Inulin-Coated ZnO Nanoparticles: A Correlation between Preparation and Properties for Biostimulation Purposes

Within the framework of plant biostimulation, a pivotal role is played by the achievement of low-cost, easily prepared nanoparticles for priming purposes. Therefore, in this report, two different synthetic strategies are described to engineer zinc oxide nanoparticles with an inulin coating. In both protocols, i.e., two-step and gel-like one-pot protocols, nanoparticles with a highly pure ZnO kernel are obtained when the reaction is carried out at T ≥ 40 °C, as ascertained by XRD and ATR/FTIR studies. However, a uniformly dispersed, highly homogeneous coating is achieved primarily when different temperatures, i.e., 60 °C and 40 °C, are employed in the two phases of the step-wise synthesis. In addition, a different binding mechanism, i.e., complexation, occurs in this case. When the gel-like process is employed, a high degree of coverage by the fructan is attained, leading to micrometric coated aggregates of nanometric particles, as revealed by SEM investigations. All NPs from the two-step synthesis feature electronic bandgaps in the 3.25–3.30 eV range in line with previous studies, whereas the extensive coating causes a remarkable 0.4 eV decrease in the bandgap. Overall, the global analysis of the investigations indicates that the samples synthesized at 60 °C and 40 °C are the best suited for biostimulation. Proof-of-principle assays upon Vicia faba seed priming with Zn5 and Zn5@inu indicated an effective growth stimulation of seedlings at doses of 100 mgKg−1, with concomitant Zn accumulation in the leaves.

Int. J. Mol.Sci.2024, 25, x FOR PEER REVIEW 2 of 23 variety of nanoparticles were employed, aiming to improve the biometric parameters as well as micronutrient deficiencies [70][71][72][73][74][75].Among the possible nanomaterials, ZnO nanoparticles (NPs) are widely used as nano-priming agents, since they can tolerate Zn soil deficiency, promote plant well-being, and can be produced at a low cost [76][77][78][79][80][81][82].Additional improvements in plant growth and well-being can be achieved by employing suitably coated NPs, where the coating acts as a mimicking agent to facilitate entry via the roots and transport through the plant.This is the case, for instance, for Fe3O4 NPs coated with Cassia occidentalis L. flower extract [83], Au NPs whose absorption was implemented by galanga rhizome extract [84], or Ag coated with layers of alfalfa extract [85].As for ZnO NPs for biostimulation purposes, currently explored coatings are cysteine [86], chitosan [87], or lignin [88].
One of the latest biostimulants employed for Vicia faba L. seedling growth includes ZnO NPs coated with inulin [89], a fructan, as sketched in Figure 1.This can be extracted from several plants such as chicory roots (Cichorium intybus L.), Jerusalem artichoke tubers (Helianthus tuberosus L.), or novel sources like globe artichoke inflorescence (Cynara cardunculus L.) [90], as well as from agave (Agave sisalana P.) [91].Size, morphology, the degree of coating, and the overall chemico-physical properties of the nanoparticles play a pivotal role in the efficacy of their action as biostimulants.Therefore, the synthetic strategy becomes crucial for the ensuing application.These combined properties, however, are often overlooked when dealing with plant administration, often relying on commercial nanoparticles that are not always fully characterized [75,[92][93][94][95][96][97][98].Engineered nanoparticles for biostimulation are usually lab prepared, although the implementation of the protocol is typically not discussed.Requirements of coated nanoparticles are uniformity in size and a sufficient degree of coating, as well as being low-cost and simple to prepare.
In the current paper, we focus on different synthetic pathways to achieve inulincoated ZnO NPs.In particular, two different approaches were explored, i.e., two-step and gel-like processes.The resulting NPs were characterized by different techniques aimed at determining differences in the structural, morphological, and electronic properties, the latter being especially important because ZnO NPs act on the antioxidant potential in seedlings [99].As for the two-step procedures, depending on the synthesis conditions Zn(OH)2 could be detected, whereas the temperature of the preparation phases could influence the size, morphology, and agglomeration of the particles, though not the bandgap.This is similar for all coated samples and in line with their uncoated counterparts.The gellike process yields larger agglomerates (on average 300 nm) of smaller nanoparticles (on average 30 nm), characterized by a significant decrease in the bandgap.Overall considerations of the uniformity of the particle sizes, the degree and uniformity of the coating, and a not unusual bandgap point identified sample Zn5@inu as the best suited for further biostimulation application.Proof-of-principle assays upon Vicia faba seed, priming with Zn5 and Zn5@inu, indicated an effective growth stimulation of seedlings at doses of 100 Size, morphology, the degree of coating, and the overall chemico-physical properties of the nanoparticles play a pivotal role in the efficacy of their action as biostimulants.Therefore, the synthetic strategy becomes crucial for the ensuing application.These combined properties, however, are often overlooked when dealing with plant administration, often relying on commercial nanoparticles that are not always fully characterized [75,[92][93][94][95][96][97][98].Engineered nanoparticles for biostimulation are usually lab prepared, although the implementation of the protocol is typically not discussed.Requirements of coated nanoparticles are uniformity in size and a sufficient degree of coating, as well as being low-cost and simple to prepare.
In the current paper, we focus on different synthetic pathways to achieve inulin-coated ZnO NPs.In particular, two different approaches were explored, i.e., two-step and gellike processes.The resulting NPs were characterized by different techniques aimed at determining differences in the structural, morphological, and electronic properties, the latter being especially important because ZnO NPs act on the antioxidant potential in seedlings [99].As for the two-step procedures, depending on the synthesis conditions Zn(OH) 2 could be detected, whereas the temperature of the preparation phases could influence the size, morphology, and agglomeration of the particles, though not the bandgap.This is similar for all coated samples and in line with their uncoated counterparts.The gel-like process yields larger agglomerates (on average 300 nm) of smaller nanoparticles (on average 30 nm), characterized by a significant decrease in the bandgap.Overall considerations of the uniformity of the particle sizes, the degree and uniformity of the coating, and a not unusual bandgap point identified sample Zn5@inu as the best suited for further biostimulation application.Proof-of-principle assays upon Vicia faba seed, priming with Zn5 and Zn5@inu, indicated an effective growth stimulation of seedlings at doses of 100 mgKg −1 [89], with concomitant Zn accumulation in the leaves.In particular, verification in relation to whether the selected coated and non-coated samples had geno-or cytotoxic effects on Vicia faba seedlings was undertaken.Furthermore, preliminary experiments on single-step synthesized samples did not yield equally promising results in terms of biostimulation, pointing to the double-step synthesis as the best strategy for the effective sampling of the nanoparticles.Since the major differences between one-and two-step processes are related to the morphology and electronic properties of the samples, their influence on the biostimulation process has also been considered.

Results and Discussion
The synthesized samples were subjected to multiple characterization techniques to determine the phase, morphology, and degree of coverage.

XRD
The XRD patterns of the samples achieved with both procedures are reported in Figure 2. As for the two-step procedure, the XRD patterns achieved with and without an inulin coating are presented.The overall comparison of the spectra points at four main features: (i) ZnO is the main product, (ii) additional Zn(OH) 2 may be present depending on the preparation conditions, (iii) inulin coating, as commonly happens in organic phases, is not singled out by XRD [100] due to the limited scattering power of the thin layer composed of light atoms, and iv) capping the NPs with inulin does not have an evident effect on the composition of the core, i.e., Zn(OH) 2 is not removed by the inulin coverage.
mgKg −1 [89], with concomitant Zn accumulation in the leaves.In particular, verifica relation to whether the selected coated and non-coated samples had geno-or cy effects on Vicia faba seedlings was undertaken.Furthermore, preliminary experim single-step synthesized samples did not yield equally promising results in term ostimulation, pointing to the double-step synthesis as the best strategy for the e sampling of the nanoparticles.Since the major differences between one-and two-st cesses are related to the morphology and electronic properties of the samples, thei ence on the biostimulation process has also been considered.

Results and Discussion
The synthesized samples were subjected to multiple characterization techniq determine the phase, morphology, and degree of coverage.

XRD
The XRD patterns of the samples achieved with both procedures are reported ure 2. As for the two-step procedure, the XRD patterns achieved with and withou ulin coating are presented.The overall comparison of the spectra points at four ma tures: (i) ZnO is the main product, (ii) additional Zn(OH)2 may be present depend the preparation conditions, (iii) inulin coating, as commonly happens in organic p is not singled out by XRD [100] due to the limited scattering power of the thin laye posed of light atoms, and iv) capping the NPs with inulin does not have an eviden on the composition of the core, i.e., Zn(OH)2 is not removed by the inulin coverage Figure 2. XRD patterns of the prepared samples, with and without coating.Non-coated: Blac cyan = Zn2, orange = Zn3, brown = Zn4, pink = Zn5; coated: red = Zn1@inu, gray = Zn2@inu, m = Zn3@inu, blue = Zn4@inu, light green = Zn5@inu, dark green = ZnOInuOP.
In more detail, the curves shown in Figure 2 can be grouped into two clusters, components share the same global common features.In the top array of XRD patter presence of a ZnO hexagonal phase (P63mc space group, found naturally in the m zincite [101]) appears evident in every curve of the group from the very characteris ture of that crystal lattice, i.e., the peak triad at 32, 34.5, and 36 degrees, correspon XRD patterns of the prepared samples, with and without coating.Non-coated: Black = Zn1, cyan = Zn2, orange = Zn3, brown = Zn4, pink = Zn5; coated: red = Zn1@inu, gray = Zn2@inu, magenta = Zn3@inu, blue = Zn4@inu, light green = Zn5@inu, dark green = ZnOInuOP.
In more detail, the curves shown in Figure 2 can be grouped into two clusters, whose components share the same global common features.In the top array of XRD patterns, the presence of a ZnO hexagonal phase (P6 3 mc space group, found naturally in the mineral zincite [101]) appears evident in every curve of the group from the very characteristic feature of that crystal lattice, i.e., the peak triad at 32, 34.5, and 36 degrees, corresponding to the reflections from 100, 102, and 101 crystallographic planes of the crystal system.Regarding the intensity of the triad peaks, the latter appears more intense in the three coated systems, Zn3@inu, Zn2@inu, and Zn4@inu (in decreasing intensity order), possibly signaling the existence of slightly larger crystal domains in the system.In the lower group of curves containing Zn1/Zn1@inu patterns (two elements), several new peaks can be pinpointed.Notably, two doublet peaks at 20-21 and 27-28 degrees appear, as well as a peak at 33 and low-intensity triplet peaks at 39, 41, and 42.5 degrees.These peaks can be attributed to the reflections of a zinc hydroxide orthorombic (ε) crystal phase naturally found in the mineral Wulfingite [102,103].The two doublets, which have a stronger intensity than the other peaks of the pattern, are originated from the reflections of 011, 002, 111, and 102 crystal planes of the orhorombic lattice.

IR
IR spectra may provide complementary data with respect to XRD patterns, i.e., information on the coating of the synthesized samples.Figure 3 displays spectra of selected samples, i.e., samples Zn1 and Zn1@inu containing Zn(OH) 2 , the one-pot sample ZnOInuOP and inulin-coated sample Zn5@inu showing inulin-related features, and the corresponding uncoated sample Zn5.
the reflections from 100, 102, and 101 crystallographic planes of the crystal system garding the intensity of the triad peaks, the latter appears more intense in the three co systems, Zn3@inu, Zn2@inu, and Zn4@inu (in decreasing intensity order), possibly sig ing the existence of slightly larger crystal domains in the system.In the lower grou curves containing Zn1/Zn1@inu patterns (two elements), several new peaks can be pointed.Notably, two doublet peaks at 20-21 and 27-28 degrees appear, as well as a at 33 and low-intensity triplet peaks at 39, 41, and 42.5 degrees.These peaks can b tributed to the reflections of a zinc hydroxide orthorombic (ε) crystal phase natu found in the mineral Wulfingite [102,103].The two doublets, which have a stronger in sity than the other peaks of the pattern, are originated from the reflections of 011, 002, and 102 crystal planes of the orhorombic lattice.

IR
IR spectra may provide complementary data with respect to XRD patterns, i.e., in mation on the coating of the synthesized samples.Figure 3 displays spectra of sele samples, i.e., samples Zn1 and Zn1@inu containing Zn(OH)2, the one-pot sam ZnOInuOP and inulin-coated sample Zn5@inu showing inulin-related features, and corresponding uncoated sample Zn5.[104].Zn1@inu displays similar features to Zn1, thus cating a coating of the particles below the detection limit.At variance with this, coated samples Zn5@inu (two-step procedure) and ZnOInuOP (one-pot procedure) sent features ascribable to inulin layers covering the nanoparticles.For comparison poses, the spectrum of the inulin used for the preparation is also reported in Figure  [104].Zn1@inu displays similar features to Zn1, thus indicating a coating of the particles below the detection limit.At variance with this, both coated samples Zn5@inu (two-step procedure) and ZnOInuOP (one-pot procedure) present features ascribable to inulin layers covering the nanoparticles.For comparison purposes, the spectrum of the inulin used for the preparation is also reported in Figure 3.The spectral features of inulin may be divided into three regions.The first one, at low wavenumbers in the 600-800 cm −1 range, is assigned to -CH 2 rocking vibrations.The region between 800 cm −1 and 1400 cm −1 carries the inulin fingerprints.In particular, peaks at 827 cm −1 and 866 cm −1 are representative of the anomeric bendings δ(C1-H) and ring vibration of 2-ketofuranose, respectively, and are indicative of the β-(2→1) glycosidic bonds.In contrast, peaks at 1016 cm −1 , 1109 cm −1 , and 1219 cm −1 are assigned to the stretching bands of the fructose-furan ring [105].A weak band at 1653 cm −1 is attributed to -OH bending of residual water.The broad band at 3304 cm −1 is due to the -OH stretching vibrations of inulin and residual water, and it is broader and shifted as compared to Zn1 due to the additional contributions.
The IR spectrum of ZnOInuOP is similar to that of inulin with some additional features, most notably the steep increase at 400 cm −1 due to the Zn-O asymmetric stretching.A larger amount of bands are observed in the 600-800 cm −1 range due to the interaction with ZnO, which causes additional -CH 2 rocking vibrations.Furthermore, a small shift in the fructosefuran ring stretching to 1034 cm −1 and 1132 cm −1 is observed, which is also ascribable to the interaction with the ZnO nanoparticles.Additional bands are present at 1649 cm −1 , 1556 cm −1 , and 1508 cm −1 , compatible with the chelation of Zn by two carbonyl groups (e.g., β-diketones), such as in Zn acetylacetonate [106].These are an indication of the interaction between ZnO and inulin.The bending mode of residual water may contribute to the band at 1649 cm −1 .Finally, the broad band centered at 3365 cm −1 is a convolution of several contributions to the asymmetric stretching from -OH groups of inulin, as well as residual water.A shoulder at 2932 cm −1 is assigned to C-H asymmetric stretching.
Since the synthesis of Zn5@inu is carried out in two steps, its IR features must be offset against the bands of the naked nanoparticles, i.e., Zn5, that are typical of pure ZnO, with a steep absorption at 400 cm −1 and a rather flat spectrum in the whole remaining absorption range.
The lack of evident ν O-H vibrations in the region around 3300 cm −1 rules out the presence of residual water traces.Regarding Zn5@inu, a larger number of peaks is observed as compared to both inulin and ZnOInuOP in the 600-800 cm −1 range, thus indicating a more pronounced or a better-detected interaction with ZnO.Shifts are observed in the bands at 889 cm −1 and 829 cm −1 , and two additional peaks are present at 1086 cm −1 and at 1205 cm −1 .Furthermore, the relative intensity of the peaks in the 1000-1300 cm −1 region is opposite with respect to ZnOInuOP.In particular, Zn5@inu has stronger asymmetric C-C and C-O stretchings, whereas ZnOInuOP diplays a more intense C-O-C asymmetric stretching, possibly indicating a larger ring deformation in the latter case, due to a weaker coordination to ZnO.The bands at 1508 cm −1 and 1551 cm −1 , assigned to the chelation of Zn by two carbonyl groups, are stronger than of ZnOInuOP.Furthermore, a strong band appears at 1730 cm −1 , assigned to the stretching vibration of a carbonyl group that can be associated to an aliphatic ketone, an aldehyde, or an alfa-beta unsaturated ester.The combination of these observation points at a more pronounced chelation effect as compared to ZnOInuOP.A broad band at 3365 cm −1 is due to the contributions of asymmetric stretching from the -OH groups of inulin, as well as residual water.Finally, samples Zn2, Zn3, and Zn4 indicate the presence of ZnO, whereas Zn2@inu, Zn3@inu, and Zn4@inu all have similar spectra to Zn5@inu, though all features are less intense, hinting at a lower coverage by inulin.
The IR bands of the samples along with their assignment are reported in Table 1.

SEM
The SEM images were analyzed in terms of the shape of the particles, average size, and size variation upon coating with inulin (for samples prepared according to the two-step procedure).In addition, EDX analysis was carried out to determine whether the coating procedure was successful.The SEM images are reported in Figure 4a-k, whereas Figure 4l is an EDS map of sample Zn5@inu.
In more detail, the sample Zn1 (Figure 4a) shows the characteristic flower-like shape of Zn(OH) 2 [24,107] with an average size in the micrometer range, that is retained upon inulin coating (Figure 4b-Zn1@inu).Furthermore, a layer of coating can be observed in the inset at low magnification, whose analysis reveals the sole presence of C and O; hence, it was attributed to an excess of inulin.Samples Zn2, Zn3, Zn4, and Zn5 are all characterized by the presence of nanoparticles of a hexagonal to round shape.The nanoparticles of Zn2 have an average diameter of 62 ± 4 nm.The subsequent coating does not affect the nanoparticles' average diameter that remains at 63 ± 4 nm.The composition of Zn2@inu is rather poor on carbon, as compared to the other coated samples.However, extensive layers of carbonaceous material are observed all around the agglomerates, all in all, pointing at a partial degradation of the inulin.Samples Zn3 and Zn3@inu both have nanoparticles with a 60 ± 10 nm average diameter, inglobated in larger structures to form agglomerates.Sample Zn4 has nanoparticles with a hexagonal shape with an average size of 52 ± 5 nm.The overall appearance becomes more aggregated and the average size of the nanoparticles is 54 ± 5 nm.Zn5 nanoparticles are roundish and with average size of 65 ± 4 nm, that decreases to 58 ± 1 nm upon inulin coating.Finally, ZnInuOP, the one-pot-synthesis sample, is characterized by the formation of nanoparticles that aggregate in irregular shapes, with sizes in the range of 100-300 nm.The size of the nanoparticles is estimated to be 30 ± 10 nm.The shape, average size, and composition of the samples are reported in Table 2. Provided that the final goal of the preparation is the achievement of well-dispersed, wellcoated nanoparticles, based on the analysis of SEM images, the most-suited preparation appears to be Zn5@inu; therefore, EDS mapping was carried out on this sample (Figure 5).The overlapped mapping of Zn, O, and C indicates a homogenous distribution of Zn, O, and C, thus indicating a uniform coating of inulin.A few considerations are in order in terms of the preparation method, based on the comparison of the images of all preparations.Although small, well-dispersed NPs are achievable at 40 • C, the coating at the same temperature does not yield a homogeneous layer, and rather, induces agglomeration.On the other hand, a full preparation at 60 • C yields a sample with a rather low carbon content, probably due to a mild denaturation.The best combination of temperature parameters of the first and second preparation stages, appears to be T P = 60 • C, T D = 40 • C. It must be noted that the sole synthesis of ZnO may provide small, regularly shaped nanoparticles [24].However, the combination with the coating procedure requires higher starting temperatures.Finally, the one-pot preparation does yield samples that are well covered with inulin, but it is likely that a multi-layer is contributing to a large degree of agglomeration.

Thermogravimetric Analysis
The degree of coverage of the samples can be estimated through thermogravimetric analysis.The weight losses of selected samples are reported in Figure 5, whereas the weight loss differences between coated and uncoated pairs of samples are reported in Table 3.For sample ZnOInuOP, the difference was calculated with respect to Zn5.As far as uncoated samples are concerned, they all present major losses of the order of 3.5% weight due to water absorption.A distinction must be made, however, between Zn1, i.e., the hydroxide and the ZnO samples, since the former loses weight in a narrower temperature range.Additional weight losses related to the inulin coatings range between 0.15% (Zn1@inu) and 1.67% (ZnOInuOP).These values have been used to estimate the average number of inulin monomeric units per particle, taking into account the average

Thermogravimetric Analysis
The degree of coverage of the samples can be estimated through thermogravimetric analysis.The weight losses of selected samples are reported in Figure 5, whereas the weight loss differences between coated and uncoated pairs of samples are reported in Table 3.For sample ZnOInuOP, the difference was calculated with respect to Zn5.Table 3. Weight loss differences and degree of coverage of the coated samples.The degree of coverage is expressed as number of polymeric units per particle, converted to number of units per particle in the most right column.

Sample Weight Loss Difference Degree of Coverage Monomers
Zn1@inu 0.15% As far as uncoated samples are concerned, they all present major losses of the order of 3.5% weight due to water absorption.A distinction must be made, however, between Zn1, i.e., the hydroxide and the ZnO samples, since the former loses weight in a narrower temperature range.Additional weight losses related to the inulin coatings range between 0.15% (Zn1@inu) and 1.67% (ZnOInuOP).These values have been used to estimate the average number of inulin monomeric units per particle, taking into account the average diameter of particles of each sample, as determined by SEM imaging.The estimate is accurate only for fairly separated nanoparticles, i.e., Zn5@inu, whereas it is more of an indication for agglomerates.In the calculation, the following formula was applied: where mass inulin = mass of inulin per particle; mass uncoated-NPs = mass of a ZnO nanoparticle or mass of a Zn(OH) 2 particle, depending on the reference sample.Furthermore, it varies with the average diameter of the particle and it is for instance, 6.2 × 10 −16 g, for the hexagonal closely packed wurtzite structure of ZnO with a 58 nm diameter.%WL inulin = percentage weight loss due to inulin; %WL uncoated-NPs = percentage weight due to residual particles.This leads to an average estimation of 4.6 × 10 −18 g inulin per ZnO nanoparticle for the Zn5@inu sample that corresponds to 17,700 inulin monomeric units, or 590 polymeric units of 30 monomers each.Overall, the degrees of coverage are in line with the EDX measurements.The weight loss difference between Zn1@inu and Zn1 is rather small, as could be expected upon EDX analysis, indicating comparatively low C and O contents and corresponding to a coverage of 53,867 30-mers.Interestingly, the ratio between the coverage of the largest nanoparticle (Zn1@inu, d = 1200 nm) and the mean coverage of the two-step procedure samples (Zn5@inu, average d = 58 nm) is remarkably smaller than that expected by the surface increase, hypothesizing a one-layer coating or less.This feature hints at the establishment of a multi-layered coating in the smaller fragments.

Reflectance Spectroscopy
Nanoparticles typically influence the redox properties of plants upon adsorption; therefore, an important issue is the determination of their electronic properties [108,109].This was carried out using Diffuse Reflectance Spectroscopy (DRS), in particular, taking into account possible variations, correlated to the coating.In general, this technique is largely employed to estimate the electronic bandgap of solid materials, especially in the field of semiconductors, like the ZnO and Zn(OH) 2 considered in this article.The DRS spectra of the synthesized samples are reported in Figure 6 (Tauc plots).

Reflectance Spectroscopy
Nanoparticles typically influence the redox properties of plants upon adsorption; therefore, an important issue is the determination of their electronic properties [108,109].This was carried out using Diffuse Reflectance Spectroscopy (DRS), in particular, taking into account possible variations, correlated to the coating.In general, this technique is largely employed to estimate the electronic bandgap of solid materials, especially in the field of semiconductors, like the ZnO and Zn(OH)2 considered in this article.The DRS spectra of the synthesized samples are reported in Figure 6 (Tauc plots).Figure 6.Tauc plots for the determination of optical bandgap Eg and associated extrapolations of the linear regions (dotted lines).Red = Zn1@inu, gray = Zn2@inu, magenta = Zn3@inu, blue = Zn4@inu, light green = Zn5@inu, dark green = ZnOInuOP.
In particular, the reflectance spectra are associated with the absorption patterns by the Kubelka-Munk function [110,111], which states that the reflectance of an infinitely thick layer of sample ( ) obeys the following formula: Figure 6.Tauc plots for the determination of optical bandgap E g and associated extrapolations of the linear regions (dotted lines).Red = Zn1@inu, gray = Zn2@inu, magenta = Zn3@inu, blue = Zn4@inu, light green = Zn5@inu, dark green = ZnOInuOP.
In particular, the reflectance spectra are associated with the absorption patterns by the Kubelka-Munk function [110,111], which states that the reflectance of an infinitely thick layer of sample ( R ∞ ) obeys the following formula: in which, K and S are K-M absorption and scattering coefficients, respectively; the former parameter is related to the intrinsic absorption coefficient α by the equation α = K/2 [112], ultimately leading to K/S ≈ α whenever S can be considered constant across the various samples.Starting from the Kubelka-Munk absorption plot, the bandgap could be estimated by applying the Tauc relation [113]: In this relation, the exponent values n = 1/2 and n = 2 apply to direct gaps (such as those of ZnO and Zn(OH) 2 herein discussed) and indirect ones, respectively, with C1 being a constant.From the extrapolation of the linear part of the curve at the lowest energy side, corresponding to the energy value where α is zero (see Figure 6), the estimated bandgap values, which are reported in Table 4, could be obtained for the six inulin-coated samples.The reference values for bulk ZnO [114] and Zn(OH) 2 [115] bandgaps are 3.37 and 2.88 eV, respectively.
As is clear from inspection of Table 4, the four samples obtained with the two-step protocol, which the IR and XRD studies demonstrated to be composed of (coated) ZnO, share approximately the same bandgap value, within the experimental and fitting errors.Such similarity complies with the analogous XRD patterns observed for Zn2@inu, Zn3@inu, and Zn4@inu.In contrast, the one-pot sample ZnOInuOP, still containing only ZnO and the coating (no hydroxide detected), shows a significantly smaller bandgap.In both cases, the values are smaller than bulk value (3.37 eV) reported in the literature.This is contrary to the generally expected bandgap opening effect upon size shrinking owing to the quantum confinement phenomena [116,117].The sizes of uncoated ZnO nanoparticles were assessed with the same experimental procedure, and similar bandgap values (3.25-3.28eV, i.e., smaller than bulk value) [24] were found.These correlated to the nanoparticle diameters observed in an SEM study (around 15 nm) with a modified version of the Brus formula [112].Further and even more relevant effects can come into play when the coating is applied to the nanoparticles' surface.For instance, it was found that the amount of coating can remarkably reduce the bandgap (up to 0.5 eV) of graphene-oxide-coated Mn 3 O 4 nanoparticles [118].However, recently, Guan et al. [119] showed that treatment of titania with carbon powder led to the formation of fiber-like structures on the TiO 2 surface, with a sizable bandgap narrowing due to the formation or modulation of defects like oxygen vacancies (Ovs), onto the titania surface.Since similar mechanisms are thought to be crucial in modulating ZnO optoelectronic properties [120], it is not at all surprising that coatings, like those applied in the present study, may have a strong effect and reduce the bandgap by almost 0.4 eV in the one-pot coated sample.Therefore, the notable bandgap reduction in ZnInuOP may be due to a combination of effects, namely the coating effect just described and the aggregation of the (uncoated) nanoparticles, which have small dimensions individually (d ≈ 30 nm), into larger aggregates (hundreds of nm).Ultimately, this reduces the quantum confinement effect that the latter would bring about.Regarding Zn1@inulin, the only coated hydroxide considered in the analysis, it was found to have a 2.61 eV bandgap, thus showing a percentage reduction from the bulk value in line with that experienced by ZnInuOP, the other sample possessing extensive inulin coating.

Plant Growth Assays-Proof of Principle
Plant growth tests were performed on the selected samples, i.e., Zn5@inu and its uncoated counterpart, in order to appraise the biostimulating effect as well as a possible adverse (cytotoxicity) response to the materials and eventually correlate the observed data with structural and optical response properties.The plants employed in the tests were Vicia faba L (faba bean) seedlings, and the quantity of ZnO absorbed by the different plant tissues was assayed through ED-XRF analyses [89], focusing on the fluorescence signal coming from the zinc metal atoms that was proportional to ZnO concentration.As proof of principle, two dosing levels were considered, namely low and high dose, i.e., 50 mg kg −1 and 100 mg kg −1 , respectively.In detail, the measurements highlighted that the nanomaterial was thoroughly absorbed by the plant tissues.The increase in Zn level was actually significant in any case, with a larger relative accumulation in the leaves at the high dose and in the roots at the low dose.This corresponded to the increment ranges +289% to +9001% in the former and to +512% to +3704% in the latter case, for the two extreme cases, i.e., plantlets cultivated with ZnO + inu 50 mg kg −1 and with Zn5@inu 100 mg kg −1 .Interestingly, the administration of inulin only (both at 0.3% and 0.6%) caused significant changes in the Zn content in the leaves but not in the roots with respect to the control seedling.The full table of ED-XRF results is reported in [89].The germination tests carried out showed that no treatment induced a sizeable change in the seed germination percentage, except for the treatment with uncoated ZnO at the maximal dose; under these conditions, an increase in the germination rate was observed, suggesting a possible role in promoting sprout development of this compound.On the contrary, in the post-germination phase, the high accumulation of zinc, especially in plant leaves, seems to result in a marked effect on plant growth; indeed, the seedling dimensions were found to increase by an impressive +103% when Zn5@inu was administered, followed by a noteworthy +84% when the mixture ZnO + inu was supplied, and by a still noticeable +53% when uncoated ZnO was given to the plant.Parallel to the high stimulating power, the zinc oxide/inulin composite shows an absence of toxicity/genotoxicity (trypan blue staining test [89]), even at the highest dose administered.In particular, for the former assessment (toxicity), very negligible differences were evidenced between each sample (coated and uncoated) and the control sample, independent of the amount supplied.A maximum cytotoxicity value of approximately 17% of the total area was reached, which was regarded as physiological, considering that 11% of the stained tissue was assessed in the CNT.Regarding the latter (genotoxicity), the frequency of micronuclei (MN) was evaluated, and comparable values between the negative control (CNT) and all the various Zn@inu-treated samples were found both after 24 h and 72 h of treatment, thus suggesting the absence of genotoxicity.This is at variance with a parallel experiment where the mutagenic herbicide maleic hydrazide, considered as the positive control, was supplied to the same plant specimens.The latter appraisal showed a marked MN frequency increase and confirmed V. faba sensitivity to mutagenic compounds.A further confirmation came from the absence of hydrogen peroxide, signaling that oxidative stress in the plant is very low.Finally, the NP treatment with Zn5@inu at a high dose led to a remarkable increase in chlorophyll a and b contents (+19% and +30%, respectively) in the seedlings, accompanied by an increase in carotenoid content (+22%), compared to the control sample.Treatment with ZnO led to slightly lower figures (+16% and +27%), whereas the opposite behavior was observed for the ZnO + inulin combination (no coating), scoring a decrease (−15%) in Chl b content.Overall, our findings reveal that the coated version of the prepared ZnO nanoparticles (maximally Zn5@Inu) can give a very large boost to plant growth in the post-germinative phase.The most plausible reason for this issue resides in the sugar coating, which renders the NPs' internalization and transport across the plant tissues more efficient, thus promoting a massive metal migration towards the leaves once absorbed by the roots, while protecting against their degradation, and ultimately increasing their bioavailability in the periphery.The accumulation of ZnO NPs in the plant leaves can counteract the oxidative stress by acting as Reactive Oxygen Species (ROS) scavengers.Moreover, owing to the optoelectronic properties, which are almost fully conserved upon inulin coating, as evidenced by the study reported in Paragraph 2.5, the material may act as a spectrally selective system that optimizes the solar energy harvesting by the underlying leaf green pigments.This is carried out by filtering out the unusable portion of the solar spectrum and minimizing unwanted system overheating [121,122], with the inulin coating playing the role of the preferential carrier of the ZnO NPs through the apoplastic and symplastic pathways [123].A summary of the biostimulation effects of Zn5@inu is reported in Figure 7 [89].
Biostimulation with ZnOInuOP at both doses was not very effective, since germination and seedling growth were similar to the control, within the error margin.This opens up a complete scenario on possible explanations.The major differences in ZnOInuOP with respect to Zn5@inu are the larger NP aggregation and the lower bandgap, which may interplay to yield a low level of biostimulation.On one hand, the large, coated aggregates might encounter difficulties entering the root sheaths, or being metabolized and transported to the leaves.On the other hand, the small bandgap might play a negative role in plant growth, in so far as the particles are absorbed and remain as a whole.
the study reported in Paragraph 2.5, the material may act as a spectrally selective system that optimizes the solar energy harvesting by the underlying leaf green pigments.This is carried out by filtering out the unusable portion of the solar spectrum and minimizing unwanted system overheating [121,122], with the inulin coating playing the role of the preferential carrier of the ZnO NPs through the apoplastic and symplastic pathways [123].A summary of the biostimulation effects of Zn5@inu is reported in Figure 7 [89].Biostimulation with ZnOInuOP at both doses was not very effective, since germination and seedling growth were similar to the control, within the error margin.This opens up a complete scenario on possible explanations.The major differences in ZnOInuOP with respect to Zn5@inu are the larger NP aggregation and the lower bandgap, which may interplay to yield a low level of biostimulation.On one hand, the large, coated aggregates might encounter difficulties entering the root sheaths, or being metabolized and transported to the leaves.On the other hand, the small bandgap might play a negative role in plant growth, in so far as the particles are absorbed and remain as a whole.
XRD patterns were collected using an X'pert pro X-ray diffractometer by Philips, operated with CuK-alpha radiation.The scans were carried out with a 2θ step size of 0.02° and a 2 s/step counting time.FTIR spectra were obtained with a Shimadzu Prestige-21 instrument equipped with an attenuated total reflectance (ATR) diamond crystal (Shimadzu Ltd., Kyoto, Japan) in the range of 400-4000 cm −1 , with a resolution of 4 cm −1 .
XRD patterns were collected using an X'pert pro X-ray diffractometer by Philips, operated with CuK-alpha radiation.The scans were carried out with a 2θ step size of 0.02 • and a 2 s/step counting time.FTIR spectra were obtained with a Shimadzu Prestige-21 instrument equipped with an attenuated total reflectance (ATR) diamond crystal (Shimadzu Ltd., Kyoto, Japan) in the range of 400-4000 cm −1 , with a resolution of 4 cm −1 .
SEM images were taken with a Zeiss Auriga Field-Emission Scanning Electron Microscope instrument SUPRA™ 35 (Carl Zeiss SMT, Oberkochen, Germany) operating at 6-8 kV.The EDX analysis was carried out, operating the SEM at 20 kV and coupling it with the EDS/EDX, using INCAx-sight, Model 7426 (Oxford Instruments, Abingdon, Oxfordshire, UK).The samples for SEM imaging were prepared either by a direct deposition on carbon tape or by drop-casting deposition of a powder dispersion in ethanol onto a silicon sample holder.Termogravimetric analysis (TGA) was also performed using a Q600 thermobalance instrument, in the presence of a dry N 2 flow (100 mL min −1 ), with a ramp of 10 K min −1 from 300 to 1073 K. Ultraviolet-visible (UV-Vis) and near infrared (NIR) studies were performed using diffuse reflectance instrumentation by Avantes.The apparatus is composed of an integrating sphere (diameter 30 mm) coated with a highly reflective material (Spectralon ® ) that is illuminated by the radiation emitted by a halogen/deuterium source (AvaLight-DH-S-BAL-Labsphere, Lafayette, CO, USA), collimated by UV-Vis lenses, and introduced into the sphere by optical fiber wires.The sphere acts as a light collector, and is then used as a diffuse light source for the measurements (diameter of the sampling port, 6 mm), and to gather all the radiation reflected back by the sample, thus approximating the ideal Lambertian reflection condition.The output light is sent through another optical fiber to a laptop-controlled spectrometer (AvaSpec-2048 X 14-USB2) that generates the final pattern in the 248-1050 nm range, at 2.4 nm resolution, through the comparison with the reference reflectance of a factory-calibrated Spectralon ® (Labsphere, North Sutton, NH, USA).The preparation of samples for DRS first includes the mixing of the nanoparticle powder (approx.20 mg) with 480 mg of BaSO 4 , an utterly reflective compound, in order to achieve a very low absorption.The specimen is then introduced into a sample holder, deep enough to approximate the reflectance spectra to that of an infinitely thick sample (R∞).In the measurements of zinc-containing nanoparticles, a total of five 5 s scans were acquired and averaged.
Vicia faba seeds from the Botanical Gardens of Rome Tor Vergata were used for plant growth, cytogenetic, and cytotoxicity experiments.Plant growth and germination experiments were conducted as previously described [89].Briefly, faba bean seeds were sterilized with 5% sodium hypochlorite, washed with sterile water, and soaked for 24 h at 4 • C in the dark before planting in sterile Microbox vessels filled with agarized Murashige and Skoog culture medium (MS, pH 5.8).MS was enriched with 50 and 100 mg kg −1 of Zn5@inu, Zn5, Zn5 + inulin, or pure inulin, taking into consideration the inulin coverage of the Zn5@inu sample.For inulin alone, this corresponds to a 0.3 and 0.6% weight for the chosen treatment dose.Pure MS was used for the controls (CNT).A total of 25 Microboxes (each containing five seeds) were cultured at 22 • C for 21 days with a 14 h photoperiod (120 µmol m −2 s −1 ).Plants were than uprooted, their roots washed with distilled water, and the total length of the plants determined (millimeter precision ruler).Germination percentages were calculated as the ratio between germinated seeds and the total number of planted seeds multiplied by 100.Plant samples not analyzed immediately were stored at −80 • C. For cytogenetic analysis, sterilized seeds were sown in agarized MS.After 4 days, when the major roots were 2 cm long, the root ends were removed by cutting about 5 mm from the distal end to promote secondary root formation.Simultaneously, seed integuments were removed and seedlings were placed in plastic basins with liquid MS.To maintain the cotyledons on the solution's surface, a plastic grid was used to support the seeds while keeping the roots completely immersed.Seedlings were housed under these conditions in the dark, as specified in the standard protocol for micronuclei analysis, [124] for four days before the nanoparticles were delivered.Each experimental point consisted of 10 germinated seedlings, including negative and positive controls.During treatment, the root systems were immersed in zinc-containing solutions for 24 to 72 h.The genotoxicity of nanomaterials and inulin was tested at a dose comparable to the maximum exposure in in vitro cells (100 mg kg −1 ).Maleic hydrazide (Merck), a highly carcinogenic herbicide, was employed as a positive control (PC) at a concentration of 10 −4 M in a water solution for 4 h, followed by a 20 h recovery period in MS.The negative control (CNT) consisted of exposing roots to pure MS.Once treatments had been accomplished, secondary roots were chopped and fixed in Carnoy solution (acetic acid:ethanol, 25:75, v:v) for 30 min before being transferred to a new fixing solution and maintained overnight at 4 • C.After Feulgen staining, root tips were crushed on microscope slides with 45% (v:v) acetic acid and permanently mounted with Eukitt (ForLab, Stezzano, Italy).MN frequencies were calculated in 25,000 cells at each experimental point (2500 cells per root tip, for a total of 10 treatments).MN frequency analysis was restricted to proliferating cell populations, as determined by the mitotic index of each examined root tip.Cytotoxicity was assessed using trypan blue staining, whereas loss of cell viability was assessed using the Duan et al. [125] methodology with minimal changes.Plant tissues were immersed in a solution of 1% (w/v) trypan blue in PBS for 45 min in the dark.After three PBS washes, the samples were gently agitated with 100% ethanol to remove any remaining chlorophyll before being rehydrated in a 50% (v/v) glycerol solution.Images were taken using an Epson Perfection V700 Photo system, and dark patches were quantified against the total surface using Fiji/ImageJ software (Version 2.14, National Institutes of Health, Bethesda, MD, USA); the results were reported as a percentage.Also, the pigment contents in the leaf samples (chlorophyll a (Chl a), chlorophyll b (Chl b), and total carotenoid (Car)) were determined by UV-Vis spectrophotometry as described previously on liquid nitrogen powdered leaf samples after extraction with 80% (v/v) acetone for 24 h in the dark at 4 • C [89].Absorbances of the centrifuged supernatants were monitored at 470, 645, and 663 nm with a Varian Cary 50 Bio UV-Vis spectrophotometer (Varian, The Netherlands).
The presence of Zn in plant samples was examined by X-ray fluorescence using an energy-dispersive X-ray fluorescence (ED-XRF) spectrometer (SPECTRO XEPOS HE XRF) that was tailored for heavy elements [89].The equipment ia outfitted with a Pd/Co alloy X-Ray tube.The results were acquired by the "TurboQuant powders and liquid" technique (XRF Analyzer Pro software, version TurboQuant II) using the SPECTRO process calibration model (a combination of the Fundamental Parameter and Extended Compton scattering models, with mass attenuation coefficient calibration).Every experiment was conducted a minimum of three times, and the outcomes are presented as means ± standard error.One-way analysis of variance (ANOVA) was performed on the data, and the post-hoc lowest standard deviation (LSD) test was used to assess the statistical differences (p values: * p < 0.05; ** p < 0.01; *** p < 0.001).PAST software, version 4.03, was utilized for this purpose.The Mann-Whitney U non-parametric test was used for the cytogenetic analysis, and values of * p < 0.05 were regarded as significant.

Two-Step Synthesis
The two-step procedure foresees the synthesis of the ZnO NPs by precipitation and ensuing the addition of inulin at 3% weight, according to the following reaction: where T P and T D are the precipitation and digestion temperatures, respectively.In more detail, 0.1 M solutions of Zn(NO 3 ) 2 and NaOH are prepared by dissolving suited amounts of reagents in deionized water.A total of 100 mL of NaOH is added dropwise to 100 mL of Zn(NO 3 ) 2 solution previously heated to the selected temperature (see Table 5), kept constant by immersion of the flask in an oil bath.A fine milky precipitate forms immediately and the ensuing suspension is kept under stirring for 24 h to allow digestion of the precipitate.Afterwards, the solution is split into two portions.The first aliquot is centrifuged at 3000 rpm for 10 min and the precipitate is rinsed with deionized water.The procedure is repeated 3 times prior to the final drying at T D temperature that yields a white precipitate.The second aliquot is added with a previously prepared dispersion of inulin in deionized water (3% w/v) and left under stirring at T D temperature for another 24 h.The resulting suspension is centrifuged at 3000 rpm for 10 min and the precipitate is rinsed with deionized water.This procedure is also repeated 3 times.Eight sets of samples are synthesized by varying the T P and T D temperatures according to the values reported in Table 5. Precipitation and digestion temperatures have been chosen on the basis of our previous studies on the low-temperature synthesis of ZnO nanoparticles in order to assure precipitation of ZnO nanoparticles and the preservation of their properties during the digestion step [24,89].

Gel-like Synthesis
The starting suspension of inulin (concentration 1% w/v) is prepared by dissolving weighed quantities of inulin into 50 mL of deionized water, through stirring at room temperature for 24 h.After the suspension of the polymer is complete (and forming a sort of gel), solid Zn(NO 3 ) 2 •6H 2 O is added to achieve the target concentration of 0.1 M. The solution is stirred for 2 h at room temperature and is then thermalized at 40 • C in an oil bath under vigorous swirling for a further 2 h.Next, 50 mL of NaOH aqueous solution is added at the same concentration (1:1 w/v ratio) dropwise and the solution is left under stirring overnight at 40 • C. The ZnO@inu nanocomposite is formed during the subsequent centrifugation phase that is performed for 10 min at 3500 rpm and 25 • C. The achieved nanocomposite is washed once with deionized water and dehydrated at 50 • C for 24 h.The sample obtained with this procedure is labelled ZnOInuOP.

Conclusions
In the present paper, we investigated the preparation procedures of inulin-coated ZnO NPs to tune the synthetic parameters for biostimulation applications.Subsequent characterizations provided information on the chemico-physical properties.The XRD study highlighted that all the samples synthesized and dried at temperatures greater than 40 • C, regardless of the synthetic strategy, contained pure zinc oxide, as previously reported, and that preparation at lower temperatures yielded zinc hydroxide nanoparticles.The infrared spectroscopy investigation confirmed the purity of the samples and reveals that the NPs were successfully coated with inulin, as it is highlighted by the presence of the fructan vibrational fingerprint.However, the binding mechanism was different in one of the cases, i.e., for sample Zn5@inu, prepared at T P = 60, T D = 40 • C, since the IR spectrum indicated the formation of Zn complexes.The SEM microscopy analysis showed that all ZnO-based samples possess nanometric sizes, whereas Zn(OH) 2 is micrometrical.Also, the one-pot procedure led to the formation of micrometric aggregates of small nanoparticles; therefore, this is less well suited for a homogeneous application in priming protocols.The coating, instead, was most homogeneous and likely did not undergo a denaturation procedure for sample Zn5@inu.The diffuse reflectance study endorsed the role of coating in modulating the material bandgap and shows that the nanocomposites prepared with the two-step procedure had similar values in the 3.25-3.30eV range.In contrast, the sample prepared with the one-pot reaction pathway showed a remarkably smaller energy difference between the ground and excited states, decreasing to 3.0 eV, likely because of the formation of lower energy defect states within the gap.The observed electronic behavior complied satisfactorily with the structural properties highlighted in the SEM study, which hinted at the presence of quite large micrometric aggregates of small nanometric particles (diameter 30 nm), externally coated by the inulin in the one-pot synthesis sample.Our study indicated that the preparation carried out in two steps at T P 60, T D 40 • C yields the most homogeneous and uniformly coated nanoparticles, with electronic properties in line with their uncoated counterparts; hence, they are more suited for biostimulation applications.Proof-of-principle assays employing the latter sample, both coated and uncoated, indicated an effective growth stimulation of Vicia faba seedlings upon seed priming, which was more pronounced at doses of 100 mgKg −1 .The biostimulation worked better when using inulin-coated particles and concomitant Zn accumulation in the leaves was observed.No inherent cyto-or genotoxicty was observed.All in all, the inulin coating played the role of preferential carrier of the ZnO NPs.Given these very encouraging premises, additional studies on other types of coating are in progress, with the aim of further improving the nanoparticle biostimulating effect while retaining the electronic features of the material, as well as its overall formulation chain.Among the many possible future applications, conceivable for our systems, we envisage a prominent role in the manufacturing of a new, low-cost, and marketable class of zinc-nanoparticle-based fertilizers with enhanced features, going from the ready absorption from plant root tissues to the transport through the plant stem up to the leaves.The use of highly bioavailable forms of ZnO like those constituting our nanoparticles, being the oxide compound with the highest metal content among all zinc compounds administered in agriculture, may give a significant boost in the treatment of zinc-deficient soils and plants, and finally increase food and nutrition security.

Figure 1 .
Figure 1.A sketch of the inulin monomer.Asterisks mark the connection points.

Figure 1 .
Figure 1.A sketch of the inulin monomer.Asterisks mark the connection points.

Figure 3 .
Figure 3. IR spectra of selected samples: violet solid line = Zn5, light-green solid line = Zn5@ dark-green solid line = ZnOInuOP, pink solid line = inulin, red solid line = Zn1@inu, black so line Zn1.The Zn1 sample displays typical features of Zn(OH)2, i.e., sharp and intense v tional peaks around 3280 and 716 cm −1 associated with O−H stretching and ben modes.The presence of zinc hydroxide is further confirmed by the characteristic abs tion for Zn−O−Zn asymmetric stretchings located at 1085 and 1030 cm −1 and Zn−O la vibrations at around 485 cm −1[104].Zn1@inu displays similar features to Zn1, thus cating a coating of the particles below the detection limit.At variance with this, coated samples Zn5@inu (two-step procedure) and ZnOInuOP (one-pot procedure) sent features ascribable to inulin layers covering the nanoparticles.For comparison poses, the spectrum of the inulin used for the preparation is also reported in Figure3

Figure 3 .
Figure 3. IR spectra of selected samples: violet solid line = Zn5, light-green solid line = Zn5@inu, darkgreen solid line = ZnOInuOP, pink solid line = inulin, red solid line = Zn1@inu, black solid = line Zn1.The Zn1 sample displays typical features of Zn(OH) 2, i.e., sharp and intense vibrational peaks around 3280 and 716 cm −1 associated with O−H stretching and bending modes.The presence of zinc hydroxide is further confirmed by the characteristic absorption for Zn−O−Zn asymmetric stretchings located at 1085 and 1030 cm −1 and Zn−O lattice vibrations at around 485 cm −1[104].Zn1@inu displays similar features to Zn1, thus indicating a coating of the particles below the detection limit.At variance with this, both coated samples Zn5@inu (two-step procedure) and ZnOInuOP (one-pot procedure) present features ascribable to inulin layers covering the nanoparticles.For comparison purposes, the spectrum of the inulin used for the preparation is also reported in Figure3.The spectral features of inulin may be divided into three regions.The first one, at low wavenumbers in the 600-800 cm −1 range, is assigned to -CH 2 rocking vibrations.The region between 800 cm −1 and 1400 cm −1 carries the inulin fingerprints.In particular, peaks at 827 cm −1 and 866 cm −1 are representative of the anomeric bendings δ(C1-H) and ring vibration of 2-ketofuranose, respectively, and are indicative of the β-(2→1) glycosidic bonds.In contrast, peaks at 1016 cm −1 , 1109 cm −1 , and 1219 cm −1 are assigned to the stretching

Table 1 .
IR bands of selected samples and their assignments.

Table 2 .
Average size and composition of the samples.
* Silicon and additional oxygen may be present as contribution from the sample holder of SiO 2 .Int.J. Mol.Sci.2024, 25, x FOR PEER REVIEW 9 of 23

Table 5 .
Summary of the preparation conditions of the samples.
* The final drying of this sample was carried out at 50 • C.