Preparation of biosorbent from avocado seeds for preconcentration and simultaneous extraction of trace parabens in environmental wastewater samples Authors: Aluwani Sharon Nephiphidi, Rabelani Mudzielwana, Shirley Kholofelo Selahle Published: 2025 ABSTRACT The identification of emerging contaminants, including parabens, in water resources and aquatic ecosystems has become an issue of environmental concern due to the associated human and ecological risks, emphasizing the urgent need for effective monitoring and remediation strategies. This study aimed to synthesize a low-cost magnetic activated carbon biosorbent from avocado seeds, achieving higher extraction efficiencies and paving the way for more sustainable water treatment methods. The Dispersive Magnetic Solid-Phase Microextraction (DMSPME) technique was used for the preconcentration and simultaneous extraction of parabens (methylParaben, ethylParaben, and butylParaben). Influential DMSPME parameters were optimized using a Fractional factorial design for screening and a central composite design for further analysis. Characterization results have proved that the desired biosorbent was successfully synthesized. The magnetic activated carbon showed a well-defined surface with exhaustive characterization of Brunauer-Emmett-Teller (BET) revealed surface areas of 30 m2/g and a pore volume of 0.1 cm3/g. The exhaustive characterization of the Scanning Electron Microscope (SEM) image revealed that the incorporation of iron oxide into the activated carbon resulted in the formation of small, well-defined pores. The Energy Dispersive X-Ray (EDS) indicated the existence of carbon, oxygen, iron, sodium, and potassium. The Fourier-Transform Infrared Spectroscopy (FTIR) analysis revealed the presence of several functional groups, including carboxylic acid, iron oxide, and aromatic compounds. This characterization confirmed that Magnetic activated carbon prepared with 30% KOH was the best biosorbent in extracting the targeted parabens. The parameters of interest were biosorbent amount, sample volume, type of elution solvent, pH, and volume of extraction solvent. The optimum parameters obtained were 22.5 mg of 30% KOH-magnetic carbon activated, 6.5 pH, and 875 µL methanol as elution solvent. These conditions lead to obtaining good relative recoveries of 80.0%–101%. Limit of detection (LOD) and Limit of quantification (LOQ) of the method were 0.12–0.19 μg/L and 0.25–0.49 μg/L, respectively. The reusability and regeneration adsorption percentage recoveries significantly decreased after the fifth cycle. Parabens extraction using the 30% KOH- magnetic activated carbon was considered a capable biosorbent for the extraction of parabens because it demonstrated a strong potential for extracting parabens pollutants simultaneously in wastewater samples. Introduction Water is an essential resource for economic activity, environmental balance, and human survival. Water is required in various industries, the agricultural sector, for drinking, and to support ecosystems. However, public health and water security are at risk due to the growing number of pollutants contaminating water bodies. Parabens, or para-hydroxybenzoates, are one of these contaminants that have drawn the most attention due to their frequent detection in aquatic environments (Derisso et al., 2020). They find their way into water bodies via household wastewater, including improper disposal of goods and industrial discharge from manufacturing operations (Vale et al., 2022). Parabens are created synthetically by esterifying p-hydroxybenzoic acid (pHBA) with an alcohol in the presence of a catalyst (Morelli et al., 2020). These compounds share a similar chemical structure, featuring a benzene ring, a hydroxyl group, and an ester group at the para position on the ring. However, they vary in their substituents, which can be either an aromatic ring or an alkyl chain. The most utilized parabens are: methylparaben (MethylP), ethylparaben (EthylP), propylparaben (PropylP), and butylparaben (ButylP) (Sousa et al., 2020). Parabens are commonly used as preservatives in food products, pharmaceuticals, and cosmetics. Due to the widespread use of parabens as the most primarily used chemical, they have been found in water bodies in concentrations varying from ng/L to µg/L (Wei et al., 2021). Parabens concentrations vary depending on location, water type, and wastewater treatment efficiency. Supplementary Table S1 lists the various concentrations obtained from several countries. In Africa, parabens were detected in South Africa and Nigeria (Leung et al., 2024; Mpayipheli et al., 2024). In both countries, they were detected in wastewater and surface water, with concentrations ranging from non-detectable to 84.7 μg/L and non-detectable to 0.52 mg/L, respectively. In Asia, the concentrations of China’s surface water and India Lake water were reported to be 8.38 ng/L and 0.0182 mg/L (Kachhawaha et al., 2021; Zhao et al., 2024). In Europe, parabens have been detected in Switzerland’s Glatt River at higher concentrations (Lincho et al., 2021). The global concentration of parabens suggests that traces of these compounds are found worldwide, making the contamination of water bodies by parabens a global concern. The most stressful problem is that parabens can be active in very low concentrations and can interact with any other organic compounds present in the water bodies (Derisso et al., 2020), and there is no prediction of what might happen after the interaction. Al-Qudaihi et al. (2025) demonstrated that the Acceptable Daily Intake (ADI) for combined intake of MethylP and EthylP, as assisted by the European Food Safety Authority (EFSA) and JECFA, is 0–10 mg/kg. As for Butyl, it was not accepted as a food additive or any other consumable additive according to the Scientific Committee on Consumer Safety (SCCS) (Bernauer et al., 2023). From the European Union, it could only be used in cosmetic products with a concentration not exceeding 0.028% in the final product because it was found not to be safe for children (0–10 years) when combined with other chemicals (Gaffet et al., 2025). Due to their endocrine-disrupting properties, parabens are associated with several adverse health effects. They can mimic estrogen and disrupt hormonal balance, which may have an impact on puberty, fertility, and reproductive development. Reduced sperm quality in men and early puberty or reproductive problems in women and children have been related to elevated levels of chemicals such as butylparaben and methylparaben. Concerns regarding parabens’ possible involvement in hormone-related cancers have been raised by their detection in human tissues, including tumors (Liang et al., 2023). They can also result in allergic reactions and skin irritation, especially in people with sensitive or damaged skin (Le et al., 2022). To mitigate these risks, adsorption-based techniques have proven to be a promising approach for extracting parabens, thereby enhancing their detection and facilitating water treatment efforts. Agricultural waste, such as nutshells (Narloch et al., 2024), plum shells (Moreno-Marenco et al., 2022), grapefruit peels (Molina-Balmaceda et al., 2025), and cork pellets (Morelli et al., 2020), has proven to be a promising source for producing carbon absorbents for extracting parabens in aqueous solutions. Avocado seeds have also been tested as an adsorbent for organic contaminants from water. Its effectiveness in adsorbing various pollutants is attributed to the bioactive compounds it contains. These bioactive substances include procyanidins, phenolic compounds, triterpenoids, acetogenins, fatty acids, amines, and weak acids, among others (García-Vargas et al., 2020). These Avocado seeds have also proven to be a great source of activated carbon for various applications, such as the adsorption of gases and the removal of contaminants in the environment (Ahmad and Danish, 2022). Siemak and coworkers (Siemak and Michalkiewicz, 2024) demonstrated that avocado seed activated carbon (ASAC) has a high specific surface area and is efficient in extracting organic compounds such as ibuprofen (Molina-Balmaceda et al., 2024), dyes (Munyegaju, 2022), and Phenols (Kaga et al., 2024). In addition, Riadi et al. further supported this, showing ASAC is effective for extracting dyes from water (Riadi et al., 2023). These studies collectively demonstrate ASAC’s efficacy as an adsorbent for various emerging pollutants in water. The research focuses on the development of magnetic activated carbon (MAC) as an improved version of traditional activated carbon, which often faces challenges with complex separation techniques (Zakaria et al., 2023). By incorporating magnetic properties, MAC enhances efficiency and simplifies the recovery process from treated water, thereby improving the material’s reusability. MAC is created by coating activated carbon with iron salts, such as FeCl3 and FeSO4, followed by the addition of a base to precipitate magnetic iron oxide. This study also examines the innovative use of MAC derived from avocado seeds to extract and concentrate parabens (methylP, ethylP, and butylP) from wastewater. The valorization of avocado seed waste not only transforms it into a valuable resource, promoting recycling and reusing, but also contributes to the circular economy by optimizing the use of natural resources for water treatment (Okuthe, 2024). The low-cost, eco-friendly material offers higher extraction efficiencies of the targeted analytes, earning it the designation of biosorbent and highlighting the uniqueness of the research. While previous studies have utilized avocado seed waste for carbon-based absorbents, few have focused on modified carbon for extracting parabens from wastewater. This study employs Dispersive magnetic solid-phase micro extraction (DMSPME) as an advanced method for extracting contaminants. Using a magnetic bio sorbent for the simultaneous extraction of parabens builds upon previous research. The effectiveness of this technique is underscored by high-performance liquid chromatography with a photodiode array detector (HPLC-PDA), which confirms high recovery rates and demonstrates the practical applicability of the developed DMSPME method for environmental monitoring. Experimental Chemicals and reagents Potassium hydroxide (KOH 98%), ferrous sulphate (FeSO4·7H2O 98%), sodium hydroxide (NaOH>98%), Ferric chloride (FeCl3.6H2O 28%), and hydrochloric acid (HCl >95%) were all purchased from Merck. HPLC-grade solvents, Acetonitrile and methanol, were also purchased from Merck. MethylP (>99%), ButylP (>99%), and EthylP (>99%) analytical standards were also purchased from Merck. Methanol solvent was utilized to prepare a model solution of 1,000 mg/L of MethylP, ButylP, and EthylP. The model solution was stored in the refrigerator at a temperature of 4 °C. Dilutions of the working solutions were prepared using ultra-pure water from the Direct-Q 3UV-R purification system. The parameters of ultrapure water were the resistivity of 18.2 MΩ cm at 25 °C and the electrical conductivity of 0.056 μS/cm. The pH was around 6.6. Prior to analysis, samples were filtered with 0.22 µm polytetrafluoroethylene filters purchased from Separations Scientific SA (Pty) Ltd. Avocado seeds were gathered from the villages surrounding Thohoyandou, Limpopo, South Africa. Instrumentation The bonding structures and key functional groups of the synthesized magnetic activated carbon biosorbent were investigated using a PerkinElmer Spectrum 100 FTIR spectrometer (PerkinElmer, United States). The functional groups of the biosorbents were analyzed using Fourier Transform Infrared (FTIR) spectroscopy in the range of 4,000–500 cm-1. The elemental composition, nanostructure, and morphology of the biosorbents were determined using scanning electron microscopy (SEM) and energy-dispersive X-ray spectroscopy (EDS) with a TESCAN VEGA 3 XMU instrument (LMH, Czech Republic). An X’Pert XRD (PANalytical BV, Netherlands) equipped with a Cu Ka (k = 0.15406 nm) radiation was used to study the crystallinity of the synthesized magnetic activated carbon. Nitrogen adsorption–desorption isotherms (Autosorb 6200 EPDN XR-AG-AG, Micrometrics Instrument Corporation, Norcross, United States) were utilized to analyze how pore diameters and the specific surface area of the bioadsorbent at a temperature of 77 K. Prior to the analysis, the samples were degassed for 5 h at 300 °C. A sonication bath (Eins-Sci Ultrasonic Cleaner, South Africa) operating at a normal setting of 25 °C was used to perform a dispersing operation on solutions containing analytical components. The sample separation process occurred through a Hermle Z366 centrifuge (Hermle Labortechnik GmbH, Germany) to separate any remaining fine particles from the sample solution, ensuring it is clean and safe for HPLC analysis. Chromatography conditions A Shimadzu HPLC (Prominence-i LC-2030C 3D Plus, Japan) equipped with a photodiode array detector was used to analyze the targeted analytes. Chromatographic outputs were recorded at wavelengths of 256 nm for the analytes. A wavelength of 256 nm was chosen to ensure high sensitivity and compatibility with the HPLC mobile phase. The Shimadzu Shim-pack GIST C18 column (150 mm × 4.6 mm, 5 µm) was operated at an oven temperature of 30 °C. The analysis employed gradient elution, which blended water (mobile phase A) and methanol (mobile phase B) in relative percentages programmed as follows: 80:20 (A: B) at 0 min, 60:40 at 1 min, 40:60 at 7 min, and 80:20 at 8 min. The HPLC system was operated at a flow rate of 1.0 mL/min. A Shimadzu Shim-Pack Velox 3.0 mm C18 Guard column (2.7 µm) was used for the analysis. Sample collection Inlet and outlet water samples were obtained from the Muledane Wastewater Treatment Plant (WWTP 1) (Thohoyandou, Limpopo, South Africa) and Tzaneen Wastewater Treatment Plant (WWTP 2) (Tzaneen, Limpopo, South Africa). The samples were collected in clean amber glass bottles and stored in a 4 °C refrigerator for up to 1 week before analysis. Synthesis procedure of activated carbon and magnetic activated carbon Synthesis of activated carbon from avocado seeds The synthesis procedure followed the method of Blachnio et al. (Blachnio et al., 2020) with modifications. Avocado seeds were grated into tiny pieces, washed with deionized water, dried overnight at 90 °C, and ground into a fine powder. The three activated carbons were synthesized by impregnating avocado seed powder with 10%, 30%, and 50% KOH solutions separately (1:4) and left for 3 h. The three solutions were then filtered and washed with 0.1 M HCl, followed by deionized water, to remove impurities. Their pH was adjusted to 7.0. The powdered avocado seed, which had been carbonized at 600 °C for 2 h, was then chemically activated at different KOH concentrations (Blachnio et al., 2020; Thuy Luong Thi et al., 2021). The obtained activated carbons were washed with deionized water and dried in an oven. The activated carbons prepared were labeled AC 10% KOH, AC 30% KOH, and AC KOH 50%. The synthesis procedure is summarized in Supplementary Figure S1. In-situ preparation of magnetic activated carbon (MAC) from avocado seeds The coating of the activated carbon with iron oxide was performed using the method described by Lestari et al. (Lestari et al., 2021) with modifications. The magnetic-activated carbon was prepared by dissolving 4 g of FeSO4·7H2O and 8 g of FeCl3·6H2O in 200 mL of deionized water. A hotplate stirrer was used to heat the solution to 90 °C. 4 g of the above-prepared 10%, 30%, and 50% activated carbons were each added to separate portions of the same solution, resulting in three individual mixtures. Subsequently, 5 M (100 mL) of sodium hydroxide was gradually added to each solution and stirred for 10 min. The concoction was then placed in a water bath at 80 °C for 90 min. The formed Magnetic-activated carbon was collected using magnetic separation, washed, and dried at 80 °C for 24 h. The biosorbents prepared were labeled 10% KOH-MAC, 30% KOH-MAC, and 50% KOH-MAC. Dispersive magnetic solid phase microextraction (DMSPME) The procedure for DMSPME was conducted as follows: A specific amount of magnetic-activated carbon bioadsorbent was weighed into a vial containing a model solution containing MethylP, EthylP, and ButylP of 100 μg/L. The resulting mixture was subjected to sonication using an ultrasound water bath for 10–30 min. The supernatant was separated from the magnetic adsorbent (with the extracted analytes) using an external magnet. The bioadsorbent was retained, while the supernatant was decanted. Then, the analytes of interest were eluted through sonication using methanol. The analysis of the analyte concentrations was performed using HPLC-PDA. This analysis was conducted in replicates. Optimization strategy Univariate optimization A suitable bioadsorbent composition and eluent solvent were determined through a univariate approach. This method aimed to identify which composition of the KOH of the bioadsorbent and eluent solvent achieved the highest recoveries for extracting the three parabens from water. Three various compositions of KOH (10%, 30%, and 50%) for chemical activation were tested. During the testing of the bioadsorbent composition process, the following experimental conditions were fixed: 6.5 pH of the sample; 5 mL sample volume; 22.5 mg of the biosorbent mass; 30 min extraction time; and 850 µL eluent volume (acetonitrile). It was only the biosorbent that was varied. The best eluent solvent was selected in a similar manner. The best bioadsorbent (30% KOH-MAC), with an 850 µL volume eluent, a 6.5 pH sample, 22.5 mg of bioadsorbent mass, and an extraction time of 30 min, was kept constant, and only the eluent solvent was varied. The tests were conducted using acetonitrile, methanol, and a 50:50 (v/v) mixture of acetonitrile and methanol. Multivariate optimization of dispersive magnetic solid phase microextraction (DMSPME) A multivariate strategy was utilized to enhance the dispersive magnetic solid-phase extraction method. Significant factors were identified during the initial screening phase through a fractional factorial design (FFD). These identified factors were then further optimized using a central composite design (CCD) to establish the optimal conditions for the preconcentration and simultaneous extraction of trace amounts of parabens in water. The parameters determined included pH of the sample, mass of adsorbent (MA), extraction time (ET), and eluent volume (EV). The selected numerical ranges for the parameter optimizations are outlined in Table 1. TABLE 1Variables that are independent and their corresponding levels in the design of experiments.ParametersMinimumCentralMaximumEluent volume (µL)8009001,000Extraction time (min)1522.530pH46.59Mass of adsorbent (mg)2022.530 Variables that are independent and their corresponding levels in the design of experiments. Point of zero charge of the 30%KOH-MAC Understanding the PZC of the magnetic activated carbon (MAC) is vital for analyzing its adsorption characteristics (Abdulkadir et al., 2025). To determine the PZC of the 30%KOH- MAC biosorbent, a method from the literature was followed (Zeng et al., 2021). Various 30 mL pH solutions were prepared in conical flasks. The pH levels were modified from 2 to 10 by using ammonium hydroxide and acetic acid. Each flask received 22.5 mg of 30% KOH-MAC and was sonicated for 48 h. After this, the final pH of each solution was recorded. The PZC was determined from a plot of (final pH–initial pH) versus initial pH. Method validation The DMSPME method was validated by assessing the linear range, limit of detection (LOD), limit of quantification (LOQ), and both intraday and interday precision. This validation was conducted based on a study conducted by (Selahle et al., 2022). For the intraday precision, a solution with 100 μg/L of each analyte (MethylP, ButylP, and EthylP) was analyzed in ten separate trials (n = 10). The analysis was conducted over 5 days (n = 5) to assess interday precision and was conducted in replicates. Additionally, a spike recovery approach was used to test the method’s trueness. Three levels (10, 50, and 100 μg/L) were added to the inlet and the outlet wastewater samples, and each sample was analyzed three times. The percentage recoveries of spiked samples were determined using the Formula 1, %R=Found−0added (1) In this formula, [Found] indicates the concentration measured in the spiked sample, [0] is the concentration of the blank sample (unspiked sample), and [added] is the concentration introduced for spiking. To evaluate the linearity of the method, a blank inlet and outlet wastewater sample were spiked to create seven standard solutions with concentrations from LOQ to 35 μg/L for each analyte, and this was conducted in triplicate. The DMSPME method was used to extract and preconcentrate these standards, facilitating the creation of a seven-point calibration curve. The LOD and LOQ were calculated using Equations 2, 3, respectively: LOD=3xSDm (2) and LOQ=10xSDm (3) In these equations, SD refers to the standard deviation of the peak areas from 3 to 10 replicates of the minimum concentration in the linear range, while m denotes the slope of each calibration plot for the target analytes. The enrichment factor (EF) was calculated as the ratio of the final concentration of analytes after extraction to the initial concentration of analytes before extraction. Adsorption capacities Under optimal conditions, the biosorbent’s adsorption capacity towards the parabens was investigated. 22.5 mg of the biosorbent was combined with 30 mL of a model solution that contained 100 mg/L of each paraben, and the pH was adjusted to 6.5. After that, the mixture was ultrasonically agitated for 32.5 min. The quantity of each paraben adsorbed onto the biosorbent surface was calculated using Equation 4. This analysis was conducted in replicates. Qe=Ci−CeMV (4) where Qe (mg/g) signifies the quantity of MethylP, EthylP, and ButylP that is adsorbed per gram of the adsorbent. Ci and Ce (mg/L) represent the initial and equilibrium concentrations of the solution, respectively. L stands for the volume of the aqueous solution, and M (g) indicates the mass of the adsorbent utilized. Adsorption kinetics Adsorption kinetics are determined to understand the rate-limiting steps and the adsorption mechanisms for MethylP and EthylP using 30% KOH-MAC. The reaction kinetics models of pseudo-first-order (PFO) and pseudo-second-order (PSO), as well as intraparticle diffusion, were employed to study the adsorption kinetics. These equations are clearly explained in Supplementary Table S2. The data for adsorption kinetics studies were obtained using a sample volume of 5 mL, a mass of adsorbent of 20 mg, an initial concentration of 10 mg/L, a sample pH of 6.5, and an interaction time varied from 10 to 35 min. The table below presents the equations used to estimate the parameters for the kinetic models. Adsorption isotherms Langmuir, Freundlich, Sips, and the Dubinin-Radushkevich (D-R) adsorption isotherm models were used to understand the relationship and interactions between the parabens adsorbed on the surface of 30% KOH-MAC. The isothermal equations are further explained in Supplementary Table S3. The data for adsorption isotherm studies were obtained using a sample volume of 5 mL, a mass of adsorbent of 20 mg, an interaction time of 30 min, a sample pH of 6.5, and an initial concentration ranging from 5 ppm to 50 ppm. Reusability studies through extraction cycles Studies were conducted on the reusability and regeneration through a series of adsorption-desorption experiments. 22.5 mg of the bioadsorbent mass was measured and added to a sample bottle. Then, 5 mL of the model solution containing MethylP, EthylP, and ButylP was added to the sample bottle. The mixture underwent sonication for 32.5 min before being eluted with 875 μL of methanol. The eluent, which contained targeted parabens, was analyzed using HPLC-PDA. This was conducted several times using a similar bioadsorbent lost its functionality to adsorb high percentages of the parabens. The method and the data obtained were used to conduct the statistical analysis. Results and discussion Characterization Fourier-transform infrared spectroscopy (FTIR) FTIR spectra were obtained to assess the functional groups of the synthesized biosorbent. Figures 1A–C indicate that the raw avocado seed powder and different percentages of KOH are used in the activation and magnetization of the synthesized bioadsorbent. According to Figure 1, the FTIR spectra showed few significant differences between the AC and MAC. However, it can be seen that the broadness of the OH peak increased after the introduction of the KOH. The sharp peak of C-O observed in Figure 1A indicated that the carbon structure was bare, not having anything attached to it. The last peak at 500 cm-1 also indicates that it is intensified after the addition of iron. The three samples exhibited nearly identical sets of vibrational FTIR peaks. A broad peak at around 3,500 cm−1 was observed, suggesting that the O-H stretching vibrations are from the intermolecular hydrogen bonding. These OH peaks were broad and wide, indicating chemical activation. Another peak was also observed at 2,905 cm-1 due to the C–H stretching. Other peaks were observed at 1701 cm-1 and 1,586 cm-1, which were assigned to the C=O stretching of carboxylic acids and the O=C=O stretching of carboxylates, respectively. This was followed by a peak at 1,439 cm-1, corresponding to the aromatic ring. The peak at 1,179 cm-1 was attributed to the C–O stretching of phenols. Subsequently, a peak at 752 cm-1 was assigned to C–H of the aromatic rings. Figure 1B indicated an additional band at 689–500 cm-1. This was due to Fe-O, which was added during the magnetization stage. This peak was expected as it confirms the successful formation of the magnetic activated carbon. FTIR spectra illustrating the functional groups present in (A) Avocado seed powder, (B) Activated carbon with 10%KOH, 30%KOH and 50%KOH, (C) Magnetic activated carbon with 10%KOH, 30%KOH and 50%KOH. Scanning electron microscopy (SEM) SEM of the AC (Figures 2A,B) reveals high porosity with a well-defined, interconnected pore network. The smooth pore walls suggest a lightweight structure, potentially beneficial for adsorption. Figure 2C did not indicate well-defined pores. After the introduction of iron particles, the pores of the AC became smaller. This is illustrated in Figures 2D–F. However, some of the pores were still visible in Figure 2E. This proved that the incorporation of the Fe3O4 into the AC was successful. However, literature has indicated that when iron fills in the pores of AC, it reduces the surface area of the sorbent formed and might affect the adsorption process (Montaño et al., 2024). Nonetheless, the incorporation of iron strengthens the AC, and the magnetization facilitates the easy recovery of the biosorbent using an external magnet, making it more stable in an aqueous environment. Using an external facilitates easy recovery of the biosorbent. This is also crucial for preventing degradation of the biosorbent over time. SEM images highlighting the surface morphology of (A) 10% AC, (B) 30% AC, (C) 50% AC, (D) 10%KOH-MAC, (E) 30%KOH-MAC, and (F) 50%KOH-MAC. Energy dispersive X-ray spectrometer (EDS) The EDS spectra in Figures 3A–C revealed the presence of all the expected elements: carbon, oxygen, and potassium. Nitrogen was also observed in Figure 3B, which was considered an impurity. This was expected because avocado seeds contain many other elements. Figures 3D–F EDS spectrum revealed the presence of iron, carbon, oxygen, potassium, chlorine, and sodium. All these elements were expected except chlorine and sodium. Cl and Na were present in the ferric chloride and sodium chloride salts used in the synthesis. This indicated that the synthesized MAC contained some impurities. The peak of Fe in Figures 3D–F indicated that the incorporation of Fe3O4 into the activated carbon was successful, indicating that the desired biosorbent was successfully synthesized. EDS spectra of (A) 10%KOH-Activated Carbon (AC), (B) 30%KOH-Activated Carbon, (C) 50% KOH-Activated Carbon, and (D) 10%KOH-Magnetic Activated Carbon, (E) 30%KOH-Magnetic Activated Carbon, and (F) 50%KOH-Magnetic Activated Carbon. X-ray diffraction (XRD) XRD spectra were obtained to evaluate the crystallinity of the synthesized biosorbent. Figures 4A,B indicate the XRD spectrum for the avocado seeds activated and magnetic activated carbon. Figure 4A indicates that AC was amorphous. These findings aligned closely with a study reported by (Lestari et al., 2021). Figure 4B indicates that introducing Fe3O4 to AC transformed the amorphous structure to a crystalline structure. Additional distinct peaks were observed with 2-Theta values at approximately 32.1°, 37.4°, 45.1°, 55.4°, 56.9°, and 62.5°. These peaks indicated the presence of Fe3O4 particles in the AC, which formed MAC. They also correspond to the JCPDF of iron oxide XRD spectra. XRD spectra of (A) 10%KOH-Activated Carbon, 30%KOH-Activated Carbon, 50%KOH-Activated Carbon, and (B) 10%KOH-Magnetic Activated Carbon, 30%KOH-Magnetic Activated Carbon, and 50%KOH-Magnetic Activated Carbon. Vibrating sample magnetometry (VSM) The magnetic property of 30% KOH-MAC was studied using vibrating sample magnetometry. The findings are presented in Supplementary Figure S2 as a plot. Based on the plot, the saturation magnetization value of 30% MAC was around 10.48 emu/g. The magnetic property of 30% activated MAC is significantly lower in comparison to other magnetic materials reported in the literature. The 30% KOH-MAC decrease in magnetic property indicated the successful incorporation of Fe3O4 particles into AC. 30% KOH-MAC has activated carbon embedded in it, hence its low magnetic property. Even so, the magnetization value of 30%KOH-MAC was sufficient for magnetic separation using an external magnetic field during the extraction process. Brunauer–emmett–teller (BET) The Nitrogen adsorption–desorption isotherm analysis of the MACs (10% KOH, 30% KOH, and 50% KOH) was performed to evaluate the change in surface area and to determine the porous nature of the biosorbents. This was done using nitrogen adsorption–desorption isotherms. From Table 2, the surface area of the AC 30%KOH exhibited the highest value at 51 m2/g, compared to the AC 10% KOH, which had a surface area of 43 m2/g. After magnetization, there was a decrease in surface area; however, the BET surface area of the 30% KOH-MAC biosorbent remained the highest at 30.156 m2/g, making it the most suitable option among the 10% KOH-MAC and 50% KOH-MAC biosorbents. A similar trend was observed in terms of pore volume, with 30% KOH-MAC exhibiting a pore volume of 0.075 cm3/g, indicating greater pore structure development. The average pore diameters for all the MACs ranged from 2.33 to 2.36 nm, which positions them at the boundary between microporous and lower mesoporous classification (Colomer, 2024). 30% KOH-MAC resulted in good porosity and surface accessibility, which can enhance adsorption performance. Figures 5A–E showcases the adsorption-desorption behavior of the synthesized ACs and MACs. In Figures 5A,B, we can see the nitrogen adsorption-desorption isotherms for AC 10% and AC 30%. Both of these plots reveal a type IV isotherm accompanied by noticeable hysteresis loops, Figure 5A shows an increase around 1.0 P/P0, which suggests the presence of mesopores. On the other hand, Figure 5B displays a more pronounced hysteresis loop, indicating that it has enhanced mesoporosity. Figures 5C–E also illustrate type IV isotherms with H3-type hysteresis loops, which further confirm that MACs have mesoporous characteristics. When comparing the MACs and ACs, there is a slight decrease in nitrogen adsorption. This decline indicates a reduction in both surface area and pore volume following the magnetization process. It appears that the magnetic nanoparticles, which are deposited during magnetization, filled some of the internal pore spaces or covered the surface of the activated carbon. TABLE 2Surface area, pore volume, and average pore diameter of 10%KOH-MAC, 30%KOH-MAC and 50%KOH-MAC derived from N2 adsorption–desorption isotherms.BET parametersAC 10%AC 30%10%KOH-MAC30%KOH-MAC50%KOH-MACBET surface area (m2/g)4351283023Pore volume (cm3/g)0.20.20.10.10.1Pore diameter (nm)2.62.52.32.12.4 Surface area, pore volume, and average pore diameter of 10%KOH-MAC, 30%KOH-MAC and 50%KOH-MAC derived from N2 adsorption–desorption isotherms. BET adsorption and desorption isotherm graphs for (A) 10% AC, (B) 30% AC, (C) 10% KOH-MAC, (D) 30% KOH-MAC, and (E) 50% KOH-MAC, illustrating surface area and porosity characteristics. Point of zero charge of 30% KOH-MAC The pH point of zero charge (pHpzc) of the 30% KOH-MAC was determined to understand the surface charge characteristics of the biosorbent to help predict the interaction between the 30% KOH-MAC 30% and the parabens (Kosmulski, 2023). The 30% KOH-MAC pHpzc is presented in Supplementary Figure S3, which was found to be 7.6, indicating that its surface has a positive charge at pH values below 7.6. The surface of the 30% KOH-MAC was positively charged (proton-rich), which is good for the electrostatic interaction of negatively charged species such as electron-rich parabens (R–COO-). Furthermore, the surface had negative charges at pH values above 7.6, potentially resulting in electrostatic repulsion between the 30% KOH-MAC and the parabens. Optimization strategy Univariate optimization (investigating the percentage of KOH used for activation, and selection of eluent solvent) Choosing the most effective biosorbent for extracting and preconcentrating parabens from water bodies was achieved by evaluating the effectiveness of different MACs at varying KOH concentrations (10%, 30%, and 50%). From the percentage recoveries shown in Figure 6A, 30% KOH- MAC had better percentage recoveries of the parabens than 10% and 50%. During the magnetization stage, Fe3O4 occupied the pores of AC, leaving some pores available for utilization during the adsorption process. This was also supported by an SEM image (Figure 2E), which indicated that after magnetization with Fe3O4, the 30% KOH-MAC still showed more pores than the 10% and 50% samples. Additionally, the BET analysis (Table 2) indicated that 30% KOH-MAC had a surface area and pore volume higher than activated 10% and 50% MAC. This meant that 30% KOH-MAC would adsorb parabens better than the other two materials, being activated at 10% and 50%. Illustration of the selection of (A) the best activation percentage of the biosorbent and (B) the best eluent solvent for preconcentration and extraction of parabens. To achieve a high recovery percentage of parabens and maximize the extraction of parabens from the biosorbent surface, the appropriate eluting solvent was selected by evaluating the effectiveness of various solvents. Using triplicate adsorption extraction studies, the extraction capabilities of various organic solvents (MeOH, ACN, and a 50:50 (v/v) MeOH: ACN mixture) were evaluated. The polarity of methanol and parabens allowed for dipole-dipole interactions and hydrogen bonding to occur between the solvent and the adsorbate, promoting the effective dissolution of parabens in methanol and enhancing their adsorption. This implied that the parabens had a high methanol solubility, which improved the compound’s separation and retention. This was supported by Figure 6B, which shows that the plots indicate a high percentage of paraben recoveries when eluted with a methanol solvent. Therefore, methanol was selected as the optimal solvent and used for the subsequent extraction process. Multivariate optimization The effect of the independent variables: adsorbent sample pH, adsorbate mass (AM), elution volume (EV), and extraction time (ET), on the extraction of parabens, was assessed using a fractional factorial design (FFD) screening process. The optimized experimental conditions are illustrated in Table 1 of the experimental section. The findings were examined and displayed in a Pareto chart utilizing analysis of variance (ANOVA), as revealed in Figure 7. It is depicted that the blue bars for pH crossed the 95% confidence level (red line) from the target analytes’ Pareto charts, suggesting that pH is the most significant factor for the extraction and preconcentration of the three parabens. The other significant factor was adsorbent mass (MA), as it also exceeded the red line, according to Figure 7C. EV was also considered significant, as its blue bar was near the red line. This can also be justified by the fact that methanol had a superior performance in elution. This may be attributed to its enhanced interaction with the sorbed analyte and superior infiltration of the biosorbent microporous network compared with acetonitrile. The ANOVA results in Supplementary Table S4 indicate that the sample pH and eluent volume significantly affect the recoveries of EthylP, MethylP, and ButylP at a 95% confidence level. This is because p-values were greater than 0.05. In addition, high F- and low p-values suggest that the Pareto models provide a significant fit to the described analytical response (%R) as a function of the studied factors. Pareto charts of experimental variables on the preconcentration and simultaneous extraction of (A) MethylP, (B) EthylP, and (C) ButylP; where MA = Mass of adsorbent; ET = Extraction time; and EV = Eluent volume. The desirability functions were utilized for the determination of the precise numerical values of the parameters that are essential to maximizing the target compounds’ extraction recovery (Bayuo et al., 2024). The extraction conditions that afford satisfactory percentage recoveries of the parabens were determined, as shown in Figure 8: pH of 6.5, an adsorbate mass of 22.5 mg, and an eluent volume of 875 µL. These parameters led to satisfactory extraction recoveries of parabens. Therefore, the desirability function successfully optimized the experimental conditions, and further analysis was conducted using these optimized conditions. Multivariate optimization was designed to simultaneously evaluate and optimize multiple experimental variables affecting the efficiency of the analytical procedure. Unlike univariate approaches, which assess one factor at a time, multivariate methods enable the identification of interactions between variables, resulting in a more robust and efficient optimization process. Optimization of independent parameters for targeted parabens using a desirability function. Method validation The evaluation of the analytical measurements of effectiveness for the DMSPME-HPLC-PDA method was conducted. This was conducted under the optimum experimental conditions as discussed in the experimental section. As depicted in Table 3, a linear dynamic range (LDR) of LOQ -35 was achieved. The reproducibility and repeatability of the extraction method were also assessed and presented as relative standard deviations (% RSD). Lower intraday and interday were obtained to be in the range of 2.80%–4.50% and 3.40%–4.20%, respectively. These results showed good precision as the RSDs were all below 5%. LODs and LOQs of the method were 0.12–0.19 μg/L and 0.25–0.90 μg/L, respectively. These obtained values conveyed that the DMSPME-HPLC-PDA method was sensitive. Therefore, conducting the analytical performance illustrated that the DMSPME-HPLC-PDA method was effective and sensitive enough for the simultaneous extraction of MethylP, ButylP, and EthylP in wastewater samples. Additionally, the parabens demonstrated comparable enrichment factors, ranging from 41 to 43, confirming the strong preconcentration capability and reliability of the DMSPME-HPLC-PDA method for determining EthylP, MethylP, and ButylP. TABLE 3Summary of the analytical performance parameters of the DMDMSPE–HPLC–PDA method.ParametersEthylPMethylPButylPLinear range (µg/L)0.1–350.1–350.1–35R2 0.950.980.96LOD (µg/L)0.120.190.17LOQ (µg/L)0.900.250.67Intraday (%RSD), n = 10)533Interday (%RSD, n= 3 days)443EF424344 Summary of the analytical performance parameters of the DMDMSPE–HPLC–PDA method. The validity of the presented method was evaluated through analysis of spiked water samples. Inlet and outlet samples from both wastewaters were spiked using three concentrations: 10, 50, and 100 μg/L. The analysis of these concentrations was conducted in triplicate. As shown in Table 4, the recovery percentage (% R) obtained ranged from 99.1% to 101%. The recovery results showed that the matrix effects did not affect the performance of the extraction method. This also indicated that the DMSPME-HPLC-PDA method was appropriate for the preconcentration and simultaneous extraction of the parabens. TABLE 4Spiking recovery results for influent and effluent wastewater samples containing parabens using the optimized method.ParabenAdded (µg/L)Measured value (µg/L) N = 3%Recovery%RSDMethyl109.87 ± 0.198.71.2​5049.6 ± 1.299.22.5​10099.8 ± 0.199.81.3Ethyl109.91 ± 0.999.13.0​5050.5 ± 1.71012.1​10099.8 ± 0.199.83.6Butyl109.76 ± 0.997.63.3​5049.8 ± 0.199.63.2​10099.7 ± 1.499.71.7 Spiking recovery results for influent and effluent wastewater samples containing parabens using the optimized method. Application of the developed method According to Supplementary Table S5, MethylP, EthylP, and ButylP were found in both wastewater treatment plants of interest. The concentrations found in the inlets for WWTP 1 and WWTP 2 ranged from 6.10 to 17.39 μg/L. Higher concentrations were anticipated in the inlet samples, as the water had not undergone any treatment. The chromatograms for the analysis are shown in Supplementary Figure S6. The outlet ranged from