Removal of aqueous Cr(VI) by magnetic biochar derived from bagasse

We prepared a novel adsorbent functionalized by bagasse magnetic biochar (BMBC). To study the removal behaviors and mechanisms of Cr(VI) by BMBC, batch adsorption experiments were conducted by modifying variables, such as pH, adsorption time, BMBC dosages, initial Cr concentration, co-existing ions, and ionic strength, and characterizing BMBC before and after Cr(VI) adsorption. BMBC was primarily composed of Fe2O3 and Fe3O4 on bagasse boichar with an amorphous structure. The specific surface area of BMBC was 81.94 m2 g−1, and the pHpzc of BMBC was 6.2. The fabricated BMBC showed high adsorption performance of Cr(VI) in aqueous solution. The maximum Cr(VI) adsorption capacity of BMBC was 29.08 mg g−1 at 25 ºC, which was much higher than that of conventional biochar sorbents. The adsorption process followed pseudo-second-order kinetics and could be explained by the involvement of the Langmuir isotherm in monolayer adsorption. The crystalline structure of Fe3O4 in the BMBC changed slightly during the adsorption process; Fe3O4 improved the adsorption of Cr(VI) on BMB. The desorption capacity of Cr(VI) was 8.21 mg g−1 when 0.2 mol L−1 NaOH was used as the desorption solution. After being reused three times, the removal efficiency is still as high as 80.36%.

Preparation of BMBC. (1) Wash and sift. Bagasse, in which sugar had fully been washed out, was dried in an oven at 85 °C, pulverized using a universal crusher, and screened using a 0.085-mm standard sieve.
(2) Impregnation with ferric solution. Ferrous sulfate solution (0.15 mol L −1 ) was added into a reagent bottle with a wide and frosted mouth along with 50 g of bagasse. After mechanical agitation, the sample underwent ultrasonic oscillation for 30 min and was left submerged for 36 h. The ammonia solution (10% v/v) was slowly added into a reagent bottle under mechanical agitation to adjust the pH of the suspended mixture to 8.5. The product was then heated to 85 °C in a microwave oven and filtered to obtain the filter cake. Lastly, the filter cake was washed with ultrapure water until the pH of the washing liquid was approximately 7.0. The filter cake was then placed in a beaker filled with 200 mL of absolute ethanol, oscillated for 30 min, and filtered to obtain the mixture of ferric hydroxide and bagasse.
(3) Drying and Calcination. The mixture of ferric hydroxide and bagasse was placed on a porcelain plate and dried at 110 °C to prepare the dry bagasse iron hydroxide mixture. The dry bagasse iron hydroxide mixture was carbonized at 450 °C for 4 h, cooled, ground, and screened using a 100-mm mesh sieve to obtain BMBC.
The preparation of bagasse boichar (BBC) without magnetic iron oxide is described in Supporting Information S1.
Batch adsorption experiments. First, 0.20 g of BMBC was weighed in a series of 100-mL polyethylene centrifuge tubes, followed by the addition of 50 mL of solution with different Cr(VI) concentrations, which were adjusted to a set pH value with 0.1 mol/L NaOH or HNO 3 . The tubes were then sealed with capsules and placed in a constant temperature water bath oscillator that was shaken for 48 h at a speed of 150 r min −1 and a temperature of 25 °C. The solution was filtered after adsorption equilibrium had been reached. A 0.45-μm filter membrane was used to filter the supernatant, and Cr(VI) concentrations were determined by ICP-EOS (PE Optima 7000DV).
Desorption and regeneration experiments. First, 0.20 g of BMBC saturated with Cr(VI) was added to a series of 100-mL plastic centrifuge tubes, along with 50 mL of the regeneration reagent (0.20 M HCl, 0.20 M H 3 PO 4 , 0.20 M NaOH, and 0.20 M NaHCO 3 solutions). The solution was then placed in a constant temperature water bath oscillator. After shaking for 24 h, the supernatant was filtered using a 0.45-μm filter, and the concentration of Cr(VI) in the filtrate after desorption was measured. Regeneration of the BMBC for reuse was studied from desorption and readsorption experiments in three coherent cycles.
Desorption efficiency (%) and regeneration utilization efficiency (%) were calculated by the following equation: www.nature.com/scientificreports/ where C d is concentration of Cr(VI) after desorption, and V d is the volume of the desorption solution; C 0 and C e (mg L −1 ) are the initial and equilibrium Cr(VI) concentrations, respectively; V a is the volume of adsorption solution.

Result and discussion
Characterization. Magnetic analysis. The hysteresis loop of BMBC is shown in Fig. 1a. The lower right corner provides a partially enlarged view of the figure. BMBC had magnetic properties, and the hysteresis loops of BMBC were similar to those of nano-magnetic Fe 3 O 4 , indicating that BMBC and Fe 3 O 4 had the same magnetic properties (Fig. 1a). The saturation magnetization of BMBC was 0.33 emu g −1 . As the intensity of the applied magnetic field increased, magnetization increased; furthermore, there was a nonlinear relationship between the saturation magnetization and the intensity of the applied magnetic field. When the applied magnetic field was increased, the strength of the applied magnetic field continuously increased, and the magnetization remained constant when saturation had been reached. This pattern can be explained by the fact that the magnetic particles were in a single magnetic domain state. The particles spontaneously magnetized to a saturated state along the easy magnetization direction. Under the action of an external magnetic field, the magnetic moments were neatly aligned. As the applied magnetic field strength increased, the number of particles arranged in the direction of the applied magnetic field increased proportionally with the strength of the applied magnetic field 19 .
SEM images. SEM images revealed that the surface structure of BMBC was relatively loose; the particle size was uniform, with some particles smaller than 100 nm; and the microporous structure was completely formed (Fig. 1b). The formation of the microporous structure can effectively increase the specific surface area of BMBC.

XRD analysis.
To confirm the existence of iron oxide in the BMBC sample, the BMBC diffraction peak was compared with the Fe 2 O 3 XRD reference code and the Fe 3 O 4 reference code of the International Diffraction Data Center (Fig. 1c) (Fig. 1d). The wavenumber at 3443 cm −1 corresponded to the stretching oscillation of -OH, and the  BET analysis. The BET specific surface area of BMBC was 81.94 m 2 g −1 , and the pore volume was 26.74 The specific surface area of BMBC was approximately 30 times the specific surface area of oak biochar 22 , six times the specific surface area of wheat straw and spruce wood biochar 23 , and approximately 2.4 times the specific surface area of rice husk biochar 24 . By contrast, the BET specific surface area of BBC was 10.19 m 2 g −1 , and the pore volume was 3.05 cm 3 g −1 , the surface of BMBC had a structure rich in pores, and this finding was consistent with the results of the SEM analysis.
Zeta analysis. When the Zeta potential was equal to zero, the point zero charge (pH pzc ) of BMBC is obtained at the intersection of the pH-Zeta potential curve and the abscissa. Thus, the pH pzc of BMBC was determined to be 6.2, and the pH pzc of BBC was 4.4 ( Fig. 2). The pH pzc of BMBC was lower than that of most types of commercial activated carbon 25 as well as the pH pzc of iron oxides 26 . This difference may have stemmed from the synthesis of BMBC, which was caused by adjusting the pH of the reaction end point with ammonia 27 .
Error analysis. The kinetic and isotherm parameters were estimated using their respective nonlinear models, which were fitted by Origin 9.0 software (Massachusetts, USA). The fitness confidence was evaluated through several common error analysis methods such as the coefficient of determination (R 2 ) 28 , mean relative error (MRE) 29 ,and sum square error (SSE) 30 [Equations S(1)-S(3)] in the supporting information file.

Major factors affecting Cr(VI) adsorption.
Effect of solution pH. The effect of initial aqueous pH on the adsorption of Cr(VI) onto BMBC was studied by varying the pH from 2 to 11 (Fig. 3a). A low pH was beneficial for Cr(VI) removal, and the adsorption capacity of Cr(VI) by BMBC decreased as the pH increased. The removal efficiency decreased from 99.9% to 89.9% as the pH increased from 2.0 to 4.0, respectively. At a pH of 11.0, the removal efficiency decreased to 9.7%. The complete removal of Cr(VI) could be achieved when pH ≤ 4.0. The pH causes changes in the surface charge of the adsorbent and the degree of protonation of the functional groups on the adsorbent, and the form of Cr(VI) in the adsorption system. When the pH ranged from 2 to 8, the main forms of chromium were Cr 3+ , Cr 2+ , Cr 2 O 7 2− , and HCrO 4 −31 . Chromium species were mainly present as CrO 4 2− in water at pH > 8. Dominant chemical reactions in the solution were the following 32 : www.nature.com/scientificreports/ At pH < 6.20 (pH pzc(BMBC) = 6.20), positive charges on the surface of BMBC were abundant, as a large number of H + accumulated on the BMBC, making the surface of BMBC highly protonated 31 and favoring the attraction of Cr(VI). In contrast, when pH was increased to 6.20, the surface of the biochar was negatively charged, and it repelled anionic Cr(VI) in the solution. In addition, the concentration of OH − increased as the pH increased, and it competitively adsorbed Cr 2 O 7 2− , HCrO 4 − , and CrO 4 2− , which resulted in a decrease in the amount of Cr(VI) adsorbed by the biochar. Experiments suggested that the optimal pH was 2, as this pH resulted in the highest removal rate (99.92%) of Cr(VI) by BMBC.
Effect of BMBC dosage. The experimental results of the effects of different dosages on the adsorption effect are shown in Fig. 3b. The removal efficiency increased from 29.9 to 99.9% as the dosage of BMBC increased from 0.001 to 0.012 g L −1 , respectively, with the adsorption capacity decreasing from 14.9 to 4.16 mg g −1 (Fig. 3b). Initially, increases in the dosage of BMBC led to increases in the total specific surface area of the adsorbent, resulting in the increased adsorption of Cr(VI) and thus a rapid increase in removal efficiency. At a BMBC dosage of 0.004 g L −1 , the removal efficiency was 98.20%, and the adsorption capacity was 9.01 mg g −1 . The removal tended to be stable when the dosage of BMBC exceeded 0.004 g L −1 , but the adsorption capacity decreased to 4.16 mg g −1 . Thus, the optimal dosage of BMBC was 0.004 g L −1 for wastewater with a Cr(VI) concentration of 50 mg L −1 .
Effect of initial Cr(VI) concentration. The absorption of Cr(VI) in BMBC increased when the initial Cr(VI) concentration increased from 5 to 300 mg L −1 at 25, 35, and 45 °C, with the uptake of Cr(VI) increasing from 1.25 to 24.83 mg g −1 , from 1.25 to 29.78 mg g −1 , and from 1.25 to 37.27 mg g −1 , respectively (Fig. 3c). Thus, more Cr(VI) was adsorbed by BMBC as the initial Cr(VI) concentration increased. This pattern may be driven by the fact that when the initial Cr(VI) concentration was increased, the concentration difference between Cr(VI) in the solution and the adsorbent increased, which might have increased the force and thus rapidity of mass transfer between the liquid and solid phases, promoted the migration of Cr(VI) to the surface of BMBC, and thus further improved the adsorption reaction. When the temperature was increased from 25 to 45 °C (Fig. 3c), Effect of adsorption time. As the adsorption time increased, the uptake of Cr(VI) by BMBC increased gradually and eventually stabilized (Fig. 3d). For the 20, 50, and 100 mg L −1 of initial Cr(VI) concentration at 25 ºC and pH 3, the uptake of Cr(VI) was sorbed within the first 2 h at an average sorption rate of 0.0354, 0.0407, and 0.0535 mg/(g min), respectively (corresponding to 71.00%, 54.26%, and 41.71% of the total amount of Cr(VI) sorbed, respectively). The equilibrium times for the adsorption of Cr(VI) on BMBC at initial Cr(VI) concentrations of 20, 50, and 100 mg L −1 were 18, 24, and 48 h, respectively. The equilibrium time lengthened as the initial Cr(VI) concentration increased. In the physical sorption process, most of the sorbate was sorbed within a short adsorption reaction time. However, strong chemical binding of the sorbate with sorbent required a longer adsorption reaction time for adsorption equilibrium to be reached. The uptake of sorbate is relatively rapid during the initial stages and shortly after; however, uptake of sorbate gradually slows as the equilibrium time is approached 33 . The initial rapid uptake of sorbate may stem from the large numbers of available vacant sites on the surface of the adsorbent 34  , CO 3 2− and PO 4 3− have different effects on the uptake of Cr(VI) (Fig. 4). The degree to which the adsorption capacity of Cr(VI) by BMBC was inhibited by these ions exhibited the following rank order: CO 3 2− > SO 4 2− > NO 3 − > PO 4 3− . The negative effect of CO 3 2− on the adsorption capacity of Cr(VI) by BMBC was relatively large. For example, when the concentration of Cr(VI) was 50 mg L −1 , the adsorption capacity of Cr(VI) decreased from 9.01 to 8.76 mg g −1 , a decrease of 2.77%. This decrease may stem from the fact that CO 3 2− competed with Cr(VI) for adsorption by BMBC. NO 3 − , SO 4 2− , CO 3 2− and PO 4 3− have varying degrees of inhibition on the uptake of Cr(VI), which ultimately stem from differences in the affinity between these anions and BMBC 35 . Adsorption kinetics. To understand the kinetic mechanism of the adsorption process, pseudo-first-order and pseudo-second-order models were used to fit the kinetic data. The linearized forms of the pseudo-first-order and pseudo-second-order model equations are provided below in Equations S(4) and S(5) in the supporting information, respectively: Data fitted with pseudo-first-order kinetic equations and pseudo-second-order kinetic equations are shown in Figures S1a and S2b in the supporting information, respectively. All corresponding kinetic parameters of the two models at different initial Cr(VI) concentrations are shown in Table 1 Figures S2 in the supporting information. The uptake of Cr(VI) was 28.82 mg g −1 and increased to 33.78 mg g −1 when the temperature increased to 45 °C. These results implied that the adsorption of Cr(VI) by BMBC was an endothermic process 37 . The Langmuir and Freundlich models have been extensively used to study isotherm data. In this study, isotherm data were simulated by the Langmuir and Freundlich models. The Langmuir model assumed that a monomolecular layer was formed when adsorption took place without any interaction between the adsorbed molecules. The linearized Langmuir isotherm model is represented by Equation S (6) in the supporting information. The linearized form of the Freundlich model is expressed by Equation S (7). The linear plots of the Langmuir and Freundlich models are shown in Figures S3a and S3b in the supporting information, and the parameters (including 1/n and K L 38,39 ) of both adsorption isotherm models evaluated from the linear plots are shown in Table 2. The adsorption capacity of Cr(VI) was improved after the modification of bagasse biochar. At 25 °C, the maximum adsorption capacity of Cr(VI) by bagasse biochar was 13.15 mg g −1 , and the adsorption capacity of Cr(VI) by BMBC was 28.82 mg g −1 , an increase in adsorption capacity of 119.16%. Yu Dezhong 40 used nanoscale iron oxide to adsorb Cr(VI) and found that the maximum adsorption capacity of Cr(VI) was only 0.3983 mg g −1 .
Langmuir isotherm parameters at 25, 35, and 45 °C showed that the adsorption process partially fitted the Langmuir isotherm model, with correlation coefficients (R 2 ) of 0.9757, 0.9853, and 0.9935, respectively. This finding demonstrated that monolayer adsorption was the most common way by which Cr(VI) was adsorbed by BMBC. The calculated maximum adsorption, Q max , from the Langmuir isotherm model was 29.08, 31.37, and 33.21 mg g −1 at 25, 35, and 45 °C, respectively. The correlation coefficients (R 2 ) of the Freundlich model were 0.8754, 0.8412, and 0.7242 at 25, 35, and 45 ºC, respectively. Comparison of the correlation coefficients between the two adsorption models revealed that the Langmuir adsorption isotherm model better fit the isotherm data for Cr(VI) adsorbed onto BMBC.
The similar adsorbents used for the adsorption of Cr(VI) are summarized in Table 3. BMBC had higher Cr(VI) adsorption capacities (29.08 mg g −1 ) than other previously examined adsorbents (4.61-25.27 mg g −1 ) that have been used to treat Cr(VI) from industrial effluents. Nevertheless, comparisons between adsorbents in their Cr(VI) removal capacities are potentially hampered by differences between the microenvironments of the different solutions 41 . www.nature.com/scientificreports/    (Table 4). At 25,35, and 45 °C, the values of ΔG° were − 12.1289, − 12.5741, and − 13.1420 kJ mol −1 , respectively ( Table 4). All values of ΔG° were less than 0, indicating that the adsorption of Cr(VI) by BMBC was spontaneous. The absolute value of ΔG° increased as the temperature increased, indicating that the spontaneous nature of the adsorption process increased as the temperature increased. ΔH° > 0 indicated that the adsorption reaction was endothermic, and increases in temperature were favorable for adsorption.
Potential adsorption mechanism. The energy spectra of BMBC and BMBC following Cr(VI) adsorption is shown in Fig. 5. The mass percentages (%) of C, O, Fe, and S in the BMBC were 60. 16, 23.73, 15.43, and 0.67 (Fig. 5a), respectively; after Cr(VI) adsorption, these percentages were 43.58, 27.93, 26.97, and 0.65 (Fig. 5b). The peak for Cr appeared on the energy spectrum, indicating that Cr existed on the surface of BMBC after Cr(VI) adsorption, and the mass percentage was 0.87%. FTIR analysis was used to characterize Cr(VI)-Cr(VI) interactions (Fig. 6a). After the adsorption of Cr(VI) by BMBC, the peaks for O-H (3443 cm −1 ), C=O (1629 cm −1 ), and C-C (1378 cm −1 ) groups changed slightly, and the O-H (1111 cm −1 ) and C-O (875 cm −1 ) groups disappeared from the FTIR of BMBC after Cr(VI) adsorption, indicating that O-H, C=O, C-O, and C-C were favorable for Cr(VI) removal. The functional groups on the surface of the adsorbent determine its adsorption properties. The content of functional groups on the surface of BMBC significantly increased, improving its adsorption effect.
XRD measurements were also made to study the composition of BMBC before and after Cr(VI) adsorption (Fig. 6b) XPS analysis was used to reveal changes in the valence of Fe, O, C, and Cr before and after Cr(VI) removal (Fig. 7). The adsorbed chromium (Cr 2p) on BMBC was analyzed on the BMBC surface at the atomic level of 3.16% (Table 5). The valences of Cr 2p were separated at 577.93 eV (Cr 2p 1/2 ) and 588.0 eV (Cr 2p 3/2 ), which was  Desorption and regeneration analysis. The regeneration of BMBC with different reagents (0.2 mol L −1 of NaOH, NaHCO 3 , HCl, and H 3 PO 4 ) is shown in Fig. 8a. The rank order of the regeneration utilization rate of saturated BMBC after Cr(VI) adsorption was saturated was NaOH > NaHCO 3 > HCl > H 3 PO 4 . The regeneration utilization rates of the alkaline regeneration reagents (0.2 mol L −1 NaOH and NaHCO 3 ) were higher than those of the acid regeneration reagents (0.2 mol L −1 HCl and H 3 PO 4 ). When 0.2 M NaOH was used as the desorption solution, the desorption capacity of Cr(VI) was 8.21 mg g −1 .
Given the above findings, 0.2 mol L −1 NaOH solution was selected for further study as the BMBC recycling regeneration reagent (Fig. 8b). A cyclic adsorption experiment was performed on a Cr(VI) solution with a pH of 2.0 and a concentration of 50 mg L −1 to calculate the regeneration utilization rate of the adsorbent. The experimental results are shown in Table 6. After being reused three times, 19.74% of the adsorption capacity of regenerated BMBC was lost, which likely stems from the higher affinity of CrO 4 2− for the BMBC surface compared with OH − ions, further exacerbating the difficulty of ion exchange 44,45 . In practical applications, high-concentration NaOH solutions could be used to increase the OH − concentration and drive the reaction backwards.

Conclusion
This study examined the removal of Cr(VI) from aqueous solutions by a novel BMBC based on bagasse biochar with magnetic iron oxide. The advantages of the proposed method for BMBC application include its low cost and high reproducibility. The highest percentage of Cr(VI) removal was observed at a pH of 2, with a dosage of 0.20 g of BMBC in a 50-mL solution. The rank order of the degree to which the adsorption capacity of Cr(VI) by BMBC was inhibited was CO 3 2− > SO 4 2− > NO 3 − > PO 4 3− . The enhanced adsorption capacity of Cr(VI) on BMBC was explained by the functional groups on the surface of BMBC. The adsorption process followed pseudo-secondorder kinetics and could be explained by the involvement of the Langmuir isotherm in monolayer adsorption. The Fe 2p peaks in the spectra of BMBC after Cr(VI) adsorption changed that suggesting that Fe 2p participated in the adsorption reaction as dot electrons. The process was enhanced at higher temperatures, and the adsorption was spontaneous and endothermic. BMBC also showed a high selectivity towards Cr(VI) and a high degree of reusability, which makes it a green material. Based on these results, the BMBC studied in this work demonstrated high efficacy in Cr(VI) removal from water through magnetically assisted chemical separation.