The antibiotics used were Amoxicillin tri-hydrate (AMX) with 95% purity, and Azithromycin (AZT), with 98% purity, both from by Sigma Aldrich (Barcelona, Spain). MilliQ water was used to prepare the AMX solution (at 50 µM), while AZT was prepared at 5 mM in 96% ethanol, and from this a 50 µM concentration was obtained by dilution with MilliQ water. All the reagents used for HPLC analyses (acetonitrile, phosphoric acid and monopotassium phosphate), and the different inorganic salts used, were of high purity analytical grade, supplied by Sigma-Aldrich (Barcelona, Spain). The most relevant physicochemical characteristics of both antibiotics are presented in Table 1.

The equipment used for the quantification of the antibiotics was an UltiMate 3000 HPLC (Thermo Fisher Scientific, United States), with a quaternary pump, an auto-sampler, a thermosstatted column compartment and an UltiMate 3000 series ultraviolet-visible detector. This equipment contains a computer with version 7 of the Chromeleon software (Thermo Fisher Scientific, United States). Chromatographic separations were performed on a Luna C18 analytical column (150 mm long; 4.6 mm internal diameter; 5 µm particle size) from Phenomenex (Madrid, Spain), and a security column (4 mm long; 3 mm internal diameter; 5 µm particle size), packed with the same material as the column.

As the quantification limits were 0.01 µM for AMX and 0.5 µM for AZT, the injection volumes were lower for the former (50 µL), while it was 200 µL for AZT. The flow rates were 1.5 ml min⁻¹ for AMX and 1 ml min⁻¹ for AZT. The temperature was kept constant at 25 ± 1°C throughout the analysis. Finally, it should be noted that between each measurement, the syringe was washed with a methanol:water solution (ratio 50:50).

The conditions used to achieve the separation of the antibiotic AZT were the following: the mobile phase was formed by acetonitrile (phase A), and by monopotassium phosphate 0.025 M (phase B). The linear gradient elution program was executed from 5 to 70% for phase A and from 95 to 30% for phase B with a time of 18 min. The initial conditions were restored in 2 min and held for 3 min. The total time for analysis was 25 min, with a retention time of 9.7 min, and the wavelength used for detection was 210 nm.

AMX was separated under the following conditions: the mobile phase was acetonitrile (phase A) as for AZT, but with phase B being 0.01 M phosphoric acid (at pH = 2). The linear gradient elution program was executed from 5 to 15% for phase A and from 95 to 85% for phase B, within 4 min. The initial conditions were restored in 2 min. The total analysis time was 10 min, with a retention time of 3.3 min, and the wavelength used was 230 nm.

Additionally, in the aqueous solutions in which photo-degradation was studied, the total organic carbon (TOC) was determined at times of 0, 2, 16, and 72 h of exposure to light, to determine the degree of antibiotic mineralization. However, it was performed just for AMX, as AZT had been prepared in ethanol, as indicated above. Specifically, the quantification of TOC was carried out by means of a Multi N/C2100 equipment (Analytikjena, Jena, Germany).

The experimental data were fitted to an exponential decay model (Eq. 1) C C 0 = e − k t (1) where C/C₀ is the fraction of the initial concentration C₀ that remains in the suspension after a given time t (expressed in min), and k (min⁻¹) is the kinetic dissipation constant. The half-life (t_1/2) of each compound was calculated as per Eq. 2: t 1 2 = ln ⁡ 2 k (2)

Figure 1 shows the dissipation results for AMX and AZT at different pH values, both under simulated sunlight and in the dark. For AMX, there was no dissipation in the dark at the three pHs studied, keeping the value of the ratio C/C₀ close to 1 (Figures 1A–C), while significant degradation occurred under simulated sunlight, although without showing a clear trend as a function of pH (Figures 1A–C), reaching C/C₀ values close to 0 after 40 h of contact/exposure.

The behavior of AZT in dark conditions was similar to that of AMX, keeping the C/C₀ values close to 1 for all the contact times considered (Figures 1D–F). However, the degradation under simulated sunlight was much lower than that of AMX, in spite of using longer times of exposure to light (Figures 1D–F), with C/C₀ values always above 0.5. In this case, the degradation was higher at pH 4.0 than at pH 5.5 and 7.2 (Figures 1D–F).

The exponential decay model used to calculate half-lives (Eq. 1) proved to be useful for this type of estimation, with R² scoring between 0.913 and 0.999 (Table 2). For AMX, the kinetic constants (k) did not show clear variation when pH went from 4.0 to 7.2, with just a slight increase at pH 7.2, with the half-life following the same trend, going from 8.8 h (at pH 4.0) to 7.7 h (at pH 7.2), and reaching a value of 9.9 h at the intermediate pH of 5.5 (Table 2).

Furthermore, changing the pH between 4 and 7.2 does not produce significant modifications in the AMX adsorption spectra (Figure 2), an aspect that could also influence its degradation, but it does not seem to be relevant in this case.

Figure 3 presents TOC data corresponding to the experiments with AMX under simulated sunlight at different pHs and at different times of exposition to light. With 2 and 16 h of exposure, TOC does not decrease significantly, despite the fact that the concentration of AMX does reduce by more than 50% at the three pH values studied (Figures 1A–C). After 72 h of simulated sunlight, AMX was below the quantification limit, but the TOC measurements only decreased between 30 and 35%, without significant differences among the three pH values studied (Figure 3).

The degradation of AZT was different from that of AMX, with much lower kinetic constants (ranging between 0.0016 and 0.0028 h⁻¹) and much longer half-lives (between 250 and 456 h), the lowest values corresponding to pH 4.0. This indicates that AZT suffers higher degradation at acidic pH (4.0) (Table 2), with a minimum degradation at pH 5.5, which corresponds to a longer half-life (Table 2).

The effects of the presence of the inorganic salts NaCl, CaCl₂, NaNO₃, Na₂HPO₄, and FeCl₃, and of soluble organic matter (humic acids) on the degradation of the two antibiotics studied, both in the dark and under simulated sunlight, are presented in Figures 4, 5.

Figure 4 shows the data corresponding to AMX. The results indicate that the degradation of AMX is dependent on exposition to light, as well as on the matrix, and on the characteristics of the antibiotic. In the dark, no effect on the degradation of AMX was found for the different inorganic salts studied, except for FeCl₃ at its highest concentration (500 mg L⁻¹), which causes a 100% dissipation of AMX; however, under simulated sunlight, AMX degradation was greatly increased in the presence of NaCl and NaNO₃ (Figures 1A,C). The inorganic salt NaCl had an effect favoring the dissipation of AMX, with C/C₀ passing from a value of 0.61 in the sample without NaCl to 0.34 in the presence of 500 mg L⁻¹ of NaCl, which means a decrease of 44% for this antibiotic (Figure 4A). The presence of NaNO₃ caused the dissipation of practically 100% of AMX for the highest concentration of the salt used (500 mg L⁻¹) (Figure 4C).

The effects of CaCl₂ and Na₂HPO₄ under simulated sunlight were slightly lower, with C/C₀ values changing from 0.60 to 0.48, and from 0.65 to 0.57, respectively, which represents 21% and 12% dissipation, respectively (Figures 4B,D). As was the case in the dark, the presence of FeCl₃ made AMX degradation very high (Figure 1E).

The effect of the presence of inorganic salts on the dissipation of AZT is shown in Figure 5. In the dark, no significant effect was observed, while under simulated sunlight the salts that had the greatest effect on increasing the degradation of AZT were NaCl, (with C/C₀ changing from 0.49 to 0.24, which means a dissipation of 51% for the antibiotic) and CaCl₂ (with C/C₀ changing from 0.52 to 0.27, which means 47% of dissipation) (Figures 5A,B); NaNO₃, Na₂HPO₄ and FeCl₃ did not cause increases in AZT degradation (Figures 5C–E). The fundamental difference of AZT with respect to AMX is the lack of effect of FeCl₃, and the effect of NaCl and NaNO₃ favoring the degradation for AZT, with respect to AMX (Figures 4, 5).

Finally, the presence of dissolved organic matter (humic acids) had very scarce effect on AMX degradation, both in the dark and under simulated sunlight (Figure 4F). There is a slight dissipation taking place both in the dark and under simulated sunlight (4%–19% of increase in dissipation) that can be related to the adsorption of AMX onto humic acids. However, for AZT the results are similar to those of AMX in the dark (6% of increase in dissipation), whereas under simulated sunlight the dissipation is very intense, reaching 68% (Figure 5F).

Regarding previous publications, there are few data on the photo-degradation of AMX in aqueous solutions without the intervention of a photo-catalyst, but the trend is that the degradation increases as the pH rises, as has been pointed out in studies using TiO₂ nanoparticles (Moosavi and Tavakoli, 2016; Li et al., 2019), or ZnO (Elmolla and Chaudhuri, 2010). However, there are also other studies indicating that changing pH does not affect AMX degradation, such as that of Dimitrakopoulou et al. (2012), who reported that the change in pH from 5 to 7.5 had not effect on AMX degradation, but it did slow down the mineralization of the antibiotic. This increase in degradation as the pH rises is usually related to the different sensitivity of the various species present, which is dependent on pH. Specifically, the zwitterion and negatively charged forms grow as the pH increases (Figure 2), which is determined by their pK_a values. The effect of increasing photo-degradation with rising pH has also been highlighted for other antibiotics, such as tetracyclines (Conde-Cid et al., 2018a), sulfonamides (Canonica et al., 2008; Conde-Cid et al., 2018b) and other families of antibiotics (Canonica et al., 2008; Rodríguez-López et al., 2021; Álvarez-Esmorís et al., 2022).

Regarding TOC analyses, the results obtained indicate that the mineralization of AMX is not complete and that intermediate molecules are formed through different degradation pathways, such as hydroxylation, hydrolysis and decarboxylation (Li et al., 2019; Zhang et al., 2019).

The behavior of presenting a maximum stability at pH between 5.0 and 6.5, has already been described for AZT by other authors (Farghaly and Mohamed, 2004; Zhang et al., 2009). In addition, Čizmič et al. (2019) described a lack of effect on the degradation of AZT due to varying pH between 3.0 and 7.0, but they did observe an increase in the degradation of AZT when pH reached 10. The greater or minor degree of photo-degradation in this case is not related to different ionic forms of the molecule, since AZT has a predominantly positive charge in the pH range studied in the current work, due to its very high pK_a values (Table 1; Figure 2D). Although it was not possible to verify this fact with TOC measurements, numerous intermediate degradation products have also been detected for AZT, generating different fragments due to modifications in the macrolide ring (Tong et al., 2011).

The effect of Fe favoring the degradation of AMX has already been described by several authors, both through the Fenton reagent (Ay and Kargi, 2010) and through Fe particles (Ghauch et al., 2009), the fundamental mechanism being the formation of OH radicals that favor the degradation of the antibiotic (Buitrago et al., 2020).

Comparing with previous publications, the results regarding the effects of different water matrices on the degradation of the antibiotics are contradictory, which can be attributed to the different experimental conditions and the difference in sensitivity of the antibiotics. In this sense, Canonica et al. (2008) indicate that the effects of photosensitizers on the degradation of antibiotics depended on the type of antibiotic and the light source. Different results are also described even within the same family of antibiotics, as is the case of the various responses due to the presence of dissolved organic matter and nitrates on the degradation of different cephalosporin molecules (Jiang et al., 2010). In general, the presence of nitrates in water has been related to an increase in the photolysis of AZT (Tang et al., 2019). In addition, an increase in ionic strength was related to a higher rate of degradation of AZT (Zhang et al., 2009). Other authors found an inhibitory effect due to the presence of anions such as Cl⁻, HCO₃ ⁻, NO₃ ⁻, and SO₄ ^2-, which has been observed for AMX (Zhang et al., 2019), while some other authors did not find neither inhibition nor promotion of degradation (Mavronikola et al., 2009).

Regarding the effect of humic acids on the dissipation of antibiotics, it probably depends on the content of these organic substances, since at low concentrations they can work as promoters of the photosensitivity of antibiotics, while at high concentrations they can act as a preventer of photo-degradation by increasing the turbidity of the liquid medium (Thiele-Bruhn, 2003; Xuan et al., 2010). This is one of the reasons why contradictory results are found regarding its effect on the degradation of antibiotic molecules. As example, an effect of increasing the dissipation of sulfonamides has been previously reported (Conde-Cid et al., 2018b), but their presence caused no effect on the dissipation of tetracycline (Chen et al., 2008).

Among the main facts to remark from the current study, it should be noted that the antibiotics AMX and AZT did not show significant dissipation in the dark. AMX presented a half-life under simulated sunlight that ranged between 7.7 and 8.7 h, with the lowest value associated to the highest pH in the range investigated (pH 7.2). On the other hand, AZT presented a higher half-life, with values that ranged between 250 and 456 h, showing the highest value for an intermediate pH (5.5). AMX completely dissipated after 50 h of simulated sunlight, but the evaluation of the TOC values indicates that the mineralization ranged between 30% and 35%, which corresponds to the formation of intermediate compounds. However, AZT did not completely dissipate even for the longest contact time used (216 h) for the three pH values tested (4.0, 5.5, and 7.2). The presence of NaCl had a favorable effect on the dissipation of both antibiotics, with an increase in degradation of 44% and 51% for AMX and AZT, respectively. The presence of CaCl₂ increased the dissipation of AMX by 21%, and 47% for AZT. NaNO₃ caused the total dissipation of AMX when this salt was present at high concentrations, while it had no effect on AZT. The presence of Na₂HPO₄ had no effect on the dissipation of the two antibiotics studied, while FeCl₃ drastically increased AMX dissipation both in the dark and under simulated sunlight, and had no effect on AZT. The presence of humic acids only slightly increased the dissipation of AMX, both in the dark (17%) and under simulated sunlight (4%), having a similar effect for AZT in the dark (6% increase in dissipation), and being more pronounced under simulated sunlight, with an increase in dissipation of 68% for the highest concentration of these substances (10 mg L⁻¹). These results can be considered relevant to understand and predict the fate of both antibiotics when they reach environmental compartments, and could help to define appropriate conditions to achieve the inactivation/degradation/removal of these emerging pollutants.