorption of CR-SO 3 Na by the HC/Fe 3 O 4 , a new characteristic band attributed to the SO 3 − group of CR-SO 3 Na was observed at 1112 cm −1 [55], confirming that CR-SO 3 Na is successfully adsorbed by the HC/Fe 3 O 4 . The absorption peak at about 3408 cm −1 corresponding to the stretching vibration of the hydroxyl group of the HC/Fe 3 O 4 was apparently weakened and red-shifted. This is because after the HC/Fe 3 O 4 adsorbs CR-SO 3 Na molecules, the H atom of the OH group on the HC/Fe 3 O 4 interacts with the O atom of the SO 3 − group or the N atom of the NH 2 group of CR-SO 3 Na; this interaction results in an elongation of the O-H bond on the HC/Fe 3 O 4 , generating the red-shift of the stretching vibration of O-H [57]. It is also found that the absorption peaks at 2919, 1352, and 1056 cm −1 keep almost invariable, indicating that the OH group on the HC/Fe 3 O 4 plays an important role in the adsorption of CR-SO 3 Na molecules.
The measurements of the point of zero-charge (pH zpc ) of the HC/Fe 3 O 4 were performed to determine its surface charge, and the initial pH dependence of the HC/Fe 3 O 4 ΔpH is shown in Fig. 11b. The pH zpc refers to the pH at which the electrical charge density on the surface of the HC/Fe 3 O 4 is zero. The pH zpc of the HC/Fe 3 O 4 is 7.0, suggesting that the surface of the HC/Fe 3 O 4 is neutral at pH = 7.0. When pH is less than 7.0, the surface of the HC/Fe 3 O 4 is positively charged, facilitating the adsorption of anionic dye CR-SO 3 Na due to the electrostatic attraction. When pH is higher than 7.0, the surface of the HC/Fe 3 O 4 is negatively charged, going against the adsorption of anionic dye CR-SO 3 Na on account of the electrostatic repulsion. This result is well consistent with the analysis above as well as the effect of initial pH of solution on the dye adsorption.  The SEM observation of the HC/Fe 3 O 4 sphere after the adsorption of CR-SO 3 Na is shown in Fig. 11c and d. It is found that the morphology of the HC/Fe 3 O 4 after the adsorption of CR-SO 3 Na is similar to that before the adsorption of CR-SO 3 Na as shown in Fig. 2a1 and a2, displaying a fluffy and porous structure, and the pore wall of the HC/Fe 3 O 4 material is covered by nano Fe 3 O 4 spheres. This suggests that the HC/Fe 3 O 4 adsorbent is stable, and not subject to deterioration in the adsorption process.
To further help understand how CR-SO 3 Na was adsorbed by the HC/Fe 3 O 4 , Fig. 12 shows the schematic diagram of the possible interaction of the HC/Fe 3 O 4 with CR-SO 3 Na. The adsorption of CR-SO 3 Na by the HC/Fe 3 O 4 primarily results from the following interactions. Firstly, the hydrogen bond interactions between the hydroxyl groups on the HC/ Fe 3 O 4 and CR-SO 3 Na molecules facilitate the adsorption of CR-SO 3 Na [57]. Secondly, the electrostatic interaction between CR-SO 3 Na molecule and the HC/Fe 3 O 4 is also an important driving force for the adsorption of CR-SO 3 Na. For example, the protonated hydroxy groups on the HC/Fe 3 O 4 under acidic conditions can generate electrostatic attraction   Na. This is why acidic conditions are favorable to CR-SO 3 Na adsorption, but not alkaline conditions (see Fig. 5a and b). Finally, the interaction between the HC/Fe 3 O 4 and the aromatic ring of CR-SO 3 Na is beneficial to CR-SO 3 Na adsorption [49].

Comparison investigation
Several review papers published in 2020-2023 have reported the adsorption capacities of various adsorbent materials for CR-SO 3 Na [58][59][60][61][62][63]. Thus, here, only the adsorbents containing cellulose are compared with the HC/Fe 3 O 4 in terms of CR-SO 3 Na adsorption from the preparation approach, environmental aspects (volatility or toxicity of raw starting material), equilibrium time, adsorption capacity and recycling performance (Table 4). Compared with the adsorbents reported in the literatures, the preparation procedure of the HC/Fe 3 O 4 is more facile and displays an efficient adsorption capacity for CR-SO 3 Na. Meanwhile, the preparation of the HC/Fe 3 O 4 does not uses volatile/toxic starting materials. Thus, the HC/Fe 3 O 4 is a competitive adsorbent for CR-SO 3 Na removal.

Conclusions
The magnetic HC/Fe 3 O 4 spherical materials with porous properties, an efficient adsorption capacity of CR-SO 3 Na and good reusability were developed, which were readily fabricated via simply mixing HC and PGDE in NaOH aqueous solution followed by subsequent introduction of Fe 3 O 4 at ambient temperature. The adsorption process could be well described by the pseudo-second-order kinetic model instead of the pseudo-first-order kinetic model, and the intraparticle diffusion of CR-SO 3 Na from the exterior surface to the interior surface of the HC/Fe 3 O 4 was the rate-controlling step. CR-SO 3 Na was adsorbed by the HC/Fe 3 O 4 mainly by a monolayer homogeneous adsorption process on a homogeneous adsorbent surface, and a multilayer adsorption process on a heterogeneous surface meanwhile also took place. The adsorption of CR-SO 3 Na by the HC/Fe 3 O 4 was an exothermic and spontaneous process as well as a decreased randomness at the solid/solution interface based on thermodynamic analysis. The adsorption of CR-SO 3 Na by the HC/Fe 3 O 4 mainly results from the following contributions such as the hydrogen bond interactions of the HC/Fe 3 O 4 with CR-SO 3 Na molecules and the electrostatic interaction between CR-SO 3 Na molecules and the HC/Fe 3 O 4 . After five desorption-adsorption cycles, the adsorption capacity of the HC/Fe 3 O 4 for CR-SO 3 Na retains 87%. It is expected that this work can provide valuable information for the facile fabrication of novel ecofriendly, low price and efficient adsorbent materials for removing dye pollutants from wastewater.

Conflict of interest
The authors declare no conflict of interest.