For decades, farms have used pharmacological concentrations of zinc oxide (ZnO) to combat post-weaning diarrhea [1, 2]. Different modes of action for the antibacterial activity of ZnO have been proposed, one of them being the partial dissolution of the ZnO particles in the enteric region and the release of Zn²⁺ [3]. When ZnO is formulated to reduce this undesirable dissociation in the stomach, less ZnO may be added to the diet [4]. This is important, since much of the ZnO given to the piglets is excreted with their waste, causing accumulation of this metal in the environment [5, 6]. The incorporation of ZnO nanoparticles in polymer matrices can facilitate their gradual and sustained release [7].

Chitosan (CH) and alginate (SA) are two of the most commonly studied biopolymers, particularly for drug delivery [8]. CH is the partially deacetylated derivative of chitin, composed of repeating units of β-(1,4)-2-amino-2-deoxy-D-glucose and β-(1,4)-2-acetamido-2-deoxy-D-glucose [9]. SA is extracted from brown seaweeds and contains carboxylic acid groups [8], presents linear chains with blocks of 1–4–linked α-L-gluronic (G-block), β-D-mannuronic acid (M-block) residues, and blocks of alternated G and M residues [10].

In the present study, CH and SA, with and without the addition of sodium tripolyphosphate, were combined to encapsulate nanostructured ZnO particles aiming for the control of their enteric release. To protect ZnO from the gastric acid environment, the composite structures were designed to maintain acid-soluble chitosan in the internal phase of the complex. The physical, chemical and rheological characteristics of the resulting products were analyzed by infrared spectroscopy, X-ray diffraction, atomic absorption spectrophotometry, thermogravimetry and scanning electron microscopy. The antibacterial properties of the composites were tested against Escherichia coli and Staphylococcus aureus and the MICs were calculated. The in vitro release profile of Zn²⁺ from the composites in both gastric and enteric simulated fluids was also investigated.

The crystallite size of nano and commercial ZnO (D_XRD = 17 nm and 0.22 μm, respectively) was calculated from Scherrer’s formula applied to the most intense reflection (101), as observed in the diffractogram displayed in Fig. 1a. It was observed that the reflections obtained for nano ZnO (trace II) were less crystalline with wider bases, characteristic of nanoparticles. Figure 1c shows that ZnO nanoparticles reduced the crystallinity when comparing to chitosan (CH) (Fig. 1b, trace II), as reflections were weaker and wide, as observed in a previous study [13]. Three strong reflections could be noticed, displayed in Fig. 1a between the angles of 2θ = 30° and 40°, corresponding to ZnO, but the diffraction corresponding to ZnO/CAT and ZnO/CA samples were very small, revealing the intense reduction in the orientation of crystallinity (Fig. 1c). It should also be noted that the ZnO/CAT samples encapsulated 10.2% of zinc and had a higher crystallinity than the ZnO/CA samples, with 12.6% Zn.

The FTIR spectrum for SA (Fig. 2a, trace I) shows intense absorption bands with maximums at 3402, 2920, 1613, 1419 and 1032 cm⁻¹. These bands were attributed to stretching vibrations of O-H, C-H bonds, asymmetrical and symmetrical stretching vibrations of the carboxylate groups and, finally, to C-O-C bonds from the polysaccharide structure. For the CH samples, the broad absorption with maximum at 3401 cm⁻¹ was assigned to -OH and -NH₂ stretching vibrations (Fig. 2a, trace II). The absorptions at 1657 cm⁻¹ and 1594 cm⁻¹were attributed, respectively, to the amide I (C = O stretching) and amide II (C–N stretching and C–N–H bending vibrations) from amide functional groups in the solid state. In this region (1650–1580 cm⁻¹), the absorption attributed to the N-H bending (scissoring) vibration would be expected to appear. The strong absorption at 1094 cm⁻¹ was attributed to the C-O-C bond from the polysaccharide structure.

Figure 2b displays the spectra for nano ZnO (trace I), the product obtained by encapsulating ZnO in the CH/SA complex (ZnO/CA, trace IV) and by encapsulating ZnO in the CH/SA complex reinforced with TPP (ZnO/CAT, trace III). For ZnO/CA (trace IV), the absorption at 471 cm⁻¹, which also appeared in the ZnO nano (trace I), in commercial ZnO (trace II) and in ZnO/CAT spectra, were attributed to the Zn-O stretching vibration. This result pointed to the incorporation of ZnO to the CH/SA and CH/SA/TPP complexes.

The dynamic rheological results are displayed in Fig. 3 with the mechanical spectra for the SA/CH (CA) and SA/CH/TPP (CAT) complexes shown in Fig. 3a. The elastic behavior (G’ > G”), observed for both products, along the range of studied frequencies, revealed the formation of a tridimensional network.

Surface morphologies by SEM analysis of the nano and commercial ZnO, ZnO/CAT and ZnO/CA are displayed in Fig. 4. The surface morphology of pure ZnO nanoparticles (Fig. 4a) showed a fine powder with strong aggregation, characteristic of nanomaterials. In the ZnO/CAT micrograph, microparticles can also be observed as influencing the crosslinking of the biopolymer and the surface structure of the material (Fig. 4b).

The presence of 78% Zn²⁺ in the new pure synthesized ZnO nano and 99% in the commercial ZnO was observed, while the ZnO/CAT and ZnO/CA composites showed 10.2 and 12.6%, Zn²⁺, respectively by atomic absorption spectrophotometry (AAS). The particle size of each composite is displayed in Table 1 and Fig. 5. In all measurements, the ZnO/CAT samples were shown to be a smaller particle than the ZnO/CA samples. When the mean values were evaluated, a value of 114.32 μm of ZnO/CAT against 162.48 μm ZnO/CA was observed. However, the poly dispersion index expressed in Span value for this sample was higher (1.80 μm), revealing a less homogeneous distribution of particles in comparison with the ZnO/CA, that had a smaller span value (1.24 μm) (Fig. 5).

Sample	d10% (μm)	d50% (μm)	d90% (μm)	d4,3 (μm)	Span (μm)
ZnO/CAT	18.59	109.63	216.46	114.32	1.80
ZnO/CA	47.92	169.11	257.16	162.48	1.24

The amount of inorganic residues (after heating to 700 °C) can be observed in the TGA curve, that indicated mineral matter content of 40%, approximately, for ZnO/CA, and 55% for ZnO/CAT (Fig. 6). The largest residue observed in the ZnO/CAT sample is due to phosphorus (P) from the TPP and Zn²⁺ present in ZnO. However, the results of the AAS Zn²⁺ quantification indicated that the ZnO/CAT samples have lower Zn²⁺ concentration (10.2%) compared to the ZnO/CA samples (12.6%). The in vitro release assay demonstrated that the ZnO samples complexed with CH, SA and TPP (ZnO/CAT) and no TPP (ZnO/CA) were effective in protecting Zn²⁺ in simulated gastric medium, as displayed in Fig. 7a. In SGF, at time zero, nano and commercial ZnO samples had the lowest release of Zn²⁺ between the 4 tested samples, however, the first 15 min revealed an increase of 11.2%, from 2.4 to 13.6% (commercial ZnO) and 7.4%, from 2.8 to 10.25% (ZnO nano). The release of ZnO/CA and release of the ZnO/CAT sample remained constant (4.5% and 6.2%) during the same period. From 15 min onwards, release by the ZnO nano and commercial sample increased and became more intense, reaching 40.4% and 27%, respectively, at 45 min. It was verified that despite the nano ZnO contain less Zn²⁺ in its formulation (78%) when compared to the commercial ZnO (99%), it released 1.5 times more at 45 min. This behavior was not observed in the protected samples, ZnO/CAT and ZnO/CA. Both samples had a low and constant release taxa as shown in Fig. 7a, 8.0 (5.3%) to 10.7 (6.9%) times lower than that observed in the ZnO nano sample. Figure 7b indicates that, at the beginning of the test using SES, nano ZnO released less Zn²⁺ (0.1%) than the ZnO/CAT (2.3%) and ZnO/CA (2.3%) samples and remained as such until the end time point (120 min) of the experiment, when a release of 4.2% and 4.5% was recorded, respectively. While commercial ZnO showed an initial release of 1.64% Zn²⁺ and a gradual increase of 2.62% tea at 120 min.

The addition of TPP (Fig. 1c) to the formulation interfered in the acquired reflections, where the ZnO/CA samples showed higher amounts of amorphous structures, a sodium alginate (SA) characteristic, as observed in Fig. 1b, line I, corroborating data from a previous study [14].

The band shift attributed to the asymmetrical stretching vibrations of the carboxylate group, acquired by FTIR, reported the composite formation, where a linkage occurs between the SA carboxylate groups and the CH amine groups, but only a slight shift (from 1613 cm⁻¹ for SA to 1612 cm⁻¹ from the ZnO/CA composite) was observed (the band is broader and more intense). Interestingly, a much more expressive shift to lower wavenumbers (from 1613 cm⁻¹ and 1419 cm⁻¹for SA to 1600 cm⁻¹, and 1407 cm⁻¹ for ZnO/CAT composite, respectively) of both bands was observed in the ZnO/CAT spectra (Fig. 2b, trace III). This clearly demonstrates the TPP contribution in increasing interactions in this composite.

Moreover, the spectra obtained for these products corroborated the FTIR data, reinforcing the role of TPP in increasing physical interactions between the gel components. The CAT matrix forms a stronger gel than CA (without TPP). On the other hand, the addition of ZnO (Fig. 3b) seems to disturb the gel structure. ZnO/CAT showed G’ and G” values superior to those for ZnO/CA.

The ZnO/CAT complex showed a smooth and fine surface, corroborating a previous report by [15], which may be due to interfacial interactions between the CH chains and nano ZnO, which could possibly act as intermolecular crosslinks. Therefore, ZnO nano can contract and restrict the mobility of CH chains, and subsequently change the expansion of the surface morphology. On the other hand, when observing Fig. 4c, a rougher and crowded surface was observed, probably due to the lack of TPP. This reveals the influence of the addition of TPP on ZnO/CAT morphology (Fig. 4b), in addition to ZnO linking with CH. The addition of TPP decreased the particle size, probably due to crosslinking of polymeric chains, resulting in greater particle compression.

The high performance of the ZnO/CAT sample in SGF can be explained by the addition of TPP, the crosslinking agent, used in the preparation of CH due to its rapid gelling ability [16], where TPP reduces the gel pore size and effectively traps Zn²⁺ in the interstices of the support matrix [16, 17]. In simulated enteric juice, the ZnO/CAT and ZnO/CA complexes released more Zn²⁺ than ZnO nano and commercial samples, as expected. The ZnO delivery systems developed in this study allow for high concentrations of Zn²⁺ in the enteric environment, which are necessary to achieve the antimicrobial effect. The partial dissolution of ZnO particles releases zinc ions (Zn²⁺), which contributes to the antimicrobial activity of the oxide [3] against the enteric microbiota of weanling pigs.

The Zn²⁺ dissociation from the ZnO composites is due to the formation of strong electrostatic but reversible bonds, even without the use of any covalent crosslinking agent [18]. Furthermore, the SA layer could have prevented the gastric degradation of CH, because the positive charges from CH reduced solubility at low pH. However, at around neutral pH, a higher amount of negative charges sequester positive CH charges, resulting in dissolution of the CH/SA complex and increase in the release of Zn²⁺ in the middle enteric region [19]. ZnO/CA and ZnO/CAT showed good release of Zn²⁺ in SEF, triggered by the change in the pH.

The lower concentrations of ZnO/CA are probably due to the absence of TPP, thus allowing closer contact between Zn²⁺ and the target microorganisms, where ZnO/CA showed a complete inhibition of E. coli growth at lower concentrations than ZnO/CAT (126 μg/mL Zn²⁺ incorporated against 229.5 μg/mL). For S. aureus, ZnO/CA prevented the growth complete at a lower incorporated concentration of Zn²⁺ (126 μg/mL), compared to the 204 μg/mL of ZnO/CAT. The CH/SA complex network, even in the absence of TPP, showed greater efficacy when compared to pure ZnO nano, which is effective in inhibiting E. coli and S. aureus at concentrations of 273 μg/mL and 195 μg/mL, respectively.

A comprehensive mechanism for the antimicrobial effect of nano-ZnO particles is still under debate. In an aqueous environment, the ZnO NPs aggregate to micrometer-sized particles and do not interact effectively with microorganisms. Thus, the antimicrobial activity has been mainly attributable to the dissolved zinc species [20–24]. The dissolution of ZnO nanoparticles and Zn²⁺ release seem to be able to generate radical oxygen species (ROS), consequently, activating a cytotoxic cascade of events that include intracellular calcium flux, mitochondrial depolarization, and plasma membrane leakage [25–28].

In the conditions of the present study, S. aureus was shown to be more sensitive to both the ZnO nano and ZnO/CAT samples in comparison to E. coli. The difference in antibacterial activity against both microorganisms can be related to differences in the chemical and structural composition of the cell membranes, particularly of the cell wall [29, 30] and to the preparation of the composites. S. aureus tends to develop defenses against oxidative stress by releasing products such as superoxide dismutase, catalase and thioredoxin reductase [31–33]. In our study, we can suggest that the new particles of ZnO were able to assemble at cellular membrane and damage it, despite the greater defense of S. aureus to the toxicity of ZnO.

Some ZnO particles can attach to the bacterial membrane surface, damaging it, and this mechanical damage has been postulated as a probable mechanism of antimicrobial inhibition [31]. According to the literature, ZnO nano bacteriostatic effectiveness can be ascribed to the disorganization of the membrane, increasing its permeability and resulting in the internalization of the nanoparticles. One possible explanation is that the antibacterial activity exhibited by ZnO composites is based on the abrasive texture of the ZnO surface. This abrasive texture, due to surface defects, fixes ZnO to the protein molecules from microorganism membranes, promotes internalization and consequent cell metabolism inhibition, killing the bacteria [34].

Several factors can interfere with the daily feed intake by animals, such as diet composition, the quality of raw materials and the sanitary conditions of farms [35, 36]. According to the health challenge of each farm and immune conditions of piglets, different concentrations of ZnO in their diets can be adopted, which usually range between 3000 and 5000 ppm (mg/kg). Considering an average daily intake of 400 g of feed by animal, established according to their growth phase (weaning- 21 days) [37–39], containing 4000 ppm (mg/kg) of the commercial ZnO, it should be expected the release of approximately 31.7 mg of Zn in the piglet gut, an amount 6 fold less than that, which corresponds to the MBC against E. coli, 198 mg Zn/0.4 kg of feed, and 8 fold less than MBC for S. aureus − 257.4 mg Zn/0.4 kg of feed. Released Zn2+ seems to be under the effective MBC threshold, particularly because considerations about the water intake and gut secretion should be done in order to calculate Zn2+ concentration in the gut.

The ZnO/CAT composites obtained in this study, at the same concentration of ZnO, led to a release of 7.4 mg of Zn, a value 12-fold lower than the MBC for E. coli (91.8 mg Zn²⁺) and 11-fold lower than for S. aureus (81.6 mg). However, ZnO/CA, released 8.6 mg Zn²⁺, which is 5.8-fold lower for the MBC for both evaluated bacteria (50.4 mg). Besides, the coating of ZnO showed a gut empowering effect due to the reduction in the release of Zn²⁺ in simulated porcine gastric fluid. This feature can reduce the amount of ZnO composites to be administered to weaned piglets.

These experiments revealed a remarkable advantage of the synthetized products compared to commercial ZnO, since the concentration of Zn in pig intestine released from those compounds should be higher. The intake dosage of 9.9 g of commercial Zn should be enough to ensure the release of 198 mg in the middle intestine, corresponding to the MBC value for E. coli. In the same way, the intake of 12.9 g of Zn might release 257.4 mg corresponding to ZnO MBC for S. aureus. These values ​​are approximately 5 times higher than necessary (2 g) of ZnO/CAT to obtain a release of 91.2 mg that are enough to inhibit E. coli and 7 times greater than the intake (1.8 g) necessary to release 81. 6 mg Zn from ZnO/CAT and inhibit S. aureus in the gastrointestinal environment. The same would be true for ZnO/CA in a more pronounced way, because in order to release 50.5 mg of Zn (MIC of ZnO/CA to both bacteria) it would be necessary for each piglet to ingest at least 1.2 g of Zn, value 8 and 11 times lower than that observed for commercial ZnO inhibit E. coli and S. aureus, respectively. Although, the intake of ZnO composites can alter the gut microbiota composition, reducing the bacterial diversity and effecting even the probiotic bacteria, the overall supplementation effect of wearing piglets is positive since it can prevent attachment of pathogenic bacteria to the intestinal villi avoiding wearing unspecific diarrhea.

We have successfully synthesized new ZnO nanoparticles - ZnO/CA and ZnO/CAT, using CH/SA and an ultrasound-assisted methodology. The novel composites showed an optimum in vitro release profile of Zn²⁺ in simulated enteric fluids assays and expressive antimicrobial activities against E. coli and S. aureus. The physico-chemical analysis and in vitro assays indicate that ZnO microparticles are a promising designed product to be applied for piglet livestock. The data from this study suggest that the ZnO nanoparticle composites may be effective in diets fed to weanling pigs, and may have the potential to significantly reduce fecal Zn excretion. Experiments with weanling pigs fed diets containing ZnO nanoparticle composites should be conducted to address the clinical evaluation of ZnO composites in piglets fed in controlled conditions.