Effects of the green cross-linking agent tannic acid and its oxidation on the properties of porcine plasma protein superabsorbent materials

Effects of the green cross-linking agent tannic acid and its oxidation on the properties of porcine plasma protein superabsorbent materials Massimo Alagia a , Carlos Bengoechea b,* , Barbara La Ferla c , Francesco Peri d , Antonio Guerrero b a Department of Materials Sciences, University of Milano-Bicocca, Via R. Cozzi 55, 20125 Milano, Italy b Departamento de Ingeniería Química, Universidad de Sevilla, Escuela Polit´ ecnica Superior, 41011 Sevilla, Spain c Department of Earth and Environmental Sciences DISAT, Universit` a degli Studi di Milano-Bicocca, Piazza della Scienza 1, 20126 Milan, Italy d Department of Biotechnology and Biosciences, University of Milano-Bicocca, Piazza della Scienza 2, 20126 Milan, Italy ARTICLE INFO Keywords: Porcine plasma protein Tannic acid Cross-linking ABSTRACT Tannic acid is a natural polyphenol capable of strongly interacting with proteins, with good antioxidant and antibacterial properties. Thus, tannic acid (TA) or oxidized tannic acid (oxTA) may be used as cross-linking agents in the development of reinforced and fully protein-based superabsorbent materials (SAMs). oxTA was produced so that reactive quinone groups were generated, which are expected to increase its reactivity. In this study, porcine plasma protein (PPP) and glycerol (gly) were used in a 50/50 PPP/gly ratio to obtain SAMs through twin screw mixing and injection molding. The results showed that both TA and oxTA increased the storage modulus and the loss tangent of blends and bioplastics due to the physical interactions established between TA or oxTA and PPP. The mechanical properties, particularly the Young's modulus and tensile strength, were generally enhanced as well. Water absorption was strongly influenced by the addition of TA, resulting in a decrease in the amount of water absorbed. However, samples containing oxTA resulted in a greater water absorption capacity, retaining a higher proportion of the superabsorbent properties of the reference composition. Moreover, systems containing oxTA generally possess better mechanical properties than those of equivalent TA formulations, especially those containing 5 % and 10 % oxTA. 1. Introduction Currently, environmental concerns are driving research into novel materials towards a sustainable development, where the utilization of natural products and features such as biodegradability and biocompatibility have become essential. One of them includes superabsorbent polymers (SAPs), also termed as superabsorbent materials (SAMs), which are highly porous, cross-linked polymers with hydrophilic domains capable of absorbing and retaining high weight percentages of water [1–4]. These structures include highly hydrophilic moieties, such as carboxylates, hydroxides and amides, which can interact with water molecules causing the stretching of polymeric chains and resulting in SAM swelling. Presently, conventional SAMs include either synthetic polymers, such as polyacrylates and polyacrylamides, or semisynthetic biopolymers, which are typically based on starch, chitosan or cellulose cross-linked with acrylic reagents [5]. Although all of these materials have shown great performance, there is still a need to develop fully biobased superabsorbent materials, where each component used in the formulation can be derived from biomass, which would represent a sustainable development for this application [6]. Among others, proteinbased bioplastics have recently been exploited. Generally speaking, bioplastics are plastics that are biobased, biodegradable, or both, and are being studied in a variety of applications, ranging from food packaging [7,8], automotive [9], cosmetic and personal care packaging [10]. In particular, protein-based bioplastics fulfil the requirements for bioSAMs and have the advantage that they can be treated with conventional polymer processing techniques, including twin-screw mixing and injection molding, by using a suitable plasticizer [11]. Moreover, the water absorption of these bioplastics is generally good due to the great hydrophilicity of both biopolymer (i.e., protein) and plasticizers (e.g., glycerol, sorbitol). In this sense, several proteins, including wheat gluten [12], potato protein [13], soy protein [14], keratin [15] and porcine plasma protein (PPP), have recently been considered. * Corresponding author. E-mail address: [email protected] (C. Bengoechea). Contents lists available at ScienceDirect International Journal of Biological Macromolecules journal homepage: www.elsevier.com/locate/ijbiomac https://doi.org/10.1016/j.ijbiomac.2025.140584 Received 11 October 2024; Received in revised form 24 January 2025; Accepted 31 January 2025 International Journal of Biological Macromolecules 304 (2025) 140584 Available online 6 February 2025 0141-8130/© 2025 The Authors. Published by Elsevier B.V. This is an open access article under the CC BY-NC-ND license ( http://creativecommons.org/licenses/bync-nd/4.0/ ). PPP is usually obtained from raw blood via a plasma fractionation process, which initially involves the centrifugation of porcine blood, which allows the plasma to be separated from the cellular components, which remain in the sediment. After that, plasma proteins can eventually be separated on the basis of their molecular weight via other techniques, including precipitation or membrane filtration. Indeed, plasma protein is composed by several different fractions, namely fibrinogen, albumin and globulin. The two most abundant subfractions are albumins (60 %) and globulins (40 %), where the former are subdivided into α -, βand γsubfractions of 66–69 kDa, and the latter are also distinguished into α -, βand γsubfractions with a wider molecular weight distribution [16,17]. PPP and more generally blood can be considered as underused materials, even if they present valuable properties such as waterbinding, thermal gelation, foaming and emulsifying abilities. For those reasons, it has been implemented for food and feed additive applications. Nonetheless, those properties are also interesting in non-food applications. For example, PPP has been considered for the production of films [18,19] and electrospun nanofibrous membranes [20]. Furthermore, the abundance of porcine plasma protein as a byproduct of the meat industry presents an opportunity for sustainable utilization and waste reduction. It is difficult to obtain certain data about porcine plasma protein production worldwide, thus making estimation necessary. According to FAO, in 2021, the world production of meat pig was set to 120,372 thousand tonnes [21], thus about 1.65 billion pigs are slaughtered annually (1 pig can generate 73 kg of meat on average). As one pig can generate 3.5 L of blood [17], annual porcine blood production can be estimated as 5.77 billion litres. Considering that the plasma fraction is 60 % of whole blood, 3.46 billion L of plasma are produced with a protein content of approximately 87 g/L concentration, finally it can be estimated that annual world production of porcine plasma protein corresponds to approximately 301 thousand tonnes, considering an ideal situation in which no mass loss takes place during the PPP production process. The use of porcine plasma protein in the production of bioplastics for SAMs thus takes advantage of its remarkable properties, as well as its biodegradability, biocompatibility and abundance as a by-product of the meat industry. Previous work suggested that it is possible to obtain SAMs from PPP using glycerol as a plasticizer [22]. This material has a water absorption capacity up to 15 to 20 times its original weight, but has the disadvantage of being soft and brittle, which could limit its applicability. Therefore, several attempts have been made to improve its mechanical properties. A common strategy involves the addition of a cross-linking agent, which can physically or chemically interact with the polypeptide chains to restrict their relative movement when a force is applied [23,24]. Unfortunately, cross-linking also tends to reduce the water uptake capacity [25]. Among protein cross-linkers, tannic acid (TA) is gaining increasing interest. In particular, tannic acid is a plant derived polyphenolic molecule that can be found in bark, leaves, fruits and seeds, and is capable of interacting strongly with proteins, mainly through pyrogallol moieties, to form high molecular weight complexes [26–28]. Additionally, TA presents other interesting properties, such as antioxidant and antibacterial properties. Tannic acid can interact via hydrogen bonds between its phenolic hydroxyl groups and the polar moieties of a protein, including carbonyls, amines and hydroxyls. Although it is not yet clear whether tannic acid is able to chemically bind protein, one option to increase that reactivity is oxidation. This latter derivatization is generally carried out in aqueous solution in the presence of active oxygen species and can be performed both by chemically and enzymatically methods. The oxidation of TA generates reactive quinone groups which presumably may be able to increase reactivity towards PPP via either Schiff or Michael base additions, as illustrated in Fig. 1. Tannin oxidation can generally be accompanied by successive Fig. 1. Tannic acid structure (A) and suggested mechanisms of cross-linking with proteins via prior oxidation (B). M. Alagia et al. International Journal of Biological Macromolecules 304 (2025) 140584 2 coupling and polymerization, which leads to increased structural complexity. TA, either unoxidized or oxidized, has been used to crosslink different biopolymers, such as casein, chitosan, gelatin or zein [29–31], mainly processed through casting or hydrogel formation. However, to the best of our knowledge, neither the use of TA as cross-linker of PPP or the production of TA-crosslinked bioplastics through injection molding has yet been reported. Injection molding is a common process in the plastic industry that permits to form precisely complex geometries with high reproducibility. Injection molding can be scaled and automated for a wide range of materials [32]. The specific objective of this study is to evaluate the effect of the chemical oxidation of TA on the properties of PPP based injection molded bioplastics. In this work, SAM bioplastics were produced by mixing porcine plasma protein and TA or oxTA in a twin screw mixer, using glycerol as plasticizer, to produce blends which were then injection molded into bioplastics. In particular, three different concentrations of either TA or oxTA (5 %, 10 % and 20 % weight percentages with respect to protein content) were prepared, and their effects on the thermomechanical, tensile and absorption properties were evaluated. 2. Materials and methods 2.1. Materials The porcine plasma protein (PPP) used in this study (AproPork, Essentia Protein, USA) was kindly supplied by PROANDA S.A. (Spain). The protein (74 wt%) and ash (9 wt%) contents of this concentrate were determined in previous works [22,33]. For all formulations, pharma grade glycerol (gly) (CAS 56-81-5, Panreac Química S.A., Spain) and ACS-grade tannic acid (CAS 1401-55-4, Sigma-Aldrich, Merck, Germany) were used. Other reagents, including Na 2 CO 3 , CAS 497-19-8, regenerated cellulose dialysis membranes, H 2 O 2 solution (CAS 772284-1) and Folin-Ciocalteu reagent (CAS 12111-13-6) were obtained from Sigma-Aldrich (Merck, Germany). 2.2. Tannic acid oxidation The oxidation of tannic acid was carried out following a reported procedure, with little variation [29]. Tannic acid was solubilized in distilled water at a concentration of 0.02 g/mL; at this point, the temperature was increased to 60 ◦C and 5 M NaOH was added until pH reached 9–10. Finally, an aqueous solution of H 2 O 2 (35 wt%) was added to reach final concentration of hydrogen peroxide equal to 0.4 wt%. The mixture was allowed to react for 30 min at 60 ◦C, and neutralized by adding 1 M HCl and oxidized tannic acid powder was obtained after freeze drying for 3 days with a LyoQuest freeze dryer (Telstar Technologies, Spain), at a collector temperature of −80 ◦C. This oxidation process was carried out at least three times and employed in every formulation to check the reproducibility of the results obtained. 2.3. Samples preparation The fabrication of bioplastics followed a two-stage process reported elsewhere [34], with modifications. The procedure comprises a first mixing step with a Haake Polylab mixing rheometer (Thermo-Scientific, Germany). Each component was blended at room temperature for 5 min and at a rotor speed of 50 rpm in a double cylindrical chamber. Torque and temperature are constantly monitored during this process. The blends were then stored for 1 day at 4 ◦C and subsequently processed through injection molding to obtain rectangular samples (60 mm ×10 mm ×1 mm). These samples were obtained with Haake pneumatic piston injector, Mini Jet II (ThermoScientific, Germany). The blends were introduced into a chamber heated to 40 ◦C and injected through a nozzle into a mold whose temperature was set at 60 ◦C. These conditions have proven to provide a remarkable superabsorbent character (i.e. absorbing water up to 21 ±4 times its dry weight) but leading to poor mechanical properties (i.e. with a Young modulus of 2.4 ±0.8 MPa) [22]. The pressure was maintained at 800 bar for the entire duration of the process (300 s). Samples were kept in a sealed container before their characterization, which was carried out in the 24–48 h period after their fabrication (Table 1). 2.4. Characterization 2.4.1. Dynamic mechanical analysis (DMA) The viscoelastic properties of blends and bioplastic were determined through small amplitude oscillatory tests using a DMA850 rheometer (TA Instruments, USA). Blends were analysed with a compression clamp geometry immediately after their preparation, whereas bioplastics were tested in tensile mode within two days after their molding. The experiments involved a frequency sweep test ranging from 0.01 Hz to 10 Hz and were performed at 20 ◦C and constant strain (0.01 % blends, 0.005 % for bioplastics). Moreover, temperature sweep tests were performed at a heating rate of 5 ◦C/min, a frequency of 1 Hz and strains of 0.01 % and 0.005 % for the blends and bioplastics, respectively. The temperature was varied from 0 ◦C to 140 ◦C for blends and from −30 ◦C to 160 ◦C for bioplastics. All the assays were performed within the linear viscoelastic region (LVE), previously determined for each composition with strain sweep tests (0.001 %10 %) at a constant frequency of 1 Hz. Data were collected via TRIOS software and at least three measurements per sample were recorded, even if only one representative measurement per sample was reported. 2.4.2. Tensile tests The samples obtained from injection molding with rectangular shapes were submitted to uniaxial tensile tests, according to the standard method ISO527-2, with some modifications [35]. Mechanical properties such as tensile strength, elongation at break and Young's modulus were estimated using an MTS Insight 10 Universal Testing Machine with a load cell of 10 kN. After the specimen was fixed into the supporting clamps, a displacement at constant rate of 10 mm/min was applied until reaching failure, so that the time elapsed for all the experiments was between 30 and 300 s in accordance with the standard. This test was conducted on at least 6 replicates per sample. Data were finally recorded, analysed with TestWorks4 software and the ThompsonTau statistical test for outlier searching was applied to the results within one sample. 2.4.3. Fourier-transformed infrared spectroscopy (FTIR) IR spectra were recorded in the 4000–400 cm −1 wavenumber range with 4 cm −1 resolution and 32 scans using a Jasco FTIR 4200 spectrometer (Easton, MD, USA) equipped with an attenuated total reflectance (ATR). Bioplastic samples were analysed directly with the ATR instrument in the transmittance mode. The data were then converted to absorbances and normalized over the amide I region (1700–1600 cm −1 ) of each sample. The resulting spectra are reported in normalized absorbance units versus wavenumbers. 2.4.4. Water uptake capacity (WUC) and soluble matter loss (SML) WUC and SML of bioplastics were estimated according to the protocol reported in a previous work [36]. Briefly, pieces of 20 mm ×10 Table 1 Formulations studied. System PPP Gly TA/oxTA REF 50 50 0 TA5/oxTA5 48.8 48.8 2.4 TA10/oxTA10 47.6 47.6 4.8 TA20/oxTA20 45.5 45.5 9.0 M. Alagia et al. International Journal of Biological Macromolecules 304 (2025) 140584 3 mm ×1 mm in size were cut from the bioplastics, kept for 24 h in an oven at a constant temperature of 50 ◦C and weighed (w 0 ). The samples were then placed in water (approximately 80 mL) for 24 h in closed vessels, weakly bound water was removed by gentle drying with filter paper, and swollen samples were weighed (w 1 ). After a final drying step in an oven (50 ◦C, 24 h), the residual bioplastics were weighed (w 2 ). WUC and SML were calculated as follows: WUC (%) = w1−w2 w2 ×100 (1) SML (%) = w0−w2 w0 ×100 (2) 2.4.5. Scanning electron microscopy (SEM) Swollen bioplastics in water were freeze-dried for 1 day, cut into small pieces and analysed with a ZEISS EVO microscope (USA) after silver coating. A beam current of 11–12 pA and an accelerating voltage of 10 kV were used for image acquisition. Images are acquired with a secondary electron (SE) sensor at 100×magnification. 2.4.6. Thermogravimetric analysis (TGA) TGA on bioplastics (20–50 mg) was performed between 50 ◦C and 1000 ◦C with constant heating rate of 10 ◦C/min, under air atmosphere introduced at 100 mL/min with a thermogravimetric analyser, model SDT Q600 V20.9 Build 20 (TA Instruments (USA)). The temperature ramp also included a pre-treatment as follows: from 30 ◦C to 130 ◦C, 20 ◦C/min, N 2 ; isotherm at 130 ◦C for 5 min, N 2 ; from 130 ◦C to 50 ◦C, 20 ◦C/min, N 2 ; isotherm at 50 ◦C for 5 min; N 2 . 2.4.7. Determination of bound phenolic species and the Folin-Ciocalteu assay A total of 150 mg of dried bioplastics (24 h in an oven at 50 ◦C) was dissolved in 5 mL of formic acid under magnetic stirring for 16 h. Once completely dissolved, water was added to obtain a final formic acid concentration of approximately 50 %. The mixture was then purified through dialysis (12–14 kDa MWCO) against distilled water for 2 days. Water was continuously replaced by fresh distilled water during the dialysis process to promote the purification. After dialysis, the solution was retrieved and freeze dried. The samples were submitted to FTIR analysis as previously described. The obtained powder was also analysed via the Folin-Ciocalteu assay, following a reported procedure, with modifications. Briefly, sample solutions at a 2 mg/mL concentration in distilled water were prepared. Then, 40 μ L of these solutions were mixed with 800 μ L of 75 mg/mL Na 2 CO 3 and diluted with 1.08 mL of water. Finally, 40 μ L of Folin-Ciocalteau reagent (2 N) were added and mixtures (2 mL) were kept in the dark for 30 min, after which the absorbance was read at 750 nm and correlated with the phenol content in the samples. A calibration curve with tannic acid or oxidized tannic acid (0.05–0.25 Fig. 2. Frequency sweep tests of blends with different concentrations of tannic acid (A); and with different concentrations of oxidized tannic acid (B). Temperature sweep of blends with different concentrations of tannic acid (C); and with different concentrations of oxidized tannic acid (D). M. Alagia et al. International Journal of Biological Macromolecules 304 (2025) 140584 4 mg/mL) was previously determined for its quantification in unknown samples. Measurements were carried out with an Asys UVM 340 Microplate reader (Biochrom, Cambridge, UK) using Greiner Bio-One Cellstar 24-well plates. 3. Results and discussion 3.1. Viscoelastic properties of blends After the mixing stage, the blends obtained were subjected to dynamic mechanical analysis in compression mode in order to evaluate the dependence of the viscoelastic properties on the frequency and temperature. The frequency sweep test of the blends (Fig. 2A) revealed that the storage modulus can be correlated with the amount of tannic acid in the sample; in particular, increasing the TA content results in an increase in the elastic modulus, E', of the blends even before the injection molding process, by which chemical cross-linking may be triggered. It is known that TA can interact with proteins via hydrogen bonding between phenolic OH groups and the free amine, hydroxyl and carbonyl groups of a protein, as well as hydrophobic interactions among aromatic rings and electrostatic forces. Additionally, the interplay of TA with glycerol could also occur through hydrogen bonding. All of these interactions result in the formation of more rigid structures with respect to reference. Moreover, introducing a third component in the blends also results in an increase in the loss tangent, tan δ, defined as the ratio of the viscous modulus to the elastic modulus (E"/E'), which indicates that, even if a strengthening takes place, the prevalence of E' over E" becomes less important in the presence of TA, indicating that physical interactions are predominant. For the oxidized tannic acid-containing blends, a similar correlation between the oxTA content and their viscoelastic properties was observed (Fig. 2B). This latter, compared with the non-oxidized compounds, are tendentially characterized by E' profiles set at higher values, except for oxTA10 sample. This is also accompanied by a lower dependency of tan δ with frequency, thus indicating more rigid structures. Notably, this effect should not be attributed to a chemical cross-linking in the blends, since torque and temperature profiles in the mixing stage (Supplementary information, Fig. S1) do not show a clear evolution of any of those parameters with mixing time, but rather to stronger hydrogen bonding due to the presence of carbonyl rather than hydroxyl groups in the chemical structure of the cross-linker. The results of the temperature sweep tests on the blends are shown in Fig. 2C-D. The reference composition not containing any cross-linking agent (REF) displays a first thermal softening effect, until it reaches a minimum E' value at 65 ◦C, followed by an increase with temperature. Indeed, the peak at around 82 ◦C in tan δ profile can be attributed to the thermal denaturation of plasma proteins, especially albumins, after which E' increases due to protein aggregation [37]. DMA measurements clearly revealed that TA (Fig. 3A) and oxTA (Fig. 3B) have an influence on the thermal transitions of the material. First, a peak in tan δ at 63 ◦C appears for TA-containing systems, accompanied with a pronounced decrease in E': this transition can be associated with the disaggregation of tannic acid interactions with an increase in monomeric TA [38]. Tannic acid can interact with itself via hydrogen-bonding of the galloyl moieties to form supramolecular structures. This peak is clearly detectable in the loss tangent curve of TA5, whereas for TA10 and TA20 there is an overlap with the adjacent peak. It is also reasonable to think that in that temperature range a chemical reaction between TA and PPP Fig. 3. Frequency sweep of bioplastics with different concentrations of tannic acid (A); and with different concentrations of oxidized tannic acid (B). Temperature sweep of bioplastics with different concentrations of tannic acid (C); and with different concentrations of oxidized tannic acid (D). M. Alagia et al. International Journal of Biological Macromolecules 304 (2025) 140584 5 can occur, from which one would expect an increase in the elastic modulus. Nevertheless, a clear increase in E' is not detected before the transition, probably due to the predominance of the thermal softening effect of plasma proteins, but it is possible to highlight the shift of E' minimum towards higher temperatures. At high temperatures, the E' profiles in Fig. 2C seem to reflect a compromise between increasing TA content and decreasing protein content. On the other hand, oxidized tannic acid shows a more complex thermal behaviour, comprising of a shoulder peak at about 40 ◦C, a broad peak of tannic acid disaggregation at 60 ◦C and PPP gelation above 80 ◦C. In particular, this last peak is shifted towards higher temperatures in proportion to the amount of oxTA added; this shift is also accompanied by a shift in the E' minimum towards higher temperatures. Yan et al. reported a similar effect of TA on egg white protein/ xanthan gum systems from differential scanning calorimetry (DSC) analysis and attributed this effect to the formation of an heat-resistant network after the reaction [39]. It is possible that oxTA, owing to its higher reactivity, gives rise to a more pronounced variation in the transition of the material. Thus, at high temperatures, the E' curves in Fig. 2D displays an inverted order; this phenomenon is in accordance with the decrease in protein content in the cross-linker containing blends. 3.2. Bioplastic viscoelastic properties As described in the experimental section, bioplastics were obtained after the injection molding process (T cyl =40 ◦C and T mold =60 ◦C for 500 s at 800 bar). The dynamic mechanical properties were evaluated in tensile mode by frequency and temperature sweep tests. The frequency sweep results revealed that all the samples had increased elastic moduli with respect to their corresponding blends after the molding process, as observed when comparing Fig. 3A-B (bioplastics) with Fig. 2A-B (blends). This can be attributed to the influence of temperature and pressure applied during the injection molding process. The frequency sweep tests of blends and bioplastics are in accordance regards the effects of increasing concentrations of TA and oxTA on the dynamic mechanical properties of bioplastics. As shown in Fig. 3A, unoxidized tannic acid displays a concentration-dependent increase in storage modulus, whereas oxidized tannic acid has a greater effect at lower concentrations (5 %), not observing an increase in its content with higher oxTA contents. Generally, for all the samples, cross-linking has the effect of increasing both E' and tan δ, just like observed in blends, except for the TA5 sample. The slopes of the E' curves were not significantly lower than those of the corresponding blends, probably due to the persistence of a high number of hydrogen bonds that tannic acid can form with all the components of the formulation. These relatively weak bonds are more susceptible to a constant deformation. Other authors found similar results when employing TA as cross-linker in casein films, obtaining higher viscoelastic moduli at greater TA contents [31]. The results of the temperature sweep tests on the bioplastics are displayed in Fig. 3C-D. Similarly with blends, E' dependence on temperature comprises a first region of E' lowering until the plasma thermal transition, followed by a subsequent increase at higher temperatures. Additionally, when comparing the temperature sweep tests of blends Fig. 4. Stress-strain curves of TA (A) and oxTA (B) samples. Tensile strength ( σ MAX ), elongation at break ( ε MAX ) and modulus (E) of the whole samples (C). M. Alagia et al. International Journal of Biological Macromolecules 304 (2025) 140584 6 (Fig. 2C-D) and bioplastics (Fig. 3C-D), there is a general reinforcement effect due to the injection molding processing conditions. As the reinforcement effect related to tannin cross-linkers was also visible before injection molding, it is reasonable to believe that cross-linking did not occur in this latter step. Additionally, in this case, the E' minimum is slightly shifted towards higher temperatures (especially for oxTAcontaining samples), suggesting a possible effect of the cross-linker on the bioplastic thermal transition. The tan δ profile of bioplastics changes when the cross-linker is present, either TA or oxTA: as already described for blends, an additional peak contribution at approximately 60 ◦C can be detected when the cross-linker has been added, the intensity of which increases proportionally with its content. 3.3. Tensile mechanical properties In addition to DMTA, tensile testing is a reliable test for measuring the strength of bioplastics. Fig. 4 shows the tensile stress vs. strain curves for each set of samples, which are correlated with the data values of relevant parameters, in particular, tensile strength ( σ max ), elongation at break ( ε max ), and Young's modulus (E). Compared with the reference formulation without a cross-linker, the addition of TA or oxTa generally improved the tensile mechanical properties of the bioplastics; however, there were large differences between the curves related to the reinforced samples. In particular, for the nonoxidized tannic acid samples, there is a correlation between TA content and E, which increases by a factor of up to 5 at a TA concentration of 20 % (TA20); The improvement can also be seen in the σ max and ε max , the latter being particularly noticeable for 5 % TA (TA5), which exhibits a value of ε max increased by a factor of about 1.5. However, this parameter decreases significantly with increasing TA content. Thus, the general trend of an increase in Young's modulus and maximum stress at the expense of a reduction in the maximum elongational strain is observed from 5 % AT onwards. On the other hand, the oxidized tannic acid-containing compositions did not show a clear correlation between the mechanical properties and oxTA content. In particular, σ MAX initially increases for the system containing 5 % of oxTA and then decreases with increasing concentration of oxTA. In contrast, parameter E always remains above the value obtained for the reference while no significant variations are observed for ε MAX , except at the highest oxTA concentration, where a significantly lower value than that of the reference can be observed. A possible explanation of this effect could be attributed to the different structures of the two cross-linkers; in particular, unoxidized tannic acid presents high amounts of phenolic Fig. 5. FTIR of bioplastics with different concentrations of tannic acid (A); and with different concentrations of oxidized tannic acid (B). M. Alagia et al. International Journal of Biological Macromolecules 304 (2025) 140584 7 groups which are able to interact with protein and glycerol by hydrogen bonding; the oxidized TA on the other hand, has orthoquinone groups, which are more susceptible to chemical reactions as depicted in Fig. 1B. In addition, unlike phenolic OH, o-quinones are hydrogen bond acceptors, altering the H-bond interactions that occur in these materials. It could be reasonable to think that oxTA could preferentially interact with glycerol, which is mainly an H-bond donor. The variation of mechanical properties with TA concentration partially agrees with the results obtained from rheological tests, showing that cross-linking increases interchain interactions and restricts their mobility, but major differences between these two analyses can be observed especially for oxTAcontaining samples. Even if the FTIR technique employed in this study has already been used previously to qualitatively describe the interactions between different components in a blend and the extent of crosslinking [23,40], additional experiments involving NMR spectroscopy would be helpful to confirm those interactions between oxTA and glycerol, as well as to quantify the extent of bonding with PPP. Moreover, elemental mapping through techniques like energy-dispersive Xray spectroscopy (EDS) would aide to track functional group distribution reactivity. Further research should be focus on this. 3.4. Structural characterization through FTIR Structural characterization of the bioplastics was carried out to determine the possible variation in protein chemical structure due to bond formation with TA/oxTA. From the FTIR spectra (Fig. 5A-B) it is possible to recognize several characteristic protein IR bands, especially amine N – H asymmetric stretching, referred to as the amide A band (≈3300–3100 cm −1 ), aliphatic CH stretching asymmetric C – – O stretching of the amide group, amide I band (1700–1600 cm −1 ) and N – H bending with contributions from C – N stretching, amide II (1580–1480 cm −1 ). Some other bands can be attributed to glycerol, especially the O – H asymmetric stretching (3600–3200 cm −1 ), which produces a shoulder in the protein amide A band, and sharp peaks at 1100, 1035 and 995 cm −1 , related to CO stretching. Supplementary Fig. S2 shows the FTIR spectra of TA and oxTA reagents, where particularly C-OC aromatic stretching and C-O-H aromatic bending at approximately 1311 cm −1 , ester O-CO stretching at 1180 cm −1 and C-OC asymmetric stretching at 1018 cm −1 can be distinguished. This bands are slightly shifted at higher wavenumbers for oxTA, whereas the shift at lower wavenumbers of O – H stretching can be related with stronger hydrogen bonding due to the formation of carbonyls in the oxidation, as previously reported [41]. Upon the addition of tannic acid, the intensity of the proteins bands decreased proportionally to the amount of TA in the bioplastic, which was expected as the protein content progressively decreased. The bioplastics containing either TA or oxTA differ with respect to the peaks of the cross-linker alone a peak centred at 1346 cm −1 appears in all the samples. It might therefore be suggested that this peak originates from C – N stretching associated with the formation of an aromatic secondary amine via Michael-type addition between PPP and TA/oxTA, as described in Fig. 1. Nevertheless, this latter could also originate from hydrogen bonding between the phenolic groups and the polar residues of PPP, causing a shift in the O – H bending frequency. Other differences can be seen in the O-CO ester stretching, which shows a shift at higher wavenumbers (1207 cm −1 ) with respect to virgin TA. CO-C asymmetric stretching is not clearly distinguishable because it overlaps with the CO stretching of glycerol. The relative intensities of these two peaks in the bioplastics are both correlated with the amount of TA contained in the formulations. With respect to the other mechanism of reaction (Schiff-base addition) the formation of secondary imines is difficult to observe in the spectra, as a peak related to C – – N should be expected in the 1680–1630 cm −1 region, which overlaps with the amide I protein signal. Indeed, there is no further peak contribution to the amide I signal from the peak deconvolution in the Supplementary Information for the samples (Supplementary Fig. S3). It could thus be hypothesized that cross-linking acts predominantly via Michael type addition of protein amines from lysine and N-terminus with the aromatic ring of TA. A comparison of the TA and oxTA samples revealed few differences in the spectra, thus suggesting small differences in the chemical structure of the bioplastics prepared with the two different cross-linkers. One possible explanation is that the differences observed in the mechanical analysis are not related to differences in the chemical reactivity of the cross-linker, but rather to changes in the physical interactions formed between the tannic acid or oxidized tannic acid and the other components of the formulation. 3.5. Water uptake capacity and soluble matter loss The absorption properties of the bioplastics are strongly influenced by the presence of either TA or oxTA, as shown in Fig. 6. In general, softer materials are expected to be associated with higher WUC, as a more flexible sample is more likely to deform as it swells during water absorption. This is evident from the results shown, as the uncross-linked reference has the highest value of WUC. The addition of the two crosslinking agents leads to a lower water absorption: this is expected as the cross-linker limits the mobility of the polypeptide chains, thus hindering the water swelling phenomenon. This effect has also been observed for gelatin films hardened with TA [42]. In this work, the residual water of the uptake test presents a light brown color for TA and dark brown/greenish for oxTA, which are not observed in the reference residual water. It is thus reported that those solution colors can be attributed to the presence of oxidized tannic acid. For these reasons, it is believed that a fraction of tannic acid is dissolved from the bioplastic Fig. 6. WUC and SML of bioplastics with different concentrations of tannic acid and oxidized tannic acid. M. Alagia et al. International Journal of Biological Macromolecules 304 (2025) 140584 8 during water swelling tests. Differences depending on whether unoxidized or oxidized TA is used in the formulation were observed. In detail, unoxidized TA reduces the WUC in a way proportional to its concentration, while oxidized TA samples show similar values and generally higher values than those of the equivalent compositions with unoxidized TA. The comparison between the TA10 and oxTA10 samples reveals that the latter has a WUC value that is approximately double that of the former; this is also the case for the comparison between TA20 and oxTA20. The WUC value of TA5 is still higher than that of the corresponding formulation oxTA5, probably due to the large differences in mechanical properties between these two compositions. For the oxTA samples, at cross-linker concentrations of 5 % and 10 %, the materials still exhibited superabsorbent properties and, in particular, the increase in concentration was not accompanied by a net decrease in WUC. In terms of soluble matter loss, there are small differences between the samples. The only significant decrease in SML occurs at the highest oxTA concentrations (10 and 20 %), suggesting a higher level of oxTA interactions with the other two components. It should also be noted that an increase in TA or oxTA content implies a reduction in PPP and glycerol content, which may contribute to the lack of significant differences for the other systems. In fact, it is suggested that these two components are mainly related to the soluble loss during water swelling, as seen in the value of SML for the reference sample. 3.6. Scanning electron microscopy The morphology of swollen bioplastics can be examined by scanning electron microscopy (Fig. 7). The effect of water uptake on the sample surface involves the generation of pores associated with glycerol diffusion into aqueous media associated with water uptake and retention. Indeed, reference sample after water absorption and freeze drying presents a surface with honeycomb-like structure, which was observed also for dextran–methacrylate swollen hydrogel [43]. This structure is formed following freeze drying and is related to the formation of ice crystallite, which phase separated from the matrix and introduce porosity [44]. The addition of a cross-linker influences the morphology of the surfaces, which is reflected by variations in water uptake properties. Especially, pore dimensions of the crosslinked samples differ from those of the reference sample, as well as their shapes, which appear to be more circular-like, and the number of pores seems to decrease. The porosity density is correlated with the water uptake capacity, as more compact structures like the cross-linked ones could hinder the penetration of water molecules inside the material, leading to an inferior number of pores [45]. Moreover, the introduction of TA/oxTA may generate inhomogeneities and rougher structures due to aggregation, which are believed to be embedded with minor WUC respect to more homogeneous ones like the reference [46]. In the SEM micrographs in Fig. 7, it is not clear whether there is a concentration-dependent effect of TA or oxTA on the morphology of the surface, but rather the situations appear to be similar. 3.7. Thermogravimetric analysis The thermal stability of bioplastics was determined by means of thermogravimetric analysis (TGA) in an air atmosphere to study the combustion phenomena. The results are shown in Fig. 8 for samples studied. A heating pre-run up to 130 ◦C was made, so that the water loss was not visible in the TGA curves shown. The first decomposition step took place at an onset of approximately 175 ◦C for the reference sample, which did not significantly change upon the addition of the cross-linker. The first weight loss step could be attributed to the first oxidative Fig. 7. SEM images of freeze dried swollen matrices of porcine plasma with different concentrations of tannic acid (TA) and oxidized tannic acid (oxTA). M. Alagia et al. International Journal of Biological Macromolecules 304 (2025) 140584 9