Targeted removal of enrofloxacin by 1-hexadecyl-3-methylimidazolium-chloride-modified attapulgite: a recyclable green adsorbent

2.1. Materials

ATP was obtained from Xuyi Sinoma Attapulgite Clay Co., Ltd. Enrofloxacin hydrochloride (C₁₉H₂₂FN₃O₃·HCl, 98%) and ciprofloxacin hydrochloride (C₁₇H₁₈FN₃O₃·HCl, 88.5%) were purchased from Shanghai Yuanye Bio-Technology Co., Ltd. 1-Hexadecyl-3-methylimidazolium chloride monohydrate (C₂₀H₄₁ClN₂O, 99%) was obtained from Shanghai Macklin Biochemical Co., Ltd. Hydrochloric acid (HCl, 36.0–38.0%) and sodium hydroxide (NaOH, >96.0%) were supplied by Nanjing Chemical Reagent Co., Ltd. Tert-butanol (C₄H₁₀O,≥99.5%) was purchased from Sinopharm Chemical Reagent Co., Ltd. All chemicals were of analytical grade and used without further purification.

2.2. Preparation of IL-ATP

Prior to modification, Attapulgite was acid-activated to remove impurities blocking its pore channels. A given mass of ATP was treated with 100 mL of 1.0 mol/L HCl under magnetic stirring for 1 h, followed by centrifugation and repeated washing with deionized water until the supernatant was neutral. The product was dried, cooled, and ground for further use. Acid-activated ATP was then dispersed in 50 mL deionized water, followed by the addition of 2–10 wt% C₁₆mimCl (relative to ATP mass). The mixture was stirred at room temperature for 12 h, centrifuged, washed, dried, and ground to obtain IL-ATP.

2.3. Characterization

Surface morphology was observed by field-emission scanning electron microscopy (SEM; SU8020, Hitachi, Japan). Elemental composition was determined by energy-dispersive X-ray spectroscopy (EDS; EMAX, Bruker, Germany). Specific surface area and pore size distribution were measured using N₂ adsorption-desorption (Micromeritics ASAP 2020, USA). Surface charge properties were analyzed by zeta potential measurement (Zetasizer Nano ZS, Malvern, UK). Crystal structures were determined by X-ray diffraction (XRD; PANalytical Empyrean, Netherlands). Chemical states and elemental composition were analyzed by X-ray photoelectron spectroscopy (XPS; PHI 5700 ESCA, USA). Functional groups were identified by Fourier-transform infrared spectroscopy (FTIR; Frontier, PerkinElmer, USA).

2.4. Adsorption experiments

Batch adsorption experiments were conducted in 50 mL centrifuge tubes to evaluate ENR removal under varying parameters: IL-ATP dosage (5-40 mg), pH (2–12), contact time (0–180 min), initial ENR concentration (0.5–50 mg/L), and temperature (15–35 °C). Each experiment was performed in triplicate with blanks. After equilibrium, suspensions were centrifuged at 4200 rpm for 5 min, and the supernatant absorbance at 271 nm was measured using a UV–Vis spectrophotometer (Agilent 8453, USA) [15]. Removal efficiency was then calculated.

2.5. Regeneration experiments

Ultrasonic regeneration experiments were performed using a custom-built setup (Figure 1a) consisting of an ultrasonic transducer rod (diameter: 60 mm, frequency: 40 kHz), a 2000 mL beaker, and a magnetic stirrer. Saturated IL-ATP (160 mg) was mixed with 400 mL deionized water and stirred while being subjected to ultrasound. Regeneration experiments were carried out under varying conditions, including regeneration time (5-25 min), ultrasonic power (60-100 W), and regeneration solution pH (2-10). After regeneration, IL-ATP was recovered and reused in subsequent adsorption tests to calculate regeneration efficiency.

Close

Figure 1. (a) Recovery device. Photographs of (b) pristine ATP and (c) IL-ATP.

Export to PPT

2.6. Data analysis

Equilibrium adsorption capacity (Q_e, mg/g) and removal efficiency (E,%) calculated using (Eq.1) and (Eq. 2), respectively:

where C₀ and Cₑ (mg/L) are initial and equilibrium ENR concentrations, V (L) is solution volume, and m (g) is adsorbent mass.

Regeneration efficiency (RE,%) was calculated using (Eq. 3):

where Q_d (mg/g) is the adsorption capacity after regeneration.

3.1. Characterization of materials

Figures 1(b and c) shows the photographs of pristine ATP (left) and IL-ATP (right). Although the macroscopic powder morphology exhibits only subtle differences (e.g., slight variations in color or degree of compactness), more distinct structural alterations can be uncovered via advanced characterization methods. Therefore, the morphology and elemental composition were systematically analyzed using SEM and EDS, as illustrated in Figure 2. For pristine ATP (Figure 2a), a typical loose and fractured microstructure is observed, characterized by irregularly shaped particles, numerous interparticle voids, and a relatively rough surface texture. In contrast, the morphology of IL-ATP (Figure 2b) exhibits pronounced changes after modification: the surface is smoother and denser, the pore structure is significantly reduced, and particle agglomeration is enhanced. These morphological transformations suggest that C₁₆mimCl was successfully deposited on the ATP surface. This can be attributed to the strong electrostatic interactions and hydrophobic associations between C₁₆mimCl and ATP, thereby effectively altering the surface characteristics of the adsorbent.

Close

Figure 2. SEM images of (a) ATP and (b) IL-ATP and (c) element mapping results for IL-ATP.

Export to PPT

Moreover, the EDS mapping results for IL-ATP (Figure 2c) provide further confirmation of the successful incorporation of C₁₆mimCl. The elemental distribution maps reveal the presence of C, N, and Cl—characteristic elements of C₁₆mimCl—uniformly dispersed across the ATP matrix, in addition to the intrinsic elements of ATP, namely O, Si, and Al. This uniform distribution not only confirms successful chemical modification but also indicates the stability and homogeneity of the ionic liquid on the ATP framework, and this plays a critical role in enhancing its adsorption performance.

The N₂ adsorption–desorption isotherms and pore size distribution curves for ATP and IL-ATP are shown in Figures 3 (a and b). Pristine ATP exhibits a typical type IV isotherm with a pronounced hysteresis loop in the relative pressure range 0.4–1.0, which is a characteristic feature of mesoporous materials. In contrast, the N₂ adsorption behavior of IL-ATP shows remarkable changes: the overall adsorption uptake decreases and the area of the hysteresis loop is reduced, indicating that the pore structure is regulated by ionic-liquid modification [16].

Close

Figure 3. (a) N₂ adsorption-desorption isotherms of ATP and IL-ATP, (b) Pore size distribution curves of ATP and IL-ATP, (c) FTIR spectra of ATP and IL-ATP, (d) Full scan spectrum, (e) C 1s spectrum, (f) N 1s spectrum, (g) XRD patterns of ATP and IL-ATP.

Export to PPT

As summarized in Table S1, ATP possesses superior pore structural parameters, and together with the concentrated mesopore peak in the range of 2–10 nm observed in the pore size distribution curve, it can be confirmed that ATP has abundant mesoporous surfaces and pore storage space. However, after modification, the S_BET of IL-ATP decreases drastically to 52.75 m²/g (only 43.1% that of ATP), the total pore volume (V_T) decreases to 0.121 cm³/g, while the average pore diameter (D_A) increases to 9.17 nm. This can be attributed to the introduction of C₁₆mimCl molecules to the mesoporous channels of ATP via hydrophobic interactions and ion exchange. The long alkyl chains preferentially occupy smaller pores (<10 nm), leading to a significant reduction in specific surface area and pore volume. Meanwhile, the residual pores exhibit an increased average pore diameter, which is consistent with the reduced proportion of small pores observed in the pore size distribution curve of IL-ATP. Such structural adjustments demonstrate that ionic-liquid modification effectively tailors the pore characteristics of ATP, thereby altering its adsorption performance.

The structural and chemical changes of ATP and IL-ATP were investigated by FTIR (Figure 3c). For pristine ATP, characteristic absorption bands were observed: a broad O–H stretching vibration around 3400 cm⁻¹ (attributed to surface hydroxyl groups and adsorbed water), the Al–OH bending vibration near 1600 cm⁻¹, and Si–O–Si framework vibrations in the region of 1000–500 cm⁻¹.

Upon modification with C₁₆mimCl, remarkable changes are detected. The weakened O–H absorption band (around 3400 cm⁻¹) indicates that partial surface hydroxyl groups of ATP are involved in hydrogen-bonding interactions with the imidazolium ring of C₁₆mimCl (rather than simple physical coverage, as supported by the simultaneous emergence of a new functional group peak). Meanwhile, a new weak band around 2900 cm⁻¹, corresponding to the –CH₂– stretching vibration of the alkyl chain in C₁₆mimCl, confirms successful grafting of the ionic liquid onto the ATP surface.

Figures 3(d-f) shows high-resolution XPS spectra. The wide-scan XPS spectra (Figure 3d) reveals the inherent mineral components of ATP, showing characteristic Mg 1s, O 1s, Ca 2p, C 1s, Si 2p, and Si 2s peaks. In contrast, IL-ATP exhibits a distinct N 1s peak, directly confirming the successful incorporation of nitrogen-containing hexadecyl-3-methylimidazolium cations, consistent with the chemical structure of the modifier.

The high-resolution C 1s spectra of IL-ATP (Figure 3e) show an enhanced C–C peak from the C₁₆mimCl alkyl chain and a newly observed C–N peak from the imidazolium ring, confirming successful grafting of the organic modifier. The N 1s spectra (Figure 3f) exhibit two distinct peaks at 402.5 eV (N⁺–C₁₆) and 400.5 eV (N⁺–CH₃), characteristic of quaternized imidazolium nitrogen species. These positively charged N⁺ sites, together with the hydrophobic alkyl chains, act as active centers for both electrostatic and hydrophobic interactions with ENR, while the weakened CO₃²⁻ signal reflects surface coverage by the modifier. These results indicate that IL-ATP possesses both chemical and electrostatic features favorable for ENR adsorption.

The XRD analysis results (Figure 3g) also show that the structural integrity of ATP is maintained during modification. The pristine sample exhibits reflections typical of fibrous silicate, along with impurity peaks corresponding to quartz (SiO₂, PDF#97-003-9830) and dolomite (CaMg(CO₃)₂, PDF#97-018-5049). After functionalization with C₁₆mimCl (IL-ATP), the major diffraction peaks are observed at nearly identical 2θ positions, but their intensities are attenuated and the peaks slightly broadened, confirming that the crystalline framework is preserved and no new inorganic phases are generated [17].

This diminished peak intensity is mainly attributed to the coverage of fiber surfaces and grooves by an X-ray–amorphous or poorly ordered ionic-liquid overlayer, which reduces the effective diffracting volume and highlights the relative contribution of impurity phases. Considering the chain–layered channel architecture of ATP, steric hindrance prevents regular intercalation of C₁₆mim⁺; hence, modification primarily occurs on the external surface, where electrostatic and hydrogen-bonding interactions yield a positively charged, hydrophobic interface.

3.2. Adsorption of ENR onto IL-ATP

3.2.1. Effect of operational conditions on adsorption performance

The adsorption behavior of ENR onto IL-ATP is influenced by multiple factors. To elucidate these effects, the removal efficiency of ENR was systematically investigated under varying conditions, including adsorbent dosage, solution pH, and the presence of coexisting ions. These experiments aimed to clarify how environmental parameters modulate the adsorption performance of IL-ATP toward ENR.

The effect of modifier concentration on adsorption performance is shown in Figure 4(a). When the concentration of C₁₆mimCl is in the range 2%–6%, the equilibrium adsorption capacity (Qₑ) remains stable at 50 mg/g, with removal efficiencies exceeding 90%. However, when the concentration exceeds 6%, excess ionic liquid molecules stack on the adsorbent surface, hindering mass transfer and reducing the removal efficiency to 65%, thereby validating the “optimal coverage” theory, i.e., moderate alkyl-chain functionalization can construct a hydrophobic microenvironment to enhance adsorption, whereas excessive modification induces steric hindrance. Therefore, in subsequent experiments, the modifier concentration was fixed at 6%.

Close

Figure 4. Influence of (a) modifier concentration on adsorption capacity and removal rate, (b) adsorbent dosage, (c) initial solution pH, (d) zeta potential of IL-ATP, and (e) influence of coexisting ions.

Export to PPT

As shown in Figure 4(b), with increasing adsorbent dosage from 5 to 40 mg, Qₑ decreases sharply from 110 to 22 mg/g, while removal efficiency initially increases and then plateaus. This indicates that although increasing dosage dilutes the pollutant loading per active site, excessive adsorbent addition induces particle aggregation, thereby reducing the effective utilization of adsorption sites [18]. Considering both adsorption efficiency and material economy, a dosage of 20 mg was selected for subsequent experiments.

As shown in Figures 4 (c and d), pH affects the adsorption performance by regulating the zeta potential of IL-ATP, the charge/speciation of the contaminant ENR, and the competitive effect of OH⁻ in solution. At pH < 4, the positively charged adsorbent surface electrostatically attracted the negatively charged ENR species under acidic conditions, thereby promoting adsorption. At pH = 6, non-electrostatic interactions (hydrophobic forces, hydrogen bonding, etc.) become dominant, resulting in the highest Q_e (46 mg/g) and removal efficiency (85%). When pH > 6, the zeta potential of the adsorbent is shifted negatively (enhanced negative charge), causing electrostatic repulsion with the negatively charged groups of ENR, while OH⁻ competes for adsorption sites and the dissociation state of ENR changes. These combined effects weaken adsorption, leading to a decline in performance.

The influence of coexisting ions on ENR adsorption was further examined (Figure 4e). Multivalent cations, especially Fe³⁺, significantly suppress removal efficiency through strong charge neutralization and site competition, with removal dropping to 25% at only 0.5 mM. Monovalent cations (Na⁺, K⁺) and divalent cations (Ca²⁺, Mg²⁺) also inhibit adsorption as their concentrations increase, with stronger suppression at higher valence and ionic strength. In contrast, the effects of anions such as SO-2 4 and CO-2 3 are relatively minor, with removal efficiency remaining relatively high, even at 5 mM. The cations primarily inhibit adsorption of clay-based adsorbents due to electrostatic interactions and site competition, whereas the effect of anions is less pronounced and more dependent on the chemical environment.

3.2.2. Kinetic and thermodynamic studies of the adsorption process

The adsorption kinetics were fitted using pseudo-first-order (PFO) and pseudo-second-order (PSO) kinetic models in their linear and nonlinear forms [19], as described by (Eq. 4) - (Eq. 7).

(4)

$\text{PFO model in nonlinear form:}\text{Qt = Qe×(1}-{\text{e}}^{-{\text{K}}_{\text{1}}\text{t}}\text{)}$

(5)

$\text{PSO model in nonlinear forms:}{\text{Q}}_{\text{e}}= \frac{{\text{K}}_{\text{2}}{\text{Q}}_{\text{e}}^{\text{2}}\text{t}}{{\text{1+K}}_{\text{2}}{\text{Q}}_{\text{e}}^{\text{2}}\text{t}}$

(6)

$\text{PFO model in linear form:}\ln (Qe-Qt) = \ln Qe-K1t$

(7)

$\text{PSO model in linear forms:}\frac{\text{t}}{{\text{Q}}_{\text{t}}}= \frac{\text{1}}{{\text{K}}_{\text{2}}{\text{Q}}_{\text{e}}^{\text{2}}}+\frac{\text{t}}{{\text{Q}}_{\text{e}}}$

Adsorption quantity at time t (min) is denoted by Q_t (mg/g), and the adsorption rate constants for the PFO and PSO models are K₁ (g/(mg·min)) and K₂ (g/(mg·min)), respectively.

As illustrated in Figure 5(a), adsorption shows a rapid initial uptake within the first 0–40 min, during which Q_e sharply increases from 0 to 50 mg/g. This stage is attributed to the abundant availability of active surface sites. The subsequent slowdown until equilibrium at 40–60 min reflects increasing diffusion resistance as the active sites become progressively occupied.

Close

Figure 5. (a) Influence of different times on adsorption capacity, (b) pseudo-first-order kinetics, and (c) pseudo-second-order kinetics plot. Influence of adsorption temperature, adsorption isotherms simulated with (d) Langmuir, (e) Freundlich models, adsorption kinetics of ENR by IL -ATP.

Export to PPT

Both the PFO and PSO models capture the general trend (Figures 5b and c), but PSO exhibits much better consistency with experimental data. According to Table 1, the PSO model yields an equilibrium adsorption capacity (Q_e,cal = 51.38 mg/g) very close to the experimental value (49.99 mg/g). Moreover, the correlation coefficient of the PSO model (R² = 0.9968) is markedly higher than that for the PFO model (R² = 0.7878), confirming the superior fitting performance of PSO. Such consistency between calculated and experimental values enhances confidence in the kinetic analysis. These findings demonstrate that the adsorption kinetics of ENR onto IL-ATP are best described by the PSO model, suggesting a significant contribution of chemisorption, potentially involving valence forces or electron sharing, while diffusion resistance also plays a role at the later stage of adsorption [20].

Table 1. Parameters of the kinetic model for ENR removal by IL -ATP.

Model	Parameter	Data
PFO	Qe,cal(mg/g)	54.7607
	K1	0.0032
	R2	0.9523
PSO	Qe,cal(mg/g)	51.3687
	K2	0.1022
	R2	0.9466
Linear-PFO	Qe,cal(mg/g)	19.3387
	K1	0.0519
	R2	0.7878
Linear-PSO	Qe,cal(mg/g)	51.3875
	K2	0.0746
	R2	0.9968

Adsorption isotherms were fitted using the Langmuir and Freundlich models in their linear and nonlinear forms [21], as described by (Eq. 8) - (Eq. 11).

(8)

$\text{Langmuir model in nonlinear form:} {\text{Q}}_{\text{e}}\text{ = }\frac{{\text{Q}}_{\text{m}}\times {\text{K}}_{\text{L}}\times {\text{C}}_{\text{e}}}{\text{1}+{\text{K}}_{\text{L}}\times {\text{C}}_{\text{e}}}$

(9)

$\text{Freundlich model in nonlinear form:}{\text{Q}}_{\text{e}}{\text{= K}}_{\text{F}}{\text{×C}}_{\text{e}}^{\text{1/n}}$

(10)

$\text{Langmuir model in linear form:}\frac{\text{1}}{\text{Qe}}=\frac{\text{1}}{\text{Qm}}+\frac{\text{1}}{{\text{K}}_{\text{L}}{\text{Q}}_{\text{m}}}\times \frac{\text{1}}{{\text{C}}_{\text{e}}}$

(11)

$\text{Freundlich model in linear form:}\ln Qe=\frac{1}{n} \times \ln Ce+\ln KF$

where Q_m (mg/g) is the maximal adsorption capacity, and K_L (L/mg) is the Langmuir constant for adsorption free energy. K_F ((mg/g)/(mg/L)1/n) indicates relative adsorption capability, while 1/n indicates intensity.

The fitting results of the two models are shown in Figures 5 (d and e) and Table 2. Both models exhibit good fitting performance, with the Freundlich model yielding slightly higher correlation coefficients (R² > 0.9905) than the Langmuir model (R² > 0.9896) across all tested temperatures, though both models provided satisfactory descriptions of the adsorption behavior. The theoretical maximum monolayer adsorption capacity (Q_m,cal​​) from the Langmuir model decreases from 457.89 mg/g at 289 K to 393.55 mg/g at 309 K, suggesting that adsorption is more favorable at lower temperatures. While the Langmuir model describes monolayer adsorption onto relatively homogeneous surface sites, the Freundlich constant 1/n (1.16–1.18) exceeds the typically favorable range (0.1–1.0), indicating a notable degree of surface heterogeneity and partially unfavorable adsorption. These results imply that both homogeneous monolayer adsorption and heterogeneous interactions contribute to the binding between ENR and IL-ATP [22]. Furthermore, comparison with literature data (Table 3) [23-26] demonstrates that IL-ATP exhibits superior adsorption performance relative to those of most reported adsorbents.

Table 2. Parameters of the isotherm model for ENR removal by IL-ATP.

Model	Parameters	289	299	309
Langmuir	Qm,cal (mg g−1)	457.8869	415.8192	393.5521
	KL (L mg−1)	0.0049	0.0055	00.0061
	R2	0.9929	0.9896	0.9934
Freundlich	KF ((mg g−1)·(L mg−1)1/ n)	3.1526	3.2845	3.4088
	1/n	1.1652	1.1795	1.1831
	R2	0.9940	0.9905	0.9941

Table 3. Comparison of adsorption amounts of various adsorbents for ENR reported in the previous publications.

Material	Qm (mg/g)	SBET (m2/g)	References
AmCOF-SO3H	628.9	8.79	[ 23]
N-PBC	78.43	265.41	[ 24]
AH-coMMt	55.4	175.28	[ 25]
MG-PPAC	322.58	513.34	[ 26]
IL-ATP	415.82	52.75	This work

The thermodynamics were analyzed in terms of standard change in Gibbs free energy ($\Delta \text{G}$), standard change in enthalpy ($\Delta \text{H}$) and standard change in entropy ($\Delta \text{S}$)m, as determined by (Eq. 12 and Eq. 13) [27].

(12)

$\Delta G = -RT\ln K = \Delta H-T\Delta S$

(13)

$\ln K=\frac{\Delta S}{R}-\frac{\Delta H}{RT}$

where K is the constant of the best isotherm model (L/mg), T is the temperature (K), and R is the universal gas constant (8.314 J∙ mol^-1∙ K^-1).

The thermodynamic parameters of ENR adsorption onto IL-ATP are summarized in Table 4

Table 4. Thermodynamic parameters of IL-ATP adsorption on ENR.

Sample	T(K)	ΔG(KJ·mol-1)	ΔH(KJ·mol-1)	ΔS(KJ·mol-1·K-1)
IL-ATP	288	-6.30446	17.13345	0.081786
	298	-7.48833
	309	-7.92344

. Negative ΔG values at all tested temperatures confirm the spontaneity of the adsorption, with increasing spontaneity at higher temperatures (ΔG from −6.30 kJ·mol⁻¹ at 288 K to −7.92 kJ·mol⁻¹ at 309 K). The positive ΔH value (17.13 kJ·mol⁻¹) indicates that the process is endothermic, while the positive ΔS (81.8 J·mol⁻¹·K⁻¹) suggests increased disorder at the solid–liquid interface, likely due to the release of water molecules during ENR adsorption. Although the process is endothermic, the experimental decrease in Q_m,cal with rising temperature suggests that higher temperatures weaken specific ENR–adsorbent interactions (e.g., hydrogen bonding and electrostatic attractions), thereby reducing overall uptake. Nevertheless, the positive entropy contribution ensure spontaneity under all tested conditions.

3.3 Regeneration of saturated IL-ATP

3.3.1. Effect of different conditions on regeneration performance

The regeneration efficiency and mechanism of saturated IL-ATP were systematically investigated under various ultrasonic regeneration conditions, including regeneration time, ultrasonic power, and solution pH.

Figure 6(a) illustrates the relationship between regeneration time (5–25 min) and IL-ATP regeneration efficiency. The results show that regeneration efficiency remains consistently high (95.51%–96.96%) as time is extended from 5 to 25 min. This indicates that, with the assistance of the custom-made ultrasonic transducer, effective regeneration of the adsorbent can be achieved quickly, while excessive prolongation of time did not result in significant improvement. This behavior is closely related to mass-transfer enhancement in the ultrasonic field, where cavitation effects promote the desorption of adsorbates within a certain time window [28]. Once dynamic equilibrium is reached, further extension of regeneration time yields limited benefits.

Close

Figure 6. Influence of different conditions on ultrasonic effects: (a) ultrasonic time, (b) ultrasonic power, (c) pH of the regenerative solution.

Export to PPT

The effect of ultrasonic power on regeneration efficiency is shown in Figure 6(b). When the regeneration power increases from 60 to 100 W, the regeneration efficiency shows a slight increasing trend while maintaining high stability (96.17%–97.01%). Ultrasonic power directly determines the intensity of cavitation: higher power intensifies microbubble collapse, generating stronger shear forces, local high temperatures, and pressures. At 60 W, cavitation effects can already break binding forces between ENR and IL-ATP, ensuring effective desorption. With increasing power (60–100 W), both the number and energy of cavitation bubbles rise, further enhancing the disruption of ENR-IL-ATP interactions. Notably, the lack of a sharp RE(regeneration efficiency) decline at 100 W implies this power range does not induce structural damage to IL-ATP, as such damage would otherwise cause a significant RE drop. Thus, values within this range balance cavitation efficiency and adsorbent integrity.

In addition, Figure 6(c) demonstrates that IL-ATP exhibits excellent regeneration adaptability over a wide pH range (2–10), with regeneration efficiency consistently above 96%. This stability stems from C₁₆mimCl-mediated surface modification: unmodified ATP exhibits pH-dependent surface charge (positively charged under acidity, negatively charged under alkalinity), which strongly influences ENR adsorption/desorption (as ENR is zwitterionic). In contrast, C₁₆mimCl introduces a dense imidazolium cation layer on IL-ATP, shielding the adsorbent’s intrinsic pH-sensitive surface charge and stabilizing hydrophobic adsorption sites. Under ultrasonic agitation, ENR-IL-ATP interactions are uniformly disrupted in both acidic and alkaline environments, maintaining high desorption efficiency across the tested pH range. This wide pH tolerance is critical for practical applications, as natural water matrices often exhibit pH fluctuations.

3.3.2. Mechanism of ultrasonic regeneration of saturated IL-ATP

3.3.2.1. Fundamental characterization of the adsorbent before and after regeneration

The structural and chemical properties of ENR-saturated IL-ATP (ENR/IL-ATP) and regenerated IL-ATP (Re-IL-ATP) were characterized via SEM, N₂ physisorption, and FTIR, as shown in Figure 7.

Close

Figure 7. SEM images of (a) ENR/IL-ATP and (b) Re-IL-ATP, as well as (c) N₂ adsorption-desorption isotherms, (d) pore size distribution curves, and (e) FTIR spectra.

Export to PPT

SEM images (Figures 7a and b) illustrate morphological evolution. ENR/IL-ATP exhibits a relatively dense surface with obscured porous features, suggesting that pores and active sites are occupied by adsorbed ENR molecules. In contrast, Re-IL-ATP shows a more porous and fibrous morphology, similar to that of pristine C₁₆mimCl-modified ATP. This morphological change indicates that ultrasonic regeneration removes adsorbed ENR and facilitates the recovery of IL-ATP’s open porous structure.

N₂ adsorption–desorption isotherms (Figure 7c) and pore size distribution curves (Figure 7d) further quantify the textural properties. ENR/IL-ATP shows a lower N₂ adsorption capacity than that of Re-IL-ATP, indicating a reduction of specific surface area and pore volume due to pore blocking by ENR. In contrast, the isotherm of Re-IL-ATP exhibits a steeper hysteresis loop, and its pore size distribution shows a higher pore volume, suggesting that ultrasonic regeneration reopens the clogged pores. Moreover, Re-IL-ATP retains the mesoporous characteristic (dominant pore diameter within 2–50 nm) of IL-ATP, while ENR/IL-ATP displays a narrowed distribution and suppressed pore volume due to ENR filling.

The FTIR spectra (Figure 7e) shed light on the chemical interactions and structural preservation during regeneration. For ENR-saturated IL-ATP (ENR/IL-ATP), characteristic peaks of IL-ATP, including those for O–H (3400–3640 cm⁻¹, stretching vibration of hydroxyl groups), –CH₂– (2850-2940 cm⁻¹, stretching vibration of methylene groups from C₁₆mimCl), and Si–O–Si (1030 cm⁻¹, stretching vibration of siloxane bonds in the ATP framework), remain observable. The H–O–H bending vibration (1630 cm⁻¹, from adsorbed water) shows subtle intensity changes, which may arise from hydrogen-bonding interactions with ENR.

In contrast, regenerated IL-ATP (Re-IL-ATP) exhibits spectral features almost identical to those of pristine IL-ATP: these characteristic peaks match well in both shape and intensity, and the H–O–H bending vibration also reverts to the original profile.

These observations indicate that ultrasonic regeneration effectively desorbs ENR while largely preserving the functional groups of IL-ATP, thereby maintaining its structural integrity for reuse.

3.3.2.2. Reverse verification of cavitation inhibition and regeneration mechanisms based on tert-butanol concentration

Figure 8 shows the relationship curve between tert-butanol (TBA) concentration and regeneration efficiency. At 0 mM, the regeneration efficiency is 95.68%, whereas at 2 mM, it is significantly decreased to 80.72%. Essentially, TBA inhibits the ultrasonic cavitation effect through cavitation energy consumption and radical-mediated desorption blockage, and this inhibition process reveals the core mechanism of ultrasonic regeneration.

Close

Figure 8. Influence of TBA concentration on regeneration efficiency.

Export to PPT

(1) TBA consumption of cavitation energy verifies the physical–mechanical effects of ultrasonication

TBA molecules possess strong diffusion ability and can rapidly accumulate at the gas–liquid interface of cavitation bubbles. When cavitation bubbles collapse, TBA molecules, due to their lower C–H bond energy, preferentially participate in pyrolysis reactions, which consume a large amount of energy released during cavitation [29]. This significantly weakens the physical-impact forces (shock waves and microjets) generated by bubble collapse. As a result, the physical stripping of ENR molecules is directly reduced, making it difficult for them to detach from the adsorbent surface. This suggests that the primary physical mechanism of ultrasonic regeneration involves microjets and shock waves generated by cavitation bubble collapse, which scour and break apart particle aggregates, thereby stripping ENR from surfaces and pores.

(2) TBA scavenging of free radicals verifies the chemical desorption effects of ultrasonication

During ultrasonic cavitation, the collapse of bubbles generates localized high temperature and high pressure [30], inducing water pyrolysis and producing abundant hydroxyl radicals (·OH), as shown in (Eq. 14):

(14)

${\text{H}}_{\text{2}}\text{O }\to \cdot \text{OH+H}\cdot$

OH is the key chemical species for breaking the bonds between ENR and the adsorbent. ·OH can oxidize the hydrophobic rings of ENR (such as the C=C bonds of the quinolone ring), weakening the hydrophobic interactions between ENR and the C₁₆mimCl alkyl chains. Furthermore, ·OH can attack the Si–OH groups on the adsorbent surface, disrupting hydrogen bonding between ENR amino groups and Si–OH. However, TBA rapidly “captures” ·OH through the following reaction, as presented in (Eq. 15):

(15)

$\cdot \text{OH+}{\left({\text{CH}}_{\text{3}}\right)}_{\text{3}}\text{COH}\to {\left({\text{CH}}_{\text{3}}\right)}_{\text{3}}\text{CO}\cdot {\text{+H}}_{\text{2}}\text{O}$

The generated (CH₃)₃CO· radical exhibits extremely low activity and cannot participate in the breaking of adsorption bonds. As the TBA concentration increases, the concentration of ·OH available for bond disruption decreases sharply, blocking the chemical desorption pathway. This suggests that an important chemical mechanism of ultrasonic regeneration is that the ·OH generated by cavitation oxidizes and disrupts the hydrogen bonds and hydrophobic interactions between ENR and the adsorbent, promoting ENR desorption.

Through the inhibitory effect of TBA, it can be clearly concluded that the core mechanism of ultrasonic regeneration of IL-ATP involves both physical–mechanical action and chemical desorption (Figure 9). Thus, the entire ultrasonic regeneration process is a dual-driven mechanism of physical and chemical actions, rather than being guided by a single thermal effect. This also clarifies the critical role of cavitation in ultrasonic regeneration.

Close

Figure 9. Ultrasound regeneration mechanism diagram.

Export to PPT

3.4. Application potential of regenerated materials

Figure 10(a) illustrates the relationship between the number of regeneration cycles (1–5) and regeneration efficiency. The initial regeneration efficiency is 95.68% and gradually decreases with increasing regeneration cycles, dropping to 75.01% after the fifth cycle. This decline can be attributed to irreversible loss of active adsorption sites during repeated ultrasonic desorption cycles, as well as the accumulation of residual adsorbates that hinder the desorption process. Such a trend is consistent with the typical regeneration behavior of adsorbents.

Close

Figure 10. (a) Regeneration efficiency after five cycles of regeneration, (b) Comparison of the adsorbent regeneration performance in this study with that of previously reported adsorbents.

Export to PPT

As shown in Figure 10(b) (red asterisk), the regenerated IL-ATP prepared in this study exhibits not only good regeneration performance over multiple cycles, but also distinct practical advantages[31-35]. Other adsorbents such as ZIF-8@MPS-HC and LDH@NA-SBA require complex syntheses of MOFs or hybrid nanostructures, with high raw material costs. In contrast, our adsorbent is derived from naturally abundant ATP modified by simple ionic-liquid grafting, which is easier and more cost-effective. Therefore, IL-ATP demonstrates superior competitiveness in terms of regeneration stability and economic feasibility, offering an efficient strategy for the removal of fluoroquinolone antibiotics such as ENR.