Grape stalk and vine strains (Vitis vinifera L.) were collected from Quinta da Alameda in the center region of Portugal (Nelas, Viseu) and dried at 40 °C in an oven until reaching equilibrium moisture for further processing. Raw materials were grounded at 1,500 rpm using a 10-mm square grid in a cutting mill (Retsch SN 300) and then sieved using a vibrating sieve shaker (Retsch AS 200 control) with an amplitude of 2 mm for 2 min to select the fraction of particles with a size between 500 μm and 2 mm. Sodium hydroxide (97%), sodium carbonate (97%), gallic acid (anhydrous), sodium nitrite (anhydrous), and aluminum chloride (anhydrous) were supplied by Sigma-Aldrich and used as received. Folin–Ciocalteu reagent (VWR chemicals) was purchased from VWR and used as received. (+)-Catechin hydrate (>97.0%) and 2,2-diphenyl-1-picrylhydrazyl free radical (DPPH, >97.0%) were provided by TCI and used as received.

The extraction of polyphenols was carried out at 70 °C in a 3-L jacketed borosilicate reactor. The extractive agent (aqueous NaOH and aqueous Na₂CO₃) with a certain concentration (0.1%, 0.5%, and 1.0% w/v) was pumped into the reactor with a flow rate of 20 mL min⁻¹ under recirculating mode, and a solid/liquid ratio of 1/5 (w/w) was kept in all cases. Distilled water was also used as an extractive agent for comparison purposes. After 1 h of extraction, the suspension was filtered through a nylon membrane (125 µm), and the liquid extract was chemically characterized (pH, solid content, FTIR, ABES, and extraction yield). In order to improve the stability, the liquid extract was dried through the spray drying process at 180 °C, using a spray drying system (SD-06, LabPlant, United Kingdom). The liquid was fed at a flow rate of 50 mL min⁻¹, and compressed air (a flow rate of 35 m³ h⁻¹) was used for dispersion. Table 1 presents the codification attributed to all experiments, according to extraction conditions.

All the experiments were evaluated in terms of extraction efficiency (EE) according to Eq. 1: E E   % = d r y   m a s s   o f   s o l i d   e x t r a c t   g d r y   m a s s   o f   r a w   m a t e r i a l   g × 100 . (1)

A bio-adhesive was produced by concentrating the polyphenolic extract up to 20% of solid content, using a rotary evaporator. The bio-adhesive was used in the manufacturing of particleboards (PBs), using grape stalk and vine strain particles (2 mm–500 µm) as raw material. Therefore, PBs with 250 × 250 mm² (8 mm of thickness) were produced through hot-pressing with 10 wt% of bio-adhesive and a target density of 850 kg m⁻³. The mixture was hot pressed at 160 °C during 10 min at a fixed thickness of 8 mm. The panels were denoted GS or VS, taking into account the raw material used for the production of PBs (grape stalk–GS or vine strains–VS).

The physical–mechanical properties of obtained composite panels were evaluated according to the European standards in terms of density (EN 323:2010), moisture content (EN 322:2010), internal bond strength (IB; EN 319:2010), bending strength (BS; EN 310:2010), and thickness swelling (TS; EN 317:2010). Composite panels were classified according to the standard EN 312:2010.

The wetting behavior of composite panels was determined with the Mobile Surface Analyzer (MSA) One-Click SFE (Kruss, GmbH, Germany) equipped with a video measuring system, high-resolution camera, and a high-performance digitizing adapter that enables instantaneous and frequent registration. Advance software (Kruss, GmbH, Germany) was used for data collection. Water (polar) and diiodomethane (non-polar) were used as probe liquids. A measure of 2.0 μL drop of each liquid was placed on the sample surface for each measurement. Three test drops of each sample were performed for all liquids. The right and left contact angles between each droplet and the sample surface were collected at intervals of 1 s during the first 10 s and intervals of 10 s for a total duration of 60 s, and then, the average of the contact angles was automatically calculated to determine the surface free energy, according to the Owens, Wendt, Rabel, and Kaelble (OWRK) method.

All determinations were conducted in triplicate, and the data are presented as the means ± standard deviations. The impact of extractive agents (NaOH and Na₂CO₃) and their concentrations (0.1, 0.5% and 1.0% w/v) on the extraction efficiency and chemical properties of the obtained extracts (TPC, TFC, and AA) were statistically evaluated (p < 0.05) by two-way analysis of variance (ANOVA) using OriginPro 8.5.2 software and Tukey test for differences of means. Similarly, the variation in wetting properties (contact angle) was also statistically evaluated through two-way ANOVA and the Tukey test. Regarding the comparison of physical–mechanical properties and surface free energy of particleboards, statistical analysis was performed by one-way analysis of variance (ANOVA) at p < 0.05 using OriginPro 8.5.2 software.

Grape stalk (GS) and vine strain (VS) particles were characterized using FTIR analysis, and the obtained spectra are presented in Figure 1. Due to the lignocellulosic nature of GS and VS, cellulose, hemicellulose, and lignin are the main components (c.a. 70%–90%), and for that reason, the normalization of the intensity values of both FTIR spectra was done using the band of the C−O, C−C, and C−C−O stretch vibration at 1028 ± 5 cm⁻¹.

As observed, GS and VS exhibit similar absorption spectra (Figure 1). The broad band at c.a. 3,400 cm⁻¹ is attributed to −OH stretching (Prozil et al., 2012; Pujol et al., 2013; Prozil et al., 2014; Santos et al., 2022a). At c.a. 2,935 cm⁻¹ and 2,873 cm⁻¹ is identified two sharp peaks associated with the asymmetric and symmetric stretch of −CH₂, respectively (Prozil et al., 2012; Pujol et al., 2013; Prozil et al., 2014; Santos et al., 2022a). An increase in the intensity of both peaks in the case of GS particles was observed, which indicates a high content of waxes and oils (Prozil et al., 2012; Pujol et al., 2013; Prozil et al., 2014; Santos et al., 2022a). Additionally, the region attributed to the presence of polyphenols (from 1300 cm⁻¹ to 1800 cm⁻¹) is also enhanced in the case of GS particles (Figure 1), corroborating its high potential for the recovery of phenolic compounds (Troilo et al., 2021; Jesus et al., 2022; Rodrigues et al., 2023).

The recovery of polyphenolic compounds from grape stalks is a widely studied topic, due to the significant amount of lignin and tannins in its composition (Pujol et al., 2013; Prozil et al., 2014; Rodrigues et al., 2022). Additionally, alkaline extraction is a well-known method for phenolics and lignin isolation from lignocellulosic biomass, with several operational advantages, such as low toxicity, safety, and interesting yield (Cavali et al., 2020; Oriez et al., 2020; Deniz, 2023). Grape stalk particles were used for alkaline extraction to obtain a phenolic-rich liquid extract showing potential as an adhesive. In order to optimize the process, two extractive agents were studied, namely, sodium hydroxide (NaOH) and sodium carbonate (Na₂CO₃). Moreover, the impact of the concentration of both agents was also evaluated, using aqueous solutions with 0.1%, 0.5%, and 1.0% w/v. Distilled water was also used as the extractive agent for comparison purposes. For each experiment that was determined, the extraction efficiency (EE) and the final extracts were chemically characterized in terms of total phenolic content (TPC), total flavonoid content (TFC), and antioxidant activity (AA), and the results are presented in Table 2.

As it can be observed, the extractive agent had a significative effect (p < 0.05) on the EE of the process (Table 2). The use of NaOH or Na₂CO₃ as extraction agents allowed improving the efficiency in comparison with water alone, increasing from 2.13% (EXT0, Table 2) to 2.44% and 3.02% of EE, respectively (EXT1 and EXT4, Table 2). However, a sharp variation was noticed in the chemical properties of the extracts. The use of NaOH (EXT1, Table 2) and Na₂CO₃ (EXT4, Table 2) as extraction agents allowed obtaining extracts with a higher content of phenols and flavonoids than the extract obtained using water (EXT0, Table 2). Furthermore, the antioxidant capacity of the extracts was also significantly increased (p < 0.05) with the change in the extraction agent, from c.a. 18% (EXT0, Table 2) to more than 40% (EXT1 and EXT4, Table 2). According to the literature, several phenolic compounds present in lignocellulosic biomass, as grape stalks, are highly soluble under alkaline conditions, which allow their isolation and recovery (Cavali et al., 2020; Oriez et al., 2020). Therefore, the use of aqueous solutions of weak and strong basis, such as Na₂CO₃ and NaOH, as a solvent in the extraction process allow to obtain highly rich mixtures of polyphenols, such as tannins, flavan-3-ols, hydroxycinnamic acids, monomeric and oligomeric flavonols, and stilbenes (Cavali et al., 2020; Oriez et al., 2020; Troilo et al., 2021).

By tuning the concentration of extractive agents, it was possible to significantly improve (p < 0.05) the extraction efficiency. With 0.5% and 1.0% of NaOH, EE was improved to 5.76% (EXT2, Table 2) and 8.92% (EXT3, Table 2), respectively. In the case of Na₂CO₃, efficiency was 1.5 and 2 times higher with 0.5% (EXT5, Table 2) and 1.0% (EXT6, Table 2), respectively, when compared to the result obtained with Na₂CO₃ 0.1% (EXT4, Table 2). In terms of chemical composition, both phenolic and flavonoid contents follow the same tendency, with a great increase in NaOH or Na₂CO₃ concentrations (Table 2). On the other hand, the variation in the antioxidant capacity of the extracts presents a quite different behavior depending on the extractive agent used. In the case of extractions with NaOH, increasing the concentration improves the extraction of phenolic compounds and flavonoids, which led to an increase in the AA of the extracts (EXT1-EXT3, Table 2). However, in the case of Na₂CO₃, the antioxidant potential decreased with the increase in the phenol derivatives and flavonoid content (EXT4-6, Table 2), which may indicate an antagonist effect between the extracted polyphenols. According to the literature, when two or more antioxidant compounds are present in the extract, an antagonist effect may occur, hindering the total antioxidant activity of the extract (Meyer et al., 1998; Pinelo et al., 2004; Hajimehdipoor et al., 2014; Skroza et al., 2022; Uduwana et al., 2023). This effect may be caused by chemical interactions between phenolic compounds, which can produce weaker antioxidant species, forming complexes and adducts, promote polymerization reactions, and neutralize oxidant radicals (Meyer et al., 1998; Pinelo et al., 2004; Hajimehdipoor et al., 2014; Skroza et al., 2022; Uduwana et al., 2023). For example, Meyer et al. (1998) proved the antagonist effect on antioxidant activity of catechin when combined with quercetin, cyanidin, and caffeic and ellagic acids, due to the chemical interaction between their functional groups and, consequently, to the reduced availability of hydroxyl groups for oxidation (Meyer et al., 1998). Furthermore, Uduwana et al. (2023) observed the same effect by combining green tea with bee honey and Citrus limonum extracts (Uduwana et al., 2023). Hajimehdipoor et al. (2014) concluded that ternary combinations of quercetin, rosmarinic acid, rutin, gallic acid, and caffeic acid also exhibit antagonist effect regarding antioxidant activity (Hajimehdipoor et al., 2014).

The impact of extraction agents (water, NaOH, and Na₂CO₃) on the chemical composition of the extracts (EXT0, EXT3, and EXT6) was evaluated by FTIR-ATR, and the obtained spectra are presented in Figure 2A. The normalization of the intensity values was done using the band of the aromatic skeletal vibration at (1600 ± 5) cm⁻¹.

As expected, grape stalk extracts are rich in flavonoids and stilbenes (Anastasiadi et al., 2012), and regardless of the extraction conditions used, all extracts showed the presence of the pattern bands due to the presence of these polyphenolic compounds (Figure 2A). However, the intensity of these bands is low in the spectrum of the extract obtained using only water as an extraction agent (EXT0). This confirms the results obtained for the characterization of phenolic and flavonoid content (Table 2). The low amount of extractable polyphenolic compounds in the extract obtained only with water is due to the fact that most of the water-extractable compounds were previously extracted in the winemaking process.

Regarding the extracts obtained using alkaline extraction agents (NaOH and Na₂CO₃), it was interesting to see that the alkaline agent used impacted the compounds present in the extracts. The spectra of the extract obtained with NaOH (EXT3, Figure 2B) show higher amount of stilbenes in the extract due to the intensity on its characteristics bands: a double peak at c.a. 1600 cm⁻¹ and c.a. 1568 cm⁻¹, corresponding to C−C aromatic double bond stretching and C−C oleophilic stretching, peaks at c.a. 1516 cm⁻¹ and c.a. 1463 cm⁻¹ due to benzene skeleton vibrations, at c.a. 987 cm⁻¹ and c.a. 956 cm⁻¹ due to the bending vibration of C=C−H (Ricci et al., 2015; Santos et al., 2022a; Santos et al., 2022b).

Furthermore, the spectrum of the extract obtained with Na₂CO₃ (EXT 6) shows a high proportion of tannins in the extract, based on the intensity of the bands at 1334 cm⁻¹, 338 cm⁻¹, and 1283 cm⁻¹ due to the C−O−H of primary and tertiary alcohol bending vibrations; 1235 cm⁻¹ of cyclic ether C−O−C stretching vibration; 1028 cm⁻¹ of aromatic cycle bending in a plane vibration, characteristic of the presence of condensed tannins, and the bands at 1728 cm⁻¹ due to the vibrations of the C=O groups (Pujol et al., 2013; Santos et al., 2022a; Santos et al., 2022b; Čabalová et al., 2023), at 1240 cm⁻¹ due to the stretching vibration of the C–O groups; and at 865 cm⁻¹, due to the aromatic C–H out-of-plane bending vibrations, characteristic of the presence of hydrolyzable tannins (Grasel et al., 2016).

The bonding properties of all extracts were evaluated through ABES. Therefore, the extracts were previously concentrated through a rotary evaporator until reached a solid content of 15%. Figure 3 presents the shear strength of each extract after hot-pressing at 120°C and different press times.

All extracts proved to have adhesive properties, reaching a shear strength over 1.0 MPa after 60 s of pressing time at 120 °C (Figure 3). However, significant differences were noted regarding the maximum of performance (Figure 3). In the case of EXT0, the maximum of shear strength was 1.5 MPa and was achieved only after 180 s of pressing (Figure 3). Using an extractive agent (NaOH or Na₂CO₃), the bonding capacity of extracts was significantly improved (Figure 3). Regarding Na₂CO₃ (EXT4-6, Figure 3), the best bonding performance was registered after 120 s (regardless of concentration), which means 33% less pressing time required to complete cure at 120°C. Additionally, higher amounts of Na₂CO₃ (0.5% and 1.0%) slightly increased the maxima of shear strength to 1.6 MPa and 1.7 MPa (EXT5 and EXT6, respectively; Figure 3). On the same way, NaOH extracts’ (EX1-3) also performed better than EXT0, not only by reducing the pressing time (in the case of EXT1) but also increasing the maximum of shear strength to values over 2.0 MPa (EXT2 and EXT3, Figure 3). These findings are closely related to the chemical composition of the extracts, particularly with the phenolic content. As it was previously discussed, hot water extraction has low efficiency to recover phenolics and flavonoids from grape stalks when compared to alkaline extraction agents (Table 2). In the literature, it is referred that the liquid extraction of lignocellulosic biomass, as grape stalk, generates a mixture of long- and short-chain polymers that can be successfully used as bio-adhesives in several industries (Muranaka et al., 2017; Ferreira-Santos et al., 2020; Sun et al., 2022; Zhang et al., 2022; Li et al., 2023; Šernek and Žigon, 2023). The bonding potential is commonly associated with the presence of phenolic compounds, and a high amount of polyphenols usually results in an enhanced bonding capacity of the extracts (Friedman and Jürgens, 2000; Benito-González et al., 2020; Oriez et al., 2020; Troilo et al., 2021; Rodrigues et al., 2023). Therefore, NaOH extracts (EXT1-3; Figure 3) exhibit improved adhesive properties when compared to Na₂CO₃ (EXT4-6, Figure 3) and water (EXT0, Figure 3) extracts possibly due to the higher TPC and TFC (Table 2).

The use of grape stalk as raw material for the production of bio-adhesives to apply in the particleboard industry was evaluated. In this way, EXT3, the best performing extract in terms of bonding capacity, was selected as adhesive, and two types of particles were used as lignocellulosic matrix, namely, grape stalk and vine strains. The objective was to evaluate a circular economy concept, with grape stalks being used for both applications (adhesive production and fiber) and also to study the effect of using other types of winery residue (vine strains) to improve physical–mechanical properties of the composite panels. Vine strains have a high content of cellulose (Benito-González et al., 2020; Senila et al., 2020; Jesus et al., 2022), which provides to this by-product increased resistance against tension forces when compared to grape stalk. In this way, EXT3 was previously concentrated to reach a solid content of 20% before producing the panels. The composite panels were hot-pressed at 160 °C for 10 min at a fixed thickness of 8 mm, with 10 wt% of adhesive. Figure 4 presents the obtained composite panels, using grape stalk (GS_P, Figure 4A) and vine strain (VS_P, Figure 4B) particles.

Physical–mechanical performance of both materials was evaluated in terms of density (D), moisture content (MC), internal bond strength (IB), bending strength (IB), and thickness swelling (TS), and the results are presented in Table 3.

As expected, both composite materials present similar (p < 0.05) thickness and density (c.a. 8.5 mm and 850 kg m⁻³, Table 3). Moreover, the moisture content was also very close in both cases (p < 0.05), with values between 4% and 5% (Table 3). Nevertheless, mechanical properties are substantially different, depending on the type of particles used. In the case of GS_P, IB was 0.25 MPa and BS was 2.2 MPa (Table 3). Using VS particles, the mechanical properties were significantly improved (p < 0.05), with an IB of 0.66 MPa and BS of 4.0 MPa (Table 3). As previously mentioned, vine strains have a higher content of cellulose than grape stalk, providing it higher resistance. Additionally, the mixture of grape stalk particles has a lower fraction of material with a size between 1 and 2 mm than the vine strain mixture, once that GS particles are less dense, and for that reason, the grinding process generates a higher amount of small particles (between 500 μm and 1 mm), which may negatively affect mechanical properties. Although BS is significantly lower regarding to the requirements of type P1 particleboards according to EN 312 (Table 3), the value of IB was 2.3 times higher (Table 3), which proves the potential application of VS_P. Regarding TS, the results showed that GS_P has much higher resistance to swelling than VS_P (Table 3), which can be attributed to the higher oleophilic characteristic of GS particles, as already proved by FTIR analyses (Figure 1), providing the composite material a lower tendency to absorb water and, consequently, to swell. Additionally, the lower content of cellulose in grape stalk particles (when compared to vine strains particles) also promotes the hydrophobic characteristic of GS_P, contributing to a lower TS (Prozil et al., 2012; Benito-González et al., 2020; Senila et al., 2020).

Taking into account the packaging purpose of these composite materials, both panels were characterized in terms of wettability, and the results are presented in Table 4.

The contact angles were very similar (p < 0.05) when water was used as test liquid (c.a. 50°, Table 4), revealing an equivalent hydrophilicity of the surface of both composite materials. However, when diiodomethane was used (Table 4), GS_P registered a higher contact angle (c.a. 70°, Table 4) than VS_P (c.a. 56°, Table 4), which may indicate a lower tendency to the penetration of oils and fats (Candan et al., 2012; Wang et al., 2017; Danish et al., 2019). Although surface free energy seems to corroborate this tendency, with VS_P reaching the highest value (c.a. 52 mJ m⁻², Table 4), the difference was not statistically significant (p < 0.05). In this way, GS_P seems to be a valuable alternative to consider in packaging solutions, not only due to the 100% bio-character but also to its mechanical and wettability properties that provide resistance to mechanical stress caused by handling and transportation and also reduce chemical migration of compounds from package to food.