Highly efficient removal of hexavalent chromium by magnetic Fe–C composite from reed straw and electric furnace dust waste

Reed straw and electric furnace dust (EFD) waste were used to prepare magnetic Fe–C composite (EFD&C) by co-precipitation and high-temperature activation method to remove Cr(VI) from water. The magnetic EFD&C owned a large specific surface (536.61 m2/g) and a porous structure (micropores and mesopores), and had an efficient removal capacity for Cr(VI). Under conditions of pH (2), the addition amount of EFD&C (1 g/L), the adsorption time (760 min), and the temperature (45 °C), the maximum adsorption capacity reached 111.94 mg/g. The adsorption mechanism mainly attributed to chemical adsorption (redox), Cr(VI) reduced to Cr(III) by Fe(II) and Fe(0) (from Fe3O4 and Fe components in EFD) and surface functional groups of -OH, C = C, C–C and O-C = O (from biochar), and secondary attributed to physical adsorption, Cr(VI) and Cr(III) (from reduced Cr(VI)) adsorbed into the porous structure of EFD&C. This study provided a feasible solution for the preparation of adsorbents for adsorbing heavy metals from iron-containing metallurgical solid waste and biomass waste, which contributed to reducing the environmental pollution and lowering the cost of adsorbent preparation.


Introduction
Water is an important resource for humans and other organisms' survival and reproduction. The rapid development of urbanization and industrialization, the surge in population, and changes in agricultural activities lead to a sharp decrease in the water resources that can be used, which seriously threatens the survival and development of human beings. The harm caused by industrial wastewater from metallurgy, electroplating, printing, and battery manufacturing is a huge problem that needs to be faced all over the world and has attracted many attentions from many international researchers. The wastewater produced by these industries contains many heavy metals (e.g., Cu, Cr, Cd, and Pb) (Song et al. 2020). Heavy metals in the untreated wastewater can infiltrate into the ground through the surface to pollute groundwater resources, and then threaten human survival (Dhal et al. 2013). Especially in some developing countries, the pollution caused by industrial wastewater is increasing year by year, limiting their sustainable development. Chromium-containing wastewater is the wastewater containing heavy metal chromium from factories including dyeing, electroplating, paint, petroleum, and tanneries Responsible Editor: Zhihong Xu (Singh and Kalamdhad 2011). Chromium exists mainly in the form of hexavalent chromium Cr(VI) and trivalent chromium Cr(III) in wastewater. The danger and toxicity of hexavalent chromium are much greater than that of trivalent chromium, and its toxicity is more than 100 times that of trivalent chromium (Huang et al. 2022). Cr(VI) is listed as one of the carcinogenic heavy metals, which is highly toxic and difficult to remove (Sun et al. 2014). The highest acceptable levels of Cr(VI) were 50 µg/L, 50 µg/L, 100 µg/L, and 50 µg/L, respectively, according to World Health Organization (WHO) (Zhang et al. 2020a, b), European Union (EU) (Directive 1998), America (Matome et al. 2020), and China (Zelmanov and Semiat 2011). Therefore, exploring efficient methods to remove Cr(VI) from the solution is crucial.
Chemical precipitation, reverse osmosis, ion exchange, and photocatalysis are conventional methods to treat water polluted by Cr(VI) (GracePavithra et al. 2019), but the above methods have some defects, such as (i) chemical precipitation method will produce heavy metal toxic sludge to cause secondary harm to the environment ); (ii) reverse osmosis method is a kind of membrane filtration technology with good removal efficiency, but the amount of water treated is smaller with expensive cost (Saleh et al. 2022); (iii) ion exchange with high cost is not conducive to the large-scale application (Ms et al. 2021); and (iv) photocatalyst preparation process in the photocatalysis is complex and expensive, which limits its development (Ihsanullah et al. 2022). Adsorption is a low-cost, easy-to-apply, and environmentally friendly method for removing heavy metals from aquatic environments (Wang et al. 2013). Its high adsorption performance, mild reaction conditions, and renewability make it a promising method for application (Abdolali et al. 2014). Commonly used adsorbents include nanomaterials (Efome et al. 2018), activated carbon , graphene, and some composite materials (Zheng et al. 2019). Biochar is derived from biomass pyrolysis or carbonization, which can avoid the natural degradation of waste biomass to produce a large amount of methane to exacerbate the greenhouse effect (Danish and Ahmad 2018). Biochar owns many excellent features such as large specific surface area, rich pore structure and functional groups, good chemical stability, and low preparation cost, so it gets a lot of attention for adsorbent preparation (Valentín-Reyes et al. 2019). However, there are some shortcomings in the adsorption and transport of heavy metal ions by raw biochar, such as low removal efficiency, poor adsorption selectivity, and difficulty in secondary recovery (Liu and Zhang 2022). But its adsorption characteristics can be improved by appropriate modification (Ling et al. 2017). Surface pretreatment of biomass with different chemical solutions can increase surface area and pore volume of the prepared biochar and generate more functional groups, which plays a decisive role in improving an adsorption rate and selectivity (Park and Jang 2002). Biomass pretreatment includes methods using acid (such as H 2 SO 4 , HCl, and HNO 3 ), alkali (such as KOH and NaOH), and oxides (Danish and Ahmad 2018). Compared with biomass first carbonization and then modification treatment, direct pretreatment can reduce the preparation time and cost, also help to increase its specific surface area, pore volume, and the number of active functional groups (Park and Jang 2002) to improve its adsorption performance (Barjasteh-Askari et al. 2021). Kharrazi et al. (2021) pretreats elm wood chips with HCl and found that the biochar after being pretreated had higher surface area and micropore volume, and the adsorption capacity for Cr (VI) also increased to 190 mg/g (114 mg/g with unpretreated elm wood chips). In recent years, Fe or iron oxides have been used in the modification of biochar. Iron oxide particles as effective adsorbents, they can be used alone or in combination with other adsorbents (Saha et al. 2011). Due to the presence of magnetic Fe or Fe 3 O 4 , they can be reused to reduce the cost of wastewater treatment (Nejadshafiee and Islami 2019). When biochar is added to a solution containing iron ions to prepare composites, iron ions can diffuse into the surface and pores of biochar and connect with functional groups on the surface of biochar, which will greatly improve the adsorption effect (Chen et al. 2007). Mortazavian et al. (2018) uses biocharsupported zero-valent iron nanoparticles to adsorb Cr(VI), and finds that iron and its oxides can be stably loaded on the surface of biochar at a lower iron concentration. Under the same conditions, more Cr(VI) are removed from solutions by the biochar loaded with nano-iron particles can effectively than the unmodified biochar. Although the above studies have achieved considerable results, the development of novel materials with high adsorption performance and environmentally friendly still requires exploration.
As a by-product of the steel making process, the electric furnace dust (EFD) is mostly used in landfill and accumulation. Because it contains a large amount of iron , it can be used as an iron source for adsorbent. Reed straw as a biomass is a renewable resource with high yield and low utilization . In this study, reed straw as carbon source and EFD recovered from iron and steel industry as the Fe source were used prepare magnetic adsorbent (EFD&C) by co-precipitation and high-temperature calcination to adsorb Cr(VI) in water. The effects of pH, adsorbent dosage, adsorption time, and initial Cr(VI) concentration on the adsorption effect were discussed, and the adsorption kinetics of Cr(VI) by EFD&C was investigated using pseudo-first-order and pseudo-second-order kinetic equations, the Elovich model, and the intraparticle diffusion model. The Langmuir model and Freundlich model were used to explored the adsorption isotherm of Cr(VI), and the adsorption mechanism of Cr(VI) was analyzed by XRD, VSM, zeta potential, FT-IR, XPS, and SEM-EDX.

Adsorbent preparation
Raw reed straw (40 g) was washed with 2 L deionized water for 5 min in a 5 L plastic container continuously for one repetition, filtered with 0.45 μm filter membrane, and dried to constant weight at 105 °C in an oven (DGG-9140B, Shanghai Senxin Experimental Instrument Co., Ltd., Shanghai). The dried reed straw was crushed, sieved (≤ 830 μm), and named reed straw powder. Electric furnace dust (dried at 105 °C for 4 h and ground ≤ 75 μm) with 4 g, deionized water (40 mL), and HCl (30 mL) were moved in a beaker (150 mL) at 75 °C for 40 min in an oil bath (DF-101S, Shanghai Lichen Bangxi Instrument Technology Co., Ltd., Shanghai) with magnetic stirring. After reaction, samples were cooled to 25-30 °C and filtered (0.45 μm), filtrate was collected and named as electric furnace dust solution. Electric furnace dust solution (50 mL), 500 mL KOH solution (1 mol/L), and 30 g reed straw powder were moved to in a 1000 mL glass beaker and mechanically stirred at 25 °C for 6 h at a speed of 800 r/min. The mixtures were filtered using a Buchner funnel (≤ 30 μm), dried to constant weight at 105 °C in an oven, and named as EFD&C precursor. About 20 g EFD&C precursor was placed in a tube furnace (SK-ES08143, Tianjin Zhonghuan Experimental Electric Furnace Co., Ltd., Tianjin), heated to 600 °C for 2 h with a heating rate of 5 °C/min under the condition of N 2 flow of 200 mL/min, and then naturally cooled to 25-30 °C. Calcined EFD&C precursor was washed with deionized water to pH = 7, dried at 105 °C to constant weight, ground (≤ 75 μm), and named as magnetic EFD&C as shown in Fig. 1.

Adsorbent characterization
Crystal characteristics of samples were detected with X-ray diffraction (XRD, D8 Advance, Bruker AXS GMBH, Karlsruhe, Germany) with CuKα radiation (40 kV and 200 mA). The morphology and element contents on the surface of samples were examined with scanning electronic microscope-X-ray energy dispersive analysis (SEM-EDX;

Batch adsorption experiment of Cr(VI)
The Cr(VI) adsorption performance of magnetic EFD&C was determined with batch experiments in a 250 mL conical flask. Raw Cr(VI) solution (1000 mg/L) was prepared with potassium dichromate, other six groups of Cr(VI) solutions (50,100,150,200,250, and 300 mg/L) were prepared by dilution using ultrapure water. The initial pH in the Cr(VI) solution was adjusted with 1 mol/L HCl and 1 mol/L NaOH.
Batch adsorption experiments were designed with initial pH in solution (1-9), adsorbent dosage (0.2-1.0 g/L), adsorption time (60-300 min), and initial Cr(VI) concentration (50-250 mg/L) to optimize Cr(VI) adsorption conditions with magnetic EFD&C. Batch adsorption experiment was conducted with 100 mL of Cr(VI) solution added at 25 ± 1 °C in a constant temperature oscillator (THZ-82A, Jiangsu Jinyi Instrument Technology Co., Ltd. Jiangsu). After adsorption, the magnetic adsorbent in the solution was separated by an external magnetic field, and the solution was filtered (0.22 μm), transferred into the colorimetric tube and diluted to 50 mL with pure water. The diphenyl carbazide method (Fang et al. 2021) was used to measure the concentration of Cr(VI) in the solution at 540 nm wavelength in an ultraviolet-visible spectrophotometer (U-T3, Summit Instrument Manufacturing Co., Ltd. Shanghai). Measured absorbance of the solution was substituted into the standard curve (y = 1.593x-0.0023, R 2 = 0.9974; x meant absorbance and y meant Cr(VI) concentration) to obtain the Cr(VI) concentration in the solution. The Cr(VI) removal rate (R) and the Cr(VI) adsorption capacity (q t , mg/g) at time t were calculated according to Eqs. (1), (2), and (3): (1) where: C 0 (mg/L) initial concentration; t (min) adsorption time; C t (mg/L) Cr(VI) concentration in solution at time t; m (g) adsorbent dosage; q e (mg/g) equilibrium adsorption capacity; C e (mg/L) Cr(VI) equilibrium concentration; V (L) volume of Cr(VI) solution.

Adsorption kinetic model
Five kinetic models including pseudo-first-order (Fxd et al. 2021) (Eq. (4)), pseudo-second-order (Ho and McKay 2002) (Eq. (5)), Elovich (Cazetta et al. 2016) (Eq. (6)), intra-particular diffusion model (Asimakopoulos et al. 2020) (Eq. (7)), and Boyd model (Boyd et al. 1947) (Eq. (8)) were selected to explore the adsorption kinetics of Cr(VI) by magnetic EFD&C. The adsorption conditions were set as follows: the initial pH in the solution of 2, the adsorbent dosage of 1 g/L, the initial concentration of Cr(VI) solution of 100 mg/L, and the temperature of 25 ℃. The equations of the five kinetic models were as follows: where: q e (mg/g) the equilibrium adsorption capacity; q t (mg/g) the adsorption capacity at time t (min); t (min) the adsorption time; k 1 (g/(mg min) the adsorption rate constant of the pseudo-first-order model; (2) k 2 (g/ (mg min) the adsorption rate constant of pseudosecond-order model; α (mg/(g min)) the initial adsorption rate; β (g/mg) the desorption constant; k id (mg/(g min) the intraparticle diffusion rate constant; C (mg/g) the constant related to boundary layer thickness; F the equilibrium fraction, F = q t /q e ; k fd the film diffusion rate constant.

Adsorption isotherm model
Adsorption isotherm model was used to predict the maximum adsorption capacity. The commonly used adsorption isotherm models were the Langmuir model (Eq. (9)) and Freundlich model (Eq. (10)) (Ms et al. 2021) as follows: where: q e (mg/g) the equilibrium adsorption capacity; C e (mg/g) the equilibrium concentration; q m (mg/g) the maximum adsorption capacity; K L (L/mg) the Langmuir model constant; K f ((mg/g)(mg/L) −n ) the Freundlich model constant; n the exponent of Freundlich model. R L was a dimensionless constant (0 < R L < 1, it indicated that the Langmuir model adsorption was favorable; R L = 0, it indicated that irreversible; R L = 1, it indicated that linear; R L > 1, it indicated that unfavorable), and the calculation method was as Eq. (11) (Thta et al. 2020).
The values of enthalpy (ΔH°) and entropy (ΔS°) were determined from the slope and intercept of the Van't Hoff plot, respectively (ln K L versus 1/T).

Results and discussion
Magnetic EFD&C was detected by XRD, VSM, zeta potential, and BET to analyze its crystal composition, magnetic property, surface potential, specific surface area, pore diameter, and pore volume (Figs. 2 and 3). Batch adsorption experiments were conducted with magnetic EFD&C to optimize Cr(VI) adsorption conditions in Fig. 4. And each level in pH, adsorbent dosage, adsorption time, and initial Cr(VI) concentration on the adsorption effect were analyzed statistically by using SPSS statistical 22.0 software. The differences among treatments were performed using ANOVA and subsequent Duncan's test (p < 0.05). p < 0.05 was considered to indicate a statistically significant difference. Adsorption kinetics, adsorption isotherm curve, and adsorption thermodynamic model were used to analyze Cr(VI) adsorbed process (Figs. 5, 6, and 7 and Tables 1, 2, and 3). The Cr(VI) adsorption mechanism by magnetic EFD&C was investigated by FT-IR, XPS, and SEM-EDX (Figs. 8, 9, and 10).

XRD
The XRD diffraction pattern of magnetic EFD&C was showed in Fig. 2(a), diffraction peaks at 30°, 35°, 43°, 57°, and 62° from Fe 3 O 4 crystal were observed, and characteristic peak of Fe crystal near 45° was found (Tu et al. 2021), which indicated that Fe in EFD was successfully loaded onto reed straw powder, and then converted into magnetic Fe 3 O 4 and Fe at high temperature calcination. The Fe 3 O 4 crystal in magnetic EFD&C was produced as follows: (i) the iron oxides in EFD were dissolved by HCl into liquid products including FeCl 2 and FeCl 3 ; (ii) excess KOH was added into the liquid products to formed precipitates including Fe(OH) 2 and Fe(OH) 3 , and these precipitates were evenly adsorbed on the surface of the reed straw powder under mechanical stirring; (iii) during the pyrolysis process at 300-500 °C, two processes of reed straw powder carbonized to form a void structure and precipitates on the surface of the reed straw powder converted into Fe 3 O 4 to begin to nucleate and grow occurred simultaneously, therefore Fe 3 O 4 crystal embedded into the amorphous (14) ΔG 0 = −RTln K d state in carbon matrix (Wen et al. 2017). The formation process was showed in Eq. (15): The synthesis process of Fe crystal in magnetic EFD&C was as follows: the reducing gas (such as CO) released by the pyrolysis of reed straw powder at 300-500 °C reacted with iron oxide to produce Fe (Gao et al. 2021). Carbon was not detected by XRD, indicating that the carbon in magnetic EFD&C existed in an amorphous form (Chen et al. 2013).

VSM
The saturation magnetization result of magnetic EFD&C was showed in Fig. 2(b), and its hysteresis loop was a symmetrical "S" curve without hysteresis, indicating that magnetic EFD&C was super magnetic compliance. The saturation magnetization of magnetic EFD&C was 11.35 Am 2 /kg due to Fe 3 O 4 and Fe in magnetic EFD&C. The recovery rate of magnetic EFD&C was tested by magnetic separation using an external magnetic field as showed in the attached figure in Fig. 2(b) with recovery rate of > 90%, indicating that magnetic EFD&C can be reused by magnetic separation.

Zeta potential
The zeta potential of the magnetic EFD&C with variation in pH of solution was showed in Fig. 2(c). The zeta potential on its surface was positively charged as pH in the solution ranged from 1.00 to 6.15, decreased rapidly as pH increased from 5.00 to 6.15, and presented negative as pH > 6.15 (pH zpc ). The results suggested that the magnetic EFD&C was

BET
The results of the specific surface area, pore size, and pore volume of the magnetic EFD&C were showed in Fig. 3. Its specific surface area was 536.61 m 2 /g and the pore volume was 0.21 cm 3 /g. The adsorption isotherm of the magnetic EFD&C was a typical type IV with an adsorption hysteresis loop, suggesting that the magnetic EFD&C had a mesoporous structure (Yuan et al. 2021). The average pore size of magnetic EFD&C was 1.58 nm (pore size ≤ 2 nm for micropore (Bashta et al. 2022), indicating that it had mixed structure of mesopores and micropores. The mesopores as the entrance and transport channel, while the micropores as the active site for adsorption, and both played major roles in the adsorption of heavy metals (Tan et al. 2012). Compared with other studies, such as Yin et al.
(2018) synthesized magnetic biochar adsorbents with straw/ FeCl 3 ·6H 2 O as raw materials at different pyrolysis temperatures, with a maximum specific surface area of 357.84 m 2 /g; Jian et al. (2020) used corn stover as raw material to synthesize a series of magnetic biochar adsorbents, with specific surface areas ranging from 10.59 to 28.47 m 2 /g and Xiao et al. (2019) prepared a magnetic chitosan biochar adsorbent by carbonization at 700 °C, with a maximum specific surface area of 337.35 m 2 /g, synthesized magnetic EFD&C adsorbent in this study owned larger specific surface area of 536.61 m 2 /g and provide more adsorption sites to reach significant adsorption capacity (Xu et al. 2019a, b). So magnetic EFD&C had more advantages in adsorption structure.

Effect of pH
The removal rate of Cr(VI) referred to the ratio of concentration of Cr(VI) removed/its original concentration. and initial Cr(VI) concentration of 100 mg/L, the effect of pH on the removal rate and adsorption capacity of Cr(VI) by magnetic EFD&C was investigated. As shown in Fig. 4(a), both the removal rate and adsorption capacity of Cr(VI) reached the maximum of 99.80% and 99.80 mg/g at pH of 1, respectively. As the pH continued to increase to 9, the removal rate and adsorption capacity of Cr(VI) reduced to 9.23% and 9.23 mg/g, indicating that magnetic EFD&C had stronger adsorption activity for Cr(VI) under acidic conditions. The main forms of Cr(VI) in solution were HCrO 4− at pH < 1, Cr 2 O 7 2− and HCrO 4− at pH increasing from 1 to 6 (Tu et al. 2021). Under acidic conditions, the OH − group in solution was neutralized by H, which reduced the adsorption competition between the OH − group and Cr(VI) (Rajput et al. 2016). The zeta potential of magnetic EFD&C in Fig. 2(c) suggested it was positively charged at pH < pH zpc (6.15). Positively charged components on the surface magnetic EFD&C came from Fe(OH) 2+ (the main form of iron oxide under acidic conditions) and positively charged functional groups (derived from biochar) can make Cr 2 O 7 2− and HCrO 4− deposited on the surface of the positively charged material by electrostatic attraction (Rajput et al. 2016), so pH < 6.15 was favorable for Cr(VI) adsorption (Lb et al. 2021). At pH > 6.3, Cr(VI) mainly existed in the form of CrO 2 4− , which were difficult to be absorbed by magnetic EFD&C as the following reasons: (i) OH − as a competitor hindered the diffusion of dichromate anions to the surface of magnetic EFD&C; and (ii) the surface of the EFD&C adsorbent with negatively charged at pH > pH zpc (6.15) produced electrostatic repulsion with CrO 2 4− (Zhong et al. 2018), resulting in a decrease in the adsorption rate and adsorption capacity of Cr(VI) (Sharma et al. 2009). Potential adsorption mechanism of Cr(VI) may involve electrostatic attraction and ion exchange (Zhou et al. 2016). In addition, the Fe leaching amount from magnetic EFD&C at pH 1 and 2 were 14.6 mg/L (higher than the industrial wastewater discharge standard of 3-10 mg/L (Lb et al. 2021)) and 0.14 mg/L (lower than the industrial wastewater discharge standard of 3-10 mg/L (Lb et al. 2021), respectively. Therefore, pH value of 2 was selected for subsequent experiments.

Effect of adsorbent dosage
Under the conditions of pH = 2, adsorption time of 180 min and initial Cr(VI) concentration of 100 mg/L, the effect of magnetic EFD&C dosage added on Cr(VI) removal rate and adsorption capacity was studied. As shown in Fig. 4(b), the adsorbent dosage was set as 0.2 g/L, the removal rate was only 20.54%. As an increase in adsorbent dosage from 0.2 to 1.0 g/L, the removal rate of Cr(VI) increased to 58.62%. Therefore, adsorbent dosage of 1.0 g/L was choosing to investigate the effect of adsorption time on Cr(VI) adsorption.

Effect of adsorption time
Under the conditions of pH = 2, adsorbent dosage of 1 g/L and initial Cr(VI) concentration of 100 mg/L in Fig. 4(c), the Cr(VI) adsorption capacity with magnetic EFD&C increased rapidly with time at the initial stage (0-60 min) to reach 54.40 mg/g within 60 min due to the diffusion effect between Cr(VI) and EFD&C and the fast physical adsorption by van der Waals forces (Karthikeyan et al. 2020). In the initial adsorption stage, the large concentration difference of Cr(VI) between the adsorbent surface and the solution as the driving force stimulated the rapid diffusion of Cr(VI) to the adsorbent surface ). As shown in Fig. 4(c), the Cr(VI) adsorption capacity of EFD&C increased from 54.40 to 62.36 mg/g (the adsorption rate was 0.03 mg/ (g min)) within 60-300 min, the adsorption rate of EFD&C decreased significantly compared with 0-60 min (adsorption rate was 0.91 mg/(g min) in 0-60 min), because prolonged adsorption time and the gradual coverage of the active sites on the adsorbent resulted in a decrease in the adsorption rate of Cr(VI).

Effect of initial Cr(VI) concentration
Under conditions of pH = 2, the EFD&C dosage of 1 g/L and adsorption time of 180 min, as the initial Cr(VI) concentration increased from 50 to 200 mg/L, the Cr(VI) adsorption capacity with EFD&C increased from 49.96 to 77.22 mg/g, and the highest value (80.94 mg/g) was reached at 250 mg/L in Fig. 4(d). Compared with the removal rate of 99.92% at the initial Cr(VI) concentration of 50 mg/L, the removal rate decreased to 32.38% at 250 mg/L. The removal rate was almost 100% at the initial concentration of 50 mg/L, indicating that under this condition, the EFD&C could completely adsorb Cr(VI) in wastewater.
In summary, the effect of pH on the Cr(VI) adsorption with magnetic EFD&C was obvious. At pH of 1, magnetic EFD&C had good adsorption performance, but a large amount of leached Fe was detected. Under conditions of pH = 2, EFD&C dosage of 1 g/L, adsorption time of 180 min, and the initial concentration of Cr(VI) of 50 mg/L, the Cr(VI) removal rate reached 100%, which provided a reference for the practical application of EFD&C adsorbent. In order to explore the maximum adsorption capacity of magnetic EFD&C, the initial concentration of Cr(VI) was selected to be ≥ 100 mg/L to study the adsorption kinetics, adsorption isotherm curve, and adsorption thermodynamics.

Adsorption kinetics
The adsorption kinetics model was conducted to study the adsorption rate and adsorption kinetics characteristics of Cr(VI) with magnetic EFD&C. In the adsorption kinetic experiment, the initial pH in the solution was set to be 2, the initial Cr(VI) concentration was 100 mg/L, adsorbent dosage was 1 g/L, and the adsorption temperature was 25 °C. The nonlinear fitting module in origin 2018 was used to fit the experimental data by given five kinetic models. As shown in Fig. 5(a) and (b), the data fitting results of the pseudo-first-order model and pseudosecond-order model were showed. The correlation coefficient (R 2 ) of the pseudo-second-order model was more than 0.99, and the maximum Cr(VI) adsorption capacity  (q e ) obtained by fitting was 65.91 mg/g, which was closed to the value (q e, exp = 66.34 mg/g) measured in the experiment. The results showed that the pseudo-second-order kinetic model can fit the kinetic process of Cr(VI) adsorption on EFD&C materials well, and the mechanism of Cr(VI) adsorption was mainly chemical adsorption. The correlation coefficient (R 2 > 0.95) of the pseudo-first-order model indicated that the adsorption process was accompanied by physical adsorption. The Elovich model assumed that the surface adsorption energy of the adsorbent was heterogeneously distributed, and there was no interaction between the adsorbates, and was used to analyze Binding energy(eV) Intensity (a.u.)

Fe 2p (C) Fe (III)
Fe (0)  the adsorption behavior on heterogeneous solid surfaces (Zhang et al. 2020a, b). In Fig. 5(c) and Table 1, the correlation coefficient (R 2 > 0.98) of the Elovich model indicated that the Elovich model can also simulated the kinetic behavior of Cr(VI) adsorbed by magnetic EFD&C (Yuan et al. 2021). Comparing the values from the adsorption rate of α (76.4084 g/(mg min 2 )) and desorption rate of β (0.1299 (mg/(g min)), the adsorption rate was higher than the desorption rate, which indicated that magnetic EFD&C had the ability to adsorb Cr(VI) efficiently. It can be seen from the fitting diagram of the Elovich model (Fig. 5(c)) that the change of q t with time (t) can be divided into three stages (Long et al. 2021): (i) the rapid adsorption stage was within 60 min after magnetic EFD&C added, which can be attributed to the polar interaction (physical adsorption) between magnetic EFD&C and Cr(VI), and the electrostatic attraction between the rich active functional groups of EFD&C and Cr(VI); (ii) the adsorption rate gradually decreased during 60-660 min because of the active sites on EFD&C gradually occupied; (iii) after 660 min, the adsorption process reached equilibrium with the adsorption sites occupied. In all that magnetic EFD&C was a heterogeneous solid adsorbent with the Cr(VI) adsorbed behavior from mainly chemical adsorption with partially physical adsorption. The Cr(VI) adsorption rate on magnetic EFD&C adsorbent was faster in < 60 min and reached equilibrium in 660 min. The dynamic boundary model was widely used to describe the adsorbed behavior of porous materials to determine the rate-limiting step of the adsorption rate (Xu et al. 2022). The adsorption experimental data (t vs. q t ) under conditions of the initial solution pH of 2, the initial Cr(VI) concentration of 100 mg/L, adsorbent dosage of 1 g/L, and the adsorption temperature of 25 °C was fitted to study the rate-limiting step of the adsorption process with dynamic boundary model equation. The intra-particle diffusion model assumed that the diffusion resistance was negligible and that the direction of diffusion was random. As showed in Fig. 5(d), it can be clearly seen that two lines (L 1 and L 2 ) with different slopes represented different adsorption processes: the first stage (L 1 ) was Cr(VI) passing the outer surface of EFD & C; the second stage (L 2 ) was the Cr(VI) diffusing from the surface or pores to the adsorption sites (Hu et al. 2011). In Table 1, it can be seen that the fitting effect of the intraparticle diffusion model in the two stages was good (both R 1 2 and R 2 2 > 0.95), which indicated that the intraparticle diffusion was the main ratelimiting step of EFD&C for Cr(VI) adsorbed. In addition, the fitting curve of the intraparticle diffusion model had not yet passed the origin (C ≠ 0), which suggested that intraparticle diffusion was not the only step controlling the adsorption rate and may also be affected by other reaction factors (Feng et al. 2018).
Since intraparticle diffusion included film and pore diffusion, Boyd (Eq. (16)) model was used to further analyze kinetic data in order to determine the actual rate-controlling step involved in the sorption process (Djeribi and Hamdaoui, 2008). Introduce Eq. (17) into Eq. (16) to acquire Eq. (18). where: F the fractional attainment of equilibrium at time t, F = qt/qe; q t (mg/g) the Cr(VI) adsorbed capacity by EFD&C at time t; q e (mg/g) the maximal Cr(VI) adsorbed capacity; B (1/min) the time constant; D i (cm 2 /min) the effective diffusion coefficient of metal ions in the sorbent phase; r (cm) the radius of the sorbent particle.
Assumed to EFD&C powder be spherical, and m was an integer that defined the infinite series solution (Vadivelan and Kumar 2005). Bt was given by the Eq. (19): Thus, the value of Bt can be computed with each value of F, and the Boyd plot was plotted according to time (t)and Bt. If the Boyd plot was a straight line passing through the origin, it meant that rate-controlling step was particle-diffusion mechanism, and if not, it was film diffusion. In Fig. 5(e), it can be observed that the relationship between t and Bt was not linear (not crossing the origin), which indicated that the film diffusion was the main rate-controlling step in the intraparticle diffusion (Hu et al. 2011). Both poor mixing and low Cr(VI) concentration led to increase adsorption resistance (Vadivelan and Kumar 2005).

Adsorption isotherms
The Cr(VI) adsorbed isothermal by EFD&C was studied with Langmuir model and Freundlich model. Langmuir isotherm model assumed that the adsorbent owned uniform surface, and a single molecule layer was formed on its surface with uniform adsorption energy. Freundlich model described multilayer adsorption with uneven affinity on heterogeneous surfaces with uneven adsorption energy (Lv et al. 2012). The experimental conditions were conducted as follows: the initial pH of 2, the initial Cr(VI) concentration of 100-300 mg/L, adsorbent dosage of 1 g/L, and the adsorption temperatures of 25, 35, and 45 °C respectively. The nonlinear fitting module in origin (2018) was used to fit with results showed in Fig. 6 (Yin et al. 2019). In addition, the Langmuir constant (K L ) reflected the affinity between the sorbent and the sorbate (Rajput et al. 2016). It was found that K L increased from 0.0580 to 0.0760 L/mg as temperature growing from 25 to 45 °C, indicating that Cr(VI) was more firmly combined with adsorption sites in the EFD&C at higher temperature. It can be preliminarily judged that the Cr(VI) adsorbed process by EFD&C was endothermic reaction. The calculated value of R L (R L at 25 °C, 35 °C, and 45 °C were 0.0543, 0.0467, and 0.0420 respectively) was between 0 and 1, which further explained that EFD&C was conducive to the Cr(VI) adsorbed (Thta et al. 2020).

Adsorption thermodynamic characteristics
Nonlinear fitting module in Origin (2018) was used to fit the thermodynamic model of Cr(VI) adsorbed with the results showed in Fig. 7. The thermodynamic parameters (R 2 = 0.9945) obtained by fitting were listed in Table 3. The experimental conditions of isothermal adsorption curve were conducted as follows: initial solution pH of 2, initial Cr(VI) concentration of 100 mg/L, EFD&C dosage of 1 g/L, and adsorption temperatures of 25, 35, and 45 °C, respectively. The ΔG° values at 25, 35, and 45 °C were negative, indicating that the Cr(VI) adsorbed by EFD&C was spontaneous process. Generally, ΔG° values were between 0 and -20 kJ/mol indicated that electrostatic interaction (physical adsorption) occurred, while ΔG° values ranging from − 80 to -400 kJ/mol indicated that adsorption involved charge sharing or to form coordination bonds (chemisorption) ). In Table 3, ΔG° at 25, 35, and 45 °C were − 1.9197, − 2.3554, and − 2.8975 kJ/mol, respectively, indicating that the adsorption process involved physical adsorption. ΔH° (11.0135 kJ/mol) was positive, indicating that the Cr(VI) adsorbed by EFD&C was an endothermic process (Zhou et al. 2021), and increasing temperature was beneficial to the adsorption process. ΔS° (0.07 kJ/(mol K)) was also positive indicating that EFD&C owned high affinity for Cr(VI) (Long et al. 2021).

Adsorption mechanism analysis
Characterizations including FT-IR, XPS, and SEM-TEM were used to analyze the Cr(VI) adsorbed mechanism by EFD&C. In Fig. 8(a, b), the FT-IR of EFD&C before and after the adsorption of Cr(VI) were showed. For the FT-IR of fresh EFD&C in Fig. 8(a), the absorption peaks at 601 cm −1 and 578 cm −1 corresponded to Fe-O bond vibrations from Fe 3 O 4 (Lee and Reucroft 1999), and the absorption peaks at 3426 cm −1 , at 1640 cm −1 , and at 1000 to 1350 cm −1 were attributed to O-H stretching vibrations, the stretching vibration of the C = O double bond (Panneerselvam et al. 2011), and the overlapping signals of the C-C and C-O groups (Wen et al. 2017), respectively. After adsorbing Cr(VI) in Fig. 8(b), its Fe-O bond vibrations weakened at 601 cm −1 and 578 cm −1 , which may be due to the surface complexation between Fe-O and Cr(VI) during the adsorption process , and the C = O, C-C, and -OH bonds at 1640 cm −1 , 1000 cm −1 , and 3426 cm −1 weakened because of the functional groups in the biochar involved in the Cr(VI) adsorbed process.
The XPS spectra of EFD&C were showed in Fig. 9A-D. There were three main peaks of C 1 s, O 1 s, and Fe 2p in the EFD&C in Fig. 9A(a), and after adsorbing Cr(VI), a typical peak of Cr 2p appeared in Fig. 9A(b). After fitting the peaks corresponding to the Cr, the peaks corresponding to Cr 2p were divided into two parts including Cr(VI) (578.3 eV and 587.82 eV) and Cr(III) (577.1 eV and 586.7 eV) in Fig. 9B (Lb et al. 2021). The areas of the Cr(VI) and Cr(III) peaks accounted for 57.99% and 43.01% of the total area of the Cr 2p peaks, respectively, suggesting Cr(VI) and Cr(III) coexisted in the EFD&C, which indicated that Cr(VI) adsorbed mechanisms including physical adsorption and chemical reduction (Zhou et al. 2019). There were three main diffraction peaks for Fe 2p in Fig. 9C, the peak at the binding energy of 711.2 eV attributed to Fe 2p 3/2 , and the peaks at the binding energy of 719.85 eV and 725.6 eV attributed to Fe 2p 1/2 . The diffraction peak at the binding energy of 711.2 eV can be divided into two peaks that attributed to Fe(III) (710.8 eV) and Fe(II) (713.8 eV), respectively, and the peak at a binding energy of 719.85 eV attributed to Fe(0) (in agreement with the XRD analysis, EFD&C contained Fe 3 O 4 , and Fe crystals). It further indicated that the iron in EFD was successfully loaded on the surface or embedded inside the biochar material during the EFD&C preparation. The peak at the binding energy of 725.6 eV can be divided into two peaks attributed to Fe(III) (724.2 eV) and Fe(II) (726.5 eV), respectively (Lb et al. 2021). After adsorbing Cr (VI), the peak area ratio occupied by Fe(III) increased to 56.62% (vs. 40.91% for fresh EFD&C), this from Fe(II) decreased to 35.94% (vs. 50.14% for fresh EFD&C), and that from Fe(0) slightly decreased 7.44% (vs. 8.45% for fresh EFD&C). The reasons were as follows: during the adsorption process, Cr 2 O 7 2− and HCrO 4− (main form of Cr (VI) in aqueous solution (Kenji et al. 2017)) were deposited on the surface of EFD&C due to electrostatic forces and were subsequently reduced to Cr(III) by Fe(II) or Fe(0) in Eqs. (20)-(23) (Lb et al. 2021;Dong et al. 2017;Qi et al. 2022).
The reductions of Cr (VI) to Cr (III) by Fe 2+ were: The reductions of Cr (VI) to Cr (III) by Fe 0 were: During the reduction process, part of Fe(II) and Fe(0) were converted to Fe(III) and part of Cr (VI) was reduced to Cr (III), thus leading to an increase in the proportion of Fe(III) and a decrease in the proportion of Fe(II) and Fe(0), and also indicated the presence of both Cr(VI) and Cr(III) in the Cr 2p peak. The Fe(II) and Fe(0) diffraction peaks of EFD&C after adsorbing Cr(VI) still presented, indicating that it can be recycled.
The C 1 s spectra of EFD&C after adsorbing Cr(VI) was showed in Fig. 9D, and the C 1 s can be divided into three diffraction peaks of C = C, C-C, and C-H (284.8 eV), C = O (286 eV), and O = C-O (288.6 eV) (Wang et al. 2021). The percentage of C = C, C-C, and C-H in EFD&C from XPS spectra was 63.60% and it decreased to 57.36% after adsorbing Cr(VI), while the percentage of oxygencontaining groups of C = O and O = C-O increased to 27.26% and 15% (24.83% and 11.57% for fresh EFD&C), respectively. The reason for this was the oxidation of C = C, C-C, and C-H on the surface of EFD&C to generate oxygen-containing functional groups (C = O and O = C-O), accompanied by partial reduction of Cr(VI) to Cr(III) (Lu et al. 2022), with the reaction process as in Eqs. (24) and (25)  . The O 1 s spectra of EFD&C before and after Cr (VI) of adsorption (Fig. 9E) can be divided into four diffraction peaks: peaks at the binding energy of 530.1 eV and 531.26 eV were Fe-O (Choi et al. 2015) and -OH (Long et al. 2021), respectively, and peaks at 531.74 eV and 533.41 eV were C = O/O-C = O and Fe-C-O , respectively. The comparison between before and after adsorption of Cr(VI) showed that the percentage and peak positions of these functional groups have changed, indicating that all functional groups in the EFD&C adsorbent were involved in the adsorption of Cr(VI). The percentage of -OH increased from 14.98 to 33.24%, probably due to the formation of chromium hydroxide on the surface of the EFD&C adsorbent (Qi et al. 2022). In addition, carbonyl (C = O) and carboxyl (-OH) groups can further served as binding sites for Cr(III) (Xu et al. 2019a, b), and the presence of hydrogen ions can greatly increase the activity of these functional groups under strongly acidic conditions (Tu et al. 2021).
where Composite-C represented the carbon-containing functional groups on the adsorbent surface and Composite-CO X H represented the new oxygen-containing functional groups generated in EFD&C. The microscopic morphology and surface element contents of EFD&C were showed in Fig. 10. The rough surface and obvious pore structure of EFD&C can be clearly (24) Composite − C − e − ↔ Composite − CO X H (25) 6CO X H + Cr 2 O 7 2− + 8H + → 6CO X + 2Cr 2 3+ + 7H 2 O seen in Fig. 10A(a). For the EDX results ( Fig. 10A(b)), the contents of C, O, and Fe on the surface of EFD&C were 74.91 wt%, 16.85 wt%, and 7.91 wt%, respectively. After adsorbing Cr(VI), Cr was detected with content of 0.82 wt% ( Fig. 10B(b)), while the contents of C and Fe decreased to 71.32 wt% and 5.74 wt%, and the content of O increased to 22.12 wt% because of the oxidation of C = C, C-C and C-H. In summary, it can be seen that the Cr(VI) adsorbed process by EFD&C attributed to chemisorption and both physical adsorption. Firstly, Cr(VI) in aqueous solution was adsorbed onto the surface of EFD&C by Fe(OH) 2+ and positively charged functional groups on the surface of biochar with electrostatic gravitational force. Then, part of Cr(VI) was converted to Cr(III) by redox reaction with Fe(II), Fe(0), and carbon-containing functional groups. The adsorption mechanism was shown in Fig. 11. And the feasibility of this study was further illustrated by comparison with the same type of carbon-based composite adsorbents. The results were shown in the Table 4. It can be seen from the table that the adsorbent prepared in this study had good adsorption performance for hexavalent chromium, indicating that the carbon-based magnetic adsorbent derived from reed straw and electric furnace dust was feasible and had potential for application.

Conclusion
Reed straw and electric furnace dust were used as raw materials to prepare magnetic EFD&C by three-step process: (1) electric furnace dust was dissolved with HCl to prepare dissolution solution; (2) reed straw was added to the dissolution solution to prepare EFD&C precursors by NaOH precipitation; (3) magnetic EFD&C was activated by high-temperature calcination. The adsorption kinetic model analysis showed that the magnetic EFD&C was a non-homogeneous solid adsorbent, and the Cr(VI) adsorbed   Ruan et al. (2015) magnetic EFD&C by reed straw and electric furnace Cr(VI) 111.94 2.0 In this study behavior was mainly chemisorption and physical adsorption. The Cr(VI) adsorbed rate by magnetic EFD&C was fast at < 60 min and reached equilibrium at 660 min with film diffusion as the main rate-limiting step. The isothermal adsorption curves and adsorption thermodynamic model analysis indicated that the Cr(VI) adsorbed process by magnetic EFD&C was monolayer adsorption, the adsorption process was a heat absorption reaction, and increasing the temperature was favorable for the Cr(VI) adsorbed process.
In the chemisorption process, Fe(II) and Fe(0) in EFD&C were oxidized to Fe(III), and C = C, C-C, and C-H on the surface of EFD&C were oxidized to produce oxygen-containing functional groups (C = O and O = C-O), accompanied by reduction of Cr(VI) to Cr(III). The results showed that the magnetic EFD&C prepared from reed straw and electric furnace dust had good effect on Cr(VI) removal, and this study provided guidance for the resource recovery of iron-containing metallurgical solid waste and solid biomass, but also provided a reference for the preparation of composite carbon-based adsorbent to achieve the purpose of "using waste to treat waste".