Aqueous norfloxacin removal by novel biochar adsorbent prepared through ethanol-combined ball milling

ABSTRACT

Ethanol combined-ball milling was used to modify corn-stover biochar. The adsorption performance and mechanisms of aqueous norfloxacin (NOR) onto the resulting biochar (C₂H₆O-BC) were evaluated. Ball-milling modulated its surface oxygen-containing functional groups and enhanced its specific surface area. The adsorption capacity of C₂H₆O-BC was 163 mg·g⁻¹, and about 72% of the equilibrium adsorption capacity could be reached within 60 min. The electrostatic interaction was not the dominant mechanism when initial pH was lower than 8.00. The antagonistic effect of Cu(II) on NOR adsorption was observed. The fixed-bed column adsorption showed its potential to remove NOR and Cu(II) spontaneously. Sequential two-stage batch reaction experiment showed that it was a fast and efficient technology to remove NOR using C₂H₆O-BC. The adsorption mechanisms were suggested as π-π interactions, H-bonding and pore-filling. In summary, ethanol-combined ball milling in this study provided an efficient adsorbent for NOR.

KEYWORDS:

• Corn-stover-based biochar
• mechanical ball milling
• batch adsorption
• fixed-bed column adsorption
• sequential two-stage batch reaction
• adsorption mechanism

1. Introduction

Although antibiotics have tremendous social and economic benefits, their environmental impact, especially the potential long-term adverse effects on human health and ecosystems, receives more and more concern. Currently, antibiotics have been included in the List of Key Emergent Pollutants to be Controlled (2023 Edition) by the Ministry of Ecology and Environment of China. Norfloxacin, a typical third-generation quinolone antibiotic, is detectable in ng·L⁻¹ to mg·L⁻¹ in wastewaters [1]. Diverse potential sources have led to its release into the environment [2–4], and the drug-resistant bacteria induced by antibiotics in environment has been reported [5,6]. Therefore, there is an emergent need to find green controlling methods of low cost and high efficiency to remove aqueous norfloxacin. Among different methods (e.g. adsorption [7], photocatalytic oxidation [8], advanced oxidation [9], and so on), adsorption is the most commonly used and is a low-cost technology in organic wastewater treatment [10]. Biochar, a porous and carbon-rich composite produced by slow pyrolysis of biomass feedstock [11], has some unique physicochemical properties such as huge specific surface area and numbers of active functional groups [12], which results in high potential for application in wastewater treatment. However, the previous literatures show that the pristine biochar had limited performance in removing antibiotics from wastewater and simple/effective modified methods are in great need to be developed to improve the adsorption capacities [13].

Physical and/or chemical modification of biochars have been investigated to further improve the adsorption performance of contaminants in a wide range of environmental applications [14]. Ball milling is one of relatively low-cost and high-effective physical modification methods for mechanically converting solid powders into ultrafine particles (nanoscale) to improve the surface properties of adsorbents [15]. The previous studies show that ball milling can improve the physicochemical properties of biochars and enhance the adsorption performance of biochars for organic pollutants [14,16,17]. For example, Naghdi et al. reported that the larger specific surface of ball-milling modified biochar improved its affinity for adsorption of organic materials and the removal ratio of carbamazepine from water was up to 95%, compared to ≤ 14% for the pristine unmodified biochar [18]. Zhang et al. reported that the specific surface area of ball-milled crayfish shells-biochar derived at 800°C was 289.7 m²·g⁻¹, and the adsorption capacity of tetracycline also increased up to 1.6 times of the pristine biochar which had the specific surface area of 127.9 m²·g⁻¹ [19].

Recently, combination of ball milling and chemicals in modification of biochars has also been investigated [20], greatly improving adsorption performance of biochar. For example, modified biochar by ball milling using H₂O₂ as a grinding aid showed greatly enhanced adsorption performance of VOCs [15]; ball-milled biochar used NaOH as a grinding aid had an adsorption capacity of more than 200 mg·g⁻¹ for methyl violet [21]. As a relatively environmentally friendly chemical, ethanol has been used to modify biochar to increase its adsorption capacity for sodium (washing with 0.1 M ethanol) and promotes the surface grafting (ethanol-assisted milling) through increasing the surface oxygen containing groups [22,23]. Similarly, methanol reaction with biochar surface enhanced its adsorption performance for tetracycline [24]. Therefore, wet ball milling of biochar using ethanol as grinding aids may be a novel and low-cost modified method to improve the adsorption capacities greatly of antibiotics by the as-prepared biochars. Moreover, batch adsorption experiment and fixed-bed column adsorption experiment are usually employed in the evaluation of biochar adsorption ability. However, adsorption using sequential two-stage or even poly-stage batch reactor which may further show the application potential of biochar were limitedly reported. Therefore, the combination of batch adsorption, fixed-bed column adsorption and sequential poly-stage batch reaction can more effectively and systematically evaluating the performance of as-prepared biochar adsorbents in practical wastewater treatment [25,26].

In this study, biochar derived from corn stover at 500°C was modified by using wet ball milling with ethanol (C₂H₆O), a relatively environmentally friendly grinding aid. Batch adsorption, fixed-bed column adsorption and sequential two-stage batch adsorption of NOR, a typical third-generation quinolone antibiotic and easily available in environmental water, onto the resulting biochar composite were investigated. The main aims of this study were as follows: 1) developing and validating a novel biochar adsorbent for NOR adsorption modified via C₂H₆O-ball milling; 2) systematically evaluating the performance and mechanisms of the resulting biochar adsorbent in treatment of aqueous NOR via static adsorption (batch adsorption) and dynamic adsorption (fixed-bed column adsorption and sequential two-stage batch adsorption); and 3) revealing the impact of pH, biochar dosage and coexisting metal cation (Cu(II)) on the adsorption performance of aqueous NOR onto the resulting biochar adsorbent. This study is helpful in developing a novel biochar adsorbent for NOR removal in wastewater.

2. Materials and methods

2.1. Reagents and biochars

Norfloxacin (98%) was bought from Shanghai Aladdin Biotech Co., Ltd. The other reagents (analytically pure) were bought from Sinopharm Group Chemical Reagent Co. Ltd. The water used was deionized water of 18.2 MΩ·cm.

Biochar derived from corn stover at 500°C was bought from Nanjing Zhironglian Technology Co., Ltd, and sieved through a 100-mesh nylon sieve (the pristine biochar, p-BC). At room temperature, 1.0 g of p-BC was placed in a ceramic grinding jar containing grinding aid solution (ethanol, 10 mL) and ceramic balls (diameter 3 mm : 6 mm = 3:2, 100 g), and then they were put into a planetary ball mill ground for 6 h at 480 r·min⁻¹ (reversing the grinding direction every half hour) according to the previous reports [7,12,13]. The biochar was then washed with pure water for 4 h. The washed biochars were dried at 80°C and stored (denoted as C₂H₆O-BC).

A CHN elemental analyzer (Elementar Vario MICRO cube, Germany) was used to analyze the C, H and N elements of the as-prepared biochars and the difference between the total mass and (C + H + N + ash) content was used to estimate O element content. Then O/C and H/C were calculated for evaluating polarity and aromaticity. Biochar ash (ashed at 550°C, 0.10 g) was digested with aqua regia for 4 h and then diluted with water to volume to 10 ml. Main mineral elements in the resulting solution were determined by using an inductively coupled plasma-optical emission spectrometer (ICP-OES, PerkinElmer, Avio 500, U.S.A.). The precision was within 2.0%, and the relative percentage deviation of parallel duplicate samples was within 8.0%. Biochar samples’ pH were determined using a pH meter (PHS-25, Ohaus Instrument Co., Ltd, China) with biochar: pure water ratio 1:10 (w:v) after balancing 30 min. Point of zero charge (pH_PZC) of the biochars was determined according to the drift method [27] (Fig. S1). The morphology of the biochars was observed by using a scanning electron microscopy (SEM) (S-3400N, Japan Hitachi). The specific surface and pore nature were evaluated with the BET method by using a surface area analyzer (ASAP2020, Micromeritics, Ltd., U.S.A.). The functional groups of the biochars were observed with a Fourier transform infrared spectroscopy (FTIR) (Nicolet Nexus870, U.S.A.). The scanning range was set from 4000 to 500 cm⁻¹ with a scanning resolution of 4 cm⁻¹ using the KBr pellet technique. An X-ray photoelectron spectroscopy was also employed to analyze the surface elements before and after adsorption (K-Alpha, Thermo Fisher Scientific, U.S.A.). The full XPS spectrum was analyzed in the range of 0–1200 eV and the high-resolution XPS spectra of C1s, O1s and Cu2p were further performed.

2.2. Static batch experiments

p-BC or C₂H₆O-BC were mixed with NOR solution and stirred in 50 mL centrifuge tubes in a constant temperature shaker (TS-80C, Shanghai Tiancheng, China) at 240 rpm. The initial pH of NOR solution was regulated with HCl (0.1 M) and NaOH (0.1 M) solution. Sample was collected with a sterile syringe and immediately passed through a 0.22 mm nylon filter. A blank group without biochar was set up as a control. The mass of NOR adsorbed onto biochars and the removal ratios were calculated through the difference between its initial and final concentrations. Adsorption isotherms and adsorption kinetics of NOR adsorption were investigated. Ball-milled bochar using deionized water as the grinding aid was also prepared for comparison (denoted as H₂O-BC), of which the pH was 8.20 ± 0.03 and pHpzc was 8.38 (Fig. S1). However, it showed a lower adsorption capacity for NOR than C₂H₆O-BC (Fig. S2). Effect of pH, coexisting Cu(II) and solid-liquid ratio on NOR adsorption were also investigated. Cu(II) concentration (60 mg·L⁻¹) was set to ensure adsorption saturation of Cu(II) according to Cu(II) adsorption isotherm (Fig. S3). Cu(II) solution pH (4.20) was not adjusted to ensure its stability, and the mixture of NOR and Cu(II) had an initial pH value of 4.85. Temperature of 25°C and shaking time of 24 h were used when they were not the variables. Triplicates were set up for each group of experiments. Aqueous NOR was determined with a spectrophotometer (U-T6, Shanghai Yipu, China), and mineral metal elements were determined with the ICP-OES mentioned above. The same analysis methods were used in the followed experiments.

2.3. Fixed-bed column experiment

About 0.2 g of biochar was wet packed in a 1.5 cm diameter acrylic column as an intermediate layer. 7.0 g of acid-cleaned quartz sand (50–60 mesh) was put at both ends to distribute the flowing solution. The thin biochar layer was 2 mm. The loaded acrylic column was then rinsed with deionized water for 0.5 h. Samples were fed with a fixed velocity (1.0 ml min⁻¹) using a peristaltic pump at the inlet (column bottom) and collected every three minutes using a fraction collector. Duplicates were set up for each set of experiments.

The experiments of NOR filtration alone and NOR/Cu(II) co-filtration were carried out, respectively. To NOR/Cu(II) co-filtration experiment, NOR and Cu(II) solutions were placed in two separate beakers and fed by using a double channel peristaltic pump for each sample, which were connected with a tee joint pipe at the inlet end of the column to feed the sample at a fixed flow rate. NOR concentration of 15 mg L⁻¹, 100 mg L⁻¹ and 200 mg L⁻¹ and Cu(II) concentration of 100 mg L⁻¹ and 120 mg L⁻¹ were used. The effluent was collected for the determination of NOR and Cu(II).

2.4. Sequential two-stage batch reaction experiment (STBR)

Sequential poly-stage batch reactor can be used to enhance removal efficiency or save adsorbents. Therefore, sequential two-stage batch reaction experiment (STBR) was designed to simulate sequential batch adsorption process for NOR wastewater treatment as the previous report with minor modification [26]. The scheme of the STBR system was shown in Fig. S4. The inflow NOR solution with volume of V and concentration of C₀ entered reactor 1 with mass of M fresh biochar, stirred with a magnetic bar. After adsorption, the solution containing the remaining NOR (volume V and concentration C₁) was filtered to enter reactor 2 with fresh biochar (mass M) to be adsorbed and then filtered to be separated again as in the first reactor. The removal ratio in stage 1 and stage 2 were denoted as R₁ and R₂, and the total removal ratio (R) was the sum of R₁ and R₂, respectively. In the experiment, simulated NOR waste (100 mg·L⁻¹) inflow was loaded into the first batch reactor containing biochar with residence time of 2 h and then filtered into the second reactor containing biochar for another 2 h residence. Solid/liquid ratio in both reactors was 1:2000 or 1:1000 (w: v). Triplicates were set up.

3. Results and discussion

3.1. Characterization

O/C increased from 0.39 to 0.42 after C₂H₆O combined ball milling (Table 1), suggesting possible increase of oxygen-containing functional groups and polarity of C₂H₆O-BC during ball milling which was consistent with the previous reports [28]. Increased H/C from 0.60 to 0.75 indicated decreasing aromaticity in C₂H₆O-BC [29]. p-BC had much higher ash content (12.3 ± 2.7%) than C₂H₆O-BC (9.20 ± 0.70%). Contents of K, Ca, Mg, Al and Cu were 26.9 ± 1.2 mg g⁻¹, 9.60 ± 0.22, 5.79 ± 0.07, 0.822 ± 0.027, and 0.018 ± 0.003 mg g⁻¹ in p-BC and 1.85 ± 0.09, 9.27 ± 0.97, 4.15 ± 0.42, 2.57 ± 0.42 and 0.022 ± 0.003 mg g⁻¹ in C₂H₆O-BC. Obvious decrease in K content and increase in Al content may be due to the release of soluble species of salts in the pristine biochar during ball mill while insoluble carbonates (and/or phosphates, silicate) and metal oxides remain in biochar. Species of K salts are generally easily soluble while Al may be in the forms of oxide due to the high-temperature pyrolysis (500°C) during biochar preparation. This may also explain the decrease of ash content in C₂H₆O-BC compared to that in p-BC. Moreover, C₂H₆O-BC had a lower pH value (8.30 ± 0.05) than p-BC (10.20 ± 0.01), indicating some alkaline materials were removed after treatment. The zero point charge indicates the specific pH with zero net charge of the material surface [30] and is an important indicatior of the acidity and basicity of the biochar surface. The pH_PZC of p-BC and C₂H₆O-BC were 9.87 and 8.82, respectively (Fig. S1). High pH_PZC of p-BC and C₂H₆O-BC suggests the positive surface of biochar in acidic and neutral solution [31].

Table 1. Elemental contents of p-BC, C₂H₆O-BC and C₂H₆O-BC after nor adsorption.

	C	H	N	O
	(%)	(%)	(%)	(%)	O/C	H/C
p-BC	54.6	2.73	1.89	28.5	0.39	0.60
C2H6O-BC	55.0	3.42	1.89	30.5	0.42	0.75
C2H6O-BC+NOR	56.7	3.22	1.99	32.2	0.43	0.68

The SEM images indicate that p-BC was almost sheet fragments with a relatively smooth surface (Figure 1a). After ball milling, the fine particles and the large aggregated particles for C₂H₆O-BC were observed obviously and they were nearly rough spherical (Figure 1b). Those indicated that sheet fragments were crushed into fine irregular spherical particles, suggesting the changes of surface and microstructure caused by ball milling. The N₂ adsorption-desorption isotherm of C₂H₆O-BC conformed to the type IV isotherm characteristic (Figure 1c). The N₂ adsorption-desorption curves of p-BC were nearly overlapped without hysteresis. The adsorption hysteresis of C₂H₆O-BC was observed at p/p₀ of 0.05, and it exhibited capillary coalescence after the relative pressure reached 0.8, and the hysteresis return line appeared in the curves, which indicated that mesoporous structure existed on the adsorbent. The specific surface area (BET), total pore volume (TPV) and mean pore diameter (BJH) of C₂H₆O-BC were 336 m²·g⁻¹, 0.544 cm³·g⁻¹ and 14.6 nm, while them of p-BC were 3.33 m²·g⁻¹, 0.014 cm³·g⁻¹ and 26.6 nm respectively. The BET surface area of C₂H₆O-BC was almost 100 times higher than that of p-BC, indicating that C₂H₆O combined ball milling had a good improvement in the properties of the as-prepared biochar and the larger specific surface area was attributed to more mesoporous structures’ formation. These physical changes occurring in biochar are expected to lead to changes in adsorption efficiency, which may be one of the mechanisms by which ball milling modification improves adsorption capacities of biochar. Figure 1d shows that FT-IR spectrums of C₂H₆O-BC and p-BC were similar, but the intensity of peaks varied slightly and some peaks shifted slightly. Peak of about 3432 cm⁻¹ was assigned to -OH/-COOH [32]. Peak of 1589 cm⁻¹ was attributable to C=C and C=O stretching vibration in aromatic ring [32] and C-O-C was indicated at 1097 cm⁻¹ [11]. The peaks at 1378, 792 and 466 cm⁻¹ were induced by -OH of phenolic groups [33], aromatic C-H [28], and Si-O vibrations of inorganic SiO₂ [32], respectively. The slight shift of various peaks indicates that the changes of oxygen-containing functional groups during C₂H₆O-combined ball milling.

Figure 1. SEM images of p-BC (a) and C₂H₆O-BC (b), N₂ adsorption and desorption isotherms (c) and FT-IR spectra of unloaded and NOR/Cu-loaded biochars (d).

3.2. Static batch experiments

3.2.1. Isothermal adsorption

The adsorption results of NOR on biochars were shown in Figure 2a. With the increasing initial concentrations of NOR, equilibrium adsorption capacities increased obviously at lower concentrations and then reached a maximum. This was attributed to the increased driving force for adsorbate transferring from water to the adsorbent surface at high concentrations. NOR molecules easily penetrated the pores of the biochar and reach the adsorption sites [31,34]. Compared to p-BC, the adsorption capacity of C₂H₆O-BC for NOR increased significantly, indicating that C₂H₆O combined ball milling greatly improved the adsorption capacities of the as-prepared biochar for NOR. The maximum Langmuir equilibrium adsorption capacity of C₂H₆O-BC was 163 mg·g⁻¹, much higher than the reported values in the literature (Table S1 and Figure 2b) [35–39].

Figure 2. Adsorption isotherms, kinetics and the related comparisons and fitting [(a) adsorption isotherms; (b) comparison of NOR adsorption capacity by C₂H₆O-BC with that by unmodified/modified hydrothermal/slow pyrolyzed biochars reported in literature [34–38]; (c) adsorption kinetics ; and (d) mass transfer process simulated by weber-morris model].

The adsorption isotherms provide information on the adsorption equilibrium behavior (mechanism and maximum adsorption), affinity, binding energy between adsorbent/adsorbate and adsorption mode (heterogeneous or homogeneous) [40]. Two-parameter and three-parameter isotherm models were used to describe the mechanism of interactions between adsorbent and adsorbate in this study (S1.1, Table 2). The biochar-NOR interactions fitted by Temkin isothermal equilibrium model showed that the heat of adsorption of NOR on C₂H₆O-BC corresponded to a Temkin constant BT of 24.7 J·mol⁻¹ (<20 kJ·mol⁻¹), which further characterized the physical nature of the process [41,42]. Freundlich-Langmuir model assumes the presence of two types of surface adsorption sites [43]. Its n-value characterizes the interaction between the adsorbent and the adsorbate; and the greater the nonuniformity of the system, the smaller the n-value [44]. A slight decrease in the n values of the biochar after ball milling indicated an increase in the nonuniformity of C₂H₆O-BC. The Redlich-Peterson model describes adsorption equilibrium over a wide range of concentrations for both homogeneous and heterogeneous systems. Its’ fitting determination coefficients for the biochars were all above 0.95, indicating heterogeneous adsorption at low concentrations of NOR and homogeneous adsorption at high concentrations of NOR, similar to the previous reports [45].

Table 2. Adsorption fitting parameters of C₂H₆O-BC and p-BC.

	Parameters						C2H6O-BC			p-BC
Isotherm model parameters of C2H6O-BC and p-BC.
Temkin	KT (L/mg)						4.9			0.600
	BT (J/mol)						24.7			4.20
	R^2						0.955			0.747
Freundlich-Langmuir	KLF ((L/mg)^1/n)						13.5			9.07
	Qmax (mg/g)						53.1			7.05
	n						0.762			0.791
	R^2						0.974			0.864
Redlich-Peterson (RP)	KRP (L/g)						98.3			3.07
	aRP (L/mg)						0.920			0.520
	β						0.900			0.630
	R^2						0.960			0.968
Kinetic parameters of C2H6O-BC and p-BC.
	PFO			PSO			Elovich			Ritchie
	K1	Qe	R^2	K2	Qe	R^2	α(	β	R^2	α	qe	n	R^2
C2H6O-BC	1.33	84.3	0.850	0.025	90.3	0.953	827	0.072	0.961	0.070	85.0	1.80	0.947
p-BC	0.406	11.8	0.928	0.032	14.0	0.967	9.40	0.310	0.984	0.140	12.9	1.36	0.967
Parameters of mass transfer process of C2H6O-BC and p-BC.
	Weber and Morris						Mathews and Weber
	Kid		C		R^2		KMW		R^2
C2H6O-BC	49.9		12.9		0.967		1.55		0.889
p-BC	3.54		0.511		0.990		0.267		0.966

3.2.2. Adsorption kinetics

The adsorption of NOR onto the biochars increased with the reaction time and then reached a plateau (Figure 2c). During 0 ~ 1 h, the adsorption capacities of NOR increased rapidly due to the rapid occupation of the surface active sites of the biochar in the rapid adsorption stage; increased gradually from 1 to 6 h in the slow adsorption stage and basically reached the equilibrium adsorption after 6 h. The equilibrium adsorption capacity of NOR was in the order of C₂H₆O-BC > H₂O-BC > p-BC (Figure 2c and Fig. S3b). Adsorption kinetics helps predict the removal efficiency of pollutants and control factors can be identified [46]. The fitted parameters of four popular kinetic models were calculated (S1.2 and Table 2). Table 2 shows that the PSO model better reflected the adsorption behavior of NOR by the biochars. It was inferred that the interaction of NOR and the biochar was a multiple process dominated by chemical reactions, including H-bonding, π-π interactions, etc [47]. The Elovich kinetic model describes a non-homogeneous diffusion process regulated by the reaction rate and diffusion factor [48]. The value of the determination coefficient R² for the Elovich kinetic model was higher than 0.90, suggesting that the adsorbent’s surface energy was inhomogeneous, consistent with the previous reports [32]. The Ritchie kinetic model is a physically meaningful empirical model related to active site adsorption, where one adsorbent ion/molecule can occupy n active sites [49]. C₂H₆O-BC (n = 1.80) > p-BC (n = 1.36) suggests that after ball milling the number of the active sites increased.

The mass transfer process is key to a deeper understanding of the rate-limiting step during adsorption process. Interactions of adsorbates on adsorbents can be divided into three steps: (1) liquid film diffusion, (2) intra-particle diffusion and (3) surface chemical reaction [50]. The third step is the fastest step in the surface adsorption process [51], and it is important to note that these three steps act separately or jointly in the reaction mechanism. The application of mass transfer modeling allows the determination of the limiting mechanism of the overall adsorption rate. The kinetic adsorption results for biochar were fitted using different membrane diffusion and intraparticle diffusion models in this study (S1.3). The exploration with Weber and Morris model confirmed that intra-particle diffusion was not the only rate-determining process (Figure 2d). Therefore, boundary layer control may be involved, consistent with the previous reports [52]. The K_id values of C₂H₆O-BC, much higher than those of p-BC, suggested that the internal diffusion process was accelerated by the increase in specific surface area of the biochar after ball milling. The exploration with Mathews and Weber (M&W) model found larger K_MW value of C₂H₆O-BC than p-BC, indicating lower resistance to the external mass transfer after ball milling and faster membrane diffusion process [53,54].

3.2.3. Effect of solution initial pH, co-existing Cu(II) and biochar dosage

Figure 3a shows that solution initial pH had slight effects on NOR adsorption onto C₂H₆O-BC at pH range of 3–8. Solution initial pH impacts not only on the surface charge of the biochar but also the ionic state of NOR, which influence adsorption process. Since NOR has two dissociation constants (pK_a1 = 6.22, pK_a2 = 8.50), the pH of solution initial pH has a great influence on NOR species [36]. When pH < 6.22, NOR species is dominated by NOR⁺; NOR^± is predominant species when 6.22 < pH < 8.50; NOR⁻ is predominant when pH > 8.50 [55]. pHpzc of C₂H₆O-BC was 8.38, indicating that the surface charge of C₂H₆O-BC was positive at pH < 8.38 and negative at pH > 8.38 [31]. When initial pH increased from 3.00 to 8.00, the final pH after sorption increased from 6.40 to 8.62. Therefore, NOR^± adsorption onto negatively charged or nearly neutral biochar surface dominated and the electrostatic interaction was not the dominant mechanism for NOR adsorption onto C₂H₆O-BC when initial pH is lower than 8.00. But it may explain why removal ratio decreased when initial pH was above 9.00 in the experiment (final pH from 8.98 to 10.6) since NOR and C₂H₆O-BC were supposed to be both negatively charged. C₂H₆O-BC showed excellent ability to buffer acidic solution.

Figure 3. Effect of solution initial pH on NOR adsorption onto C₂H₆O-BC (bichar 2 g L⁻¹, 24 h)(a), co-existing Cu(II) (bichar 1 g L⁻¹, 24 h) (b) and solid-liquid ration (1: 0.5 g L⁻¹; 2: 1 g L⁻¹; 3: 2 g L⁻¹; 4: 4 g L⁻¹, 24 h) (c).

Figure 3b shows the removal of NOR or Cu(II) by C₂H₆O-BC from NOR solution, Cu(II) solution and their mixture. The comparison among different curves indicated that the removal of NOR and Cu(II) was affected by each other. C₂H₆O-BC removed 86.7% of NOR and 77.4% of Cu(II) when alone in solution, while 66.9% of NOR and 58.2% of Cu(II) were removed when they were mixed in solution. This indicates an antagonistic effect on adsorption of NOR and Cu(II) onto C₂H₆O-BC and the removal of NOR was impaired by the co-existing metal cations.

Figure 3c shows that the maximum adsorption efficiency of biochar increased with the increasing spiked dosage of biochars. The higher the spiked dosage of biochar, the more the number of surface active sites presented in the system, so the interaction of the biochar particles with the pollutants increases. Therefore, the maximum removal of NOR reached 98.2% at the biochar dosage of 2 g·L⁻¹. However, the remaining of NOR in the solution was so dilute that the biochar might not be able to reach its maximum adsorption capacity. Therefore, the dosage of 1 g·L⁻¹ was relatively cost-effectiveness.

3.3. Fixed-bed adsorption

Figure 4a shows that the adsorption retention and penetration were roughly divided into three phases for different influent concentrations of NOR. 1) The initial phase, in which the adsorption of NOR was far from saturated and the C_t/C₀ was almost zero; 2) the rapid adsorption phase, in which the adsorption amount decreased and the C_t/C₀ increased rapidly and 3) the equilibrium phase, in which the C_t/C₀ remained nearly constant. With the increasing time, the C_t/C₀ increased and the penetration curve was shifted upward. This indicates that the diffusion of NOR increased and the mass transfer resistance decreased with the increasing mass of NOR, which made it possible to reach the adsorption saturation phase [56]. With the increasing influent concentration, the amount of NOR adsorbed increased from 3.22 to 17.3 mg g⁻¹ with the increasing influent concentration, showing C₂H₆O-BC’s potential for application in fixed bed adsorption. Since pH of the effluent varied between 6.33 and 7.33 during the filtration processes as influent NOR increased from 15 mg L⁻¹ to 200 mg L⁻¹, the adsorption was hardly affected by the solution pH. Thomas and Yoon-Nelson models were used to fit the penetration curves over 60 min in this study and the value of determination coefficient R² were above 0.9, showing consistence of the data with the two models (S1.5, Figure 4b,c and Table S2). The Thomas rate constant (K_TH), which is the rate of NOR transfer from solution to biochar, decreased with the increasing initial concentrations of NOR, which was consistent with previous studies [57,58]. The increasing initial NOR concentrations decreased the adsorption rate constant and shortened the 50% penetration time predicted by the Yoon-Nelson model, indicating that accelerated adsorption and shortened breakthrough time with the increasing feed concentrations.

Figure 4. Fixed bed adsorption of NOR adsorption by C₂H₆O-BC [(a) sorption of NOR with influent concentration of 15 mg L⁻¹ (NOR1), 100 mg L⁻¹ (NOR2) and 200 mg L⁻¹ (NOR3); (b) Thomas model fitting; (c) Yoon-nelson Model fitting; (d) effects of high concentration of Cu(II) (15 mg L⁻¹ NOR +100 mg L⁻¹ Cu); (e) effects of Cu(II) preloading (200 mg L⁻¹ NOR after preloading of 120 mg L⁻¹ Cu); and (f) co-transport of 200 mg L⁻¹ NOR +120 mg L⁻¹ Cu as a comparison for preloading experiment].

The values of C_t/C₀ of NOR increased obviously at the presence of 100 mg L⁻¹ of Cu(II) in 15 mg L⁻¹ of NOR influent, consistent with the result of batch experiment that Cu(II) had an inhibitory effect on the adsorption of NOR onto C₂H₆O-BC (Figure 4d). In fact, NOR removal rate decreased from 71.2% to 57.4%. The experiment of sequential filtration of Cu(II) solution and NOR solution further confirmed that preloading of Cu(II) advanced the breakthrough time of NOR and resulted in the decrease of NOR removal ratio by 2.19% (Figure 4e). During the replacement of Cu(II) solution by successional NOR influent in the column, the relatively low pH (about 5.20) of pore solution and competition of Cu(II) were both unfavorable to the adsorption of NOR, responsible for the advanced breakthrough time. Compared with that in the preloading process (Figure 4e), Cu(II) adsorption was also slightly inhibited by NOR during co-filtration of 200 mg L⁻¹ NOR +120 mg L⁻¹ Cu, showing lower removal ratio (Figure 4f, Table S3). NOR showed similar lower removal ratios in sequential filtration and co-filtration (Figure 4e,f), compared with single filtration of influent of 200 mg·L⁻¹ of NOR. However, the decrease of removal ratio was much smaller than that in the filtrations with low influent NOR concentration (15 mg L⁻¹ of NOR), indicating concentration dependent inhibition. Anyway, the modified biochar showed potential to removal NOR and Cu(II) spontaneously through fixed-bed adsorption.

The two-site kinetic model (TSM) of the nonequilibrium convective diffusion model [59] was used to simulate the NOR/Cu(II) penetration profile in the column experiments as the thin biochar layer was assumed not to effect the pore volume of the column (S1.5, Table S3). The two-site kinetic model simulated the NOR/Cu(II) penetration curves with R² >0.99. The values of the blocking factor R_d obtained from the simulation were all greater than 1, showing excellent removal ability for NOR. On the other hand, the R_d values decreased with Cu(II) presence, indicating inhibition of NOR adsorption onto C₂H₆O-BC. Through calculating the partition coefficient K_d, it was found that the variation of K_d was basically consistent with the trend of the observed results, and the value of K_d in single NOR filtration was larger than that under the rest of the conditions, which further proved that the adsorption of NOR on C₂H₆O-BC at this time was more preferred.

3.4. NOR removal via STBR

The removal ratios of NOR in the STBR system were shown in Table 3. When the mass of biochar adsorbent in both stage 1 and stage 2 was 0.5 g·L⁻¹, the two-stage adsorption system achieved a removal ratio of 81.3% for NOR (100 mg·L⁻¹) after 4 h (2 h each stage), which was a little higher than that of the single-stage adsorption system (1 g·L⁻¹ biochar). When the biochar mass in each stage was increased to 1 g·L⁻¹, the removal ratio of the two-stage adsorption system reached 98.6%, which was similar to that of the single-stage adsorption system (2 g L⁻¹ biochar). This may be attributed to the fact that the NOR in the wastewater was adsorbed nearly completely because of the high dosage of biochar. The results indicated that at relatively low biochar does, two-stage reaction system may have higher removal efficiency than one-stage system. As solid-liquid ratio was large enough, one-stage reaction system had similar efficiency to two-stage reaction system, with only half of the residence time of the later.

Table 3. Removal ratios of nor by C₂H₆O-BC via STBR.

STBR				Batch adsorption
Biochar (g/L)	t (h)	Removal ratios of stage 1(%)	Removal ratios of stage 1 + 2(%)	Biochar (g/L)	t (h)	Removal ratios (%)
0.5	2	40.4 ± 0.6		1	2	69.8 ± 0.4
0.5	2		81.3 ± 0.4		4	78.4 ± 0.6
1	2	70.7 ± 0.4		2	2	98.6 ± 0.1
1	2		98.7 ± 0.1		4	98.7 ± 0.1

This may be explained by the model scheme of STBR. If each stage reaches equilibrium, the NOR loading on the bichar for each stage Qi could be predicted as K_fC_i^1/n according to the Freundlich equation. Then removal ratio R₁ in stage 1 equals (C₀-C₁)/C₀ or MK_fC₁^1/n/(VC₀), and in stage 2 removal ratio R₂ equals (C₁-C₂)/C₀ or MK_fC₂^1/n/(VC₀). Therefore, the total removal ratio R equals (R₁+ R₂), which is MK_f (C₁^1/n+ C₂^1/n)/(VC₀). Since C₁ equals (1-R₁)C₀ and C₂ equals (1-R₁-R₂)C₀, R equals MK_f C₀^1/n ((1-R₁)^1/n+ (1-R₁-R₂)^1/n)/(VC₀). In this process, a total of 2 M bichar was used. If a one stage reactor is used for the same inflow NOR solution and biochar, the removal ratio R’ equals 2MK_f C₀^1/n (1-R’)^1/n/(VC₀). When biochar mass M is very large, R is expected to be near 100% and similar to R₁ (much larger than R₂) and R’ according to Figure 3c. When biochar mass M is very small, biochar is easy to be saturated. R is expected to be very small and similar to R₁, R₂ and R’ as NOR amount is too large to remove due to the large volume relative to the biochar mass. When biochar mass M is between them (e.g. biochar of 0.5 to 1 g L⁻¹), (1 - R₁)^1/n is larger than (1 - R’)^1/n as R’ > R₁ > R₂. If R’ ≥ (R₁ + R₂), (1 - R₁ - R₂)^1/n ≥ (1 - R’)^1/n, indicating a contradict result that MK_f C₀^1/n ((1-R₁)^1/n + (1-R₁-R₂)^1/n)/(VC₀) is larger than 2MK_f C₀^1/n (1-R’)^1/n/(VC₀). Therefore, R’ is less than (R₁ + R₂). It suggests that the STBR system may have a higher or similar removal ratio to the one stage system.

Although the two adsorption stages were far from equilibrium in the current experiment with residence time of only 2 h, the STBR system had a higher or similar efficiency to that of the single-stage adsorption system for 100 mg L⁻¹ NOR waste water. As the two reactors in STBR system can run spontaneously in practice, it will greatly save reaction time at rational biochar dosage. Furthermore, the separated biochar could still be used until they were exhausted, since it had not reached its adsorption capacity in both adsorption systems. This experiment shows that the STBR system using C₂H₆O-BC is an efficient, fast and convenient system to remove the aqueous NOR. However, adsorption of NOR is a complex process and depends on many factors (e.g. initial concentration, residence time, and solid-liquid ration and so on), so further investigation should be done to seek the effective and economical operating conditions for practical application in wastewater treatment via this two-stage batch reaction process.

3.5. Adsorption mechanisms

Diverse adsorption mechanisms are believed to be involved in the adsorption of organic pollutants in water [60]. In order to investigate the potential mechanisms, the relevant characterization of C₂H₆O-BC after the adsorption of NOR and Cu(II) was given. Combined with the fitting and analysis of the observed data mentioned above, possible adsorption mechanisms involved in the adsorption process could be suggested.

Isotherm and kinetic analysis and effect of coexisting Cu(II) on NOR adsorption suggest the physisorption and chemisorption may proceed at the same time. Effect of solution initial pH on NOR adsorption mentioned above shows that the electrostatic interaction was not the dominant factor for NOR adsorption onto ball-milled biochar when pH was below 8.00. The adsorption mechanisms of NOR onto C₂H₆O-BC may include π-π interactions, H-bonding and pore-filling, etc. NOR can act as a strong π-acceptor containing fluorine groups with high electron-withdrawing capacity [36]. The previous study shows that ball milling can improve the aromatic π-π structure of biochar, thus enhancing the removal of organic pollutants [61]. The aromaticity (H/C) of C₂H₆O-BC was lowerthan p-BC and increased after adsorption of NOR (Table 1), which may be due to the stacking of aromatic compound of NOR on C₂H₆O-BC. The oxygen-containing functional groups were slightly reduced after NOR adsorption (Figure 2d), which further demonstrated that they may involve in the surface adsorption through hydrogen bonding interactions. In the full XPS spectra after adsorption, a characteristic peak of F1s at 688 ± 0.2 eV was observed, indicating that norfloxacin was adsorbed on the surface of biochar (Figure 5A). In the high-resolution XPS fits of C1s, O1s, the characteristic peaks of C-C/C=C and C-O were at 284.5 ± 0.2 eV and 285.7 ± 0.2 eV, respectively (Figure 5B) [34,35]. Before adsorption, the O1s peaks of C₂H₆O-BC at 530.5 ± 0.2, 531.7 ± 0.2 and 534.3 ± 0.2 eV can be attributed to C=O, C-O and COOH [Figure 5B(e)] [62]; and after adsorption of NOR, the C=O and COOH are shifted, which may be attributed to the π-π interaction between C₂H₆O-BC and NOR [Figure 5B(f,h)] [63].

Figure 5. The XPS spectrum of C₂H₆O before and after adsorption [(a) the full XPS spectrum and (b) the high-resolution XPS spectrum of C1s, O1s and Cu2p].

Coexisting Cu(II) inhibited the adsorption process of NOR on C₂H₆O-BC [Figures 3b and 4(d,e,f)]. As for C₂H₆O-BC, the intensity of the peak of phenol-OH at 1378 cm⁻¹ was enhanced after adsorption, suggesting there may be surface complexion by C₂H₆O-BC (Figure 1d). After simultaneous adsorption of NOR and Cu(II), a new Cu-O peak was generated in XPS spectrum (Figure 2h), indicating that Cu(II) was adsorbed on C₂H₆O-BC through oxygenated functional groups [64]. As for adsorption of Cu(II) alone, Cu(II) was present in both Cu²⁺ and Cu⁺ forms on C₂H₆O-BC surface, but the amount of Cu⁺ was reduced in simultaneous adsorption of NOR/Cu(II) [Figure 5B(k,l)], suggesting reduction of Cu²⁺ and possible competition between NOR and Cu(II). Cu(Nor^±)²⁺, Cu(Nor^±)₂²⁺, Cu²⁺ and NOR⁺ may occur at pH of 4.0–6.0 value in co-adsorption of NOR and Cu(II) [65]. Competition between NOR⁺ and Cu(II) may occur during adsorption. Static repulsion between the cations and the positive charged biochar surface may also be strengthened when monovalent NOR⁺ was transformed to divalent Cu(Nor^±)²⁺ and Cu(Nor^±)₂²⁺. Complexation may result in persistence in the solution of the two pollutants or co-adsorption onto biochar surface as complexes. Moreover, Surface complexation of heavy metal ions may also alter the surface chemistry or pore structure of biochar, thereby affecting organic adsorption on biochar [66]. Cu(II) ions underwent hydration in aqueous solution, holding dense waters in the hydration shell. The held ‘hard’ water may enter into the neighboring biochar surface, thereby clogging the pores of the biochar surface and thus inhibiting NOR adsorption on C₂H₆O-BC [66]. Therefore, reciprocal inhibition of the two pollutants in some extent were observed in our experiment.

4. Conclusion

In this study, a simple and easy method for the preparation of biochar through combined ball milling and chemical modification (ethanol) was developed and the adsorption performance for NOR was evaluated via static adsorption (batch adsorption) and dynamic adsorption (fixed-bed column adsorption and two-stage batch adsorption). The C₂H₆O-combined-ball-milling method has a significant influence on the physicochemical properties such as particle size, inorganic element content and pH of the biochar. The Langmuir maximum adsorption of NOR onto C₂H₆O-BC was 163 mg·g⁻¹. The adsorption kinetics data was fitted well with the pseudo-second-order model, which indicated that the adsorption of NOR by C₂H₆O-BC was a multiple adsorption process. Initial solution pH almost did not affect NOR adsorption when pH was below 8.0. NOR removal ratio was nearly 100% when biochar dosage was larger than 2 g L⁻¹. Although coexisting Cu(II) inhibited the adsorption of NOR by C₂H₆O-BC, it showed potential to removal NOR and Cu(II) spontaneously. Moreover, C₂H₆O-BC showed excellent NOR removal ability in both fixed-bed column experiments and two-stage batch reactor. The process was mainly influenced by the combined effects of π-π interactions, H-bonding, and pore-filling; Cu(II) inhibited the adsorption of NOR by C₂H₆O-BC through competitive adsorption and complexation. In summary, this study provides an adsorbent with high adsorption capacity, good environmental adaptability and potential for practical application for the removal of NOR.

Disclosure statement

No potential conflict of interest was reported by the author(s).

Data availability statement

Data will be made available on request.

Supplementary material

Supplemental data for this article can be accessed online at https://doi.org/10.1080/26395940.2024.2311675.

Additional information

Funding

This work was partly supported by the National Key R&D Program of China [No. 2018YFC1800600] and the National Natural Science Foundation of China [No.21677075]. Ding Z. thanks for the support of “Jiangsu 333 High Level Talent Training Project [2022-2026]”. Yang S. also thanks for the support of “Postgraduate Research & Practice Innovation Program of Jiangsu Province [KYCX23_1484]”.