3.1. Morphological Characterization
The formation of mesostructured nanoparticles as well as their nature and morphology were studied by SEM and TEM analyses. This combination of techniques allows the detection of individual nanoparticles and associated particle-aggregate morphologies.
Figure 1 reports electron microscopy images of bare silica nanoparticles (SiO
2 NPs) and of all hybrid systems synthesized in the presence of different amounts of tannic acid.
Both microscopy analyses revealed a spherical shape for SiO
2 NPs (
Figure 1a), with a mean diameter of about 180 nm. The 30TA-SiO
2 sample appears quite similar to the bare silica (
Figure 1b), but the TEM image points out a fuzzy connection between spherical particles in the typical pearl-necklace structure, which is probably due to the presence of the organic component (
Figure 1b). A change in size and morphology can be detected when increasing the TA amount (
Figure 1c–e). In particular, 60TA-SiO
2 shows a double population of particles: more dense spherical NPs (~250 nm in diameter) and aggregates made of small core-shells NPs. We assume that larger spherical nanoparticles are mostly made of silica, while the different densities evidenced by the core-shell structure might be related to the presence of an organic component mainly located in the core, surrounded by a shell characterized by a predominant silica phase. Larger spherical nanoparticles are no longer visible in the micrographs of the 120TA-SiO
2 sample, mainly composed of clusters of small core-shell NPs. Finally, the 150TA-SiO
2 sample appears characterized by co-continuous hybrid structures with no well-defined shape (
Figure 1e). In view of the above considerations, a possible mechanism related to NPs formation may be proposed (
Scheme 1).
Considering the nucleation theory, the first step of NPs generation involves the condensation reaction of hydrated monomers which form primary particles of ~5 nm in diameter. These later aggregate up to the stationary critical size, forming larger nanoparticles [
61]. The small number of TA-APTS species used in the synthesis of 30TA-SiO
2 NPs does not affect the formation mechanism resulting in an NPs population very similar to that of bare silica. In the case of 60TA-SiO
2, a stoichiometric ratio between amine (-NH
2) and quinone groups (1:1—mol:mol) is achieved, therefore the amino groups of APTS are all engaged in binding with tannic acid molecules. According to this result, the larger spheres are presumably made of pure silica formed by hydrolysis and condensation reaction of TPOS, whereas it is possible to assume that the concentration of tannic acid is such as to generate coupled TA-APTS hybrid monomers, that act as nucleation sites leading to a second population of core-shell type particles. The hypothesis that there is a limiting concentration of tannic acid to permit the formation of nucleation sites TA-APTS can be supported by the absence of core-shell structures, when a smaller amount of tannic acid was used, keeping constant the other reagent amount. A higher amount of TA (120TA-SiO
2), resulted in an increasing number of hybrid coupled monomers (see
Scheme 1). Therefore, a growing number of homogenous nucleation sites in the starting solution is expected. This behaviour drives towards the formation of solely core-shell particles, formed by the condensation reaction of TPOS, which preferably occurs at coupled TA-APTS species. Finally, at the highest TA content (150TA-SiO
2), both coupled and free molecules of TA, are expected in the starting solution. Hence, the organic phase can act as a binder and constraining agent in forming organic–inorganic components, limiting NPs growth and producing hybrid clusters with irregular shapes.
The DLS distribution presented in
Figure 2, points out the significant role of the supposed mechanism in regulating particle size and properties. The curves showed that a decrease in particle size was related to an increase in tannic acid content. For low-content samples (bare SiO
2 and 30TA-SiO
2), the curves can be considered almost overlapped with a very low polydispersity value (PDI about 0.02) and higher size value. These data are in good agreement with the hypothesis of spherical monodispersed nanoparticles. In the case of 60TA-SiO
2 and 120TA-SiO
2 samples, the curves shifted towards smaller sizes, but the polydispersity index increased (about 0.15). This last observation could indicate the presence of a multicomponent system, which cannot be completely distinguished in its individual parts because of the aggregation effects for smaller particles. This event seems to occur, in particular, for the 60TA-SiO
2 sample, where electron micrographs show a double population that could be covered by an average aggregation. Finally, for the 150TA-SiO
2 sample, the average size is even lower, but the presence of a high PDI value could confirm the production of clusters with irregular shapes. The size discrepancy between electron microscopy analysis and DLS measurements is a well-known pitfall, but in our case, the comparison of both results showed the same nanosystems behaviour [
62].
3.2. Physical–Chemical Characterization
In
Figure 3, FT-IR spectra of hybrid systems are reported together with those of both SiO
2 and pure TA. The FT-IR spectrum of SiO
2 NPs (
Figure 3a) showed the typical bands of silica gel: (i) 1100 cm
−1 due to the antisymmetric Si–O–Si stretching vibration mode; (ii) 950 cm
−1 attributed to Si-O terminal non-bridging vibration; (iii) 800 cm
−1 and 460 cm
−1 due to Si–O–Si bond vibration and bending, respectively [
63,
64]. The NH
2 bending vibration 1559 cm
−1 (triangles), emerged in the TA30-SiO
2 NPs spectrum together with 3190 cm
−1 adsorptions for NH
2 groups symmetric stretching [
65]. Furthermore, the bands at 2940 cm
−1 and 2880 cm
−1 (spectrum b
Figure 3, circle) were assigned to the antisymmetric and symmetric stretching of non-hydrolysable alkyl groups of APTS [
64]. FT-IR investigation on hybrid samples highlighted that a rising amount of TA (
Figure 3c–e) led to the disappearance of primary amino groups signals, coupled with the presence of new bands at 1350, 1505, 1615 and 1707 cm
−1. The band at 1350 cm
−1 (squares) was associated with the double bond in the aromatic rings of TA. As reported in the literature, when the hydroxyl groups of the polyphenols were oxidized, the aromatic rings become prone to nucleophile attack by the nitrogen of the primary amino groups [
66,
67]. This fact resulted in the formation of a secondary amine, as proved by the presence of the new bands at 1505, 1615 and 1707 cm
−1 (stars), that are mainly visible in the 120TA-SiO
2 and 150TA-SiO
2 FT-IR spectra. The absence of the characteristic amide peaks ultimately confirmed that the chemical interaction between TA and the inorganic component occurred through the formation of secondary amines. The presence of TA is detected by the broad band at 3500 cm
−1 (rhombs) related to the stretching vibration of hydroxyl groups. Commonly, this pronounced band is also present in silica sol–gel derived particles. However, in this case, TA-SiO
2 samples were synthesized by using two different silica precursors (i.e., APTS and TPOS). This fact results in a lower amount of -OH surface groups because, probably, replaced by non-hydrolysable aminopropyl ones from APTS [
45]. Finally, in the hybrid samples spectra, from b to e, the characteristic Si–O–Si stretching vibration in 1000–1300 cm
−1 region shifted to a higher wavenumber indicating a more crosslinked SiO
2 network.
TGA analysis was performed to assess the organic content of the hybrid platforms and thermal stability. In
Figure 4, TGA curves of all hybrid nanoparticles are reported together with the ones of SiO
2 nanoparticles. Moreover, the TGA curve of tannic acid is also shown as an inset in
Figure 4. In the TGA curve of pure TA, a first weight loss is observed below 200 °C and attributed to the degradation of tannic acid [
68]. The TA decomposition, via decarboxylation, causes a pronounced weight loss in the temperature range of 230–400 °C [
69]. TGA curves of SiO
2 showed a first weight loss of about 5 wt%, at temperatures lower than 150 °C, attributed to the loss of physically adsorbed water. A second weight variation, observed at higher temperatures, was ascribed to the decomposition of residual alkoxide groups and/or dehydroxylation. The TGA curves of 30TA-SiO
2 and 60TA-SiO
2 samples resemble that of bare SiO
2 with a slight increase in weight loss. The trend of the TGA curves is different for 120TA-SiO
2 and 150TA-SiO
2 hybrid NPs. It exhibits a continuous weight loss up to 700 °C. This fact seems to suggest higher thermal stability of the organic phase, probably due to the chemical coupling with tIe inorganic component [
70,
71,
72]. The increase in TA in the 120TA-SiO
2 and 150TA-SiO
2 samples can account for the greater weight loss.
Elemental analysis was performed on dried nanoparticles to confirm their chemical composition. In
Table 2, the elemental weight percent obtained through EDS analysis are reported for all the investigated samples.
It is possible to observe that, the 150TA-SiO2 sample showed the largest amount of carbon, due to the highest content of TA in this formulation. This fact is in good agreement with the weight loss data of TG measurements.
Hybrid TA-silica samples have been characterized in terms of textural properties. The bare SiO
2 and 30AT-SiO
2 nanoparticles are not shown in
Figure 5 due to their non-porosity behaviour [
73]. In contrast, the reported isotherms, for 60TA-SiO
2, 120TA-SiO
2 and 150TA-SiO
2 NPs systems, show textural properties which appear to be related to the content of tannic acid deployed during the synthesis procedure. The increase in the organic component leads to the formation of porous structures, which exhibit a higher surface area (15 and 237 m
2∙g
−1, for 60TA-SiO
2 and 150TA-SiO
2 NPs, respectively) and also a rise in the total volume of pores (0.054 and 1.005 cm
3∙g
−1, for 60TA-SiO
2 and 150TA-SiO
2 NPs, respectively) (
Table 3). The isotherm shape of 60AT-SiO2 resembles type III without hysteresis. In this case, the monolayer formation is not identifiable, indicating a non-microporous solid, with a pronounced condensation in inter-particle voids [
74]. The shape of the pore size distribution by the BJH method (
Figure 6 grey curve) does not show the presence of mesopores, whereas the absence of macropores has been evaluated by the DFT method (data not shown). As far as the 120TA-SiO
2 and 150TA-SiO
2 samples are concerned, the fact that the shape of the isotherms seem to be a combination of type III and type II, does not appear to be of simple interpretation. The common type II isotherm, shows increasing adsorption at low relative pressures, indicating the occurrence of micropore filling. As the pressure increases, the monolayer adsorption evolves to multilayer and the adsorbance sharply rises when the
p/
p° value is close to 1 (macropore filling). In particular, the adsorption curves rise with a slightly convex shape at values of about
p/
p° = 0–0.03, which reveals monolayer physical adsorption and micropore filling. The adsorbance continues to increase slowly until relative pressures
p/
p° attain a value close to 1, where a sharp rising occurs, thus indicating the presence of macropores and capillary condensates in the surface of the nanoparticles. The above considerations suggest that the samples 120TA-SiO
2 and 150TA-SiO
2 have a complex pore system that includes mesopores and macropores not excluding however micropores. For these samples, moreover, the presence of a type C hysteresis could indicate the presence of open-wedge pores [
75]. The pore size distribution, obtained by the BJH method on the desorption branch, is reported in
Figure 6. As far as samples 120 and 150 AT-SiO
2 are concerned, a sharp peak at about 3.8 nm denotes that a large part of their pores are low-range mesoporous. However, the presence of larger pores may be detected, although to a lower extent. Combining these observations with the type of hysteresis found, it is possible to further support the hypothesis of particle formation around a tannic acid core.
3.3. Evaluation of Adsorption Capacity
The adsorption capacity was measured by determining the copper amount trapped by nanoparticles using ICP-MS. The results were reported as the amount of adsorbed copper ions per gram of sample (Cu
2+/TA-SiO
2 NPs (mg/g)) in
Table 4. The data show a comparable content of Cu
2+ in the case of low TA samples, whereas a remarkable increase in Cu
2+ is observed for high TA content samples. The general trend suggests a combined effect exerted by both specific surface area and TA component [
69,
70,
71].
To support ICP-MS data, the adsorption behaviour of TA-SiO
2 NPs toward Cu
2+ ions was also qualitatively evaluated by UV-vis spectroscopy measurements. In particular, after the adsorption experiments, all supernatants were recovered to verify the presence of residual copper ions, i.e., the copper not adsorbed by nanoparticles. To this aim, the formation of a benzotriazole–copper complex (BTH/Cu
2+ complex) was detected by UV-vis analysis and reported in
Figure 7 [
76,
77,
78]. In fact, before UV measurement, the samples were diluted two times in a BTH solution (20 mg/L). The curves shown in
Figure 7 were baseline corrected.
In the same figure, the spectrum of pure BTH (black curve) was compared with the spectrum of BTH/Cu
2+ complex obtained with an initial Cu
2+ ions concentration of 20 mg/L (dotted curve). The formation of the BTH/Cu
2+ complex resulted in a modification of the pristine BTH curve together with a remarkable reduction in the absorption peak at 257 nm, which is characteristic of the BTH ligand. Moreover, the presence of residual Cu
2+ ions in the supernatants results in a reduction of this peak. Thus, the larger the reduction, the higher the concentration of residual Cu
2+ and, hence, the smaller the amount of adsorbed copper onto the nanoparticles [
79].
Overall, the trend observed in
Figure 7 is in good agreement with ICP-MS results, highlighting that 120TA-SiO
2 and 150TA-SiO
2 exhibited the best adsorption performances.
As far as the samples exhibiting the best adsorption ability (120TA-SiO
2 and 150TA-SiO
2 NPs) are concerned, the effect of the different Cu
2+ initial concentrations was investigated and reported in
Table 5.
The adsorption capacity significantly increased when the initial copper ions concentration increased up to 20 mg/L. This could be due to the saturation of available active sites on the hybrids nanoparticles above 20 mg/L of copper (II) ions concentration.
In addition, the effect of solution pH on the adsorption of copper (II) ions, for an initial concentration of 20 mg/L, was evaluated and reported in
Table 6.
Both 120TA-SiO
2 and 150TA-SiO
2 samples showed that the adsorption capacity is enhanced by the increase in pH. This behaviour could be explained by the different NPs surface charge as a function of pH. It can be seen in
Figure 8 that the surface charge for both 120TA-SiO
2 and 150TA-SiO
2 NPs was quite different when varying the pH value. The copper ions adsorption is probably favoured by negatively charged surfaces [
80]. At low pH, the reduced Cu
2+ ions adsorption could be attributed to the positively charged surface of TA-SiO
2 NPs. When the pH value increased from 4 to 7, for both samples, an enhancement in the adsorption capacity was observed, which was larger in 150TA-SiO
2 NPs, probably due to the higher negative surface charge already observed at pH 7. A further increase in pH value does not influence much the outgrowth in the case of sample 150TA-SiO
2, while the increase is clear in the case of sample 120TA-SiO
2, where the charge value moves from about −10mV (at pH = 7) to about −40mV (at pH = 10).
All these results proved that the adsorption of copper ions onto synthesized NPs was highly dependent on the solution pH. Furthermore, the good performances at about pH 7 proved that our systems can be effectively used in the treatment of the real wastewater, which always has pH values close to neutral.
Finally, XRD analysis was performed on the 150TA-SiO
2 NPs after adsorption (
Figure 9), to verify the presence of eventual crystalline phases related to the copper component, in view of a secondary application of the hybrid systems. No characteristic crystalline Cu bands were observed, but only a weak signal at 2θ value about 23°, ascribed to amorphous sol–gel-derived silica [
81].