Adsorption Mechanism of a Perfluorinated Carboxylic Acid on Charged Polystyrene Surfaces in Aquatic Environment: A Combined Experimental and Computer Simulation Investigation

Abstract

Interactions between per- and polyfluoroalkyl substances (PFASs) and small plastic particles (PPs) in the micro- and nanosize ranges raise serious concerns due to the possible joint harmful physiological effects of these emerging contaminants on the ecosystem. In this study, the adsorption of undecafluorohexanoic acid (UFHA), a C6 representative of PFASs, on amidine-functionalized polystyrene PP is investigated by using experimental and computer simulation methods. Electrophoretic light scattering measurements reveal that UFHA adsorbs strongly on oppositely charged PP leading to charge neutralization and overcharging at appropriate concentrations. Charged UFHA–PP adducts form stable colloids, while rapid aggregation was observed at the isoelectric point. Simulation results indicate that, at least at isoelectric conditions, a H-bond is formed between the oppositely charged groups of the PP and UFHA, providing an additional driving force, besides charge pairing, for this adsorption. This strong interaction results in an association of the PP chains and the majority of UFHA molecules, although some of the UFHA molecules can still stay isolated from these associates. It is found that the ionic strength applied in the solutions influences the adsorption process through charge screening of the PP surface, while it does not affect the charge neutralization conditions. Accordingly, the spatial distribution of the Na⁺ and Cl^– ions is found to be largely uncorrelated with that of the polymeric chains and UFHA molecules at every salt concentration studied. The findings obtained provide new insights into the adsorption mechanism of UFHA onto a charged polystyrene surface and the features of the new interfaces formed, which play a key role in the joint migration and toxicity of these emerging contaminants.

1. Introduction

The issue of environmental plastic pollution has become an increasingly pressing contemporary problem. − Owing to the industrial and the consumer–social activities, tons of plastic debris are being released to the environment each year due to the improper waste management. − The plastic objects that enter and accumulate in nature undergo fragmentation processes leading to the formation of micro- and nanosized plastic particles (PPs), as a result of various environmental effects, like UV radiation, mechanical impacts, or biochemical degradation. − The smaller size fractions of PPs easily enter into living organisms through nutritional and respiratory activities, so thereby they are able to cross the biological barriers and, hence, already were detected in human tissues.

PP pollutants are often found in association with per- and polyfluoroalkyl substances (PFASs), synthetic industrial compounds of high chemical stability and considerable ecological concerns. − As PPs may serve as vectors for their transport and accumulation in nature, − understanding the adsorption mechanisms of PFASs onto plastic surfaces and the features of the new interfaces formed upon such an adsorption process is crucial for assessing their environmental fate and potential ecological impacts. −

Recent research has highlighted the complex physicochemical interactions that govern the adsorption of PFASs on PPs, influenced by factors, namely, particle size, surface chemistry, and environmental conditions. For instance, studies have shown that smaller PPs exhibit a higher sorption capacity for PFASs due to an increased surface area and altered surface properties resulting from weathering processes. , In addition, adsorption kinetics data indicate that PFASs can rapidly partition between plastic surfaces and bulk solutions, reaching equilibrium within hours to days. The role of the perfluoroalkyl chain length in the adsorption process was revealed with light scattering techniques using polystyrene PP and it was found that a longer chain length leads to stronger adsorption, which also influences the colloidal stability of the PP–PFAS systems. In addition, it was demonstrated that biofilms forming on plastic surfaces in the environment play a decisive role in PFAS–plastic interactions and significantly enhance the vector potential of PP for other toxic pollutants too.

Experimental investigations can be very efficiently supplemented by computer simulations, as in a simulation such a deep, atomistic insight into the structure, interactions, and dynamics of a suitably chosen model of the system of interest is obtained that cannot be provided by any experiment alone. Indeed, adsorption at various surfaces, such as clay minerals, , metal oxides, , zeolites, , self-assembled organic monolayers, , various carbonaceous materials, , protein crystals, icy surfaces, − and also nanosized plastic particles, has been studied several times by computer simulation methods. Clearly, experimental and computer simulation investigations have a synergistic effect, as the combination of these approaches provides deeper insights into the system of interest than any of them could alone. However, the adsorption kinetics and corresponding surface properties have, to the best of our knowledge, never been assessed simultaneously by experimental and theoretical methods in the same PP–PFAS systems.

Therefore, this study aims at elucidating the interactions of the PFAS molecule undecafluorohexanoic acid (denoted here as UFHA) with the amidine-functionalized, positively charged polystyrene PP surface through a systematic combination of experimental and theoretical approaches. Surface charge properties were investigated in a wide UFHA concentration and ionic strength range by the electrophoretic light scattering technique, while the UFHA–polystyrene adsorption mechanism was studied in simulations at the level of intermolecular interactions. By integrating these methodologies, the results aim to contribute valuable knowledge to the field of surface and environmental chemistry, particularly regarding the implications of joint pollution of similar emerging contaminants in aquatic systems.

The remainder of this article is organized as follows. The methodological issues, including materials and methods used in experiments as well as details of the molecular dynamic simulations, including adjustment of the potential model employed, are described in Section . The obtained results are then presented and discussed in detail in Section . Finally, in Section , the main conclusions of this study are summarized.

2. Methods

2.1. Experiments

2.1.1. Materials

All chemicals were used as received without further purification. Undecafluorohexanoic acid (UFHA) (≥97.0%) was obtained from Sigma-Aldrich. Laboratory-grade HCl (35%) and NaCl (∼99.8%) were bought from VWR. The 2,2′-azobis­(2-methylpropionamidine)­dihydrochloride (≥98.0%) and styrene (99%) were purchased from Acros Organics, while the Hellmanex III cleaning agent was bought from Hellma. The PP stock was synthesized by emulsion polymerization, as detailed elsewhere. For sample preparation, ultrapure water was used, which was obtained using an Adrona purification system. The water and salt solutions were filtered with PVDF-based 0.1 μm syringe filters from Millex-VV. The pH was set to (4.0 ± 0.2) throughout all the experiments by the addition of an appropriate amount of HCl and was checked with a combined glass electrode connected to a benchtop pH meter (WTW inoLab). The ionic strength was adjusted with NaCl solutions.

2.1.2. Electrophoretic Light Scattering

To determine the electrophoretic mobility (μ_e) of PPs, a Litesizer 500 device (Anton Paar) was used, which is equipped with a 658 nm wavelength laser source. The applied voltage was in the range of 0.1–200 V depending on the charge of the particles, and the temperature was set to (25 ± 0.2) °C in all experiments. The primarily measured μ_e data were converted into zeta potentials (ζ) with the Smoluchowski equation:

$$\zeta =\frac{{\mu }_{\mathrm{e}}\eta }{\epsilon {\epsilon }_0}$$ 1

where ε₀ is the permittivity of the vacuum, ε is the relative permittivity of the medium, and η is the dynamic viscosity of the solvent. The samples were prepared in a total volume of 4.0 mL, and the measurements were carried out in omega cuvettes (Anton Paar). The particle dose was fixed to 25 mg/L and the concentrations of NaCl and UFHA were systematically varied in the range of 10–1000 mM and 0.01–10 mM, respectively. During sample preparation, the following order was applied. First, calculated amounts of UFHA, NaCl, and water were mixed in a total volume of 3.6 mL, and then, 0.4 mL of previously sonicated PP stock dispersion (250 mg/L) was added to the samples, which were equilibrated at room temperature for 24 h. Eight consecutive measurements were implemented, and the average of the determined values is reported with the experimental error obtained.

2.1.3. Dynamic Light Scattering (DLS)

The particle size was determined with the same Litesizer 500 device and in the same samples as in the electrophoretic measurements discussed above using DLS. The correlation function was collected for 20 s and fitted with the cumulant method to obtain the translational diffusion coefficients, which were converted to hydrodynamic radii using the Stokes–Einstein equation. Five runs were performed, and the average values are reported.

2.2. Computer Simulations

2.2.1. Setting the Systems to Be Simulated

To complement the light scattering measurements, computer simulations of the UFHA–PP system have also been performed in an aqueous environment with different NaCl concentrations. The simulation system has been set in such a way that the monomer of the positively charged amidine-functionalized polystyrene polyelectrolyte was 4-ethenylbenzene-1-carboximidamidium, while the UFHA molecule is in its negatively ionized form due to its high acidity (pK _a = −0.16). The PP has been mimicked by an atactic 15-mer styrene model. To reproduce the experimental circumstances, this 15-mer has been in its fully dissociated, cationic form. General Amber Force Field (GAFF) , has been used to generate the necessary charge parameters for the polyelectrolyte and UFHA species, as GAFF had been proven to be successful in describing systems containing polyelectrolytes. According to this protocol, geometry optimization for the 15-mer styrene model has been carried out using the B3LYP/6–31G­(d) level of theory, employing the compatible polarizable continuum model (PCM) to achieve the solvation effects of water implicitly. We have repeated this calculation by including Grimme’s D3 empirical dispersion corrections; however, both the geometry of the polymer and the fractional charges assigned to the atoms turned out to be largely insensitive to this correction. The necessary quantum chemical calculations have been carried out using the Gaussian 09 package. Atomic charges have been calculated from the optimized geometries using the Restrained Electrostatic Potential (RESP) fitting scheme with the Antechamber program.

It should be emphasized that the main challenge here was to assign appropriate charges for modeling the polyelectrolyte 15-mer styrene molecule. Although we are aware of several recently proposed force fields to model perfluorinated compounds, , having an appropriate model of the polyelectrolyte nanoparticle developed, UHFA has needed to be described in a consistent way, and therefore the same procedure has also been adopted here.

2.2.2. Molecular Dynamics Simulations

Molecular dynamics simulations of the system consisting of amidine-functionalized polystyrene oligomers and UFHA molecules in aqueous NaCl solutions of various concentrations have been performed on the isothermal–isobaric (N,p,T) ensemble at a temperature of 300 K and pressure of 1 bar. The cubic basic simulation box has consisted of 10,000 water molecules, 2 oligomeric chains of functionalized polystyrene, each consisting of 15 residues, and 30 UFHA molecules. To address the effect of the salt concentration, the number of NaCl ion pairs has been varied from simulation to simulation, being 0, 5, 18, 54, and 200 in the different systems, corresponding to the salt concentrations of roughly 0 mM, 25 mM, 100 mM, 300 mM, and 1000 mM, respectively. In accordance with the experimental situation, all residues of the polymers have been in their cationic form, while all UFHA molecules in their anionic form; hence, the simulation box has been neutral. The schematic structure of the monomeric units of the polymer and of UFHA in their fully charged forms is illustrated in Figure . Polymers, UFHA molecules, as well as Na⁺ and Cl^– ions have been described by the standard GAFF, while water molecules have been modeled by the rigid, three-site, AMBER-compatible , SPC/E potential. According to these potential models, the intermolecular energy of the system has been pairwise additive, consisting of charge–charge Coulomb and Lennard-Jones terms.

1.

Schematic structure of the residues of the amidine-functionalized polystyrene PP and the UFHA molecule in their charged forms. Atoms considered later in radial distribution function calculations are marked by red asterisks.

Simulations have been done with the GROMACS 2024.3 program package. Equations of motion have been integrated by the leapfrog algorithm in time steps of 2 fs. Standard periodic boundary conditions have been applied. The geometry of the water molecules has been kept fixed by means of the SETTLE algorithm, while all other chemical bonds involving an H atom have been constrained using the LINCS algorithm. All interactions have been truncated to zero beyond the cutoff distance of 12 Å based on the center of the molecules and monomeric units. The long-range part of both the Coulomb and Lennard-Jones interactions has been taken into account by the Particle Mesh Ewald (PME) method. , Temperature coupling has been applied using the velocity rescaling thermostat with a stochastic term with a coupling constant of 1 ps, while pressure has been controlled by the Parrinello–Rahman barostat with the coupling time of 5 ps.

To prepare the initial configurations, the two 15 residues long polymer chains have been inserted into a cubic simulation box with 68.3 Å long edges, along with the 30 UFHA molecules in random positions. The basic box has then been filled by the water molecules and required the number of NaCl ion pairs. All systems have undergone the steepest descent minimization, followed by a 500 ps long equilibration on the canonical (N,V,T) ensemble and an at least 100 ns long equilibration on the isothermal–isobaric (N,p,T) ensemble. In the production phase of the simulations, a 200 ns long trajectory has been generated for each system, in which 10,000 equilibrium sample configurations, separated by 20 ps long trajectories each, have been dumped for the analyses. An equilibrium snapshot of each system simulated is shown in Figure .

2.

Equilibrium snapshot of the systems containing 0 mM, 25 mM, 100 mM, 300 mM, and 1000 mM NaCl (from left to right), as obtained from our simulations. Atoms of the polymer chains and UFHA molecules are shown by blue and red color, respectively, Na⁺ and Cl^– ions are represented by orange and green balls, respectively, while water molecules are only shown by faded light-blue sticks for clarity.

3. Results and Discussion

3.1. Experimental Results

To explore the interaction between UFHA and polystyrene PP, zeta potentials were determined in electrophoretic light scattering measurements. Note that UFHA is negatively charged due to its low pK _a value; hence, its carboxylic group is deprotonated under the experimental conditions applied, while PP possesses a net positive charge owing to the presence of protonated amidine surface functional groups. The effect of UFHA adsorption on the oppositely charged PP surface has been followed by zeta potential measurements, which reflects the surface charge of the particles (see Figure ).

3.

Zeta potential values (calculated from the electrophoretic mobility values with eq ), measured for amidinylated polystyrene PPs as a function of the UFHA concentration (bottom axis) or dose (top axis) at various ionic strengths adjusted with NaCl. The chemical structure of UFHA in its deprotonated form is shown in the inset.

The generic tendency in the data can be described by a continuous decrease of the zeta potentials by increasing the UFHA concentration at all ionic strengths studied. The potentials are positive at low UFHA doses due to the positively charged bare PPs and are around zero at intermediate concentrations owing to the charge neutralization of the surface upon adsorption of UFHA. The isoelectric point is found to be the same within the experimental error at each ionic strength, indicating that UFHA adsorbs much stronger than chloride anions on PP. Such a strong adsorption leads to charge inversion at higher concentrations, as indicated by the negative zeta potential values determined. Similar charge reversal has already been reported with plastic particles and PFAS in other systems. ,

These results clearly confirm the adsorption of UFHA on the PP surface. As discussed above, the ionic strength did not influence the charge neutralization points; nevertheless, the zeta potentials at low and high UFHA concentrations change significantly by varying the background NaCl level in the samples, as seen from Figure a. At a high UFHA dose (10 mM), the particles are negatively charged at all ionic strengths investigated. However, a clear minimum in the potential data can be observed in the intermediate salt concentration regime. This can be explained by the electrokinetic effect. Accordingly, the Debye screening length, which determines the extent of the electrical double layer around charged particles in electrolyte solutions, is decreasing with the ionic strength leading to more compressed double layers at high salt concentrations. Thus, the double layer is rather broad at low ionic strength, giving rise to the appearance of friction forces, which slow charged particle movements in an electric field, like the situation in the electrophoretic light scattering measurements performed in the present work. Such a decrease in the velocity of the PPs results in a lower magnitude of the electrophoretic mobility and, subsequently, to smaller zeta potentials, which gradually decrease by further increasing the salt concentration. In this regime, the electrical double layer is compressed, and the friction forces are negligible. The intermediate minimum in the data obtained for negatively charged particles (Figure a) originates from this phenomenon. At NaCl concentrations beyond this minimum, the surface charge is screened by the dissolved ions, and the zeta potential increases; i.e., the magnitude of the surface charge decreases.

4.

(a) Zeta potentials measured in the PP–UFHA systems at 0.01 mM (red circles) and 10 mM (violet squares) UFHA concentrations at different NaCl levels to vary the ionic strength. (b) Hydrodynamic radii as a function of the UFHA concentration at 20 mM ionic strength. The solid lines serve as guides to the eye.

The effect of the ionic strength on the zeta potentials is also visible for positively charged particles at a 0.01 mM UFHA concentration in the data shown in Figure a. However, the intermediate maximum, which can be explained in the same way as the minimum for the negatively charged particles above, is not that pronounced, while potentials close to zero were recorded at the highest NaCl concentrations studied. Accordingly, at elevated ionic strengths, the magnitude of the zeta potentials is significantly lower in this case compared to the negatively charged particles. This result clearly indicates different interfacial features upon adsorption of UFHA on PP in comparison to the bare PP surface. From these data, one can conclude that PFAS adsorption led to a higher charge magnitude, which cannot be screened to zero by the dissolved ions. This finding has significant environmental relevance, since it indicates that PP–PFAS compounds may remain highly charged even in electrolyte media and, hence, this affects their migration and accumulation in aqueous compartments.

The hydrodynamic radii of the particles were measured in the same UFHA concentration regime as above with DLS. In general, the size values were close to the one determined for the individual PPs at UFHA doses below and above the isoelectric point, while large radii of a few micrometer dimensions were measured close to the charge neutralization condition (around 1 mM UFHA concentration), as illustrated in Figure b. Such a charge–size relation can be qualitatively explained by the theory developed by Derjaguin, Landau, Verwey, and Overbeek (DLVO) to describe the stability of charged colloidal particles. This model assumes the presence of attractive dispersion forces and electrostatic repulsion between charged surfaces. The latter one originates from the overlapping electrical double layers and its strength strongly depends on the surface potential. The DLVO theory states that the nature of the overall interaction is governed by the sum of these forces. Therefore, in the present systems, the PPs form stable dispersions away from the isoelectric point, at which conditions they possess significant charge, and thus, electrical double layer repulsion exceeds the attractive forces. However, at charge neutrality due to UFHA adsorption, double-layer repulsion vanishes and attractive dispersion forces become predominant leading to rapid aggregation of the particles and to large hydrodynamic radii around the isoelectric point (Figure b).

3.2. Computer Simulation Results

3.2.1. Polymer–UFHA Interaction

To investigate the interaction between the polymer nanoparticles and (deprotonated) UFHA molecules at the atomistic level, we have calculated the radial distribution functions (RDFs) between the amidine group C atoms of the PP and carboxylic C atom of UFHA, g _CC(r), between the N atoms of PP and O atoms of UFHA, g _ON(r), and between the amidine H atoms (attached to N) of the PP and O atoms of UFHA, g _OH(r), in all systems simulated. The obtained RDFs are plotted in Figure .

5.

Radial distribution functions of the amidine C atoms of the PP and carboxylic C atoms of UFHA (bottom panel), N atoms of the PP amidine group and carboxylic O atoms of UFHA (middle panel), and H atoms of the PP amidine group and carboxylic O atoms of UFHA (top panel), as obtained from the simulations of the systems of 0 mM (black), 25 mM (red), 100 mM (green), 300 mM (blue), and 1000 mM (orange) NaCl concentrations. The inset shows the amidine C–carboxylic C RDF as obtained in the salt-free system on a larger scale. The dashed horizontal line corresponds to the RDF value of unity. Atoms corresponding to these radial distribution functions are marked by red asterisks in Figure .

The most striking feature of the obtained functions is that they do not damp to unity in the distance range compatible with the size of the simulation box; instead, they drop below unity around 25 Å and remain less than one in the distance range accessible from the simulations. This is illustrated in the inset of Figure on the example of g _CC(r) in the salt-free system. This behavior suggests a strongly inhomogeneous distribution of the UFHA molecules within the basic box, namely, that they are strongly accumulated at the vicinity of the polymer chains, while they only occur with small probability at large distances from them. It is also seen that none of the three RDFs considered are sensitive to the salt concentration. These findings are in clear accordance with the experimental results, namely, that (i) deprotonated UFHA molecules are adsorbed at the oppositely charged polymer surface and (ii) at the isoelectric point (which corresponds to the situation in the simulations), this adsorption is independent from the salt concentration.

The g _CC(r) function exhibits a marked maximum around 4.5 Å. Integration of this peak up to the first minimum position at 5.3 Å results in the first-shell coordination number value of 1 ± 0.1 in all systems simulated, indicating that, on average, each amidine group is in close contact with one UFHA COO^– group. This is understandable, considering that this way, the positive charge carried by the amidine group of a monomeric unit is neutralized by the negative charge of the nearby UFHA COO^– group.

The g _ON(r) and g _OH(r) functions both exhibit a very sharp peak at short distances, i.e., at 2.8 and 1.8 Å, respectively, followed by a marked second peak at about 2 Å away from the first one. The integration of the first peak results in a coordination number value of roughly 0.5 for g _ON(r) and 0.25 for g _OH(r), while the coordination number of the first two peaks turns out to be about 2 and 1, respectively. The position of the first peaks indicates that, besides the obvious charge pairing, H-bonding also occurs between the two oppositely charged ionic groups, in which a NH₂ hydrogen atom of the PP is donated to one of the carboxylic oxygens of UFHA. The first peak coordination number suggests that such a H-bond is formed, on average, by one out of two N atoms and one out of four NH₂ hydrogens. Considering that each amidine group contains two N and four H atoms, this result reveals that, on average, each amidine group of the PP forms one H-bond with the closely coordinated UFHA molecule. This H-bonding is illustrated in Figure on the example of a few monomeric units, UFHA pairs taken out from equilibrium configurations of our simulations.

6.

Illustration of the H-bonding occurring between a monomeric unit of the polymer and an UFHA molecule on pairs taken out from equilibrium configurations of our simulations. C, F, O, N, and H atoms are shown by light-blue, pink, red, dark-blue, and white balls, respectively. H-bonds are indicated by dashed black lines.

The presence of the clear second peak of the g _ON(r) and g _OH(r) functions simply reflects the fact that the amidine and carboxylic groups have more than one NH₂ group and O atom, respectively. Thus, the second peak is given by the atoms of the non-H bonding NH₂–O pairs of the amidine groups and the closely attached UFHA molecules.

3.2.2. Competition of the Charged Groups with the Na⁺ and Cl^– Ions

To address the competition of the charged groups of the PP and UFHA with the Na⁺ and Cl^– ions, respectively, for charge pairing, we have also calculated the RDFs between the amidine N atoms and Cl^– ions, g _NCl(r), and between the carboxylic O atoms and Na⁺ ions, g _ONa(r), in the salt-containing systems simulated. The obtained RDFs have again turned out to be practically independent from the concentration of NaCl (apart from the obvious increase in the statistical noise with decreasing salt concentration).

The g _NCl(r) and g _ONa(r) functions, obtained in the system of the largest (i.e., ∼1 M) salt concentration, are shown in Figure . For reference, the g _ON(r) function obtained in the same system is also shown. As is seen, the g _NCl(r) and g _ONa(r) functions have a rather similar shape than g _ON(r), i.e., a sharp first peak is followed at about 2 Å larger distances by a marked second one. This shape suggests that when approaching the charged amidine or carboxylic group, the Cl^– or Na⁺ ion gets as close as possible to one of the two N or O atoms, respectively, rather than being accommodated between them. The relative positions of the respective peaks of the three RDFs simply reflect the size difference between the Na⁺ and Cl^– ions, as well as the positions of the O and N atoms.

7.

Radial distribution functions of the carboxylic O atoms of UFHA and Na⁺ ions (red), and amidine N atoms of PP and Cl^– ions (blue) as obtained from the simulation of the system of 1000 mM NaCl concentration. For reference, the amidine N–carboxylic O RDF in the same system is also shown (black). The scales at the left and right sides correspond to the g _NCl(r) and g _ONa(r) and to the g _ON(r) functions, respectively.

More importantly, the peaks of the g _ON(r) function are about 2 orders of magnitude higher than those of g _NCl(r) and g _ONa(r) (note the different scales used for these functions in Figure ). Correspondingly, the coordination numbers of the first as well as the first two peaks are 1–2 orders of magnitude larger for g _ON(r) than for g _NCl(r) and g _ONa(r) (the exact ratio obviously depends on the salt concentration). Finally, the coordination number of the first two peaks of g _NCl(r) and g _ONa(r) increases more or less linearly with the salt concentration, indicating that the position of the Na⁺ and Cl^– ions in the system is largely uncorrelated with that of the PPs and UFHA molecules. These findings indicate, in clear accordance with the experimental findings, that the amidine and carboxylic groups of the PP and UFHA, respectively, strongly outcompete the Na⁺ and Cl^– ions for charge pairing with each other even at the largest salt concentration considered.

3.2.3. Associates of the Polymer Chains and UFHA

The strong interaction between the UFHA molecules and the polymer chains suggests a large degree of association of these particles, which might well include self-association of the polymeric chains. This view is supported by the experimental finding that coagulation occurs under isoelectric conditions, the shape of the RDFs describing the PP–UFHA interaction, and also by the snapshots shown in Figure . To investigate this point in more detail, we have calculated the radius of gyration, R _g, of (i) the individual polymer chains, (ii) the two polymer chains together, and (iii) the two polymer chains and 30 UFHA molecules altogether. The distribution of R _g is shown in Figure as obtained in these three cases in the five systems simulated, while their mean values and standard deviations are collected in Table .

8.

Distribution of the radius of gyration of the single polymer chains (black solid lines), the two polymer chains together (red dashed lines), and the two polymer chains and 30 UFHA molecules together (green dash-dotted lines), as obtained from the simulation of the systems consisting 0 mM (top panel), 25 mM (second panel), 100 mM (third panel), 300 mM (fourth panel), and 1000 mM (bottom panel) NaCl.

1. Mean Value and Standard Deviation of the Radius of Gyration (in Å Units) of the Single Polymer Chains, the Two Polymer Chains Together, and the Two Polymer Chains and 30 UFHA Molecules Together in the Five Systems Simulated.

NaCl concentration	single polymer chains	two polymer chains together	polymer chains and UFHA together
0 mM	10.81 ± 0.26	12.43 ± 0.24	15.98 ± 2.76
25 mM	10.78 ± 0.30	12.55 ± 0.21	15.72 ± 2.15
100 mM	11.05 ± 0.21	12.20 ± 0.19	16.77 ± 3.05
300 mM	10.88 ± 0.24	13.47 ± 0.35	17.00 ± 2.78
1000 mM	10.79 ± 0.52	12.34 ± 0.16	15.26 ± 2.63

As seen, the radius of gyration of the individual polymer chains is around 11 Å in every system. Considering that the distance of the two chain-end C atoms of these 15-mers is about 37 Å in the fully stretched conformation of the chain, this mean value indicates that the polymer chains have a rather compact, folded shape in these systems. More interestingly, the mean radius of gyration of the two polymer chains together is only 1–2 Å larger than that of the individual chains. This finding clearly shows that polymer chains are associated in these solutions. Considering that all of the monomeric units bear the same net positive charge, this folded conformation and self-association implies the presence of the negatively charged UFHA particles in these associates. This is confirmed by the fact that the maximum position of the P(R _g) distribution considering both the polymers and UFHA molecules occurs only about 1–3 Å larger values than that when only the polymer chains are considered. However, the R _g distribution of the polymers and UFHA molecules together is rather broad and extends to larger distances; thus, unlike the maximum position, the mean value of R _g is already 5–7 Å larger here than in the distributions considering only the polymers (see also Table ). The considerable broadening of the P(R _g) distribution and their extension to large radii when including also the UFHA molecules in its calculation clearly indicates that although the majority of the UFHA molecules are part of the polymer–UFHA associates, some of them are still unbound, move rather freely in the system, and hence can be rather far from these associates.

Finally, to characterize the general shape of the association of the two polymer chains, we have calculated the ratio of the mean values of its three principal moments of inertia, I ₁ ≤ I ₂ ≤ I ₃, in all systems simulated, and plotted the values of I ₂/I ₃ against I ₁/I ₂ in Figure . For comparison, the same analysis has also been performed for the individual 15-mer chains. For spherical-like objects, I ₁ ≈ I ₂ ≈ I ₃ and, hence, both I ₁/I ₂ and I ₂/I ₃ are close to unity. On the other hand for prolate-like, cylindrical objects I ₁ ≪ I ₂ ≈ I ₃, while for disk-like oblate shapes I ₁ ≈ I ₂ ≪ I ₃. Thus, in the former and latter case, the ratios I ₁/I ₂ and I ₂/I ₃, respectively, are close to zero, while the other ratio is still close to unity. In other words, prolates, oblates, and spherical objects appear in the upper-left, lower-right, and upper-right corner, respectively, of the I ₂/I ₃ against I ₁/I ₂ morphology map.

9.

Morphology map of the individual polymer chains (green) and their association (red), as obtained in the simulated systems of different salt concentrations. Error bars are in the order or below 0.05.

As seen, the individual polymer chains give their contribution to the upper-left corner of the morphology map, with their I ₁/I ₂ and I ₂/I ₃ values scattering around 0.2 and 0.95, respectively. This is a clear indication that the individual polymer chains are, not surprisingly, elongated, prolate-like objects. The points corresponding to the association of the two chains, however, are substantially shifted toward the upper-right corner of the map with respect to those of the individual chains; as both the I ₁/I ₂ and I ₂/I ₃ values scatter between about 0.55 and 0.8 in every case. This finding clearly indicates that the association of the two elongated chains is a much more spherical object than the individual chains constituting it.

4. Conclusions

In this paper, we studied the interaction of two increasingly abundant and harmful environmental contaminants, namely, small amidine-functionalized polystyrene model plastic particles and UFHA, a typical representative of perfluoroalkyl substances at various UFHA-to-PP ratios and in the presence of salts. To combine the advantages of the experimental and computer simulation approaches, we performed both electrophoretic and dynamic light scattering measurements, as well as molecular dynamics simulations.

Measurements of the zeta potential revealed a strong adsorption of UFHA at the PP surface. Namely, the zeta potential, being positive at low UFHA concentrations due to the positive charge carried by the protonated amidine functional groups of the PP, decreases with increasing UFHA concentration as UFHA stays deprotonated under the experimental conditions studied. The increase in the UFHA concentration thus leads to charge neutralization and a subsequent overcharging. Computer simulations revealed that this strong adsorption is caused not only by the obvious charge pairing but also by H-bonding between the amidine groups of the PP and the COO^– groups of UFHA.

We also found that the UFHA concentration corresponding to charge neutralization is independent of the ionic strength. This is explained by the considerably stronger affinity of the UFHA COO^– and PP amidine groups to each other than that of them to the Na⁺ and Cl^– ions, respectively, of the background salt. Far from the charge neutrality conditions, however, we found that the zeta potential varies with the salt concentration and exhibits an extremum. This behavior can be explained by the electrokinetic effect, namely, by the opposite effects of the decreasing dimension of the electrical double layer and increasing screening of the surface charge, both induced by the increasing salt concentration, on the zeta potential.

Dynamic light scattering results revealed low colloidal stability and aggregation of PP around the isoelectric point, while stable dispersions were observed below and above it, in accordance with the DLVO theory. Correspondingly, strong association involving both polymeric chains and the majority of the UFHA molecules was found in the simulations, performed always under an isoelectric condition.

Finally, it should be emphasized that, in addition to their general interest in colloid chemistry, the present results also have certain environmental relevance. In particular, the H-bonding that occurs between the amidine groups of the PP and the COO^– group of UFHA in addition to their charge pairing considerably increases the thermodynamic driving force of the association of these harmful pollutants. Further, the fact that the surface of the PP–UFHA associates can remain highly charged even in systems of high ionic strengths has important consequences on their accumulation and migration in salty aqueous environments, such as in seawater. These findings are particularly alerting in the light of the recent finding that PPs and PFAS have a strong synergistic effect concerning their toxicity on living ecosystems.

⊥.

K.B. and H.T.U. are co-first authors.

The authors declare no competing financial interest.

Published as part of The Journal of Physical Chemistry B special issue “Physical Chemistry of Microplastics and Nanoplastics”.