Adsorption characteristics and mechanism of norfloxacin in water by γ-Fe₂O₃@BC

In recent years, the types and concentrations of antibiotics detected in the aquatic environment have been increasing day by day, which has caused potential ecological risks and endangered human health (Liu et al. 2011). Norfloxacin (NOR) is a third-generation fluoroquinolone antibiotic, and it is the most commonly used fluoroquinolone antibiotic (FQs). Because it can effectively inhibit the synthesis and replication of DNA, it has been widely used to treat intestinal, urinary, respiratory and other diseases of sensitive bacterial infections. By 2010, global sales of fluoroquinolones had exceeded $18 billion, accounting for 17% of the global antibiotic market, and annual consumption in China alone had reached 1,820 t. However, the metabolism of norfloxacin in humans and animals is limited; usually only 20–80% can complete metabolism and degradation. As a result, a considerable part of the norfloxacin ingested by humans and animals is discharged without degradation, and most of these unmetabolized drugs enter urban sewage treatment plants along the drainage pipes (Chahm et al. 2019). However, traditional sewage treatment plants lack the corresponding sewage treatment process for antibiotics, and eventually these undegraded antibiotics are returned to natural water. For example, the norfloxacin concentration detected in the discharge water of a sewage treatment plant in Hong Kong reached 2,500 ng/L (Minh et al. 2009). The effluent concentration of norfloxacin in a sewage treatment plant in Croatia also reached 1,035 ng/L (Senta et al. 2008). The residual sludge from sewage plants is one of the main enrichment sources of norfloxacin. For example, Chen et al. (2013) investigated 45 sewage plants distributed in 20 cities in China and found that the detection rate of norfloxacin was above 90%, and the highest detected concentration even reached 5,280 µg/L. Norfloxacin discharged into natural water bodies has huge potential risks, such as harming human health through drinking water, inducing bacterial mutation and producing drug-resistant bacteria (Yan et al. 2017).

In previous studies, municipal wastewater treatment systems have paid little attention to the removal of trace organic pollutants such as norfloxacin. Conventional methods for norfloxacin removal include activated sludge (Zorita et al. 2009), photocatalytic degradation (Sturini et al. 2012), chemical oxidation (Akmehmet Balcıoğlu & Ötker 2003) and adsorption (Huang et al. 2015). The adsorption method is widely used in sewage treatment due to its advantages of easy operation, low cost, high removal efficiency and less harmful by-products (Yan & Niu 2018). Biochar is one of the commonly used adsorbent materials. Biochar is a carbon-rich solid substance formed by pyrolysis of biomass at high temperature (<700 °C) under anoxic or anaerobic conditions, with abundant pores, large specific surface area and rich functional groups on the surface. However, it is difficult to separate biochar with small particle size after absorbing pollutants in water. The commonly used methods for recovering biochar include centrifugation or filtration, which is not only time-consuming and complicated to operate, but also has a large loss (Wang et al. 2015). In addition, the cyclic adsorption performance of biochar on pollutants is still unknown, and the selection of carbon sources also affects the adsorption performance of biochar. Therefore, further research is needed.

To solve the problem that biochar is difficult to recover after absorbing pollutants, some scholars proposed the preparation of biochar with magnetic properties by chemical methods. This magnetic modification not only increased the adsorption capacity of the biochar, but also gave the biochar excellent magnetic separation ability. Biochar can be rapidly recovered from water by simply applying an external magnetic field (Jiang et al. 2019). In the related research on magnetic modified biochar, the magnetic material loaded is mostly Fe₃O₄, but Fe₂O₃ as a magnetic material is rarely studied. Therefore, this experiment chose to load γ-Fe₂O₃ as a magnetic material onto the surface of biochar.

Grapefruit peel is a kind of porous biomass rich in cellulose and pectin (Methacanon et al. 2014). The annual output of pomelo peel is huge, but only a small amount of pomelo peel is reprocessed into medicine or used as chemical raw material. Most pomelo peel is treated as agricultural waste, which not only causes great waste of resources, but also causes pollution to the environment (Zhang et al. 2018). Therefore, in this experiment, pomelo peel-based biochar (BC) was prepared from discarded pomelo peel during pyrolysis for 2 h at 400 °C, and γ-Fe₂O₃ was loaded onto the surface of pomelo peel-based biochar by chemical precipitation without oxygen limit. The properties of BC and γ-Fe₂O₃@BC are characterized. The effects of biochar dosage, time, concentration, temperature, pH and ionic strength on the adsorption of norfloxacin by BC and γ-Fe₂O₃@BC are investigated, and the adsorption mechanism of BC and γ-Fe₂O₃@BC on norfloxacin is discussed to enrich the iron carrier types of magnetic biochar and related studies on the treatment of antibiotic wastewater.

Norfloxacin (properties shown in Table 1), FeSO₄·7H₂O, FeCl₃·6H₂O, NaCl and NaOH reagents were purchased from Aladdin Pharmaceutical Company (Shanghai, China), and they were analytical reagents (AR). The experimental water was ultrapure water (UP). The norfloxacin sample was weighed at 0.10 g in a beaker, dissolved with an appropriate amount of hydrochloric acid and transferred to a 1,000 mL volumetric bottle. The sample was configured into a 100 mg/L NOR storage solution diluted to the appropriate concentration when used. To prevent biodegradation, the norfloxacin storage solution used in the experiment was prepared when it was needed.

Experimental instruments: UV-1810 UV-visible spectrophotometer (Youke, China); PHS-3C digital pH meter (Shanghai Thunder Magnetic, China); SHA-B digital display constant temperature water bath oscillator (Changzhou Guohua Electric Appliance Co. Ltd, China).

A certain amount of discarded grapefruit peel was rinsed repeatedly with ultra-pure water, and after removing sundries on the surface, it was put into a drying oven at 70 °C to dry to constant weight. The dried pomelo peel was wrapped with aluminum foil, compressed, and put into a muffle furnace. The temperature was raised to 400 °C at a heating rate of 20 °C/min, and the pyrolysis temperature was maintained for 2 h. The pomelo peel biochar was taken out after pyrolysis, ground and passed through a 60-mesh sieve, sealed in a plastic bottle and labelled ‘BC’ for later use.

Five grams of the prepared pomelo peel biochar was taken, soaked in 100 mL of FeSO₄ of 0.125 mol/L and 0.25 mol/L of FeCl₃ solution, and mixed evenly on the magnetic mixer. Then 100 mL of 1 mol/L of NaOH solution was gradually added. After the reaction was complete, stirring continued for 1 h, then separation by magnetic field and washing with ultra-pure water until the pH of the solution system was close to neutral. Then it was put in a drying oven and dried to constant weight at 70 °C, ground and sealed in a plastic bottle with the label ‘γ- Fe₂O₃@BC’ for later use.

The surface structure and morphology of the biochar were observed by scanning electron microscope (SEM, MLA650F, USA). The contents of C, O, S, N, K and Fe were determined by XPS (ESCALAB250). The crystal structure was analyzed by X-ray diffraction (XRD, scan from 10° to 80°). The surface area of the biochar was determined by the Brunauer–Emmett–Teller (BET) adsorption method (ASAP2460, USA). The surface functional groups were determined by Fourier transform infrared spectroscopy (FT-IR, Nicoletis5, USA), with a wavelength scanning range of 400–4000 cm⁻¹. The magnetization of the samples was measured at room temperature using a vibrating sample magnetometer (VSM, Lake Shore 7410, USA). The pH of the biochar was determined as follows: the biochar was mixed with ultrapure water at a mass ratio of 1:10, stirred magnetically for 0.5 h and then placed for 1 h. Then the pH of the biochar was measured by PHS-3C pH meter (Thunder Magnet, Shanghai, China). The zero point charge of the biochar was measured by titration in accordance with Bastami & Entezari (2012).

A certain amount of BC, γ-Fe₂O₃@BC, and 25 mL NOR solution were weighed in a 50 mL centrifuge tube, placed in a constant temperature oscillator in a water bath and shaken in the dark at 150 rpm (to avoid possible norfloxacin photodegradation). After a certain period of oscillation, the absorbance of the solution was measured at the wavelength of 278 nm by UV-visible spectrophotometer after filtrating by 0.45 um membrane, and the concentration was converted into the established NOR standard curve. In the single-factor experiment, the effects of biochar dosage (0.1, 0.2, 0.3, 0.4, 0.5, and 0.6 g/L), adsorption time (10, 20, 30, 60, 90, 120, 150, and 180 min), initial NOR concentration (1, 3, 6, 10, 15, 20, 35, 45, 70, and 100 mg/L), reaction temperature (25, 35, and 45 °C), pH (3–10), and Na⁺ concentration (0, 0.2, 0.4, 0.6, 0.8, and 0.10 mol/L) on the adsorption of norfloxacin by BC and γ-Fe₂O₃@BC were investigated. In order to test the reusability of γ-Fe₂O₃@BC, adsorption and desorption experiments were designed as follows: 0.01 g of γ-Fe₂O₃@BC was weighed in a 50 mL centrifuge tube, and 25 mL, 100 mg/L norfloxacin solution was added and shaken for 2 h. Then γ-Fe₂O₃@BC was separated with a magnet and the concentration of supernatant was determined. The separated γ-Fe₂O₃@BC was soaked in an appropriate amount of ethanol, washed and dried overnight, before continuing to the next adsorption test.

Figure 1(a) and 1(b) show the scanning electron microscopy of BC and γ-Fe₂O₃@BC. The layered structure on the surface of the BC expands into fractures and there are many pores on the surface, which are conducive to providing more adsorption sites; γ-Fe₂O₃@BC not only retains the pore structure of 400 °C BC, but also loads nano-γ-Fe₂O₃ particles on the surface. However, the loaded nano-γ-Fe₂O₃ particles block some of the pores, which may cause the specific surface area of the biochar to decrease. The partial physicochemical properties of BC and γ-Fe₂O₃@BC are shown in Table 2. Compared with BC, the modified γ-Fe₂O₃@BC reduced the content of the C element, increased the content of the O and Fe elements, and significantly decreased the surface pH of the biochar. The total pore volume increased significantly, from 0.003 cm³/g to 0.110 cm³/g, and the pore diameter expanded. Since the specific surface area of the Fe₂O₃ is up to 106 m²/g, some iron oxide particles embedded in the surface of the BC do not reduce the specific surface area of the γ-Fe₂O₃@BC, but significantly increase the specific surface area of the biochar from 1.706 m²/g to 20.732 m²/g.

Figure 1(c) is the XRD characterization of the BC and γ-Fe₂O₃@BC. BC showed an obvious wide peak at 2θ = 24°, corresponding to the characteristic peak of the locally ordered carbon layer structure (002) and surface (101) (Li et al. 2016), indicating that there was an obvious aromatic carbon structure on the surface. The wide diffraction peak of γ-Fe₂O₃@BC at 2θ = 22.4° corresponds to graphite structure (Tian et al. 2011), which is caused by the retention of natural cellulose in the biomass. The diffraction peak of 2θ = 26.4° corresponds to the characteristics of the graphite carbon layer (Hu et al. 2017). The seven diffraction peaks at 30.3°, 32.9°, 35.5°, 43.3°, 53.7°, 57.2° and 62.7° are consistent with the characterization of γ-Fe₂O₃ (Liu et al. 2014; Han et al. 2016). The results show that the nano-γ-Fe₂O₃ particles have been loaded onto the surface of the biochar.

The infrared characterization of BC and γ-Fe₂O₃@BC is shown in Figure 1(d). The infrared spectrum of BC and infrared spectrum of γ-Fe₂O₃@BC are similar. The characteristic peak of 3450 cm⁻¹ is related to the O-H stretching vibration. It can be found that the O-H stretching vibration peak of γ-Fe₂O₃@BC is significantly weakened, which means that in the process of magnetization modification, Fe is not mainly loaded on biochar in the form of hydroxide. The characteristic peaks of aromatic group -CH₂- (2,921, 1,436 cm⁻¹), C-O (1,100 cm⁻¹) and C-H (830 cm⁻¹) are weak and do not change significantly. The C = O stretching vibration characteristic of γ-Fe₂O₃@BC is enhanced compared with that of BC, indicating that more oxygen elements are introduced in the magnetization process, and the surface of the BC and γ-Fe₂O₃@BC is rich in oxygen-containing functional groups. At 568 cm⁻¹, there is a stretching vibration of Fe-O, and the characteristic peak of the stretching vibration of γ-Fe₂O₃@BC is significantly higher than that of BC, indicating that Fe has been successfully loaded onto the surface of the biochar.

Figure 1(e) is the hysteresis curve of γ-Fe₂O₃@BC. The hysteresis loop of γ-Fe₂O₃@BC is symmetric about the origin, so it has superparamagnetism (Rong et al. 2019). The saturation magnetic intensity was up to 30.60 emu/g, much higher than that of similar magnetic biochar. The lower right corner of Figure 1(e) shows the separation states of BC and γ-Fe₂O₃@BC after one minute of external magnetic field, and the results show that γ-Fe₂O₃@BC has excellent magnetic separation ability. Table 3 is the comparison between biochar and magnetic biochar prepared in previous literature and this study.

The effect of biochar dosage on norfloxacin adsorption is shown in Figure 2. With the increase of the dosage of the two adsorbents, the removal rate of norfloxacin also increased gradually, from 17.15% and 18.45% at the dosage of 0.1 g/L to 27.89% and 52.27% at the dosage of 0.6 g/L, respectively. When the concentration of norfloxacin in solution is constant, the increased biochar can provide more active adsorption sites, thus increasing its adsorption capacity for norfloxacin. In contrast, with the increase of biochar dosage, the unit adsorption capacity of norfloxacin by the two adsorbents decreased gradually. The reason is that when the concentration of norfloxacin is constant, the dosage of biochar is increased, and the unused free active sites on the surface of the biochar increase accordingly. The aggregation and overlap of the active sites lead to a reduction of effective adsorption area and a decrease of the unit adsorption amount of norfloxacin.

The effect of adsorption time on the adsorption of norfloxacin by BC and γ-Fe₂O₃@BC is shown in Figure 3. As can be seen from the figure, the adsorption of norfloxacin by BC and γ-Fe₂O₃@BC can be divided into three stages: fast adsorption stage, slow adsorption stage and adsorption equilibrium stage. In the initial 30 min, the adsorption capacity of norfloxacin by BC and γ-Fe₂O₃@BC could reach more than 90% of the total adsorption capacity, and then the adsorption rate slowed down and reached equilibrium around 120 min. This phenomenon of ‘first fast, then slow’ occurs because there are a large number of adsorption sites and functional groups on the surface of the biochar at the initial stage of adsorption. Norfloxacin quickly occupies functional groups with high adsorption affinity on the surface of the biochar, so the adsorption rate is fast. As the adsorption process continues, the limited adsorption points on the surface of the biochar gradually saturate, resulting in a decrease in the adsorption rate.

The adsorption process of biochar to solute is very complicated, including physical adsorption and chemical adsorption. In order to further explore the adsorption process of norfloxacin by biochar, the pseudo-first-order and pseudo-second-order kinetic equations, Elovich equation and intraparticle diffusion equation were used to fit the adsorption dynamic data. The equations are as follows.

The fitting effect of the pseudo-first-order dynamic equation and intraparticle diffusion equation is poor, so the fitting curve is not drawn in the figure. The fitting curve is shown in Figure 3, and the relevant fitting parameters are shown in Table 4. The pseudo-second-order kinetic equation has the best fitting effect on the adsorption of norfloxacin by BC and γ-Fe₂O₃@BC, and the correlation coefficient is above 0.95. The results indicate that chemical adsorption may be a rate-limiting step in the adsorption process, and that there may be ion exchange between norfloxacin and BC and γ-Fe₂O₃@BC (Zhou et al. 2019). The fitting effect of the Elovich equation is also good, indicating that the adsorption of norfloxacin on biochar is a process from fast adsorption to slow adsorption. NOR adsorbs rapidly on the surface of the biochar, while the diffusion inside the biochar is slow. The theoretical value I of the intraparticle diffusion fitting is not zero, indicating that internal diffusion is not the only rate control step, but the external diffusion process (such as external liquid–film diffusion and surface adsorption) also controls the adsorption rate.

At different temperatures, the adsorption isotherms of BC and γ-Fe₂O₃@BC were drawn by plotting the equilibrium concentration and adsorption capacity respectively, as shown in Figure 4. In the initial stage, with the increase of equilibrium concentration, the adsorption capacity of norfloxacin by BC and γ-Fe₂O₃@BC also increased. However, when the equilibrium concentration reached a certain value, the adsorption increment of norfloxacin by biochar gradually decreased until it stabilized. At the same time, the adsorption capacity of biochar for norfloxacin increased significantly with the increase of temperature. The equations of Langmuir, Freundlich, Sips, Temkim and Dubinin–Radushkevich (D-R) are used to fit the adsorption isothermal curve. The equations are as follows.

The R_L values of BC and γ-Fe₂O₃@BC ranged from 0.1488 to 0.9705, and decreased with the increase of norfloxacin concentration, indicating that the adsorption of norfloxacin by biochar was favorable.

The effect of pH on BC and γ-Fe₂O₃@BC adsorption of NOR is shown in Figure 5. With the increase of solution pH, the adsorption capacity of BC for NOR decreased gradually, but it did not change much under the condition of close to neutral pH. This is because norfloxacin has two pKa values for the acid dissociation constants, 6.22 and 8.51, respectively. When the pH of the solution is less than 6.22, norfloxacin mainly exists in the NOR⁺ state. When the pH of the solution is between 6.22 and 8.51, norfloxacin mainly exists as ampholytic NOR^+,−. When the pH of the solution is greater than 8.51, norfloxacin's main existing form is NOR⁻ (Yang et al. 2012). The zero point charge of BC, pHpzc = 10.98, is larger than the experimental pH range, so the surface of the BC is positively charged. The adsorption force between BC and NOR in an acidic environment is mainly cation exchange and hydrogen bonding. In the range of close to neutral pH, the main form of NOR is the molecular state, and the main force for BC to adsorb NOR is hydrophobicity. In a strongly alkaline environment, although NOR⁻ can be combined with the surface of BC by electrostatic attraction, too much OH⁻ will also form competitive adsorption with NOR⁻, resulting in a decrease in NOR adsorption. When pH increases from 3 to 5, the adsorption capacity of γ-Fe₂O₃@BC to NOR increases significantly. When pH is 5–9, the adsorption capacity of γ-Fe₂O₃@BC to NOR remains roughly the same. When the pH increases to 10, the adsorption capacity of γ-Fe₂O₃@BC to NOR decreases significantly. The zero point charge of γ-Fe₂O₃@BC pHpzc = 5.86, so when the pH of the solution rises from 3 to 5, the surface of the γ-Fe₂O₃@BC carries a positive charge, while the rise of pH will lead to the decrease of NOR⁺ concentration and the reduction of electrostatic repulsion with the γ-Fe₂O₃@BC, thus increasing the adsorption capacity for NOR. When the pH is 5–8, there are more NOR molecular states. In this case, the γ-Fe₂O₃@BC mainly relies on hydrophobicity to adsorb NOR in the molecular state. In a strongly alkaline environment, the negative charge on the surface of the γ-Fe₂O₃@BC will form an electrostatic repulsion with NOR⁻, while too much OH⁻ will also form competitive adsorption with NOR⁻, resulting in a decrease in the adsorption capacity for NOR. However, it can be found that γ-Fe₂O₃@BC has a wider pH adaptability than BC, with great application potential.

The effect of Na⁺ on BC and γ-Fe₂O₃@BC adsorption of NOR is shown in Figure 6(a). The results showed that the presence of Na⁺ inhibited the adsorption of NOR by BC, and the inhibition increased with the increase of Na⁺ concentration (0.02–0.10 mol/L). This is because Na⁺ can participate in cation exchange and compete with NOR to adsorb active sites on the surface of BC. At the same time, the added Na⁺ will partially replace the protons on the surface of the BC, so as to weaken the hydrogen bond between BC and NOR. Different from BC, the addition of Na⁺ will increase the adsorption capacity of γ-Fe₂O₃@BC for NOR. The reason is that the added NaCl will reduce the ionic form of NOR in water and precipitate into the NOR molecular state, which is called ‘a salt-out effect’. Then the NOR of the molecular state is combined with the γ-Fe₂O₃@BC through hydrophobic action to enhance its adsorption capacity. Through the addition-of-Na⁺ experiment, it can be found that the prepared γ-Fe₂O₃@BC has stronger ionic adaptability than the BC, which is more suitable for the treatment of complex sewage under actual conditions. Figure 6(b) is the adsorption–desorption diagram of γ-Fe₂O₃@BC. The results showed that after four cycles, the adsorption capacity of γ-Fe₂O₃@BC for NOR could still reach 61.43% of the initial adsorption capacity, with good recycling performance. The decrease of adsorption capacity may be due to the repeated desorption and washing process leading to the decrease of active sites on the surface of the γ-Fe₂O₃@BC and the loss of γ-Fe₂O₃. Therefore, in future experiments, how to load γ-Fe₂O₃, which is stable and not easy to lose, will be the focus of research.

The adsorption mechanism of BC and γ-Fe₂O₃@BC for norfloxacin is shown in Figure 7. In the pH range of 3–10, the surface of the BC is positively charged, so there are electrostatic repulsion and electrostatic attraction between BC and NOR⁺, NOR⁻ under acid or alkaline conditions. Infrared spectra showed that there were oxygen-containing functional groups such as -COOH and -OH on the surface of the BC, which could combine with F, O and N atoms in the NOR molecule through hydrogen bonding. Under acidic conditions, the adsorption capacity of BC for NOR was large, indicating that there was cation exchange between BC and NOR. In addition, the more acidic the solution, the higher the adsorption capacity of BC for NOR, indicating that the main forces for BC to adsorb NOR are cation exchange and hydrogen bonding, which can be confirmed in the addition-of-Na⁺ experiment. Electrostatic attraction, repulsion, pore-filling effect and hydrophobicity are not the main forces for BC to absorb NOR.

The γ-Fe₂O₃@BC zero point charge, pHpzc, is 5.86, so the pH of the solution significantly affects the surface electrification. Under acidic or alkaline conditions, the surface of γ-Fe₂O₃@BC is positively and negatively charged, and electrostatic repulsion exists with NOR⁺ and NOR⁻ respectively. The adsorption capacity of NOR is the largest under the condition of close to neutral pH, and the addition-of-Na⁺ experiment shows that the ‘salting out effect’ can promote the adsorption of NOR by γ-Fe₂O₃@BC, indicating that γ-Fe₂O₃@BC has more affinity for NOR molecules, that is, hydrophobicity is the main force for the adsorption of NOR by γ-Fe₂O₃@BC. The reason for the stronger hydrophobicity of γ-Fe₂O₃@BC may be the introduction of γ-Fe₂O₃ particles. In addition, the F atom adjacent to the benzene ring on the NOR structure has a strong property of adsorbing electrons, which is shown as π-electron acceptors. The -OH, C = C, C-H, C-O and other functional groups on the surface of γ-Fe₂O₃@BC can be used as electron donors to form π-π electron donor receptors and enhance the adsorption capacity of γ-Fe₂O₃@BC for NOR.

Compared with BC, γ-Fe₂O₃@BC has a larger specific surface area, larger pore volume, stronger hydrophobicity, and stronger adsorption capacity for NOR. The adsorption process of BC and γ-Fe₂O₃@BC for NOR conforms to the pseudo-second-order kinetic model. The adsorption isotherms of BC and γ-Fe₂O₃@BC for NOR mainly conform to the Sips model, indicating that the adsorption of NOR was a heterogeneous system. The pH and ionic strength of the solution have great influence on the adsorption capacity of BC and γ-Fe₂O₃@BC for NOR. The pH and ionic strength of the solution mainly affect the adsorption capacity by influencing the ion and molecular morphology of NOR in water. By studying the mechanism of adsorption of NOR by BC and γ-Fe₂O₃@BC, it is found that cation exchange and hydrogen bonding are the main forces of adsorption of NOR by BC. However, γ-Fe₂O₃ particles enhance the hydrophobicity of biochar, making hydrophobicity become the main force for γ-Fe₂O₃@BC to adsorb NOR. Compared with the literature (Table 3), γ-Fe₂O₃@BC has a larger adsorption capacity for NOR and strong magnetism (M_s = 30.60 emu/g), which makes γ-Fe₂O₃@BC have great application value.

This work was supported by the National Natural Science Foundation of China (Grant No. 51808001, 51409001) and Anhui Provincial Natural Science Foundation (1808085QE146, 1708085QB45).

The authors have declared no conflict of interest.