The optimum pH, solvent concentration, and elution volume for sample extraction were determined using a synthetic solution containing 1 ppm of efavirenz and levonorgestrel. The extraction was performed using an Oasis HLB cartridge. Before extraction, the cartridges were preconditioned by adding 5 mL of 10% methanol and then rinsed with 5 mL of ultra-pure water at a 1 mL/min flow rate (Abafe et al., 2018). The optimization was performed by varying one parameter and keeping others constant. Briefly, the effect of solution pH in the extraction of efavirenz and levonorgestrel was accomplished by changing the solution pH from 2 to 12 using 0.1 M of NaOH and HCl. A volume of 100 mL of the synthetic solution containing 1 ppm of the EFA and LVG was then loaded into the HLB cartridges under vacuum. After extraction, the cartridges were rinsed with 5 mL of 10% Methanol and 5 mL ultra-pure water to remove the untargeted compounds from the surface of the cartridge. Thereafter, the adsorbed analytes were eluted with 6 mL of 80% Methanol. The eluted samples were then dried under nitrogen at 50°C and reconstituted using 1 mL of Methanol. Before analysis, the samples were filtered using nylon syringe filters (0.22 µm). To elucidate the effect of elution solvent and solvent concentration, acetonitrile and Methanol were used at concentrations varying from 50, 80, and 100% to elute efavirenz and levonorgestrel. The above procedure was repeated to evaluate the effect of elution volume, except that the volume of eluent solvent varied from 3, 4, 5, and 6 mL using 100% Methanol. The summary of the method is shown in Figure 2.

LC-20 Prominence HPLC system with a photodiode array detector (Shimadzu, Japan) was used for analysis. The column type, size, and analytical parameters are summarized in Table 1.

Seven-point calibration curves were constructed for EFA and LVG development to determine the linear working range, with concentrations ranging from 0.5 to 30 ppm. The calibration curve was plotted using the linear regression peak area vs. concentrations. The respective standards were analysed in triplicates. The calibration curve’s correlation coefficient (R²) values were >0.98, implying method precision. The limit of detection and quantification were calculated using Equations 1, 2 (Nicolay et al., 2011) to evaluate the linear working range, recoveries, and method sensitivity. The lowest and highest known concentrations (0.5 and 30 ppm) and the blank matrix were analysed to evaluate the method’s accuracy. L O D = 3.3 * σ S (1) where σ is the standard deviation of the calibration curve, and S is the slope of the calibration curve. L O Q = 10 * σ S (2)

The analytes recoveries were calculated using Equation 3: %   R e c o v e r y = S a m p l e   p e a k   a r e a S t a n d a r d   p e a k   a r e a   100 % (3)

The chromatographic variations on the RSD, peak area, resolutions, tailing factor, and theoretical plates of the mixed standard working solution were used to assess the method’s robustness. The validated method was employed to quantify EFA and LVG in environmental samples. The concentration ( C f ) of the efavirenz and levonorgestrel were calculated using the following expression (Equation 4) (Pindihama and Gitari, 2020): C f = M e a s u r e d   c o n c e n t r a t i o n * v o l u m e   o f   e x t r a c t   u s e d     E x t r a c t e d   v o l u m e (4) where, C_f is the final concentration (µg/L), measured concentration (µg/L), volume extracted in (L), extracted volume in (L).

The method was applied for the quantification of EFV and LVG in wastewater collected from Thohoyaṋdou, Malamulele, Giyani, Makhado, Nkowankowa, Tzaneen, Kgapane, and Siloam wastewater treatment plants located in Limpopo Province, South Africa (Supplementary Figure S1). For sample collection, bottles were soaked in 10% Methanol, rinsed with de-ionized water, and oven-dried for 2 h at 80°C to avoid contamination. The samples were kept in ice and transported to the laboratory. In the laboratory, samples were subjected to a solid phase extraction process using optimized conditions at pH 2, 100% Methanol, and an elution volume of 4 mL. The eluted samples were dried under nitrogen flow at 50°C. Thereafter, the samples were reconstituted with 1 mL of 100% Methanol and analysed.

To validate the obtained results from HPLC, the EFA and LVG were analysed using liquid chromatography–quadrupole time-of-flight tandem MS instrument (LCMS-9030 qTOF, Shimadzu Corporation, Kyoto, Japan). Chromatographic separation of the analytes was done using a Shim-pack Velox C18 column (100 × 2.1 mm, 2.7 μm) (Shimadzu Corporation, Kyoto, Japan) maintained at 40°C. An injection volume of 10 μL was used. The analytes were separated using a mobile phase gradient elution method at a flow rate of 0.4 mL min⁻¹. The mobile phase composition of A and B was 0.1% (v/v) formic acid in ultrahigh purity water and acetonitrile, respectively. The mobile phase composition was 5% mobile B from 0 to 1.5 min. At 1.5–4 min, the composition of mobile B was increased to 95% and kept constant until 4 min. The gradient was changed to 5% of mobile phase B at 5 min and maintained at this composition until 6 min. Mass spectral analysis was performed using a QqQ mass spectrometer with an electrospray interface (ESI) in positive (LVG) and negative (EFA) modes. Parameters were set as follows: nebulization 3 L min⁻¹, heating gas, and drying gas flow 10 L/min, interface voltage of 4.0 kV, interface temperature of 300°C, dissolving temperature 526°C, DL temperature of 250°C, heat block temperature of 400°C, detector voltage of 1.8 kV and the flight tube temperature at 42°C.

Figure 5 shows the variation of % recoveries for LVG and EFA with the change in solution pH. The highest percentage recoveries for EFV and LVG were observed at pH of 2 (67.3% and 70.82%), respectively. A decrease in the respective analytes recoveries was observed with increasing pH. LVG and EFA are weak acid compounds that are easily protonated and become soluble at low pH, hence their substantial recoveries at low pH. At alkaline pH, they form strong bonds with HLB carboxylic, divinyl benzene, and N-vinylpyrrolidone, reducing their mobility. Hence, low EFA and LVG recoveries with increasing pH. Based on the obtained results, it was observed that sample pH can significantly affect the recoveries of EFA and LVG. pH of 2 was further used for subsequent experiments.

Figure 6 depicts the effect of solvent type and concentration on the recovery of efavirenz and levonorgestrel. Methanol gave higher recoveries for EFV (71%) and LVG (84.89%) compared to acetonitrile (50.17% and 73.63%). High recoveries for Methanol could be attributed to the introduction of methoxy groups on the surface of the HLB cartridges, enhancing their reactivity (Cortés-Lagunes et al., 2024). Comparatively, methanol is a proactive solvent with high polarity compared to acetonitrile. Thus, the OH functional group dissociate from methanol through hydrogen transfer (Jordaan and Shapi, 2017; Ngwenya and Mahlambi, 2023) and infused into the HLB matrix, improving the solubility of the analytes and enhancing their leaching efficiency. Due to its strong nitrile group, acetonitrile is a non-aprotic solvent that suggests lower polarity than methanol, affecting its reactivity due to its nitrile functional group, which undergoes hydrolysis in acid pH to form carboxylic acid through protonation of nitrogen taking time to infuse the HLB matrix (Zarzycki et al., 2010; Gooßen et al., 2008). An increase in recoveries of the target compounds was observed as the solvent concentration increased for both solvents. This could be attributed to an improved reactivity of the analytes and increased solvent purity (Hu et al., 2009; Jing et al., 2020). Low recoveries at low concentrations might be attributed to reduced reactivity due to increased solvent impurity. Based on the results, 100% of Methanol achieved high EFA and LVG recoveries and was used for subsequent experiments.

Figure 7 shows the effect of elution volume towards efavirenz and levonorgestrel recoveries. EFA and LVG increased from 3.5% to 83% and 1.08% to 94.61% with an increasing volume from 3 mL to 6 mL, respectively. The higher EFA and LVG recoveries were observed with an elution volume of 3–4 mL. This could be attributed to improved dissociation from surface charges, possibly due to increased methoxy ions (Koekemoer, 2016; Sturm, 2005). This phenomenon effectively returns the pollutants to their original state. A slight decrease in the recovery of LVG with increasing elution volume could be due to the degradation of these compounds during prolonged drying. A significant reduction in EFA recoveries was observed when using an elution volume of 5–6 mL, which could be linked to the degradation of EFAs during prolonged drying periods (Yao et al., 2021). Furthermore, since methanol was used as an elution solvent, its increased volume increases the methoxy ions in a solution, which initiates the EFA leaching process from the HLB sorbent. This further increases the solution’s electronegativity, reducing the solute’s leaching activity (EFA) since it mainly comprises electronegative charges, implying less EFA recoveries.