Insights into the Roles of Surface Functional Groups and Micropores in the Sorption of Ofloxacin on Banana Pseudo-Stem Biochars

The sorption kinetics of OFL on BS biochars before and after de-ashing treatment were investigated at the initial concentration of 50 mg L⁻¹. As shown in Figure 6, the sorption of OFL on all investigated biochars (except OBS5) increased fast at an initial stage, and then leveled off when the sorption approached the equilibration after 50 h. Three typical kinetic models, namely, PFOM, PSOM, and Two-PFOM, were employed to describe these curves, with the regression curves plotted in Figure 6 and fitting results listed in Table 2. PFOM and PSOM can afford a satisfying fitting to the OFL sorption on OBS3, WBS3, and WBS5, indicated by the high r²_adj values (>0.978). However, neither PFOM nor PSOM could precisely describe the kinetic data of OFL sorption on OBS4 and WBS4 because of the underestimation of the slow sorption at longer time points. By contrast, Two-PFOM, among the three models, exhibited the best fitting performance for all the kinetics (except WBS5), as evidenced by the r²_adj values (0.979–1.000). In addition, the experimental sorption capacity q_e,exp and the predicted q_e,cal at equilibrium based on Two-PFOM are almost identical, implying that Two-PFOM may provide more reliable fitting parameters compared with the other two models. More importantly, Two-PFOM could separately evaluate the sorption rate constant of the fast and slow compartments and simultaneously estimate the contribution of two compartments to the total sorption. Therefore, the present data were discussed based on the fitting results of Two-PFOM.

The two compartments could be distinguished by two varied sorption rate constants (k_fast and k_slow). As shown in Table 2, the k_fast and k_slow values of OFL sorption substantially decreased when the pyrolytic temperature was elevated from 300 °C to 400 °C, for both the original and de-ashed biochars, probably due to the decrease in the number of sorption sites on the surface of biochars prepared at 400 °C. According to the above characterizations, the oxygen content, O/C, as well as the surface oxygen functional groups on biochars remarkably decreased at temperatures above 300 °C, implying that the surface oxygen functional groups of biochar may play a key role in the sorption of OFL. As the pyrolytic temperature further increased to 500 °C, the k_fast and k_slow values of OFL sorption on WBS5 increased up to 15.43 and 0.108 h⁻¹, respectively, which could be attributed to the increased SSA and V_t of WBS5 as analyzed above. Based on the kinetics data shown in Figure 6 and fitted equilibria q_e,cal in Table 2, the sorption capacity of OFL significantly decreased with the rise in pyrolysis temperature for both OBS and WBS biochars, again revealing that the surface oxygen functional groups may be crucial for OFL sorption, which will be discussed extensively in the follow-up sections.

Unlike the other kinetic curves, the OFL sorption on BS5 exhibited a unique multi-stage process. Specifically, the sorption of OFL increased rapidly in an initial 24 h, followed by a remarkable decline in a later 48 h, and then reached re-equilibrium after 72 h. Therefore, the above three kinetic models failed to describe these data. In Zuo’s study [42], they reported a similar multi-stage process regarding the sorption between dibutyl phthalate (DBP) and chicken-feather-derived biochars, and attributed this phenomenon to the strong inter-molecular interactions of DBP. In the present study, a similar speculation could be proposed. Specifically, a strong negative charge-assisted hydrogen bond ((−)CAHB) may form between two OFL⁻ molecules under an alkaline condition (pH = 9.4 for OBS5 system). Thus, the rapid sorption of OFL at an initial stage was attributed to inter-molecular interactions of OFL, which may shield the negative charges of OFL and facilitate their sorption. However, the OFL sorption decreased when it reached a new equilibrium with the interaction time. Extensive discussions about this unique phenomenon are out of the scope of the present work, which will be further investigated in another work of ours.

The sorption isotherms of OFL on OBS biochars (Figure 7a) and WBS biochars (Figure 7b) which varied with the pyrolysis temperature were investigated. FM and LM were selected to fit the sorption isotherms, and the fitting results are shown in Table 3. In general, both models showed a satisfying fitting performance with r²_adj values mostly in the range of 0.90–1.00. However, FM was more appropriate for describing the actual sorption process of OFL on BS biochars, such as BS3 and WBS4 (shown in Figure 7), implying a multilayer sorption process [43,44] and an energetic preference of OFL for the heterogeneous sorption sites of BS biochars. The N values obtained from the FM were generally below 0.563, suggesting that the sorption energy of OFL on BS biochar was highly heterogeneous. In addition, for both OBS and WBS biochars, the N values significantly decreased with the rise in pyrolysis temperature, indicating increased sorption nonlinearity after pyrolysis [34]. This was attributed to the irregular pores and rigid aromatic structures formed at higher pyrolysis temperatures, which provided heterogeneous sorption sites for OFL.

As summarized in Table S1, the maximum adsorption capacity is in the range of 0.31–218.29 mg g⁻¹. The sorption capacity of OFL on BS biochars in the present study is comparable to most results in the literature. However, the sorption capacity cannot be directly compared due to different solute concentrations or sorption conditions used in these studies. In addition, the K_F values calculated by FM could also not be directly compared because the unit of K_F is determined by the nonlinearity of the isotherm. Therefore, the single-point sorption coefficient, K_d, was calculated at C_e = 0.001 C_s and 0.01 C_s (shown in Table 3) using the fitting parameters of FM to further study the sorption mechanisms of OFL. From the sorption isotherms shown in Figure 7 and the calculated K_d values listed in Table 3, it can be found that the sorption of OFL on BS biochars significantly decreased with increasing pyrolysis temperature either before (Figure 7a) or after (Figure 7b) water washing treatment, consistent with the above kinetic results, further confirming that the surface oxygen-containing groups of biochars played a vital role in OFL sorption. At C_e = ~3–5 mg L⁻¹, the K_d value of OFL sorption by WBS3 (22,880.1 L kg⁻¹) was at least ~5 times that by Ulva prolifera-derived biochar (4743 L kg⁻¹) [45] and ~31–497 times that by cassava-residue-derived biochar (46–732.49 L kg⁻¹) [22], indicating the excellent sorption performance of WBS3 to OFL. As previously discussed in the literature [30,46,47], various mechanisms could be involved in OFL sorption, such as electronic interaction, cation exchange, the hydrophobic effect, the pore-filling effect, π–π EDA interaction, ordinary hydrogen bond (OHB), and (−)CAHB, which will be discussed in detail in the follow-up section.

Previous studies indicated that the sorption of OFL on sorbents with various functional groups (e.g., biochar, activated carbon, or carbon nanotubes) is highly pH-dependent [24,27,30,48]. In the present study, the biochar surface is negatively charged due to the deprotonation of various oxygen functional groups (e.g., –COOH and –OH) at the current pH levels (7.22–9.4). In addition, having two pK_aS (pK_a1 = 6.10, pK_a2 = 8.28), the distribution of three species (OFL⁺, OFL^±, and OFL⁻) of OFL also varied with different pH conditions. As shown in Figure 7c, 82.3% of OFL existed as OFL^± in the sorption system of OBS3 (pH = 7.5), while 93.4 and 91.8% of OFL presented as OFL⁻ in those of OBS4 (pH = 9.4) and OBS5 (pH = 9.5), respectively. Hence, the OFL sorption was facilitated by the electronic attraction (EA) between the positive charge of OFL^± and the negative charge on the OBS3 surface, while it could be restricted by the electronic repulsion between OFL⁻ and negatively charged surface of OBS4 and OBS5. This is probably why the sorption of OFL by OBS3 was much greater than those by OBS4 and OBS5. Moreover, water washing treatment removed most potassium salts in OBS biochars, leading to a visibly declined solution pH in the WBS4 (pH = 7.4) and WBS5 (pH = 7.2) systems, and, thus, an increased fraction of OFL^± (84.9 and 83.8% in the WBS4 and WBS5 systems, respectively) (Figure 7c), thereby improving OFL sorption. Nevertheless, electronic interaction only partly explained the OFL sorption because OFL sorption markedly decreased with pyrolysis when the species distribution of OFL was identical in each sorption system of WBS biochars (Figure 7c).

The cation exchange between OFL and oxygen functional groups may be involved in OFL sorption (Equation (7)) and it could decrease with pyrolysis owing to the decreased acidic functional group. It was suggested that the cation exchange of cationic OFL was much greater than that of amphoteric OFL [45,49,50]. Thus, the contribution of cation exchange to the differential sorption behaviors of OFL on the investigated sorbents could be minimal because OFL⁺ only accounted for 3.83–6.21% of overall OFL concentrations in the sorption systems. Moreover, the phenomenon of H⁺ releasing after OFL sorption as indicated in Equation (7) was not observed in any batch sorption system of the current study, again confirming the above speculation.

Furthermore, the above analysis assumed that increased SSA and V_t either with pyrolysis or after water washing treatment might provide more sorption sites for OFL. However, the order of OFL sorption was just reversed to the order of the SSA and V_t of biochars with pyrolysis, and the K_d value (C_e = 0.001 C_s) of OFL sorption by WBS3 was nearly 30 times that by WBS5, indicating that SSA and pore-filling were not the dominant factors determining the OFL sorption on BS biochars. This speculation was further supported by the results shown in Figure 7d,e, where the K_d values of OFL decreased with pyrolysis even more significantly after SSA normalization. Previous studies [51,52,53] suggested that the target molecules can effectively penetrate the pores only when the pore diameter is 1.7 times larger than the molecule’s second-widest dimension. The calculated molecular size of OFL is 1.43 nm × 8.4 nm [54], so the pores with a diameter smaller than 1.43 nm would be inaccessible for OFL. Although the SSA of WBS5 was up to 390 m² g⁻¹, some micropores smaller than 1.43 nm, especially the inner pores of WBS5, may not be accessed by OFL due to the size exclusion effect.

π–π EDA was usually involved in the interaction between ionic compounds and biochar because both of them have various functional groups [5,19]. OFL could be identified as a strong π-acceptor due to the strong electron-withdrawing ability of the fluorine on the benzene ring [46], while the biochars with electron-donating groups (e.g., –OH and –NH₂) can be a π-donor. With the rise in pyrolysis temperature, decreased –OH and –NH₂ on biochars weakened the π–π EDA interaction between OFL and biochars, leading to decreased OFL sorption. Although the increased polyaromatic structure at high temperatures may provide rich π electrons for the interaction with OFL [55], the inner surface cannot be accessed by OFL because of the size exclusion effect mentioned above.

The formation of hydrogen bonds between the functional groups of OFL and the oxygen-containing groups of biochar may facilitate the sorption of OFL on biochar. It is worth noting that the carboxyl groups of OFL molecules were mostly deprotonated at the current pH condition in the systems of OBS and WBS biochars. Thus, OHB with low-energy berries could not be involved in the carboxyl group of OFL. Instead, (−)CAHB may play a crucial role in OFL sorption [46,47]. Previous studies [47] indicated that, compared with OHB, (−)CAHB was a stronger hydrogen bond, which is generated when the hydrogen bond donor and hydrogen bond acceptor have similar pK_a values (|Δ pK_a| = |pK_a, hydrogen bond donor − pK_a, hydrogen bond acceptor| K_a| approaches 0. According to the previous study, the pK_aS of –COOH on biochars could cover a wide pH range of 2–7, and the pK_aS of –OH could be around 9–11 [56]. Thus, the two pK_aS of OFL might be located in these pH regions. Thus, (−)CAHB may be involved in the sorption of OFL by biochars (Equations (8)–(11)). As discussed earlier in this work, surface oxygen-containing functional groups (e.g., –COOH and –OH) significantly decreased with pyrolysis, which could weaken the (−)CAHB between the OFL and biochar, and, thus, OFL sorption. Additionally, as shown in Figure 7f, an elevated pH was observed in the systems of WBS4 and WBS5 after the sorption of OFL, further confirming the formation of (−)CAHB (Equations (8) and (10)). Although the (−)CAHB between OFL and WBS3 could be stronger than the other systems, the buffer effect of abundant functional groups of WBS3 may counterbalance the pH variation caused by (−)CAHB. Similarly, the pH variation caused by (−)CAHB could not be seen in the OBS biochar systems because the hydroxide ions were consumed by potassium bicarbonate (Equation (12)) or other potassium salts.

As discussed above, the OFL sorption was significantly affected by the solution pH owing to the varied OFL speciation and surface charge of biochars caused by pH variation. To further elucidate the effect of pH on the sorption of OFL by BS biochars, we selected two biochars, WBS3 (with abundant functional groups) and WBS5 (with a developed micropore structure), and simultaneously investigated the role of the surface functional groups and micropore structure in OFL sorption under different pH condition. The pH levels were set to obtain different distributions of three OFL species. Generally, a significant reduction was observed for OFL sorption by both WBS3 and WBS5 with the pH increase from 6.0 to 9.0. According to the calculation shown in Figure 8b, with the rise in pH, the fraction of OFL⁺ declined from 54.5% to 0.02%; that of OFL^± first increased to 85.0% at pH 7, and then decreased to 15.1% at pH 9; and that of OFL⁻ gradually increased from 0.25% to 84.9%. Based on the above discussion, EA, cation exchange, π–π EDA interaction, OHB, and (−)CAHB were involved in the sorption of OFL by WBS3 and WBS5. These proposed interactions, except for (−)CAHB, all decreased with an elevated pH, ultimately leading to a decreased OFL sorption. In addition, as can be seen from the sorption isotherms of OFL at different pH levels (Figure 8a,c) and K_d values obtained from the FM fitting (Figure 8d), the sorption of OFL by WBS3 was extraordinarily higher than that by WBS5 at each investigated pH level, revealing that surface functional groups other than micropores governed the sorption of OFL. Furthermore, even when the OFL⁻ molecules accounted for 85.0% at pH 9, the K_d values of OFL sorption on WBS3 were 9 and 19 times those on WBS5 at a low concentration and high concentration, respectively, suggesting that (−)CAHB was the dominant sorption mechanism for OFL by WBS3 at an alkaline condition, which could overcome the electron repulsion between OFL⁻ and negatively charged surface of biochars. These results demonstrated that, compared with the micropore structure, the surface functional groups were more important for the OFL sorption on BS biochars.

Although Huang et al. concluded that the pore-filling effect played a major role in OFL sorption, their present data were not sufficient to support this opinion. For example, they did not provide the O content of the investigated biochar and pH condition of the batch sorption experiments [22]. Both of them are essential for OFL sorption. As indicated by the characterization [61], the ash content of cassava-residue-derived biochar made at 750 °C (CW750, the same sorbent used in Huang et al.’s study) was as high as 30.56%. However, the investigators did not analyze the inorganic composition of CW biochars and discuss their influence on OFL sorption [22]. Therefore, their conclusions are not convincing. Furthermore, many studies found the importance of the surface functional groups of biochar in OFL sorption [1,23,24], again confirming our findings and conclusions.

As discussed earlier, OBS biochars are abundant in K-containing salts, mainly including KCl, KHCO₃, and a small amount of K₂C₂O₄. The above relevant results indicated that these inorganic compositions hindered OFL sorption by covering the sorption sites or blocking the inner pores of OBS biochars, as well as releasing OH^– into solution. Moreover, the K⁺ and anions (Cl⁻ and HCO₃⁻) should be taken into consideration due to their co-existing with OFL in the sorption system, which may affect the sorption of OFL. According to the quantification (Figure 9), the concentration of K⁺ in the solution of OBS biochar systems increased from 266 to 310 mg L⁻¹ with the pyrolysis temperature, due to the accumulation of K salts at higher temperatures. After water washing treatment, the concentration of K⁺ decreased to 2.34 mg L⁻¹, indicating that minimal K⁺ co-exists with OFL in the systems of WBS biochars. Herein, two major K-containing salts KCl and KHCO₃ with varied concentrations (300, 600, and 1000 mg L⁻¹ quantified depending on K⁺ concentrations) were added to the systems of WBS biochars to investigate their influences on OFL sorption.

The impact of ions on FQs has been previously discussed. Huang et al. suggested that no significant effect on OFL sorption was observed for the K⁺ at a concentration of 0.01 M (390 mg L⁻¹) which is comparable with the concentration detected in the systems of OBS biochars. As can be seen from Figure 10, the addition of KCl and KHCO₃ arouses an adverse trend in the pH variation. Specifically, with the increasing amount of KCl, a general decline in the solution pH was observed, which was attributed to the cation exchange between K⁺ in the solution and H⁺ released from the acidic functional groups of biochars. Clearly, with an increasing pyrolysis temperature, the extent of reduction in the pH became weaker because the acidic groups or cation exchange capacity of biochars decreased with pyrolysis. The addition of KCl slightly increased OFL sorption by WBS3, which was caused by the decreasing solution pH. However, the addition of KCl to WBS4 and WBS5 caused little change to their sorption affinity to OFL, consistent with the results reported by Huang et al. [22]. He et al. also suggested that K⁺ had a relatively small effect on OFL sorption due to its negligible competition for the sorption sites [62]. In contrast, the solution pH generally increased with the increasing addition of KHCO₃, owing to the hydrolysis of KHCO₃, and, thus, released OH⁻ into the solution. Thus, an obvious decline in OFL sorption by WBS3 and WBS4 was caused by an elevated solution pH. However, the addition of KHCO₃ did not significantly affect the sorption of OFL on WBS5. A previous study also indicated that the hydrolysis of carbonate slats could produce a large amount of OH⁻ in the solution [63], thereby inhibiting the ciprofloxacin sorption. These findings suggested that the co-existing K⁺ did not significantly affect OFL sorption on BS biochars, while HCO₃⁻ may increase the solution pH, and thus decrease the OFL sorption.