Antibiotics have become ubiquitous aquatic pollutants owing to their increasing use in agriculture and medicine. As a large portion is excreted unmetabolized, antibiotics enter water bodies through hospital, pharmaceutical, and municipal wastewater streams [1,2]. Conventional biological treatments fail to mineralize the stable structures of antibiotics, leading to their persistent presence in the environment [3,4]. Consequently, sub-inhibitory concentrations exert sustained selective pressure on aquatic microbiomes, accelerating antimicrobial resistance (AMR), a critical global health threat [5-7]. Effective post-treatment technologies are therefore urgently needed.

Ciprofloxacin (CFX; C₁₇H₁₈FN₃O₃; M = 331.34 g mol⁻¹), a widely prescribed broad-spectrum fluoroquinolone, is frequently detected in environmental waters at nanogram- to microgram-per-liter levels. CFX resists biodegradation because it targets bacterial DNA gyrase and topoisomerase IV. The environmental fate of CFX is governed by its amphoteric character (pK_a1 ≈ 6.1; pK_a2 ≈ 8.7), where pH-dependent speciation yields a large fraction of CFX⁺ at pH < 6.1, CFX^± at pH 6.1–8.7, and CFX⁻ at pH > 8.7 [8]. This speciation significantly affects CFX interactions with adsorbent surfaces, an important consideration when designing removal systems.

Despite achieving high CFX removal rates, advanced oxidation and membrane technologies remain limited by prohibitively high capital costs and operational complexity. This limitation is especially relevant in resource-constrained settings in developing countries, where antibiotics are widely used and contribute significantly to pollution. Sorption using inexpensive solid sorbents therefore remains a feasible option from both economic and efficiency standpoints. Among various solid sorbents, biochar has emerged as a promising material for carbon sequestration and antibiotic removal; however, feedstock and pyrolysis conditions significantly influence its sorption characteristics. At moderate pyrolysis temperatures (e.g., 400 °C), biochar retains a substantial amount of oxygen-containing functional groups that act as active sites for hydrogen bonding and electrostatic interactions with polar adsorbates while also developing incipient aromatic domains capable of π–π electron donor–acceptor interactions [9-11].

Among various biomasses, lotus (Nelumbo nucifera) is abundant in Asia and serves as an important precursor for biochar production [12]. Lotus-derived biochar exhibits exceptional versatility in remediating diverse environmental contaminants, effectively eliminating heavy metals [13], adsorbing organic dyes such as methyl orange [14], capturing carbamate pesticides [15], and catalyzing heavy oil degradation [16]. However, no reports exist on using unmodified lotus stem biochar to remove CFX, despite its demonstrated effectiveness against a variety of other pollutants. Consequently, it remains unclear how factors in solution chemistry, such as pH and background cations, influence CFX sorption onto this material, particularly given that pH-modulated antibiotic speciation and biochar surface charge can alter removal pathways and binding affinity [10,17].

CFX adsorption has been investigated on pinewood biochar [11], chitosan/biochar composites [18], and surface-modified biochars [3, 19-20], with widely varying capacities reported. Adsorption is known to be highly sensitive to CFX speciation [4], and divalent cations exhibit competitive electrostatic interactions and chelation with the carboxylate and amine groups of CFX [21,22]. No research to date, however, has examined these factors for unmodified lotus stem biochar.

Therefore, this study aimed to investigate the adsorption behavior of CFX onto biochar (BC) derived from Nelumbo nucifera stem petioles produced by slow pyrolysis at 400 °C. Sorption kinetics and equilibrium isotherms were evaluated at three pH levels (3, 7, 10), three temperatures (20, 26, 32 °C), and in the presence of coexisting mono- and divalent cations (Na⁺ and Ca²⁺). Thermodynamic parameters were calculated using the Van’t Hoff equation. This study represents the first systematic investigation of CFX adsorption by unmodified lotus stem biochar, integrating solution speciation and competitive ionic matrix effects within a unified mechanistic framework and contributing to the valorization of agricultural residues as cost-effective adsorbents for antibiotics.

2.1. Reagents and Materials

Ciprofloxacin (CFX; purity ≥ 98%, Acros Organics) was used as the target adsorbate without further purification. Stock solutions of 1000 mg L⁻¹ were prepared by dissolving CFX in doubly distilled water and stored at 4 °C in the dark. Working solutions were prepared daily by serial dilution. Sodium hydroxide (NaOH, AR grade, Xilong Chemical) and hydrochloric acid (HCl, AR grade, Xilong Chemical) were used to adjust the pH. Sodium chloride (NaCl, AR, Xilong Chemical) and calcium chloride (CaCl₂, AR, Xilong Chemical) were used in background ionic strength experiments. All solutions were prepared with doubly distilled water.

2.2. Preparation of Biochar

Lotus stems (Nelumbo nucifera) were collected from local ponds in the Mekong Delta region of Vietnam. The raw material was washed thoroughly with tap water, followed by doubly distilled water, to remove surface impurities; cut into small pieces (~3–5 cm); and oven-dried at 105 °C for 24 h. The dried biomass was placed in a muffle furnace (without N₂ gas protection) and pyrolyzed at 400 °C at a heating rate of 10 °C min⁻¹, then held at the target temperature for 2 h. Pyrolysis in a muffle furnace provides a simple, cost-effective approach for producing biochars with sorption performance comparable to biochars produced using tube furnaces and traditional kilns [23,24]. The resulting biochar was allowed to cool naturally to room temperature and was then ground and sieved. The product was washed until the solution reached a pH of 7 and subsequently stored in a desiccator until use.

2.3. Characterization of Biochar

The specific surface area (SSA), total pore volume, and average pore diameter of BC were determined by N₂ adsorption–desorption at 77 K using a surface area analyzer (Quantachrome NOVA 1000e, USA). Surface morphology and elemental composition were examined by field-emission scanning electron microscopy coupled with energy-dispersive X-ray spectroscopy (FESEM/EDX; HITACHI S-4800, Japan; HORIBA EDX H-7593, UK). Surface functional groups were identified by Fourier-transform infrared spectroscopy (FTIR; PerkinElmer NIR/MIR Frontier) in the range 400–4000 cm⁻¹ at 1 cm⁻¹ resolution and 50 scans per spectrum. FTIR spectra were collected for both pristine biochar and samples recovered after CFX adsorption.

The point of zero charge (pH_pzc) was determined by the pH drift method [25]. A series of 25 mL aliquots of 0.1 M KCl solution was adjusted to initial pH values (pH⁰) of 2–11 using 0.1 M NaOH and 0.1 M HCl. To each, 0.1 g of biochar was added, and suspensions were agitated at 120 rpm for 24 h at room temperature. Final pH (pHᶠ) was recorded after filtration, and ΔpH (= pHᶠ − pH⁰) was plotted against pH⁰. The pH_pzc was identified as the intercept of the ΔpH versus pH⁰ curve with the x-axis.

2.4. Adsorption Experiments

2.4.1. Sorption Kinetics

Adsorption kinetics were studied by preparing a CFX solution with an initial concentration of 16 mg L⁻¹ in a 250 mL beaker. The solution pH was adjusted to 7 using 0.1 M NaOH and 0.1 M HCl solutions. Subsequently, 0.5 g of biochar was added to the beaker, and the mixture was stirred with a magnetic stirrer at 120 rpm at 26 °C (room temperature). The biochar dosage was fixed at 2.0 g L⁻¹ [26]. Aliquots were withdrawn at predetermined time intervals from 0 to 1440 min, immediately filtered through 0.45 µm syringe filters, and analyzed by UV–Vis spectrophotometry (Jasco V-730, Japan). For aqueous solutions at pH 7, absorbance was measured at 271 nm [27] and quantified using the calibration equation (R² = 0.9999). Kinetic data were fitted to the nonlinear pseudo-first order (PFO), pseudo-second order (PSO), and two-stage intraparticle diffusion model (two-stage IPD) [28] (Equations (1)–(4)):

where q_max and q_e (mmol kg ⁻¹) are the sorption capacities of the PFO and PSO models, respectively.

1. q_t (mmol kg ⁻¹) is the sorption capacity at time t;

2. k₁ (min⁻¹) and k₂ (kg mmol⁻¹ min⁻¹) are the rate constants for PFO and PSO, respectively.

3. k₁^stage1 and k₂ ^stage2 (mmol kg⁻¹ min^−1/2) are the rate constants of the two-stage IPD;

4. t₁ is the time required for quasi-equilibrium of solute concentrations to be reached.

Several criteria have been established to select the appropriate model, including the second-order Akaike Information Criterion (AICc), the Shapiro–Wilk test, the p-value, and Adjusted R-squared (Adj-R²) [4].

2.4.2. Adsorption Isotherms

Equilibrium isotherm experiments were performed in centrifuge tubes containing 10 mL of CFX solution at initial concentrations of 12–80 mg L⁻¹. A 0.02-g portion of biochar was added, and the tubes were shaken at 26 °C. Samples were agitated at 120 rpm in a temperature-controlled shaker at the target temperature until adsorption equilibrium was confirmed (≥8 h, based on kinetic results). After equilibration, suspensions were filtered through 0.45 µm syringe filters and analyzed by UV–Vis spectrophotometry. Equilibrium data were described by the nonlinear Langmuir isotherm (Equation (5)):

where Q_max (mmol kg⁻¹) is the maximum monolayer adsorption capacity, K_L (L mmol⁻¹) is the Langmuir affinity constant, and C_e (mmol L⁻¹) is the equilibrium liquid-phase concentration. The Freundlich model was also applied. Parameter estimation was performed by nonlinear least-squares regression.

2.4.3. Effect of Solution pH

To evaluate the effect of solution pH on adsorption capacity, isotherm experiments were conducted at pH 3, 7, and 10, following the general procedure described above. pH was adjusted using 0.1 M NaOH and 0.1 M HCl and verified with a calibrated pH meter (Hanna Instruments HI 2210, USA). Experiments were conducted at 26 °C in the temperature-controlled shaker. For aqueous solutions at pH 3 and 10, absorbance was measured at 277 nm and 272 nm, respectively [27].

2.4.4. Effect of Background Cations

The influence of coexisting cations on CFX adsorption was investigated by separately adding 1 mL of a 0.1 M NaCl or CaCl₂ solution to each adsorption system, yielding final cation concentrations of 10 mM for monovalent (Na⁺) and divalent (Ca²⁺) species. Experiments were conducted at pH 7 and 26 °C. CFX concentrations were quantified using matrix-specific calibration curves at 270 nm for the Na⁺-containing solutions and 272 nm for the Ca²⁺-containing solutions. A control system without added cations was run in parallel under the same adsorption conditions and was quantified using the CFX calibration curve established at pH 7 (271 nm).

2.5. Thermodynamic Calculations

Thermodynamic parameters were derived from thermodynamic equilibrium constants obtained at three temperatures (20, 26, and 32 °C; pH 7) using the Van’t Hoff equation (Equation (6)):

where K_c (dimensionless) is the thermodynamic equilibrium constant obtained by converting the Langmuir K_L (L mmol⁻¹) to a dimensionless form, R is the universal gas constant (8.314 J mol⁻¹ K⁻¹), and T is the absolute temperature (K).

The estimated K_L (L mmol⁻¹) was converted to an approximate K_c value (Equation (7)) using the activity coefficient (γ_e) of the adsorbate and the molar concentration of the standard reference solution (C_s), which is equal to 1 mol L⁻¹.

According to the Debye-Hückel law, the activity coefficient γ_e ≈ 1 for neutral or weakly charged adsorbates and K_c ≈ 1000 K_L.

The standard Gibbs free energy change (ΔG°, kJ mol⁻¹), standard enthalpy change (ΔH°, kJ mol⁻¹), and standard entropy change (ΔS°, J mol⁻¹ K⁻¹) were calculated at each temperature as (Equation (8)):

3.1. Characterization of BC

As shown in Figure 1a, the SEM image revealed a mixed porous and compact surface morphology for biochar produced at a pyrolysis temperature of 400 °C. EDX analysis confirmed that carbon was the dominant element (Figure 1b), accounting for 87.65 ± 0.11 wt%, while oxygen accounted for 11.30 ± 0.11 wt% (Table 1).

The low O/C ratio (0.13) suggests significant removal of oxygen-containing functionalities during pyrolysis. Small amounts of sulfur (0.32 ± 0.02 wt%) and calcium (0.73 ± 0.05 wt%) were also detected. The O/C ratio of BC (0.13) is markedly lower than that of other biochars pyrolyzed at comparable moderate temperatures (300–500 °C with N₂ gas), including coconut shell biochar CN400 (O/C = 0.17) [28], corncob biochar CC400 (O/C ≈ 0.27) [29], and cassava peel biochar CP500 (O/C = 0.18) [4].

This suggests that lotus stem biomass undergoes more extensive thermal decomposition of oxygenated functionalities than conventional lignocellulosic and marine feedstocks, likely reflecting the hemicellulose-rich nature of aquatic macrophyte stems. The FTIR spectra further confirmed that pyrolysis significantly changed the surface chemistry (Figure 1c). In the raw biomass, a broad and intense band centered at 3447 cm⁻¹ was observed, which weakened and shifted to 3416 cm⁻¹ in BC, indicating a reduction in hydroxyl-related structures during carbonization. The band at 1633 cm⁻¹ in the raw lotus stem shifted to 1612 cm⁻¹ after pyrolysis, suggesting structural reorganization of the carbon matrix and increased contributions from unsaturated or aromatic C=C vibrations. A more noticeable change was seen for the C–O-related band, which moved from 1108 cm⁻¹ in the raw material to 1163 cm⁻¹ in BC. This reflects modifications to oxygen-containing functionalities after thermal treatment.

As shown in Figure 1d, the N₂ adsorption–desorption data were further analyzed using the Langmuir equation to determine the surface area. The linear plot showed excellent linearity with y = 65.50 x + 13.69 (R² = 0.997). The Langmuir surface area was estimated to be 53.175 m² g⁻¹, indicating that the biochar possesses a measurable accessible surface. The Langmuir-derived SSA is higher than that of other biochars pyrolyzed at 400 °C, including coconut shell biochar (4.06 m² g⁻¹) [28], corncob (4.8 m² g⁻¹) [29], and Sargassum polycystum (11.7 m² g⁻¹) [30], and even exceeds that of cassava peel (29.45 m² g⁻¹) pyrolyzed at a higher temperature of 500 °C [4]. The pH_pzc of the biochar was determined to be approximately 6.0 by the pH-drift method (Table 1). Overall, lotus biochar produced at moderate-temperature pyrolysis exhibited a carbon-rich composition with a low O/C ratio, and the Langmuir surface area indicated a moderately developed pore structure. FTIR analysis confirmed that pyrolysis reduced hydroxyl-related groups while promoting aromatic C=C and C–O groups. The coexistence of residual oxygen-containing functional groups and accessible porosity suggests that this biochar may support both surface-driven interactions and pore-filling mechanisms.

3.2. Sorption Kinetics at pH 7 and 26 °C

Table 2 presents the nonlinear kinetic fitting results for CFX sorption. The PSO model provided a better fit than the PFO model, with the higher adjusted coefficient of determination (Adj-R²: 0.904 vs. 0.738). However, the two-stage IPD model provided the best overall fit, with the highest Adj-R² (0.984) and the lowest AICc (−14.598). This suggests that CFX uptake did not proceed through a single kinetic step. The fitted IPD parameters showed a much faster first stage (k₁ = 2.775 ± 0.112 mmol kg⁻¹ min^−1/2) followed by a slower second stage (k₂ = 0.117 ± 0.010 mmol kg⁻¹ min^−1/2), with a transition time t₁ of 11 min. This pattern indicates a rapid initial uptake, likely associated with pore diffusion, followed by a slower diffusion-controlled stage.

3.3. Effect of pH on the Adsorption Isotherm

The adsorption isotherms of CFX were evaluated at pH 3, 7, and 10 to analyze the effect of solution pH (Figure 2a). The equilibrium data were fitted using both the Langmuir and Freundlich models, and the fitted parameters are summarized in Table 3. Both models described the data well, with Adj-R² values ranging from 0.874 to 0.986. The Freundlich model gave a slightly better fit at pH 3 and 10, whereas the Langmuir model performed better at pH 7 (Adj-R² = 0.974). The Langmuir maximum adsorption capacity, Q_max, varied with pH and followed the order pH 7 (22.96 mmol kg⁻¹) ≈ pH 10 (22.88 mmol kg⁻¹) > pH 3 (20.70 mmol kg⁻¹); notably, BC maintained high CFX uptake across acidic, neutral, and alkaline conditions, a promising result for practical pollutant removal applications. At pH 7, lotus stem biochar exhibited a Q_max that exceeded the capacities reported for conventional biochar derived from agricultural waste such as groundnut shell (6.9 mmol kg⁻¹) [31], cassava peel (10.5 mmol kg⁻¹) [4], and rice straw (11.2 mmol kg⁻¹) [32].

According to Table 3, K_L was also highest at pH 7 (65.16 L mmol⁻¹), suggesting stronger overall affinity at near-neutral pH. By contrast, the lower uptake at pH 3 can be attributed to less favorable surface–solute interactions under acidic conditions, interpretable in relation to the pH-dependent speciation of CFX and the surface charge of BC. At pH 3, CFX is expected to occur mainly in its cationic form (CFX⁺, 99%), while the BC surface is also positively charged; under these conditions, electrostatic attraction is reduced. The adsorption capacity nonetheless remained substantial, suggesting that non-electrostatic interactions still contributed significantly. At pH 7, CFX is expected to be predominantly zwitterionic (CFX^±, 87.2%), a condition that may provide a more favorable balance between electrostatic effects and other interactions. FTIR analysis of BC after CFX sorption at pH 7 revealed shifts in the –OH stretching band (3416 → 3438 cm⁻¹), aromatic C=C vibration (1612 → 1618 cm⁻¹), and C–O stretching (1163 → 1178 cm⁻¹), accompanied by a reduction in overall peak intensity. This is attributed to hydrogen bonding, π–π interactions, and oxygen-containing functional groups in CFX sorption.

3.4. Effect of Temperature on Adsorption

Adsorption isotherm experiments were conducted at 20, 26, and 32 °C to investigate the thermal dependence of CFX uptake on BC (Figure 2b, Table 3). At constant pH 7, Q_max showed a non-monotonic response to temperature. Q_max increased from 19.75 mmol kg⁻¹ at 20 °C to 22.96 mmol kg⁻¹ at 26 °C, then decreased to 16.26 mmol kg⁻¹ at 32 °C. The initial increase from 20 to 26 °C is consistent with the endothermic characteristic of adsorption. The elevated temperature provides additional thermal energy to overcome activation barriers for surface site occupancy and intraparticle diffusion. The subsequent decline at 32 °C is attributed to the increasing desorption rate at higher thermal energy. The inverse trend between K_L and Q_max at 32 °C suggests that elevated temperature strengthened individual binding interactions, while simultaneously reducing the number of weak sorption sites. Across the tropical temperature range (20–32 °C), Q_max exhibited a moderate coefficient of variation of 17.1%, yet remained consistently above 16 mmol kg⁻¹ under all conditions tested.

3.5. Sorption Thermodynamics

Thermodynamic parameters derived from Van't Hoff analysis at pH 7 are presented in Table 4 and Figure 2d. Specifically, ΔG° ranged from −26.24 kJ mol⁻¹ (20 °C) to −29.11 kJ mol⁻¹ (32 °C). ΔG° was negative at all temperatures examined, confirming that CFX adsorption onto BC is spontaneous, and the increase in |ΔG°| with temperature confirms that higher temperature enhances the thermodynamic driving force for sorption. The maximum adsorption capacity, however, increased from 20 to 26 °C and then decreased at 32 °C. This can be explained because ΔG° is governed by K_c [28-30], which varies with temperature, and its trend may not coincide with that of Q_max, which is instead influenced by specific interactions inherent to the heterogeneous surface [28-30]. Increased thermal energy, coupled with entropy gain, facilitates desorption, thereby reducing Q_max. Similar phenomena have been observed for other materials, such as micronized oil shale, ferrite-doped lignin biochar, hematite, and humic acid, used to adsorb antibiotics or other substances [33-36]. The standard enthalpy ΔH° was positive (+44.02 kJ mol⁻¹), establishing the endothermic nature of CFX adsorption onto BC. Endothermic adsorption arises when the thermal energy required to desorb water molecules from both the adsorbent surface (hydration sphere disruption) and the CFX hydration shell exceeds the enthalpy released upon CFX–surface bond formation. The ΔH° value of 44.02 kJ mol⁻¹ exceeds the typical physisorption threshold (<20 kJ mol⁻¹), indicating the involvement of stronger, chemically specific interactions in addition to physical forces, consistent with a mixed physico-chemical adsorption mechanism involving simultaneous electrostatic, π–π stacking, and hydrogen-bonding interactions [9,19]. ΔS° was large and positive (+239.66 J mol⁻¹ K⁻¹); large positive ΔS° values are attributed to structural rearrangement at the solid–liquid interface. This large positive TΔS° term overcomes the energetic cost of ΔH° > 0, rendering ΔG° = ΔH° − TΔS° negative and adsorption spontaneous.

3.6. Effect of Coexisting Cations

Figure 2c and Table 3 summarize the influence of background mono- and divalent cations on CFX adsorption at pH 7 and 26 °C. In the absence of added cations (control), Q_max was 22.96 mmol kg⁻¹. Addition of monovalent Na⁺ caused only modest reductions, to 20.91 mmol kg⁻¹, corresponding to a capacity decrease of ~9%, while K_L decreased by 42%. This suppression by Na⁺ suggests competitive adsorption at the surface sites of biochar; however, the minor effect observed is also attributable to ionic strength effects, since increased ionic strength compresses the electrical double layer around the biochar surface, partially screening the electrostatic attraction between the piperazinyl nitrogen of CFX^± and negatively charged surface sites [21,22]. Our observed 9% capacity loss under higher Na⁺ concentrations is consistent with previous studies [37], which similarly report that higher electrolyte levels weaken the attraction between the negative surface of the adsorbent and the antibiotic's amino groups.

In sharp contrast, divalent cations caused severe and differentiated suppression of Q_max and K_L (Figure 3). In the presence of Ca²⁺, Q_max decreased by 49.0%, while K_L decreased by up to ~68%. The overall inhibition order was Control > Na⁺ > Ca²⁺, consistent with competitive interactions reported for fluoroquinolone adsorption on carbonaceous materials [21,22]; divalent cations are known to cause much stronger inhibition, as seen, for example, in bamboo biochar, where Ca²⁺ disrupted adsorption far more severely than Na⁺ [38].

The strong inhibition by Ca²⁺ is attributable to the substantially larger hydration sphere (hydrated ionic radius ~ 4.12 Å for Ca²⁺ vs. 3.6 Å for Na⁺) and its stronger Lewis acid character, which confer two mechanistically distinct inhibitory pathways. First, Ca²⁺ can form stable bidentate chelates with the adjacent 3-carboxyl and 4-keto (oxo) groups on the quinolone core of CFX in solution [39,40], converting free CFX into poorly reactive Ca–CFX complexes and reducing the effective CFX concentration accessible to the adsorbent surface. Second, Ca²⁺ competes directly for the same –COOH and –OH surface sites on the biochar that would otherwise engage in electrostatic attraction and hydrogen bonding with CFX.

Because a divalent cation can occupy two adjacent functional sites simultaneously, its site-blocking efficiency is inherently greater than that of the monovalent Na⁺ [41]. The stronger affinity of Ca²⁺ for both surface functional groups and CFX functional groups explains its greater inhibitory capacity compared with sodium ions (Na⁺). Pretreatment to remove divalent cations or the use of biochar with a higher degree of aromatization, thereby reducing reliance on functional group-mediated mechanisms, may be necessary to maintain adequate performance under such conditions.

3.7. Proposed Sorption Mechanism

A mechanistic framework is proposed here, integrating evidence from biochar characterization, sorption kinetics, pH- and temperature-dependent isotherms, thermodynamic analysis, and ionic competition experiments. The carbon-rich composition (87.65 wt%) and the Langmuir surface area of 53.175 m² g⁻¹ indicate a moderately developed pore structure. Sorption kinetics were rapid during the first stage of the IPD model (k₁ = 2.775 mmol kg⁻¹ min^−1/2), corresponding to external surface adsorption and diffusion through macropores, indicating that CFX sorption onto lotus biochar (BC) is governed by surface-driven interactions.

Electrostatic attraction and hydrogen bonding are considered the dominant mechanisms for BC. At pH 7, the negatively charged BC surface (pH > pH_pzc = 6.0) attracts the positively charged piperazinyl nitrogen of zwitterionic CFX^± (CFX^±, 87.2%) via Coulombic attraction. The Langmuir model yielded the highest Q_max (22.96 mmol kg⁻¹) and K_L (65.17 L mmol⁻¹). The FTIR shifts observed after CFX sorption at pH 7—specifically the blue shift of the –OH band (3416 → 3438 cm⁻¹), the shift of the aromatic C=C vibration (1612 → 1618 cm⁻¹), and the displacement of the C–O stretching band (1163 → 1178 cm⁻¹)—collectively confirm the involvement of hydrogen bonding and π–π stacking. The positive ΔH° (+44.02 kJ mol⁻¹) established the endothermic nature of CFX adsorption, indicating that energy is required to disrupt the hydration shells surrounding both the CFX molecules and the biochar surface.

The large positive ΔS° (+239.66 J mol⁻¹ K⁻¹) compensated for this enthalpic cost, rendering the process spontaneous (ΔG° < 0) across all temperatures. Na⁺ reduced Q_max by only 9% while divalent Ca²⁺ caused severe suppression of Q_max (49%). The stronger inhibition by Ca²⁺ compared to Na⁺ confirms that functional-group-mediated interactions (electrostatic attraction and hydrogen bonding) constitute a significant fraction of the total adsorption capacity.

Table 5 compares the Q_max of lotus stem biochar for CFX removal against values reported for other lignocellulosic and marine-derived biochars. At neutral pH, lotus stem biochar exhibited a Q_max of ~23.0 mmol kg⁻¹, exceeding the capacities reported for conventional agricultural-waste biochars such as corncob (1.2 mmol kg⁻¹) [42], macadamia nutshell (7.4 mmol kg⁻¹) [8], groundnut shell (6.9 mmol kg⁻¹) [31], cassava peel (10.5 mmol kg⁻¹) [4], and rice straw (11.2 mmol kg⁻¹) [32]. The Q_max of lotus stem biochar was also comparable to that of Cupuaçu bark AC (H₃PO₄, 88%) (18.2 mmol kg⁻¹) and KOH-modified rice straw biochar (26.4 mmol kg⁻¹) [32], and it is noteworthy that the lotus stem biochar was produced without any chemical activation or metal impregnation. Even when benchmarked against brown-algae biochars, which are known for their abundant nitrogen- and oxygen-containing surface functionalities, lotus stem biochar outperformed the 400 °C and 600 °C seaweed biochars (11.4–15.2 mmol kg⁻¹) [30]. The robustness of BC across a wide pH window (Q_max of 20.6, 23.0, and 22.9 mmol kg⁻¹ at pH 3, 7, and 10, respectively) further differentiates it from most reported biochars. These comparisons are for reference only, given significant differences in experimental conditions such as temperature, particle size, and ion concentration. The purpose of this comparison is, therefore, only to demonstrate the potential of lotus stem biochar, and not to assert absolute superiority.

This study investigated the adsorption of CFX onto unmodified biochar derived from lotus stem pyrolyzed at 400 °C. Adsorption kinetics followed the Two-Stage IPD model, with k₁ = 2.775 ± 0.112 mmol kg⁻¹ min^−1/2, and the adsorption isotherms could be described by the Langmuir model, with Q_max reaching 22.96 mmol kg⁻¹ at pH 7 and 26 °C. Solution pH governed adsorption performance through the interplay of CFX speciation and biochar surface charge, and unmodified biochar adsorbed CFX across the entire studied pH range (3, 7, 10). The highest capacity, observed at pH 7, was attributed to electrostatic attraction, hydrogen bonding, and π–π electron donor–acceptor interactions. Thermodynamic analysis revealed that CFX sorption onto BC is a spontaneous (ΔG° < 0), endothermic (ΔH° = +44.02 kJ mol⁻¹), and entropy-driven (ΔS° = +239.66 J mol⁻¹ K⁻¹) process. The presence of divalent Ca²⁺, however, suppressed Q_max through competitive sorption with CFX. Overall, unmodified lotus stem biochar stands out as an affordable adsorbent that performs effectively across a broad pH range (3–10) and under tropical conditions. Nonetheless, while Q_max is comparable across this range, the underlying interactions differ, and long-term stability at extreme pH warrants further study. These findings provide a sustainable pathway for valorizing lotus agricultural residues into biochar for the removal of antibiotics from water.

Conceptualization, Writing—original draft, Writing—review & editing: P.D.T., L.Q.N.K.; Investigation, Formal analysis, Data curation, Writing—original draft, Writing—review & editing: V.N.L.U., V.D.H., D.T.M.K.; Methodology, Software, Visualization, Validation, Resources, Supervision, Project administration, Writing—original draft, Writing—review & editing: B.-T.D.; Writing—original draft, Writing—review & editing: N.L.T.Q., N.T.B., P.T.H. All authors have read and agreed to the published version of the manuscript.

The data supporting the findings of this study are available from the corresponding author upon reasonable request.

The authors declare no conflicts of interest.

The study did not receive any external funding and was conducted using only institutional resources.

The authors acknowledge Ho Chi Minh City University of Technology (HCMUT) and Vietnam National University Ho Chi Minh City (VNU-HCM) for supporting this study.

The authors acknowledge the use of Grammarly^® for language editing to improve the clarity and readability of the manuscript during the preparation of this work. The authors reviewed and edited the content as needed and take full responsibility for the accuracy, originality, integrity, and scientific content of the manuscript.