Role of polyamide microplastics as vector of parabens in the environment. An adsorption study

Environmental Technology & Innovation 32 (2023) 103276 Contents lists available at ScienceDirect Environmental Technology & Innovation journal homepage: www.elsevier.com/locate/eti Role of polyamide microplastics as vector of parabens in the environment: An adsorption study Carmen Mejías, Julia Martín, Juan Luis Santos, Irene Aparicio, Esteban Alonso∗ Departamento de Química Analítica, Escuela Politécnica Superior, Universidad de Sevilla, E-41011 Seville, Spain article info Article history: Received 2 February 2023 Received in revised form 3 July 2023 Accepted 4 July 2023 Available online 12 July 2023 Keywords: Adsorption Parabens Polyamide Environmental factors Environmental water matrices abstract The prevalence of microplastics in the aquatic environment has become a global problem. Their capacity to adsorb pollutants may influence in their environmental fate, bioavailability, and toxicity to biota. This work provides a systematic study to characterize the role of polyamide (PA) as vector of parabens (PBs) in the environment. The effect of PA and influence of environmental factors in the adsorption process were investigated. The amount of PBs adsorbed onto 50 µm PA at 30 mg/L of PBs follows the order: butylparaben (1.440 mg/g) > propylparaben (1.321 mg/g) > ethylparaben (0.995 mg/g) > methylparaben (0.543 mg/g), which is positively correlated with their log Kow and length of the ester alkyl chain. Physical adsorption forces such as hydrophobic interaction, pore filling and hydrogen bond dominated the adsorption mechanism The size of the PA particles has been resulted in a significant factor; a higher adsorption capacity was remarkable when decreasing PA particle size (from 10.3% for 3 mm to 79.5% for 50 µm (0.5 mg/L PBs)). The adsorption percentage increases with pH until the PBs pKavalue and decrease significantly at pH>pKavalues due to repulsion forces. Salinity increases adsorption capacity until at 2% NaCl content while the dissolved organic matter negatively affects and leads to low adsorption. Our results reveal that PBs were potentially adsorbed onto PA in real water matrices (average of 82% in tap water, 72% in surface water, 76% in effluent wastewater, and, 53% in influent wastewater (0.5 mg/L PBs)), which may have important environmental implications. ©2023 The Author(s). Published by Elsevier B.V. This is an open access article under the CC BY-NC-ND license (http://creativecommons.org/licenses/by-nc-nd/4.0/). 1. Introduction Plastic pollution is an issue of global concern and receives ever-increasing attention. Microplastics (MPs), plastics with a particle size less than 5 mm, are likely the most numerous plastic debris in the environment. MPs can reach the aquatic environments through three ways: (1) plastic waste littering; (2) discharge from industrial and domestic wastewater treatment plants (WWTPs), and (3) plastic waste of sources, including fisheries, agriculture, and other industries (Zandaryaa,2021;Elgarahy et al.,2021). Polystyrene (PS), polyethylene (PE), polyvinyl chloride (PVC), polyamide (PA), and polypropylene (PP) (Ding et al.,2022) are the plastic polymers most widely used in numerous applications and hence more frequently found in the environment. Recent studies have also proved how MPs can be accumulated in living organisms from fish, molluscs, and crustaceans to smaller marine biota such as plankton or larvae (Martín et al.,2022; Hanun et al.,2021). ∗Correspondence to: Departamento de Química Analítica, Escuela Politécnica Superior, Universidad de Sevilla, c/ Virgen de África, 7, 41011 Seville, Spain. E-mail address: [email protected] (E. Alonso). https://doi.org/10.1016/j.eti.2023.103276 2352-1864/©2023 The Author(s). Published by Elsevier B.V. This is an open access article under the CC BY-NC-ND license (http://creativecommons. org/licenses/by-nc-nd/4.0/). C. Mejías, J. Martín, J.L. Santos et al. Environmental Technology & Innovation 32 (2023) 103276 The physico-chemical properties of MPs, particularly their tiny size and large surface area provide them with a high adsorption potential and the ability to affect the fate and toxicity of adsorbed contaminants in the aquatic environment and the biota (Gatidou et al.,2019;Chae and An,2017;Paul-Pont et al.,2018). Montero et al. (2022) demonstrated that PP MPs and the benzophenone-3 act synergistically to generate inflammation in the intestine of European sea bass. Similarly, He et al. (2021) exposed zebrafish to PP MPs and triphenyl phosphate and found that PP aggravated the estrogenicity of male fish caused by triphenyl phosphate. Furthermore, triclosan accumulation in liver and gut of zebrafish was found to be higher when it is adsorbed to PE, PP and PVC MPs (Sheng et al.,2021). The investigation of association of MPs and emerging pollutants is a rapidly evolving field with a growing number of publications (Martín et al.,2022). The adsorption of organic compounds on MPs is influenced by many factors including the physico-chemical characteristics of the polymer (crystallinity, dosage, and size of the plastic particles), the characteristics of the organic compounds (polarity and pKa), and environmental factors (dissolved organic matter (DOM), ionic strength, salinity, and pH) (Concha-Graña et al.,2022; Martín et al.,2022). Polyamide (PA) or nylon represents about 44.7% of polymers discharged in the marine environment (Mofakhami et al., 2020;Hamidian et al.,2021;Wang et al.,2018). Its porous structure plays a major role in the adsorption capacity of emerging compounds. As a class of emerging contaminants, parabens (PBs) have received increasing attention due to their wide application in pharmaceutical, food, and cosmetic for its antimicrobial properties (Bolujoko et al.,2021). PBs are inevitably discharged into environmental media and are considered as ubiquitous contaminants. These compounds have been detected in natural water samples worldwide (Haman et al.,2015;Malvar et al.,2019;Garrido et al.,2016) from 30000 ng/L in raw wastewater to 103 ng/L in effluent wastewater (Haman et al.,2015) or up to 37.4 ng/L in surface water from the Antarctic (Emnet et al.,2015). PBs have been associated to different toxicological effects such as endocrine disruption, carcinogenic effect, telomere shortening, thyroid inhibition, among others (Bolujoko et al.,2021). Yamamoto et al. (2011) reported that the toxicity of seven PBs using Oryzias latipes and Daphnia magna increased with their alkyl chain length. To date, the adsorptive behaviour of PBs on MPs under the influence of environmental conditions has not been comprehensively characterized or reported. As far as we know, there are only two studies that reported the adsorption of PBs onto MPs (Arvaniti et al.,2020;Chen et al.,2022). Arvaniti et al. (2020) carried out preliminary experiments to investigate the adsorption potential of four PBs onto PS and PE, but their investigation has been focused to examine the removal efficiency of these MPs during coagulation experiments. In another attempt, Chen et al. (2022) studied the adsorption of methylparaben (MeP) onto PE, PS, PVC, and polyethylene terephthalate (PET) MPs. This work provides a systematic study to characterize the adsorptive behaviour of four PBs (MeP, ethylparaben (EtP), propylparaben (PrP) and butylparaben (BuP)) onto PA MPs. The role of dosage and size of PA MPs and the influence of environmental factors (pH, salinity, and DOM) in the adsorption process were investigated using batch adsorption experiments. Adsorption experiments were also conducted using real water matrices including influent and effluent wastewater, surface water and tap water. 2. Materials and methods 2.1. Chemicals and reagents Standards of MeP, EtP, PrP and BuP (≥99%) were purchased from Sigma-Aldrich (Steinheim, Germany) (physicochemical properties in Supplementary material, Table S1). Humic acid and ammonium acetate were provided by Sigma-Aldrich (Madrid, Spain). Sodium hydroxide, sodium chloride, hydrochloric acid, and distilled water were supplied by Panreac (Barcelona, Spain). All reagents were of analytical grade. The chromatographic solvents methanol and water (LC-MS-grade) were purchased from Biosolve BV (Valkenswaard, the Netherlands). The different particle size of PA (50 µm, 55 µm and 3 mm) were supplied by Goodfellow (Hamburg, Germany). Stock standards solutions of PBs (1000 mg/L) were done in methanol. A mixture solution of all compounds was obtained by dilution in water to reduce (<0.1%) any cosolvent effect. 2.2. Microplastic characterization PA MP was characterized before and after the adsorption using scanning electron microscopy (SEM), Fourier transform infrared spectroscopy (FT-IR), X-ray diffraction (XRD) and Kr adsorption–desorption isotherms and porosimetry system. SEM analysis was conducted with a FEI-TENEO scanning electron microscope (FEI, USA). For that, samples were previously coated by a layer of platinum using a Sputter Coater Leica EM ACE600. FT-IR analyses were performed using a Cary 630 FT-IR (Agilent, USA) with Diamond attenuated total reflection. The spectral range and resolution were 4000–650 cm−1 and 4 cm−1, respectively. XRD analyses were carried out by a Bruker D8 Advance A25 diffractometer (Bruker, Germany) equipped with a Cu Kαradiation source operating at 30 mA and 40 kV. An accelerated surface area and porosimetry system (ASAP 2420) (Micromeritics Instrument, United Kingdom) was used for the measurement of the surface areas using the Brunauer–Emmett Teller (BET) equation. The Barrett–Joyner–Halenda (BJH) method was applied to measure the pore-size distributions from the desorption of N2. The pH of zero-point charge (pHZPC) was determined by pH drift method (Gatabi et al.,2016) using a pHmeter BASIC 20 (Crison Instruments, Barcelona, Spain). For that, we prepared 50 mL of distilled water solution adjusted to pH 2, 4, 7, 9 and 11 by adding HCl or NaOH. Then, 0.5 g of PA was added, and the mixture was shacked for 48 h at 25 ◦C. After that, the pH was measured again. The pHZPC was obtained by plotting the initial pH value vs. the final pH value. Intersection point of the resulting curve with the line passing origin (pHfinal =pHinitial) gives pHZPC. 2 C. Mejías, J. Martín, J.L. Santos et al. Environmental Technology & Innovation 32 (2023) 103276 2.3. Batch experiments Batch adsorption experiments were performed using glass vials containing 0.5 g of PA and 30 mL of distilled water solution of PBs (0.5 mg/L) (Wang et al.,2020;Mejías et al.,2023). An interval of time from 1 to 4320 min was first tested for kinetic studies. Then, the adsorption isotherms were studied in the concentration range from 0.1 to 30 mg/L of PBs. All experiments were made by triplicate using a thermostatic shaker at 350 rpm and 298 ±2 K. Procedural blanks were simultaneously prepared without PA under the same conditions to evaluate possible adsorption of PBs to the glass bottles or for loss due to volatilization. To study the influence of PA particles, three different sizes (50 µm, 55 µm and 3 mm) and different dosages (from 10 to 80 g/L) of PA were tested. To assess the influence of environmental parameters, pH was tested from 1 to 13, which was adjusted with sodium hydroxide or hydrochloric acid, as needed. Salinity (from 0 to 3.5% of NaCl, w/v) was used to evaluate ionic strength. Moreover, DOM was examined through the presence of humic acid (from 0 to 25 mg/L using commercialized humic acid in distilled water). The applied conditions in each batch of experiment are summarized in Table S2. Aqueous samples were collected at the start and the end of the experiment, and they were filtered to 0.22 µm previous the analysis. PBs were quantified in the samples by liquid chromatography-tandem mass spectrometry (LC-MS/MS). Chromatographic conditions were previously reported by Martín et al. (2017) and summarized in the supplementary material (Table S3). Adsorption using real water samples was also evaluated. For that purpose, in February 2022, 2 L of water from the Guadalquivir River (Seville, Spain), 2 L of influent and effluent wastewater from a WWTP located in Seville (Spain) and 2 L of tap water samples from the same city during that period were collected. Before the experiments, samples were filtered through a 1.2µm glass-microfibre membrane filter. The adsorption experiments were carried out using 0.5 g of PA of lower particle size (50 µm) and 30 mL of each matrix. Fortified (0.5 mg/L) and blank samples were prepared for each type of matrix. The quantification was carried out in each case by matrix-matched calibration curves and by direct injection. Data analysis is described in the supplementary material. Different mathematical models including linear, Langmuir and Freundlich were evaluated to determine the adsorption performance to experimental isotherms, and pseudo-first order (PFO) and pseudo-second order (PSO) were used to determine the kinetic model (Table S4). Statistical correlation analysis was used to set up relations between the physicochemical characteristics of the water samples and the adsorption % of PBs (Statistical 10.0 software for Windows). A correlation matrix was performed using the adsorption capacity of each PB at 50 µm particle size in each type of aqueous samples and the physicochemical characteristics of water samples as variables. A correlation coefficient of 1 shows a perfect positive relationship, 0.8 shows a fairly strong positive relationship and 0.6 shows a moderate positive relationship while 0 shows a no relationship, so significant differences were considered when correlation coefficient is higher or equal than 0.8. A correlation coefficient of −1 shows a perfect negative relationship, −0.8 shows a fairly strong negative relationship and −0.6 shows a moderate negative relationship while 0 shows a no relationship, so significant differences were considered when correlation coefficient is lower or equal than −0.8. Student’s t-test was applied to evaluate the experimental data comparing values of calculated (tcal) and tabulated (ttab). 3. Results and discussion 3.1. Polyamide characterization Surface morphology of PA MPs was visualized by SEM and the images are shown in Fig. 1. PA particles can be seen as pellets-like shape with differences in their sizes. The BET surface area is of 0.46 and 0.41 m2/g for 50 and 55 µm PA MPs, respectively (Table S5). A high surface area provides more active sites to retain the contaminants. The BET surface area of the 3 mm particle could not be measured by ASAP due to its large size and low surface area. The surface of PA is rough and exhibits characteristics of porous polymer (average pore diameter of 37.5 and 28.1 nm for 50 and 55 µm PA MPs, respectively). Micropollutants could be successfully retained by these pore spaces by pore-filling mechanism. No differences were observed in the SEM analysis in the shape and size of the particles after the adsorption process (Fig. 1), only a decrease in the specific surface area (BET of 0.23 and 0.12 m2/g for 50 and 55 µm PA MPs, respectively, after the adsorption process). The FTIR spectroscopy was used to identify the functional groups of PA particles and to observe the changes in functional groups or the creation of new ones in MPs after adsorption process (Figure S1). FTIR spectrum obtained in Fig. S1 (upper) fully fits of PA spectrum (Mejías et al.,2023). The band at 3300 cm−1fits with N–H stretching signal. Bands at 2850 and 2950 cm−1adjust to symmetric and asymmetric CH2, respectively. Bands at 1650 cm−1and 1550 cm−1fit with amides (C==O +C–N and C–N +N–H bending, respectively). Bands at 1400 cm−1correspond to amide (C–C stretching +N–H band +C==O bending). Bands at 1200 cm−1could be associated C–N stretching. Bands at 1000 cm−1fit with CCH bending (symmetric)/CH2twisting, 800 cm−1band are correlated with C–C stretching. After adsorption, the clearest changes in the spectrum were that the peaks at 750, 850 and 1100 cm−1were stronger in intensity than those picks before the adsorption process. However, any new peaks were formed after adsorption for PBs, implying that no new chemical bonds were formed, and that the chemical adsorption was not the main mechanism in the adsorption. Thus, FTIR confirmed that physical adsorption might play a crucial role in the adsorption (Ma et al.,2019;Liu et al.,2019b). Additionally, the degree of crystallinity was detected using XRD (Figure S2). XRD analysis fits basically with nylon and after the adsorption it is basically the same compounds but more amorphized. 3 C. Mejías, J. Martín, J.L. Santos et al. Environmental Technology & Innovation 32 (2023) 103276 Fig. 1. SEM images of PA MP: before the adsorption process (magnification 600x (A) and 5000x (B)) after the adsorption process (magnification 600x (C) and 5000x (D)). 3.2. Kinetic and isotherm studies In Figure S3 the influence of the equilibrium time on the adsorption of PBs is represented. The adsorption of PBs by PA MPs gradually increased until reaching the equilibrium at 24 h. The contact period between the adsorbent and the adsorbate is critical in adsorption process. Significant differences in the adsorption capacity were observed between 12 h to 24 h (Student’s t-test: tcal =4.99, 5.43, 5.89 and 5.91 for MeP, EtP, PrP and BuP, respectively; ttab =4.6041, p <0.05). However, no significant differences were obtained between days (Student’s t-test: tcal =3.31, 3.42, 3.55 and 3.98 for MeP, EtP, PrP and BuP, respectively; ttab =4.6041, p <0.05). This fact indicates that 24 h was time enough to achieve the equilibrium and to get reproducible results. Similarly, Chen et al. (2022) reported that almost 2000 min were needed to reach reproducible results when assessing the effect of the contact time on the adsorption of MeP onto PS. To clarify possible adsorption mechanisms of PBs by PA MPs the widely used adsorption kinetics models, PFO and PSO models, were further investigated (Table S6). The first model assumed that the adsorption process was dominated by chemical adsorption while PSO assumed that it was dominated by physical adsorption. The fitting plots of the kinetic models can be seen in Fig. S4. In comparison to the PFO, the PSO model provided the higher correlation coefficient, which demonstrated that the adsorption process of PBs by PA MPs was dominated by the chemisorption and was supplemented by physisorption. In some adsorption studies, the adsorption mechanism was determined according to the kinetic data. However, this approach was criticized by some authors (Hacıosmanoğlu et al.,2022) because the adsorption mechanism cannot be directly determined according to the kinetic model fit and adsorbent characterization studies are necessary. Fig. 2 shows the adsorption isotherms of PBs on PA particles at room temperature (298 ±2K). PBs displayed different adsorption capacity, highly influenced by their structure (Table S1). Enhanced adsorption capacity was remarkable when increasing the length of the ester alkyl chain of PBs (BuP >PrP >EtP >MeP). PBs are esters of p-hydroxybenzoic acid. As the ester chain length increases, water solubility decreases and the octanol/water partition coefficients (log Kow) increase (Table S1). Our results showed that PBs with longer ester alkyl chains possess higher adsorption onto PA suggesting that hydrophobicity interactions can be the main mechanism of adsorption of PBs onto PA. As can be seen in Figure S5, log Kow values were positively correlated with the length of the ester alkyl chain (R2>0.97; p <0.05). This result is in good agreement with previous findings by Lara et al. (2021) who found a positive relationship between the hydrophobicity of 4 C. Mejías, J. Martín, J.L. Santos et al. Environmental Technology & Innovation 32 (2023) 103276 Fig. 2. Adsorption isotherms of solutions of MeP, EtP, PrP and BuP onto PA. PA MPs and the Log Kow of endocrine estrogenic compounds suggesting the important role of hydrophobic interactions in the adsorption process. The same effect has been reported for perfluoroalkyl substances and PA MPs (Mejías et al.,2023). 5 C. Mejías, J. Martín, J.L. Santos et al. Environmental Technology & Innovation 32 (2023) 103276 Langmuir, Freundlich and Linear models were performed to fit the adsorption data using different PA MP sizes (Table S7). The equilibrium adsorption data were well represented by the three models. For example, at 50 µm of PA size the correlation coefficients (R2) obtained for the four PBs were in the following ranges: Langmuir (0.955–0.996), Freundlich (0.937–0.994) and linear (0.938–0.988). It seems that the linearity of the adsorption isotherms worsens with the increase of the PBs chain length. BuP and PrP were better fitted to Langmuir or Freundlich models than to the linear model, especially for higher MP particles, although differences were not significant. Thus, the interaction between PBs and MPs mainly includes the adsorption and partition (Chen et al.,2022). Similarly, Yu et al. (2019) reported the adsorption of tetracycline on PE was adjusted to linear (R2=0.948–0.997) and Freundlich isotherms (R2=0.912–0.988). The separation factor (RL) based on the Langmuir isotherm model was used to assess the adsorption process. The value of RLindicated the isotherm shape as follows: RL>1 unfavourable, RL=1 linear, 0 <RL<1 favourable and RL =0 irreversible (Mohanty et al.,2023). The RLranged from 0.146 (BuP) to 0.870 (MeP), which displayed the favourable adsorption process of PBs by the PA MPs. In addition, a value of 0 <n<10 also revealed favourable adsorption conditions (from 0.534 (BuP) to 1.621 (MeP) (Mustapha et al.,2019)). Overall, a slightly better adjustment was observed at lower PA MPs sizes although, except for BuP at 3 mm particle size, the values obtained were >0.9 for all sizes. The parameters KD, KL, qmax, and KFincrease with the alkyl chain of PBs and with the decrease of MP size. For example, qeof MeP at 30 mg/L of PBs increase from 0.31 to 0.54 mg/g for 3 mm and 50 µm PA size, respectively, while BuP increase from 0.67 to 1.44 mg/g for 3 mm and 50 µm PA size, respectively. These values were similar (or even higher) in magnitude with respect to other adsorbents, indicating the worrying adsorption properties of PA (Gao et al.,2022;Chen et al.,2017). As an example, Gao et al. (2022) reported a maximum adsorption capacity of 0.6015 mg/g for the adsorption of MeP onto magneticnanoparticles. Arvaniti et al. (2020) reported no adsorption of PBs onto PS and PE. Thus, the polarity and functional groups of MPs seem to play a crucial role in the adsorption capacity. According to our results, the linear relationship between log Kow and qmax provided evidence that the hydrophobicity of the contaminant itself also had a high effect on the adsorption capacity of PA particles. Other mechanisms such as pore-filling could be implicated. The decrease in surface area can be explained by pore filling mechanism. Before adsorption, PA MPs had a BET surface area of 0.46 and 0.41 m2/g for 50 and 55 µm particle size, respectively, and a pore volume of 0.003 and 0.001 cm3/g, for 50 and 55 µm particle size, respectively. But after reached saturation adsorption in 24 h, both decreased. The BET surface area decreased to 0.23 and 0.12 m2/g for 50 and 55 µm, respectively and the micropore volume nearly disappeared. This fact could indicate that PB molecules occupied pore channels as adsorption sites by occupying the external surface and penetrating the inner micropore spaces. Previous studies described that the pore filling is the main mechanism for adsorption by glassy polymers, such as PS, PVC, and PA (Hüffer and Hofmann,2016). Liu et al. (2019a), Guo et al. (2019a) and Li et al. (2019) found that some groups of emerging pollutants such as bisphenol A and antibiotics are strongly adsorbed onto PA. In addition, besides pore filling and hydrophobic interaction, PBs present some groups in their structures such as -OH or –CO, which could involve other physical interactions such as complex hydrogen bonding between the amide groups (proton donor) present in PA and the carboxylic and OH groups (proton acceptor) of PB structures; or n-πinteractions between the aromatic ring of PBs and –CO group of PA. The electrostatics interactions play a minor role in the adsorption mechanism. As described below in Section 3.4, the highest adsorption occurred at pH <9 in which the PBs are in neutral form with no electric charge. The proposed adsorption mechanism for PBs adsorption onto PA particles can be seen in Fig. 3. 3.3. Effect of polyamide dosage and size MPs amount and size are key parameters in the adsorption process. As expected, an increase in the adsorption % was observed with the dosage (Figure S6). This effect was significant for all PBs studies (tcal =13.47 (BuP)-7.11 (MeP) ttab =2.78; p <0.05; when comparing 10 g/L and 80 g/L PA dosages) although more pronounced in the case of BuP (from 14% to 67% at 10 and 80 g/L of PA dosage, respectively). The same behaviour was found in the adsorption of five pesticides on PE (Wang et al.,2020). When MPs dosage is low, the adsorption sites can be completely occupied by the pollutants with large amounts in aqueous solution. An increase in MPs dosage can provide more available adsorption sites which results in a substantial growth of the adsorption efficiency. Previous reports have stated that MPs with the same chemical composition increased their adsorption capacity with decreasing MPs size, probably due to its differences in adsorption surface area (Elizalde-Velazquez et al.,2020). The results obtained agree with this, and as can be seen from Fig. 4 the size of the PA particles resulted an important factor, the adsorption capacity was notable as the size of the PA particles decreased. Adsorption of BuP increased from 29% for 3 mm particle size to 99% for 50 µm particle size. However, adsorption % were similar in the case of 50 and 55 µm due to the low difference in the particle size. Similar findings were recently reported by Cormier et al. (2021), Moura et al. (2022) or Ma et al. (2019) for perfluorooctanesulfonate, microcystin analogues or triclosan onto PE, PP and PVC, respectively. 6 C. Mejías, J. Martín, J.L. Santos et al. Environmental Technology & Innovation 32 (2023) 103276 Fig. 3. Possible mechanisms for PBs adsorption on PA MPs. Fig. 4. Adsorption (%) of MeP, EtP, PrP and BuP on PA MPs in the evaluation of MPs particle size. 3.4. Effect of pH, salinity, and dissolved organic matter Overall, the adsorption process exhibited a pronounced pH dependence due to the speciation of the compounds and the surface charge of the MPs particle. The adsorption of PBs onto PA as a function of pH was investigated in the range from 1 to 13. The adsorption percentage tends to increase as the pH increases from 1.0 to 9.0 (Fig. 5A), but at a very extreme alkali pH decreases. This can be explained by repulsion forces since PBs exist partially as an anion at pH >pKa values (Table S1), and the surfaces of PA tend to be also negatively charged at pH >pHZPC values (6.0, Figure S7). A similar effect was reported by Chen et al. (2022) when assessed the pH effect on the adsorption of MeP onto PS. At pH =1 no adsorption was observed, i.e., the calculated removal efficiency for MeP was practically nil. This fact could be due to the higher polarity and lower Log Kow value of MeP in comparison to the other PBs. In natural marine or surface water (6 < pH <9), PA MPs will exhibit relatively stable adsorption of PBs. Similarly, Guo et al. (2019b) reported that the adsorption of sulfamethazine on PA tends to increase as the pH increases from 3.0 to 7.0 but at alkali pH the adsorption decreases (pKaof sulfamethazine is 7.4). 7 C. Mejías, J. Martín, J.L. Santos et al. Environmental Technology & Innovation 32 (2023) 103276 Fig. 5. Adsorption (%) of MeP, EtP, PrP and BuP on PA MPs in the evaluation of (A) pH, and (B) salinity. Salinity is an environmental factor that could influence the adsorption of PBs onto PA. In this work, a NaCl concentration gradient was reached to simulate the salinity environment of oceans and rivers. As is presented in Fig. 5B, the results showed that adsorption of PBs was improved as NaCl concentrations increased up to a 2% content value. For example, the adsorption of PrP increased from 20% in distilled water to 37% at 2% of salinity, but at higher NaCl contents the adsorption decreases to 29%, the same pattern was found with MeP on PS by Chen et al. (2022). Salting-out effect of NaCl could be the key to enhance the adsorption affinity of organic pollutants by decreasing their solubility. Salting-out is a physicochemical phenomenon based on electrolyte–non-electrolyte interactions, in which at high salt concentrations (or high ionic strength) some solutes precipitate due to increased hydrophobic interactions between them. Furthermore, salinity increases the agglomeration of MPs and, hence, the number of sites for organic pollutant adsorption (Velzeboer et al.,2014). The electrostatic repulsion between MPs and the pollutants may decrease with the increase of the ionic strength as the added salts could neutralize the negative charges on the surface of MPs (Vilar et al., 2005). Similar results were reported by Arvaniti et al. (2022) and Chen et al. (2022) who suggested that the adsorption capacity of valsartan and MeP on PS increased proportionally to the increase of salt concentration. Nevertheless, contradictory results have been reported in the literature when evaluating the influence of salinity on MP adsorption (Liu et al.,2019b;Wang et al.,2020). This fact suggests that the impact of salinity can differ across the types of MPs and organic pollutants. Wu et al. (2016) and Lin et al. (2020) reported that salinity does not affect the adsorption of carbamazepine, triclosan, 4-methylbenzylidene camphor, 17α-ethinylestradiol, and tetracycline on PE and PA MPs, respectively. Mejías et al. (2023) concluded that salinity has a negative effect in the adsorption of perfluoroalkyl compounds onto PA as the increase of the ionic strength can lead to a competition between the organic compounds and the cations for the adsorption sites of the negatively charged MPs. Similarly, Liu et al. (2019b) observed that salinity decreases the adsorption of ciprofloxacin onto PVC and PS MPs. They explained that fact by an increase of the cohesive density between the polymer chains by chloride ions, reducing the mass transfer, whereas sodium ions, in the case of PS, reduce the cation exchange mechanism by competition. The presence of DOM, commonly found in environmental samples, can represent a competition for adsorbent binding sites conditioning the adsorption process. In this work, humic acid (0–25 mg/L) was used to represent natural DOM. It is an ordinary class of natural organic matter, which carries oxygen-functional groups, and usually exists with a negative surface charge (Chen et al.,2018). As can be seen from Fig. 6, the presence of DOM causes a negative effect on the adsorption. This effect was most marked for MeP and EtP. For example, the adsorption capacity of MeP decreased from 10% to 5% when the concentrations of humic acids increased from 5 to 25 mg/L, respectively. Two possible explanations are the competition effect of DOM for the adsorption active sites of PA or that the humic acid molecule could preferentially cover the PA surface, thereby inhibiting the pollutant adsorption active site (Li et al.,2021). The humic acid can bind through n-πelectron donor–acceptor interactions between CO groups of the PA and the aromatic ring of the fulvic acid, thereby reducing the availability of adsorption sites to the PBs (Tang et al.,2021;Lara et al.,2021). In a previous study, Tang et al. (2021) corroborated the adsorption of PA by fulvic acid (qmax up to 1.44 mg/g) by electrostatic attraction, hydrophobic interactions, H-bonds and n-πelectron donor–acceptor. These findings are in concordance with those reported by Lara et al. (2021). They reported that endocrine disrupting compounds decrease their adsorption capacity onto PA MPs in the presence of DOM. Similarly, Yu et al. (2020) reported that fulvic acid exert a negative competition impact in the adsorption of levofloxacin on PVC MPs. In another study, Atugoda et al. (2020) reported that ciprofloxacin adsorption capacity was decreased from 3.0 mg/g to 0.25 mg/g when the concentrations of humic acids increased from 0.5 to 2.5 mg/L, respectively. 3.5. Adsorption in environmental real matrices To achieve a greater environmental realism and model complexity of real samples, the adsorption test was repeated using four fortified (0.5 mg/L of PBs) environmental waters including tap water, surface water, and influent and effluent wastewater. The concentration levels of PBs measured in blank samples (0.17 ng/mL (EtP)–0.72 ng/mL (MeP) in influent 8 C. Mejías, J. Martín, J.L. Santos et al. Environmental Technology & Innovation 32 (2023) 103276 Fig. 6. Adsorption percentage (%) of MeP, EtP, PrP and BuP on PA MPs in the evaluation of DOM. Fig. 7. Influence of environmental real matrices on the adsorption of PBs onto PA. wastewater, 0.17 ng/mL (EtP)–0.43 ng/mL (BuP) ng/mL in effluent wastewater and none in tap water and surface water) were subtracted to calculate the removal efficiency (Table S8). Water sources characterization (conductivity, pH, chemical oxygen demand, PT, and NT) is presented in Table S9. Adsorption capacity in real water matrices decreased slightly with matrix complexity (Fig. 7) in comparison with distilled water experiments at the same particle size. Influent wastewater possesses the minor adsorption capacity of PBs which can be explained by the higher presence of DOM. For example, EtP presented an adsorption of 73% in distilled water, 78% in tap water and 68% in influent wastewater. The differences were higher for BuP. Its adsorption onto PA decreased from 94% in distilled water and 95% in tap water to 29% in influent wastewater. Arvaniti et al. (2022) also observed that the adsorption of valsartan and losartan was inhibited in PS MPs by increasing matrix complexity. Table S10 shows the correlation analysis of PBs adsorption onto PA MPs with the physico-chemical characteristics of water 9