Magnetic Fe2O3/biochar composite prepared in a molten salt medium for antibiotic removal in water

The large-scale use of antibiotics is causing serious water pollution problems, and it is of great significance to develop new technologies to remove antibiotics from water. As an environmentally friendly and economical adsorption material, carbon derived from biomass is a low-cost and feasible material for removing antibiotics in sewage, but the current removal efficiencies are not high enough for large-scale practical application. In this study, poplar wood chips are used as raw material, and a magnetic biochar is prepared by co-pyrolysis of poplar wood chips and FeCl3/CaCl2 mixed molten salt. X-ray diffraction (XRD), Fourier transform infrared spectroscopy (FTIR), N2-isothermal adsorption and desorption, X-ray photoelectron spectroscopy (XPS) and scanning electron microscope (SEM) techniques showed the successful synthesis of Fe2O3/bioC composite. In-situ formed Fe2O3 makes the biochar possess the characteristic of superparamagnetic, which is conducive to the recycling of biochar. Due to the etching effect of the molten salts, rough surface was formed on Fe2O3/bioC, resulting in a maximum norfloxacin (NOR) adsorption ability up to 38.77 mg g−1 at pH 6.0. The NOR adsorption behavior on Fe2O3/bioC followed the pseudo second order kinetic model and the equilibrium data was best fitted the Langmuir model. In addition, the adsorption process of NOR on Fe2O3/bioC was thermodynamically spontaneous. The results show that this low-cost and reusable magnetic biochar has the potential for rapid and efficient removal of antibiotic from aqueous solution. A magnetic Fe2O3 modified porous biochar composite was prepared in a molten-salt medium. Molten salt made a higher proportion of micropores in Fe2O3/biochar during the carbonization. Fe2O3/biochars composite exhibited higher removal efficiency for NOR than original biochars. A magnetic Fe2O3 modified porous biochar composite was prepared in a molten-salt medium. Molten salt made a higher proportion of micropores in Fe2O3/biochar during the carbonization. Fe2O3/biochars composite exhibited higher removal efficiency for NOR than original biochars.


Introduction
Norfloxacin (NOR), a third-generation quinolone antibiotic is widely used in daily life due to its broad-spectrum antibacterial properties, low toxicity and good oral absorption (Liu et al. 2021a). In 2015, Chinese Academy of Sciences published the "Antibiotic Pollution Map of Rivers in China", showing that the average concentration of antibiotics in major rivers was about 303 ng L −1 , and the concentration of NOR in the Pearl River (Guangdong province) was even higher than 1000 ng L −1 (Zhang et al. 2015). Antibiotics in water and soil can cause multiple threats to human body, such as toxic damage, allergic reactions, reduced immunity, and it can also lead to the evolution of drug-resistant bacteria (Luo et al. 2020;Fang et al. 2021a;Li et al. 2021). Therefore, it is of great significance to develop low-cost and highly efficient technologies to remove antibiotics from wastewater.
There are many ways to remove the antibiotics from water, such as chemical oxidation, bio-degradation, photodegradation, electrochemical degradation, and adsorption (Langbehn et al. 2021;Wu and Hu 2021;Du et al. 2021;Hung et al. 2021;Liu et al. 2021b). Among them, the adsorption method has several advantages over other methods, including high removal efficiency, low energy consumption, easy operation, and environment-friendly. Carbon nanotubes (Ncibi and Sillanpää 2015;Soran et al. 2017), graphene (Wang et al. 2021a) and carbon aerogels Liu et al. 2019a) with high specific surface area and hierarchical pore structure, are commonly used as the adsorbent. However, the complicated synthesis process and high cost of these carbon materials impede their largescale application. Therefore, it is a promising and meaningful work to develop low-cost carbon materials with superior adsorption capability.
Currently, biochars have been developed via carbonizing the biomass and then used as electrode materials for energy storage and adsorbents for antibiotics Zhou et al. 2021a). However, the adsorption capacities of unmodified biochars are usually quite low, and their shortcomings such as small particle size and low density are making them difficult to separate from the treated solution. Combining the magnetic medium with biomass not only enhances the adsorption performance of pollutants, but also can be easily isolated by using external magnet (Thines et al. 2017;Li et al. 2020). Shan et al. 2016 prepared magnetic Fe 3 O 4 /biochar by ball milling, which exhibited high removal efficiency for carbamazepine. However, the bond between Fe 3 O 4 and biochar was weak, 74: which made it difficult to ensure the reusability of magnetic biochar. Fe 0 /biochar composite prepared by liquid phase reduction-deposition method almost completely adsorbed pentachlorophenol from industrial effluent (Devi and Saroha 2014). However, nano zero valent ion was unstable in the environment, which may lead to a sharp decline in adsorption performance. Yap et al. 2017 prepared a magnetic biochar via impregnating the coconut shell with FeCl 3 and subsequent microwave pyrolysis. The results showed that the optimum adsorption capacity of magnetic biochar was achieved at the FeCl 3 /biomass ratio of 0.5, due to the best pore stucture. Although there are many reports on magnetic biochar, the complex synthesis methods (impregnation calcination and coprecipitation) hinder its further large-scale production. Preparing porous carbon materials in molten salt media has advantages of simple process, low synthetic temperature, easy separation, uniform chemical composition, and low cost over other methods (Pang et al. 2021;D´ıez et al. 2021;Guo et al. 2020). The molten salt is a kind of ionic liquid, which can avoid the desolvation of ions in traditional chemical synthesis process, reduce activation energy, and enable rapid and efficient synthesis of materials . By adjusting the ratio of the mixed molten salt, the pore structure of the obtained carbon materials can be controlled precisely (Pang et al. 2021;D´ıez et al. 2021). In addition, the high-energy ions in the molten state have certain etching effect, as well as recrystallization and embedding effect in the cooling stage, resulting in the formation of a hierarchical porous structure including micro-, meso-, and macro-pores in the obtained carbon material Yang et al. 2020).
In this study, a magnetic Fe 2 O 3 /biochar composite derived from poplar wood was synthesized in mixed molten salt of CaCl 2 /FeCl 3 at high temperature and then used as highly efficient adsorbent to remove NOR from water. The physical and chemical properties of biochars in the molten salt medium were investigated by XRD, TEM, and BET. The adsorption properties of the obtained magnetic biochar for NOR under different reaction conditions were evaluated, and the adsorption fitted isotherm and kinetics were also applied to study the adsorption properties. The recyclability and reuse of the as-synthesized adsorbent for NOR in water were also investigated in this study. The adsorption capability under various conditions, such as different pH, was also investigated. Also this work explored the model of adsorption isotherms and kinetics, and the adsorption mechanism of NOR.

Materials and chemicals
Poplar wood (PW) was collected from local wood processing plants and further crushed and passed through a 100-mesh sieve. Anhydrous FeCl 3 , CaCl 2 and HNO 3 were purchased from Sinopharm Chemical Reagent Co., Ltd. All the reagents were of analytical grade and used without further purification. NOR (98%) was obtained from Macklin Reagent Co., Ltd. The physicochemical properties of NOR are shown in Table 1.

Preparation of magnetic Fe 2 O 3 /biochar composite
The PW sawdust was washed thoroughly with deionized water, dried and stored in an oven at 60 °C. Clean PW chips were mixed with CaCl 2 /FeCl 3 (molar ratio = 1:2) and the ratio of PW/salts was 1:10. The mixture was put in a crucible and placed in a tubular furnace. The sample was heated up to 700-900 °C with a heating ramp of 5 °C min −1 in Ar atmosphere and kept at this temperature for 1 h. The sample was taken out after cooling, washed with 0.1 M HNO 3 aqueous solution (10 mL) and deionized water, filtered out and dried in a vacuum oven. The obtained carbonized sample was labeled as CPW-T, in which T stands for calcination temperature. To investigate the effect of molten salt medium, the carbonized PW sample was also prepared without adding any salt under identical conditions. The final product was denoted as PWC-T.

Physical characterizations
The obtained biochar was mixed with deionized water (mass ratio of water and biochar was 20:1), and pH of the solution was measured using pH meter (PHSJ-4F, INESA, China). The specific surface area and pore size distribution were investigated by N 2 -isothermal adsorption and desorption (Autosorb iQ, Quantachrome, USA). The vibrating sample magnetometer (PPMS-9 T, 7410 QuantunDesign, USA) was used to measure the magnetic properties of the sample. The X-ray diffraction (XRD) patterns were recorded on Shimadzu XD-3A instrument using filtered Cu-K α radiation (λ = 0.15418 nm) at 40 kV and 30 mA. Field emission scanning electron microscope/energy-dispersive X-ray spectroscopy (SEM-EDX) (FEI QuantaTM 250, FEI, USA) was applied to observe the morphology and surface element distribution. The surface properties of biochars were characterized by Fourier transform infrared spectroscopy (FTIR) (VERTEX 80 V, Bruker, Germany). X-ray photoelectron spectroscopy (XPS) (ESCALAB 250Xi, Thermo Fisher, USA) was used to characterize the chemical state of biochars, and the binding energy was calibrated by C 1s at 284.8 eV. Thermogravimetric analysis (TGA) was carried out on a TG analyzer (SDT Q600, TA, USA) at a heat rate of 10 °C min −1 between 25 °C and 800 °C in air.

Adsorption experiments
The stock solution of NOR (200 mg L −1 ) was prepared by mixing NOR powder with deionized water. Prior to the tests, the stock solution was diluted to a desired concentration to form working solution. All adsorption experimental solutions were shaken at 200 rpm min −1 in dark box. The initial solution pH was adjusted by 0.01 mol L −1 HCl and 0.01 mol L −1 NaOH solutions. The adsorption isotherm experiments were carried out by vigorously shaking 20 mL of NOR solution (5 to 200 mg L −1 , pH 6) mixed with 10 mg biochar for 24 h to reach equilibrium. The adsorption kinetics was investigated by adding 10 mg of biochar into 20 mL NOR solution (50 mg L −1 , pH = 6). The supernatant was filtrated via centrifuging and the concentrations of NOR were measured by UV-vis spectrophotometer (UV-5500PC, Metash, China) at the wavelength of 273.0 nm. The used adsorbent was washed to neutral by deionized water and ethanol and then dried at 60 °C for the cycle test. All the experiments were performed under identical conditions. Adsorption capacity and removal efficiency were calcualated by the following Eqs. where, q e is the equilibrium adsorption capacity of the corresponding biochar when the adsorption reaches equilibrium (mg g −1 ), R is the removal efficiency (%), C 0 is the initial concentration of NOR (mg L −1 ), C e is the concentration of NOR when the adsorption reaches equilibrium (mg L −1 ), V is the volume of the solution participating in the adsorption reaction (L), and m is the mass of added biochar (g).

The physical and chemical properties of the obtained biochars
The material composition and crystal structure of CPW-700 and PWC-700 were firstly characterized by XRD, shown in Fig. 1a. In the XRD pattern of PWC-700, there were only 26.5° and 42.2° diffraction peaks, which corresponded to (002) and (100) (012), (104), (113), (024), (116), (214), (300) and (1010) (206), (119), (0012) and (4012) crystal planes of γ-Fe 2 O 3 (JCPDF#25-1402) respectively were also observed in the XRD pattern of CPW-700. The results showed that PW was carbonized in the molten salt media and Fe 2 O 3 was formed in situ during the pyrolysis process. Figure 1b shows the hysteresis curve measured using vibrating sample magnetometer in room temperature. It can be seen that the specific saturation magnetization of CPW-700 is 47.0 emu g −1 , the remanence magnetization Mr is 6 emu g −1 , and the coercivity Hc is 180 Oe, indicating its typical superparamagnetic behavior. The specific saturation magnetization of PWC-700 is zero, suggesting that it does not have magnetism. CPW-700 has good magnetic properties, which facilitates the separation from aqueous solution.
The porosity of CPW-700 and PWC-700 was evaluated by N 2 adsorption/desorption isotherms. As shown in Fig. 1c, there was an obvious uptake at relatively lowpressure region for both samples, implying that there were plenty of micropores in both samples. A hysteresis loop was exhibited at the relative high-pressure region of the curve of CPW-700, indicating mesopores existed in CPW-700 (Mercy et al. 2012). The hierarchical porous structure of CPW-700 was further confirmed by pore size distribution (inset of Fig. 1c). Such hierarchical porous structure not only has high surface area but also provides the pathway for efficient mass transfer (Dong et al. 2021). Table S1 (Additional file 1) shows the porous parameters of two samples. The Brunauer-Emmett-Teller (BET) specific surface area of CPW-700 is 431.6 m 2 g −1 , slightly higher than that of PWC-700 (422.1 m 2 g −1 ). The average pore size of CPW-700 calculated by DFT method is ca. 2.1 nm, smaller than that of PWC-700 (3.2 nm). In the molten salt system, the eutectic point of mixed molten salts of CaCl 2 /FeCl 3 is only 100 °C, when the molar ratio of CaCl 2 to FeCl 3 is 1:2 (http:// www. facts age. cn/ fact/ docum entat ion/ FTsalt/ FTsalt_ list. htm). The liquid molten salt can penetrate into the pores of biomass, etching the biomass during the carbonization, forming more micropores. Simultaneously, Fe 2 O 3 nanoparticles were also formed in-situ and embedded in carbon material, resulting in relatively low narrower pore sizes.
FTIR spectra of CPW-700 and PWC-700 (Fig. 1d) showed many absorption peaks, which represented different functional groups in the samples. The peaks in the 3691, 1715 and 1140 cm −1 bands were caused by stretching vibrations of -OH, C = O, and -COOH, respectively, indicating that both samples contained abundant oxygen-containing functional groups. Peaks at 1500-1550 cm −1 resulted from C = C stretching vibration . For CPW-700, vibration peaks of Fe-O bonds were also found at 640, 565 and 471 cm −1 , which were attributed to the deformation vibration and stretching vibration of Fe 2 O 3 , further confirming that Fe 2 O 3 was successfully loaded on the obtained biochar (Pardeep et al. 2014). The absorption peak at 887 cm −1 indicates the presence of aromatic functional groups on the surface. These aromatic groups can provide electron donors, adsorb and bind NOR through the interaction between the electron donor and NOR. Therefore, it is expected that the CPW-700 with aromatic groups on the surface will exhibit a better NOR adsorption capacity.
The surface composition and chemical state of the obtained biochars were evaluated by XPS analysis. The XPS spectra of CPW-700 showed the presence of iron on its surface (Fig. 1e). The high-resolution C 1 s spectra were fitted into three peaks (Additional file 1: Fig. S1a). The peak at 284.8 eV corresponds to C-C/C = C, and the two peaks between 286 and 290 eV ascribes to oxygen-containing functional groups (C-O/C = O) and π-π* vibration satellite peaks respectively (Jo et al. 2020). The high-resolution Fe 2p spectrum of the TMBC-700 sample (Fig. 1f) was fitted into 4 different peaks, in which the peaks at 710.9 and 724.1 eV are attributed to Fe 3+ 2p 3/2 and Fe 3+ 2p 1/2 orbitals, while 712.9 and 726.5 eV correspond to satellite peaks (Zhou et al. 2021b). No Fe signal was detected in the XPS of PWC-700. In the fitted O 1 s XPS spectrum (Additional file 1: Fig.  S1b), the peaks at 530.4 and 532.6 eV in TMBC-700 corresponded to Fe-O bond (Shi et al. 2021), the peak at 531.2 eV ascribed to the −OH or −OOH bonds , and the peak at 533.0 eV resulted from the -OH bond of adsorbed water molecules (Liu et al. 2019b), indicating that the Fe 2 O 3 was loaded on CPW-700. Figure 2a and c exhibit that both CPW-700 and PWC-700 remain the original honeycomb-like structure of the PW after carbonization. As shown in Fig. 2b and d, the main difference between CPW-700 and PWC-700 is that the surface of CPW-700 is quite rough and coated with a layer similar to the accumulation of nanoparticles, while the surface of PWC-700 was very smooth. The elemental distribution of CPW-700 and PWC-700 was analyzed by EDS. Figure S2a (Additional file 1) shows that the C, O, N and Fe elements are evenly distributed in CPW-700, while only C, O, and N elements were detected in PWC-700 (Additional file 1: Fig. S2b), which further implied that the Fe 2 O 3 was evenly distributed in CPW-700. Intensity / a.u.

The effect of carbonizing temperature on the adsorption performance of NOR
Under the fixed ratio of molten salt and PW, the effect of carbonizing temperature on the adsorption of NOR was investigated. As shown in Fig. 3, when the dosage of biochar was 1 g L −1 , as the carbonizing temperature increased, the removal efficiency of NOR by biochar showed an upward trend, which could result from the moisture. Notably, volatile substance of biomass gradually escaped with the increase of the temperature, forming plenty of porous structures in the biochars. Compared with PWC-700, there is a significant enhancement of the removel rate of NOR for CPW-700 (Ge et al. 2021). Due to the etching effect of the molten salts, rough surface was formed on CPW-700, which could provide more adsorbing sites for NOR. The BET results showed that the specific surface of CPW-700 was higher than that of PWC-700, especially there was a higher micropore ratio in CPW-700, which can facilitate the mass transfer, resulting in a better adsorption capability of NOR.

The effect of pH and biochar dosage on the adsorption capability of NOR
Antibiotics are usually polar and hydrophobic materials, and thus electrostatic attraction, hydrophobic distribution, π-π conjugation, cation exchange, and hydrogen bonding may be involved in the mechanism of their adsorption. The pH value of NOR solution may affect the surface charge of the biochar. In addition, NOR has two acid dissociation constants in aqueous solution, pKa 1 = 6.30 (related to the deprotonation of the carboxyl group (-COOH)) and pKa 2 = 8.38 (related to the protonation of the amino group (−N 2 C 4 H 9 ) on the piperazinyl group). Therefore, At pH ≤ 6.34, NOR 6.34, NOR molecules are mainly present in cationic form (NOR + ); at pH > 8.75, NOR molecules are mainly present as NOR − in solution; Proton dissociation occurs and NOR mainly exist in the form of neutral molecules or amphoteric NOR. pH value is directly related to the existed form of NOR molecules (Fu et al. 2021).
Therefore, the NOR adsorption ability of CPW-700 was investigated in the pH range of 3.0-9.0, as shown in Fig. 4a. When the dosage of CPW-700 was 0.5 g L −1 and the initial concentration of NOR was 20 mg L −1 , the removal efficiency of NOR first increased and then decreased as the pH increased, and reached the maximum value of 38.77 mg g −1 at pH 6. When the pH is low, the protonation of the biochar surface makes it have a large amount of positive charge, and the electrostatic repulsion between H + and NOR + is inevitable. The plenty of surface adsorption sites are occupied by H + , so the adsorption capability of NOR is low in the strong acid condition. As the pH value increases, the adsorption of H + on biochar decreases, resulting in more NOR + adsorbs on biochar. Under alkaline conditions, the NOR − in the solution increases with the pH. In such solution, the hydrogen bonding and cation exchange are inhibited, and there is also the repulsion of OH-and NOR-, which leads to the decline of NOR adsorption (Gege et al. 2018).
It was clear that, the maximum adsorption capacity of CPW-700 for NOR was 38.77 mg g −1 when the pH value was 6. Compared with other biochars, the adsorption performance of CPW-700 was competitive, as shown in Table 2.  The effect of biochar dosage on the percent removal of the studied antibiotic is depicted in Fig. 4b in the range of 0.25-4.0 g L −1 . As the dosage increased from 0.25 to 0.5 g L −1 , the removal efficiency of antibiotics increased. After the dosage of CPW-700 or PWC-700 increased more than 1.0 g L −1 , the change of adsorption efficiency for NOR was not obvious. The changing trend of removal efficiency could be due to the increased dosage of biochar providing more active adsorption sites and functional groups, resulting in stronger interaction between biochar and antibiotics. However, when too much adsorbent was used, the adsorption and desorption of antibiotics reached a balance, and the removal efficiency almost did not increase (Amin et al. 2017). Compared with CPW-700, despite the continuous increase in the dosage, the NOR adsorption of PWC-700 did not exceed 70%. These results demonstrated that the FeO 3 had a significantly positive effect on NOR adsorption. 0.5 g L −1 of dosage was selected for subsequent experiments.

Adsorption kinetic studies
The adsorption kinetic equation can reflect the reaction pathways and adsorption mechanism (Plazinski et al. 2009). As shown in Fig. 5a, the NOR adsorption capacity increased with the adsorption time on both samples. At the initial stage (the first 30 min), the adsorption rate of both samples was quite high. After that, the adsorption rate rapidly decreased and quickly reached adsorption equilibrium. In the beginning, there are plenty of adsorption sites available on these samples for NOR, so the adsorption rate is quite high. As the time increases, the effective adsorption sites on the sample surface gradually decrease. NOR needs to diffuse into the micropores to reach the adsorption sites, causing adsorption rate to slow down, and finally, the adsorption reaches a dynamic equilibrium (Lyu et al. 2017;Mohan et al. 2007).
In this study, the commonly used pseudo-first-order, pseudo-second-order kinetics, and intra-particle diffusion model equations were used to fit the adsorption data (Garzo et al. 2021).
Pseudo-first order dynamics model, as Eq. (3): Pseudo-second order dynamics model, as Eq. (4): Intra-particle diffusion model, as Eq. (5): in which t is the adsorption time (min), q e is the equilibrium adsorption capacity of biochar when the adsorption reaches equilibrium (mg/g), q t is the adsorption capacity of biochar corresponding to time (mg/g); k 1 is the rate constant of the quasi-first order kinetic equation (min −1 ), k 2 is the rate constant of the quasi-second order kinetic equation (g/mg/ min), k 3 is the diffusion constant (g/mg/min 1/2 ), and c is the thickness constant of the interface layer. Table S2 (Additional file 1) shows that the correlation coefficients fitted by the pseudo-second-order kinetic equations (R 2 = 0.9991, 0.9977) are all larger than those fitted by the pseudo-first-order kinetic equations (0.9936, 0.9041) and the intra-particle diffusion model equations (0.6136, 0.5863), and the fitted adsorption value is closer to the actual, indicating that the pseudo-second-order kinetics fits the NOR adsorption data more closely. The adsorption rate of biochar is mainly determined by chemical adsorption mechanism. NOR is mainly adsorbed on the surface of biochar via interacting with the π-π electron acceptor/ donor, electrostatic attraction, and hydrogen bonding. It is a result of the combined effect of the above various influencing factors (Tan et al. 2008). Figure 5b shows that the fitting straight line of the intra-particle diffusion model does not pass the coordinate origin, and the c value is not zero, showing that the intra-particle diffusion is not the only ratedetermining step in the adsorption process. A linear fitting was performed on the latter stage of the model, and it was found that the diffusion constants ki were not much different, indicating that the boundary layer effect and mass transfer resistance of the diffusion increased to the peak value, and then gradually reached a balance (Zhou et al. 2017).

Equilibrium isotherm studies
The adsorption isotherm can reflect the affinity of the adsorbent to the adsorbate and describe the interaction between the adsorbate and the adsorbent (Xia et al. 2019). The adsorption isotherms of biochar in NOR solutions with different initial concentrations at 25 °C are shown in Fig. 6. Langmuir and Freundlich isotherm adsorption models were used to simulate the NOR adsorption performance of the two samples (Leandro-Silva et al. 2020). Their equations can be expressed as follows: Freundlich isotherm model, as Eq. (6): Langmuir isotherm model, as Eq. (7): In Freundlich isotherm model, Q e represents the equilibrium adsorption capacity (mg g −1 ), C e represents the concentration of adsorbate when the adsorption reaches equilibrium (mg L −1 ), K f represents the Freundlich affinity coefficient (mg 1−n g −1 L −n ), n denotes an empirical parameter of adsorption strength.
In Langmuir isotherm model, Q m (mg g −1 ) is the maximum adsorption capacity, and K L (mg L −1 ) is the Langmuir adsorption constant.
The adsorption rates of both CPW-700 and PWC-700 increased with the initial concentration of NOR, and the NOR adsorption amount rate on CPW-700 was higher than that on PWC-700. From the shape of the isotherm, the two samples are in line with the Freundlich and Langmuir isotherms, but the Langmuir model fits better, indicating that the adsorption of NOR on biochar is a single-layer adsorption process. The maximum adsorption capacity (Q m ) calculated via the model for CPW-700 (379.85 mg g −1 ) was higher than PWC-700 (275.69 mg g −1 ), shown in Table S3 (Additional file 1). The value of n in the Freundlich model is related to the properties of the adsorption material and the adsorption system. Generally, the value of n indicates the difficulty rate of adsorption process, in other words, the smaller 1/n value represents the easy adsorption process. The 1/n value of CPW-700 is smaller than that of PWC-700, indicating that NOR adsorption is easier to occur on CPW-700 (Ding et al. 2013).

Adsorption thermodynamics
Thermodynamic study was used to describe the driving force and to estimate whether the process of the adsorption of NOR onto biochars was spontaneous or not. The equations are as follows: where ΔG is the Gibbs free energy change (kJ mol −1 ), ΔH is the enthalpy change (kJ mol −1 ), ΔS is the entropy change (J mol −1 k −1 ), T is the absolute temperature (K) and R is the universal gas constant (8.314 J mol −1 K −1 ); K d is the equilibrium distribution coefficient, which is the ratio of the amount adsorbed on the adsorbents (mg g −1 ) to the equilibrium concentration in solution (mg L −1 ).
The thermodynamic parameters of CPW-700 and PWC-700 for the NOR adsorption process are shown in Table S4 (Additional file 1), according to Van't Hoff equation (see supporting information). Through calculating the data at three different temperatures, it was found that the obtained ΔG was less than 0, indicating that the adsorption process was spontaneous, and the absolute value of ΔG increased with the temperature, showing that the temperature promoted the spontaneous NOR adsorption. The obtained ΔH of both samples was greater than zero, indicating that the NOR adsorption process was endothermic, so heating up

Adsorption reusability
As well known, the reusability of adsorbent is very important in practical application. After each run, the biochar was collected, washed by distilled water and ethanol for three times, and dried in vacuum oven at 60 °C for 6 h. Subsequently, the clean biochar after each adsorption experiment was remixed with NOR solution. As shown in Fig. 7, the removal efficiencies of CPW-700 and PWC-700 in the second time were 87.06 and 51.44%, 537: dropping to 83.62% and 49.32% in the third time, and 80.21% and 45.07% in the fourth time, respectively. Although the removal efficiency of both samples decreased, the removal efficiency of CPW-700 was still considerable, higher than 80%, much higher than that of PWC-700. Thus, CPW-700 had a better reusability than PWC-700.

NOR adsorption mechanism
In order to explore the adsorption mechanism of biochars for NOR, the FTIR and N 2 isothermal adsorption desorption curves before and after adsorption were investigated. As shown in Fig. S3 (Additional file 1), it was clear that some changes took place in the characteristic peaks of FTIR before and after adsorption. The stretching vibration peaks at 1562 cm −1 , 640 cm −1 , 565 cm −1 and 471 cm −1 shifted slightly after adsorption. It can be inferred that the corresponding functional groups were involved in the adsorption reaction of NOR. The weakening of -OH characteristic peak intensity proves the importance of hydrogen bonding in the reaction process. Solis et al. (Solis et al. 2021) confirmed that hydrogen bonds can promote the adsorption of polar organic compounds on biochar.
-COOH, -OH and other oxygen-containing functional groups can interact with O, N, and F atoms in NOR molecule in the form of hydrogen bonds. The shift of C = C and the weakening of peak intensity of O-C = O show that π-π interaction plays a role in the adsorption process of composites and NOR. Due to the presence of strong electron-attracting element F in the NOR, the electron density on the benzene ring was reduced, as a π-electron acceptor, while the aromatic functional group on the surface of the biochar can be used as a π-electron donor to form π-π electron donor-acceptor interaction of NOR, which could be another adsorption effect between NOR and bichar (Wang et al. 2010). At the same time, the peak position of Fe-O stretching vibration peak also shifted to a certain extent, indicating that Fe-O may provide more adsorption sites for the adsorption of NOR. From Fig. S4 (Additional file 1), it can be seen that the volume of CPW-700 and PWC-700 after adsorbing NOR decreases significantly. In addition, the pore size distribution also changed significantly (as shown in the illustration of Fig. S4 (Additional file 1)). Only mesopores of 3.7 nm existed in PWC-700, while 2-8 nm mesopores can hardly be deteced in CPW-700. Moreover, BET specific surface area, average pore diameter and pore volume of CPW-700 and PWC-700 after adsorbing NOR decreased (Additional file 1: Table S5), implying that pore filling may be one of the forces for biochar adsorbing NOR. Chun et al. (Chun et al. 2004) suggested that the specific surface area and pore structure are important factors that affect the adsorption of organic pollutants by biochar, and the organic pollutants are adsorbed on the biochar through pore-filling mechanism. It is clear to see from the Fig. 4a that the adsorption efficiency of NOR on CPW-700 is the highest when the pH value is close to neutral, indicating that hydrophobicity is the main force between CPW-700 and NOR. Therefore, as shown in Fig. 8, we believe that the adsorption process of biochar and NOR molecules may be a synergistic effect of hydrogen bonding, π-π electron donor and acceptor interaction, and pore filling.

Conclusions
Fe 2 O 3 nanoparticles modified magnetic porous biochar derived from PW was prepared in a molten-salt medium. The obtained CPW-700 exhibited a superior NOR adsorption capability due to high surface area, pore volume, and the presence of functional groups on the surface. The obtained bichar followed the pseudo-second-order kinetic model, and the main forms of adsorption could be liquid film diffusion and surface adsorption. The adsorption isotherm data of CPW-700 is more in line with the Langmuir model. The adsorption process of NOR on CPW-700 is determined by multiple factors, including pore filling, hydrogen bonding, π-π electron donor-acceptor interaction, electrostatic repulsion, and hydrophobic interaction. The results show that magnetic porous biochar developed in this study is a promising feasible adsorbent for removing antibiotics from water.