3.1. Interaction between PFOA and ZnO-NPs
Particle size: Alterations in the particle size and surface charge of the ZnO-NPs were observed following interaction with various concentrations of PFOA. With the addition of various concentrations of PFOA to the ZnO-NPs (after 1 day of interaction), the particle size of the ZnO-NPs at the peak of the particle size distribution curve (PSDC) remained between 157 and 186 nm, and the particle size ranged from 106–955 nm (
Figure 1a and
Figure S1a). The slight variations in the particle size at the peak of the PSDC and within the particle size ranges after 1 day of interaction may be due to less interaction time between the nanoparticles and the organic compound (PFOA). It is suggested that interactions between ZnO-NPs and PFOA as organic pollutants could be time-dependent after their coexistence was ascertained.
A significant increase in the particle size of the ZnO-NPs was observed after 1 week of interaction with PFOA. The size of the ZnO-NPs at the peak of PSDC increased from 166 nm to 203 nm, and similar trends were observed in the presence of PFOA up to 50 µg/L PFOA (
Figure 1a and
Figure S1b). However, the size ranges of ZnO-NPs in the presence of PFOA from 100 µg/L to 50 mg/L were not measurable via a Zetasizer, with a maximum size ranging from 0.3 nm to 10 microns. It could be assumed that the size of the ZnO-NPs after 1 week of interaction with PFOA (from 100 µg/L to 50 mg/L) is more than 10 microns. The alteration in the size distribution of the ZnO-NPs was influenced by the presence of PFOA through aggregation and the magnitude of the surface coating/adsorption of organic substances [
14,
22,
36]. Overall, an increase in the particle size of the ZnO-NPs was observed alone and in the presence of PFOA from 1 day to 3 weeks of interaction. This increase might be due to the presence of large (agglomerated) and/or sedimenting particles resulting from particle–particle interactions, electrostatic interactions, and hydrophobic interactions. The ZnO-NPs were monodispersed (particles of uniform size) from the time of the nanoparticle suspension preparation to a few days later. However, after 1 week of interaction, nonuniform (polydisperse) behaviour of the nanoparticles was observed. Overall, particles from 1 week to 3 weeks were found in their polydispersed form, which was also reflected by their surface charge.
Zeta potential: The zeta potential (
Figure 1b) of the ZnO-NPs did not significantly change after 1 day of interaction with the addition of different concentrations of PFOA, while a significant decreasing trend was observed for samples after 1 and 2 weeks of interaction. The alterations in the charge potentials suggested that PFOA coating/adsorption on the surface of the ZnO-NPs decreased in the magnitude of surface charge values. The zeta potential values also suggested that the ZnO-NPs became less stable in solution and tended to agglomerate (increasing size,
Figure 1a and
Figure S1b) with increasing concentrations of PFOA. This aggregation might be attributed to the higher molecular weight and greater surface coating of PFOA on the ZnO-NPs via hydrophobic and van der Waals interactions [
36,
37,
38]. The hydrophobic tail of PFOA has a decreased tendency to interact with water molecules, while the hydrophilicity and charged head of PFOA could lead to a greater tendency to attach to the ZnO-NP surface. The aggregation suggested the unstable nature of the ZnO-NPs after interactions with PFOA [
15,
16].
Adsorption: The adsorption analysis of PFOA in solution confirmed the interaction between ZnO-NPs and PFOA after 1 day and 2 weeks of interaction.
Figure 1c shows the increased adsorption of PFOA molecules on the surface of the ZnO-NPs after 2 weeks of interaction compared to after 1 day. Various interaction mechanisms could be involved in the process of PFOA adsorbing onto the surface of ZnO-NPs, which increased with time. An increase in the particle size and an overall decrease in the net charge potential of the surface nanoparticles were observed (
Figure 1a,b). Once the ZnO-NPs are in water, the formation of hydroxide layers (
on the surface of the nanoparticles due to hydrolysis is a common process, as water molecules can be adsorbed (both chemically and physically) onto the surface of the particles [
22,
36,
39]. This resulted in the formation of a quantity of positive charges on the surface of the nanoparticles, attracting PFOA
− molecules to adsorb to the surface of the ZnO-NPs by electrostatic interactions. After that, the hydrophobic tail of PFOA
− combined on the surface of the ZnO-NPs adsorbed the free PFOA
− molecule in the solution by hydrophobic interactions (hydrophobic PFOA molecules tend to accumulate more on the surface of the ZnO-NPs in aqueous media) [
40] and further increased the adsorption amount of PFOA after 2 weeks’ time interval compared to that after 1 day. This can also be corroborated by the decreased surface charge of the ZnO-NPs and the increase in particle size (
Figure 1a,b).
Dissolution: The dissolution of ZnO-NPs in terms of dissolved zinc (mg/L) after 1 day and 2 weeks of interaction with PFOA was determined via ICP–OES (
Figure S3). The concentration of dissolved zinc in the 0 µg/L sample (i.e., ZnO-NPs in buffered water) slightly increased after 2 weeks of interaction compared to that in the 1 day interaction. The dissolved zinc concentration increased in the presence of PFOA, which further decreased after 2 weeks of interaction compared to that for 1 day. This could be due to the aggregation/agglomeration and settlement of particles in the test tube, where coprecipitation occurred and reduced the dissolved amount of zinc in solution. The dissolution rate is often claimed to be directly proportional to the specific surface area of the material, meaning that the smaller the particles are, the faster dissolution occurs [
41]. Agglomeration has been proposed as a rationale for slow dissolution [
42]. Moreover, aggregation/agglomeration may have resulted in decreased dissolution [
42] as the particle size increased and the settlement of particles at the bottom of the test tube was observed. The surface properties of ZnO are quite complicated due to the presence of polar and nonpolar crystallographic planes [
43]. The interaction of PFOA molecules with nonpolar planes leads to agglomeration and less dissolution. A similar dissolution trend was observed in the presence of electrolytes and HA, as shown in the next sections.
TEM, XRD, FTIR, and Raman analysis: Alterations in the morphology of ZnO-NPs were observed after 1 day of interaction with PFOA using TEM (
Figure 2a). The original (virgin) ZnO-NPs (
Figure 2a) were in an agglomerated/spongy form with a particle size ≤ 100 nm. Both rod and spherical morphologies were observed. The lattice pattern and glittering spots/rings suggest the crystalline structure of the ZnO-NPs [
15,
44]. The dissolved zinc concentrations in Milli-Q water at various pH values for the same batch of ZnO-NPs were reported in our previous research [
15], which indicated minimal dissolution at pH ≥ 7. The changes in crystallinity after the interaction of ZnO-NPs with water were also examined using TEM analysis. TEM showed a diffraction pattern (
Figure 2a) and the crystalline/lattice arrangement of the original ZnO-NPs [
15,
44]. The mixture was dull and cloudy after interacting with buffer containing Milli-Q water. A more disordered/amorphous structure of the ZnO-NPs (
Figure 2a) was observed after interaction with PFOA. Similarly, compared with those of the pure powder, the diffuse and cloudy bright spots (
Figure 2a) represented the impure morphology of the ZnO-NPs after interaction with PFOA. The elemental composition obtained from TEM analysis is shown in
Figure S2a. The structural parameters, such as lattice spacing and the crystallite size of the ZnO-NPs after 1 day of interactions (ZnO-NPs in Milli-Q water and ZnO-NPs + 50 mg/L PFOA) were also measured using TEM analysis (
Figure S2b). Lattice spacing of 0.26 nm was found in the lattice structure of ZnO-NPs. However, the crystallite size ranging from 4.29 nm to 8.39 nm were observed after 1 day of ZnO-NPs interactions with 50 mg/L PFOA (
Figure S2b).
X-ray diffraction (XRD) analysis of powder ZnO-NPs and ZnO-NPs in buffered Milli-Q water and with 50 mg/L PFOA after 1 day of interaction is shown in
Figure 2b. Sharp peaks at 2θ values of 31.84°, 34.6°, and 36.5° are observed for the ZnO-NPs, which represent the (hexagonal wurtzite) crystal structure of the nanoparticles with three perfect alignments: (1 0 0), (0 0 2), and (1 0 1). These alignments match the defined standard powder diffraction (JCPDS, No. 36-1451) [
22,
45]. This result revealed no change in the crystal phase of the ZnO-NPs after 1 day of interaction. However, after 3 weeks of interaction (
Figure 2b), the intensities of the peak alignments at (1 0 0), (0 0 2), and (1 0 1) were suppressed. Two new peaks at 2θ values of 9.62° and 19.35° were observed, which could be due to the formation of new compounds, such as zinc hydroxide dihydrate Zn
5(OH)
10·2H
2O [
46] and zinc phosphate nanocrystalline materials [
47], respectively. The intensities of the peaks at 2θ values of 9.62° and 19.35° were lower for the ZnO-NP sample with 50 mg/L PFOA than for the ZnO-NP sample in buffered water, which could be due to the coating of PFOA molecules onto the surface of the ZnO-NPs (
Figure 2b).
FTIR analyses of the ZnO-NPs, PFOA, and ZnO + PFOA were also conducted (
Figure 2c). The peak at 450 cm
−1 suggested the presence of Zn–O [
15,
48], which is in the range of metal oxides (400–600 cm
−1) [
15,
49]. A very low-intensity peak at 450 cm
−1 was also observed for the sample of ZnO-NPs + PFOA after interaction, indicating the presence of ZnO-NPs after interaction with PFOA molecules. The peak at 640 cm
−1 is due to characteristic bands of organic halogen compounds (such as C–F stretching) [
50], which were identified in both PFOA and ZnO-NPs + PFOA after interacting with PFOA. The vibrational peaks at 1102, 1149, and 1204 cm
−1 may be due to C–F stretching, with the peak at 1102 cm
−1 being identified in the spectrum of the ZnO-NPs after interaction with PFOA. C–H stretching, O–H stretching, and C=O stretching was also identified at 1369, 1461, and 1623 cm
−1, respectively [
50,
51]. These peaks were identified in both the PFOA and ZnO + PFOA samples, indicating the link between PFOA and ZnO-NPs after their interaction. The peak at 2360 cm
−1 could be due to carbon dioxide from the atmosphere. The peak at 3424 cm
−1 is due to stretching of the water band [
15,
50]. The peak detected between 750 and 1050 cm
−1, at 950 cm
−1, could be due to K-potassium and P-phosphorous (from a buffer solution used to maintain pH 7) stretching with O and C in the ZnO-NPs after interaction with PFOA [
50]. The FTIR results revealed the presence of bonds in the ZnO-NPs after interaction with PFOA, suggesting that there was an association between PFOA and the ZnO-NPs. This is consistent with the elemental analysis results indicating the presence of F and C in the ZnO-NPs.
The Raman spectrum (
Figure 2d) also confirmed the presence of ZnO-NPs (such as at 430 cm
−1) in the pure ZnO-NPs samples in buffered water and in the presence of PFOA after 1 day of interaction. After 3 weeks of interaction, new peaks at 587, 934, 992, and 1370 cm
−1 were observed for both samples (such as ZnO-NPs in water and with PFOA) including ZnO-NPs with PFOA after 1 day of interaction (
Figure 2d). Zn−O stretches fall in the region between 350 and 600 cm
−1. The spectral range of Zn−OH bonds (which are also called OH linkages) is between 600 and 1200 cm
−1. The asymmetric stretches (with a high infrared intensity) are in the range 470−550 cm
−1, whereas the symmetric stretches were observed at 368 and 382 cm
−1 in the Raman spectrum. Bands below 350 cm
−1, as observed in the Raman spectrum, are attributed to lower-energy lattice modes [
46].
3.2. Influence of Electrolytes on PFOA and ZnO-NPs’ Interaction
The salinity of surface water and groundwater can vary considerably. Salinity is one of the most significant abiotic factors affecting the growth, metabolism, immunity, and survival of aquatic species in farming environments. Due to global climate change, evaporation of seawater, variations in local rainfall, and the whereabouts of ocean currents, environmental salinity in coastal areas alters frequently and violently. Under environmental stresses, physiological mechanisms are adaptively modulated to sustain body homeostasis, which can further impact the normal biological functions, comprising the immunity of the aquatic species [
52]. The presence of salts could influence the interaction between ZnO-NPs and coexisting contaminants, which determines the environmental fate of ZnO-NPs. The ionic strength, pH, and other organic materials present in the solution could influence the surface charge and stabilization of ZnO-NPs. Fixed concentrations (i.e., 5 mM) of monovalent and divalent salts (NaCl and CaCl
2) were used to investigate the effect of electrolytes on the stability of ZnO-NPs alone and in the presence of PFOA. The changes in the particle size and zeta potential of ZnO-NPs after various time intervals, i.e., after 1 day, 1 week, 2 weeks, and 3 weeks of interaction, are presented and discussed below.
Particle size: Alterations in the size of ZnO-NPs were observed alone (i.e., ZnO-NPs in buffered Milli-Q water and PFOA) and in the presence of salts with PFOA after 1 day, 1 week, 2 weeks, and 3 weeks of interaction (
Figure 3a and
Figure S4). The particle size of the ZnO-NPs in buffered water increased after 1 week of interaction compared to that after 1 day, e.g., 162.4 nm after 1 day to 206 nm after 1 week of interaction; these findings are consistent with the results shown in
Figure 1a (166 nm after 1 day and 203 nm after 1 week of interaction) and
Figure S1a,b. Similarly, an increase in the size of the ZnO-NPs was observed in the presence of 10 and 500 µg/L PFOA. A similar effect on the correlation between particle size and the surface charge of ZnO-NPs was observed for some samples, such as ZnO-NPs with 10 µg/L PFOA and ZnO-NPs with 500 µg/L PFOA (
Figure 3a,b).
For instance, in
Figure 1a, the size at the peak of the PSDC of ZnO-NPs with 10 µg/L PFOA is 173 nm after 1 day and 298.5 nm after 1 week, with a surface charge of −45.4 mV after 1 day and −1.9 mV after 1 week. The size of ZnO-NPs was also increased affecting the surface charge of ZnO-NPs at the same concentration (
Figure 3a,b); the larger the particle size, the greater the decrease in the magnitude of the surface charge values and vice versa. A similar observation (such as a drop in magnitude of surface change with the increase in the particle size) with surface charge was observed for different batches of samples of the same concentration (such as ZnO + 10 and 500 µg/L PFOA,
Figure 1a,b and
Figure 3a,b). After 2 and 3 weeks, the instability trend was similar, with the particle size being outside the machine range. This result showed that, initially, the stability of ZnO-NPs could vary based on Brownian motion (the random movement of particles due to bombardment by the solvent molecules that surround them). Normally, dynamic light scattering involves the measurement of particles suspended within a liquid and their shape. If the shape of a particle changes in a way that affects the diffusion speed, then the hydrodynamic size and surface charge may also change. This trend remained consistent with these concentrations (such as ZnO + 10 and 500 µg/L PFOA) in the next sections.
An increase in the particle size with a decreased surface charge was also observed in the presence of 5 mM NaCl with ZnO-NPs and with ZnO-NPs and PFOA after various time intervals (
Figure 3a,b). The ions in the medium and the total ionic concentration may affect the particle diffusion speed by altering the thickness of the electric double layer (the Debye length, K
−1). The resulting extended double layer of ions around the particles due to electrostatic interactions leads to a reduction in the diffusion speed and results in a larger, apparent hydrodynamic diameter. However, the presence of hydrophobic PFOA molecules, which can accumulate more easily around the surface of nanoparticles due to their hydrophobic nature than when suspended in water, may also alter the size and charge of the nanoparticles.
Generally, the stability of nanoparticles in the presence of electrolytes strongly depends on the capping agents used for stabilization [
53,
54]. Variations in the size of the ZnO-NPs with the co-occurrence of PFOA and electrolytes could be linked to adsorption, electrostatic interactions, hydrogen bonding, van der Waals effects, and cation bridging (Schematic 7). For instance, in the case of 5 mM NaCl, few alterations in the particle size were observed at the peak of the PSDC (
Figure 3a). However, the presence of 5 mM monovalent salts (NaCl) balanced the net electrostatic interactions between highly electronegative fluorine atoms and salt due to charge screening/shielding effects and London interactions [
54,
55].
The influence of divalent cations on the size of ZnO-NPs was different in the presence of PFOA. Alterations in the size and surface charge of the ZnO-NPs were observed by interacting the ZnO-NPs with 5 mM CaCl
2 alone or in the presence of PFOA after 1 day, 1 week, 2 weeks, and 3 weeks of interaction (
Figure 3 and
Figure S4). This increase in the size of the ZnO-NPs could be due to the accumulation of divalent cations, which results in a decrease in the diffusion speed and generation of larger particles compared to those of NaCl. It could also be expected that in the presence of 5 mM CaCl
2, the electrostatic repulsion between the negatively charged ZnO-NPs surface and PFOA molecules (which have a negative charge due to the anionic nature and high electronegativity) was reduced because of bridging interactions between the negatively charged surfaces of ZnO-NPs, divalent cations, and PFOA molecules. Consequently, the gathering of positively charged divalent cations enhanced the nanoparticle size due to bridging effects and made the nanoparticles more unstable, leading to agglomeration.
Cations, such as Ca
2+, could be the cause of the bridging phenomenon between carboxyl groups. Consequently, the adsorption of PFOA may be hindered in certain aqueous environments enriched with the aforementioned cations due to the decrease in electrostatic interactions between PFOA and the protonated surface [
56,
57,
58]. The same behaviour was observed for the ZnO-NPs in the following experiments.
Zeta potential: The zeta potentials of the ZnO-NPs (0.1 g/L nanoparticle dispersion, experimental batch one) were −47.2 and −35.7. −1.1, and −0.2 mV after 1 day, 1 week, 2 weeks, and 3 weeks of interaction, respectively (
Figure 3b). The decrease in the surface charge of the ZnO-NPs could be caused by the aggregation of nanoparticles resulting from van der Waals forces, hydrogen bonding, and hydrophobic interactions based on the aging factor. A similar decreasing trend in the magnitude of the surface charge was observed in the presence of PFOA (
Figure 3b). The presence of mono- and divalent salts also altered the surface charge of the ZnO-NPs after various durations of interaction with PFOA.
The presence of 5 mM NaCl did not significantly change the zeta potential of the nanoparticles after 1 day of interaction alone or in the presence of PFOA (
Figure 3b). It could be argued that the large extent of PFOA adsorption on the surface of nanoparticles balanced the overall electrostatic interaction forces between PFOA molecules and the monovalent salts based on shielding effects and London interactions [
54,
55]. However, the surface charge of the ZnO-NPs (in the presence of 5 mM NaCl) decreased to −0.5 mV after 3 weeks of interaction. This may indicate that aging affects the surface charge of nanoparticles, allowing more attachment of monovalent cations to the negatively charged surface of ZnO-NPs via electrostatic forces of attraction and van der Waals interactions. However, the same decreasing trend after various numbers of interactions was also observed in the presence of PFOA molecules (
Figure 3b).
Similarly, a decrease in the magnitude of the zeta potential of the ZnO-NPs was observed in the presence of divalent cations (5 mM CaCl
2) alone and in the presence of 5 mM CaCl
2 at various concentrations (such as 10 and 500 µg/L) of PFOA (
Figure 3b). The electrostatic repulsion between the ZnO-NP (negatively charged) surface and PFOA molecules was reduced in the presence of 5 mM CaCl
2. The gathering of positively charged ions on the surface was responsible for the decrease in zeta potential (less negative zeta potential) (
Figure 3b), which increased the particle size due to bridging effects and increased instability, leading to agglomeration (
Figure 3a and
Figure S4). The aggregation behaviour of ZnO-NPs in the presence of salts suggested that aging during the interaction of ZnO-NPs with salts affects the particle diffusion speed by changing the thickness of the Debye length due to the gathering of ions, resulting in agglomeration alone and in the presence of organic pollutants.
Adsorption: PFOA was analyzed to investigate the effects of salts (such as 5 mM NaCl and 5 mM CaCl
2) on the adsorption (interaction) of PFOA (10 and 500 µg/L) on ZnO-NPs after 1 day and 2 weeks of interaction (
Figure 3c). Increased adsorption (interaction) of PFOA molecules was identified with increasing concentration and interaction time (more sorption after 2 weeks than after 1 day). A similar trend with increased adsorption of PFOA was observed in the presence of 5 mM NaCl. A possible explanation for this result could be that the increase in the ionic strength of monovalent ions (Na
+) due to NaCl caused an increase in electrostatic attraction between the negatively charged ZnO-NP surface and the negatively charged PFOA molecules due to the presence of monovalent ions (Na
+) in between serving as a bridging carrier to support bridging interactions. However, less adsorption of PFOA on ZnO-NPs was observed with a 5 mM CaCl
2 concentration in the solution for both time intervals. Both CaCl
2 and NaCl affected the adsorption of PFOA on ZnO-NPs, potentially due to the electrostatic force of attraction. However, in the case of CaCl
2, the bridging effect of divalent (Ca
2+) cations between ZnO-NPs and PFOA may further lead to a reduction in PFOA adsorption on the ZnO-NPs [
57,
59]. This finding is consistent with one study showing that the adsorption of PFOA decreases with increasing ionic strength [
57].
Dissolution: The dissolved zinc (mg/L) in the ZnO-NPs in buffered water and after the interaction of PFOA with the ZnO-NPs in the presence of salts (such as 5 mM NaCl and 5 mM CaCl
2) were measured using ICP–OES (
Figure 3d). The particle size increased due to agglomeration and sedimentation after several weeks of interaction, decreasing the specific surface area and resulting in restrained dissolution. However, in the case of CaCl
2, less dissolution was measured than in all the other samples, which may be related to more agglomeration due to the bridging effect of Ca
2+. The smaller the size of the ZnO-NPs, the more easily dissolution occurred compared to the dissolution of larger particles [
18]. The attachment and penetration of nanoparticles inside the pores of low-density polyethylene tubes cannot be ignored.
3.3. Influence of HA on PFOA and ZnO-NPs’ Interaction
Particle size: The sizes of the ZnO-NPs in buffered water, treated with various concentrations of PFOA (such as 10 or 500 µg/L), treated with various concentrations of HA (1, 5, or 10 mg/L), and mixed with each of the other substances, were analyzed after 1 day, 1 week, 2 weeks, and 3 weeks of interaction at pH 7 (
Figure 4a and
Figure S5). Overall, an increase in the particle size of the ZnO-NPs was observed alone and in the presence of PFOA from 1 day to 3 weeks of interaction. This increase might be due to the presence of large (agglomerated) and/or sedimenting particles resulting from particle–particle interactions, electrostatic interactions, and hydrophobic interactions. The ZnO-NPs were monodispersed (particles of uniform size) from the time of nanoparticle suspension preparation to a few days. However, after 1 week of interaction, nonuniform (polydisperse) behaviour of the nanoparticles was observed. Overall, particles with sizes ranging from 1 week to 3 weeks were obtained in their polydispersed form, which was also reflected by their surface charge (
Figure 4b).
The size and distribution range of the ZnO-NPs in the presence of 1, 5 and 10 mg/L HA decreased after 1 week of interaction compared to after 1 day (
Figure 4a and
Figure S5). However, after 2 and 3 weeks of interaction, the samples might be very polydispersed, and the particle size was not suitable for measurement by a dynamic light scattering analyzer (a scattered fraction of the samples was observed). Similar behaviour of the ZnO-NPs was observed in the presence of various concentrations of PFOA (i.e., 10 and 500 µg/L) with HA (
Figure 4a and
Figure S5). It is quite possible that HA (a large aliphatic network of carbon molecules) capped the effective edges of the nanoparticles, which ultimately caused their dispersion.
The presence of both PFOA and HA altered the size of the ZnO-NPs differently than the presence of individual PFOA or HA.
Figure 4a and
Figure S5 illustrate the size of the ZnO-NPs at the peak of the particle size distribution curve (PSDC (d, nm)) in the presence of various concentrations of HA and PFOA after various durations of interaction. The particle size and range of the ZnO-NPs increased (agglomerated particles) alone and in the presence of PFOA and decreased (polydispersed) at various concentrations (1, 5, and 10 mg/L) of HA. This dispersion behaviour of the ZnO-NPs may be associated with the presence of organic acids (i.e., HA), which may impact engineered ZnO-NPs by reducing their aggregation behaviour [
19,
60].
The number of specific affinity sites and the affinity coefficient of specific sites for organic pollutants are deemed to be the main influential parameters on the adsorbent capacity to deal with pollutants. Enhanced nanoparticle (ZnO) stability in suspension media by adsorbed dissolved organic matter can increase the total number of specific affinity sites, which supports the adsorption of organic pollutants (PFOA) on the surface of dissolved organic materials rather than on the nanoparticle surface. Simultaneously, the adsorbed organic matter may also produce new affinity sites and/or block the nanoparticle affinity sites to alter their capacity to adsorb pollutants. Dissolved organic matter, which is not adsorbed by nanoparticles, may also first adsorb pollutants and, second, curtail further adsorption of pollutants on the nanoparticle surface [
15,
38].
Zeta potential: The zeta potentials of the ZnO-NPs in buffered water, with PFOA, and with 1, 5, and 10 mg/L HA were measured after 1 day, 1, 2, and 3 weeks of interaction, respectively (
Figure 4b). The magnitude of the surface charge of the ZnO-NPs alone and in the presence of PFOA decreased (from 1 day to 3 weeks) from −47.7 to −3.0 mV for the ZnO-NPs, from −48.6 to −3.1 mV for the ZnO-NPs with 10 µg/L PFOA, and from −47.9 to −3.5 mV for the ZnO-NPs with 500 µg/L PFOA. This behaviour confirmed the increase in the size of the nanoparticles due to agglomeration/sedimentation, which resulted in a reduced net charge (less negative) on the surface of the nanoparticles. However, the presence of HA decreased the magnitude of the change in the surface charge of the ZnO-NPs compared to that of the non-HA-containing samples. However, a high concentration of HA remained dominant in restraining the decrease in the magnitude of the zeta potential of the samples compared to that of lower HA concentrations (such as 1 mg/L HA) (
Figure 4b) [
18].
The aforementioned electrical potential data revealed that the aggregation behaviour of the pure ZnO-NPs in aqueous systems could be due to electrostatic interactions, van der Waals forces, and hydrophobic interactions. The environmental aging of nanoparticles alone and in the presence of organic pollutants, such as PFOA, could decrease the surface charge of the nanoparticles, increasing their sedimentation in environmental waters by decreasing their stability in aqueous systems. The presence of HA altered the surface charge in the reverse pattern compared to that of pure ZnO-NPs with and without the presence of PFOA. The HA substances covered the surface/effective sites of the nanoparticles because of their high aliphatic carbon content, which resulted in the least possibility of PFOA adsorbing on the nanoparticle surfaces. This also leads to the dispersion of the nanoparticles.
Adsorption: The PFOA in solution was measured to examine the sorption of PFOA with the ZnO-NPs alone and in the presence of various concentrations of HA after two weeks of interaction (
Figure S6). An increase in the amount of adsorbed PFOA (10 and 500 µg/L) was calculated for ZnO NPs without HA after 2 weeks of interaction. However, the adsorption of PFOA decreased with the addition of high HA concentrations (such as from 1 to 5 and 10 mg/L HA). Dissolved humic acids can foul the adsorption of organic chemicals to microporous activated carbon through direct competition for adsorption sites and pore blockage [
61].
Perfluoroalkyl acids, such as PFOA, contain a negatively charged hydrophilic head group and a hydrophobic–oleophobic perfluoroalkyl chain. Accordingly, a variety of mechanisms might be involved in the adsorption of PFOA in response to the surface properties (such as the charge and hydrophobicity) of adsorbents. The surface of HA is dominated by graphitic carbons, which are highly hydrophobic and have large electronic polarizability. Adsorption of PFOA molecules to HA is expected to be driven mainly by hydrophobic effects, which are combinations of entropic gradients and van der Waals (mainly dispersion) interactions between the adsorbate and adsorbent, whereas electrostatic forces play only a minimal role here. The low adsorption affinity of high concentrations of HA for PFOA is likely due to the low electronic polarizability of these molecules, thus decreasing potential van der Waals interactions despite the large electronic polarizability of graphitic carbons [
62].
It could also be assumed that the highly aliphatic structure of HA dispersed the ZnO NPs, providing fewer active sites for the attachment of PFOA molecules. This also supported the results obtained (such as decreased zeta potential values for ZnO NPs alone and in the presence of 10 and 500 µg/L PFOA compared to samples with HA) in
Figure 4a,b and
Figure S6.
Dissolution: The dissolution of ZnO-NPs alone or in the presence of various concentrations of HA was observed in Milli-Q water at pH 7 controlled by using buffer solution (
Figure S7). The presence of zinc in its dissolved or ionic form is potentially toxic to microorganisms, such as microflora [
63,
64]. The dissolution of ZnO-NPs can be influenced by the presence of other compounds in water [
65], such as HA. The dissolved zinc concentration (mg/L) from ZnO-NPs alone or from ZnO-NPs combined with PFOA was calculated with the addition of various concentrations of HA in this study (
Figure S7). After 1 day of interaction, the dissolved zinc concentration was greater in the presence of various concentrations of HA than in the absence of HA, and this trend was observed even after 1 week of interaction. It could be assumed that HA dispersed the nanoparticles after a long interaction time (such as 2–3 weeks), decreasing this dispersion effect by dominating the electrostatic forces, van der Waals forces, and hydrophobic interactions. This could also be caused by complexation (for zinc ions) with anionic HA followed by its large complex structure. Therefore, our findings are the same as those hypothesized by [
63], i.e., that HA binds zinc ions.
TEM and XRD analysis: The samples from ZnO-NPs alone or in the presence of PFOA or HA were analyzed using TEM morphology and elemental mapping after immediate preparation (such as after 0 h of interaction) and after 1 day of interaction (
Figure 4c). ZnO-NPs were more aggregated after 1 day of interaction than after 0 h, which is consistent with previous findings (
Figure 1a). TEM revealed an increase in the size of the ZnO-NPs in the presence of 10 mg/L PFOA (only this concentration was selected for TEM analysis to confirm the presence PFOA, such as fluorine in mapping) after 1 day of interaction compared to 0 h. Elemental mapping further confirmed the presence of Zn, O, F, P, and K (
Figure S8). The dispersion patterns of ZnO-NPs, alone and in the presence of PFOA, due to the presence of highly aliphatic and complex structures of HA, can be observed (
Figure 4c) when comparing images from 0 h and 1 day. The SAED images did indicate less crystallinity on the nanoparticles after 1 day of interaction, which matches the results obtained from the zeta analysis. Elemental mapping further confirmed the presence of the expected elements.
Figure S8 shows the elemental composition comparisons of the ZnO-NPs with contaminants at different intervals.
Post-photocatalysis characterization, namely XRD (
Figure S8b), of ZnO-NPs, HA, and ZnO-NPs with and without the presence of 10 mg/L PFOA and 10 mg/L HA was performed to further examine the impact of PFOA and HA on the crystallinity and purity of ZnO-NPs. It was observed that there was no change in the crystal phase of the ZnO-NPs after 1 day of interaction. However, after 1 week of interaction (
Figure S8b), the intensities of the peak alignments at (1 0 0), (0 0 2), and (1 0 1) were suppressed. Four new peaks at 2θ values of 9.68° and 19.36°, 22.58°, and 25.6° were observed, indicating signs of alterations into the crystallinity and purity of the ZnO-NPs influence by the adsorption of co-contaminants and aging factors.
3.4. Influence of Electrolytes and HA Together on PFOA and ZnO-NPs’ Interaction
A mixture of salts and dissolved organic matter could influence the particle size and surface charge differently, which was investigated and explained in this section (
Figure 5).
Particle size: The sizes of the ZnO-NPs at the peak of PSDC (d, nm) without and with the presence of PFOA and with 5 mM concentrations of monovalent and divalent electrolytes and various concentrations of HA after 1 day, 1 week, 2 weeks, and 3 weeks of interaction are illustrated (
Figure 5 and
Figure S9). With the increasing size of ZnO-NPs alone or with PFOA and the addition of 5 mM concentrations of monovalent salt (NaCl) and divalent CaCl
2, the size of the ZnO-NPs increased as the interaction time increased, such as from 1 day to 1 week. The size was out of the machine range (range:
) after 1 week due to the aggregation of nanoparticles, and only a few fragments were measured after 1 week of interaction. It is assumed that the particle size at the peak of the PSDC would be larger than 10,000 nm. In the presence of HA, initial dispersion was observed with increasing HA concentration in the presence of salts and PFOA; however, after a few weeks, the same behaviour of aggregation/agglomeration/sedimentation was observed, with the particle size being nonuniform and not measurable, while there was a decreasing zeta potential.
These findings revealed that the influence of humic substances on the interactions between PFOA and ZnO-NPs is somewhat complicated, especially when electrolytes are present. Counteractions for PFOA between HA and ZnO-NPs could reduce the effective interactions of ZnO-NPs by decreasing the amount of PFOA available for sorption. Moreover, natural organic matter can cover the surface of nanoparticles and thereby reduce their affinity for organic pollutants [
38,
57,
66].
Zeta potential: A decrease in the magnitude of the surface charge of ZnO NPs alone or in the presence of PFOA or salt was observed after 1 day to 3 weeks of interaction, as shown in
Figure 5b. This indicated that the ZnO-NPs exhibited a similar agglomeration (size increase) behaviour (
Figure 5) with a diminished surface charge, as described in the aforementioned sections. The addition of various concentrations of HA in the presence of PFOA did not significantly change the zeta potential compared to that of samples with HA. This confirms that interactions between HA molecules (the dispersion of HA-coated ZnO-NPs) are more dominant at high concentrations (such as 10 mg/L HA) than the electrostatic interactions between negatively charged nanoparticles surrounded by monovalent cations (Na
+). Conversely, divalent cations interacted more strongly with negative surface charges, and the overall zeta potential decreased in magnitude in the presence of both divalent cations (Ca
2+) (
Figure 5b). However, the dispersion effect of HA on ZnO-NPs was also observed when PFOA and divalent electrolyte (CaCl
2) salts were present (
Figure 5b), as noted in the findings. It could be inferred that salinity and natural organic substances play significant roles in the transport of ZnO-NPs and their associated organic pollutants (PFOA) from fresh water to the ocean, especially in estuary regions.
Adsorption: Adsorption of PFOA was examined for ZnO-NPs alone or in the presence of various concentrations of HA in the presence of 5 mM NaCl and CaCl
2 after two weeks of interaction (
Table S1). An increase in the amount of adsorbed PFOA (10 and 500 µg/L) was calculated for ZnO NPs without HA after 2 weeks of interaction. However, the adsorption of PFOA decreased with the addition of high HA concentrations (such as 1 and 10 mg/L HA) (
Table S1). It could be assumed that the highly aliphatic structure of HA dispersed the ZnO NPs, providing fewer active sites for the attachment of PFOA molecules. This result also supported the results obtained, i.e., decreased zeta potential values for ZnO NPs alone and in the presence of 10 and 500 µg/L PFOA compared to samples with HA in
Figure 5b.
Dissolution: The dissolved zinc concentration (mg/L) from ZnO-NPs alone, with PFOA and salts, and with the addition of various concentrations of HA, was measured in this study (
Figure 5c). After 1 day of interaction, the dissolved zinc concentration was greater in the presence of various concentrations of HA than in the absence of HA. It could be assumed that HA dispersed the nanoparticles; however, after a long interaction time (such as 1, 2, or 3 weeks), the dispersion effect decreased due to the dominant electrostatic forces, van der Waals forces, and hydrophobic interactions.
TEM analysis: The morphological behaviour of ZnO-NPs with PFOA, HA, and CaCl
2 after 0 h and 1 day intervals was examined via TEM (
Figure S10). The particles were aggregated after 0 h in the presence of 5 mM CaCl
2. However, after 1 day of interaction, shaded (due to CaCl
2) and dispersed (due to HA) patterns of the nanoparticles can be observed. The bright diffraction signals are due to the crystalline ZnO-NPs, including the presence of CaCl
2 crystals. To examine the morphological changes associated with high concentrations of CaCl
2, 10 mM CaCl
2 (after 0 h of interaction) was added, which generated clusters/agglomerates. However, after 1 day of interaction, the nanoparticles were dispersed by coating them with large HA molecules (
Figure S10d).
FTIR: The interactions of ZnO-NPs with PFOA in the presence of HA and electrolytes were investigated via FTIR analysis (
Figure 6). The metal oxide (ZnO) absorbance ranged from 600 to 400 cm
−1 [
15,
49,
67], which indicated the presence of interacting ZnO-NPs. This difference was detected in all the samples, while the intensity of the absorbance peak depth and location varied. Peaks at 1800 and 600 cm
−1 represent carboxylate functional groups and C−F [
51], C−C, and C−H stretching, respectively [
50]. In particular, the vibrational peak at approximately 1102 cm
−1 appeared in all samples with PFOA, representing the presence of C–F stretching bonds. This indicated the interaction between PFOA and the ZnO-NPs. The absorbance at 1645 cm
−1 is due to H−O−H bending. The infrared band at 1010 cm
−1 in all the examples except for ZnO-NPs and ZnO + PFOA is due to the stretching of N–H bonds [
50] present in an organic compound, such as HA. The peaks between 1050 and 750 cm
−1 could be due to K-potassium and P-phosphorous (from a buffer solution used to maintain pH 7) stretching with O and C. The absorbance at 3490 cm
−1 is due to O-H stretching [
50]. The peaks at approximately 933 and 871 cm
−1 could be due to triatomic inorganic molecules (calcium chlorine), while 670 cm
−1 could be due to CO
2 from the atmosphere [
50].