Valorization of Grape Stalk Rachis for Particleboard Manufacturing: Chemical Characterization and Performance Assessment for Sustainable Interior Panel Applications

Abstract

The wine industry generates large volumes of grape stalk residues that remain largely underutilized and are frequently disposed of by open-field burning, contributing to greenhouse gas emissions and the loss of biomass valorization opportunities. This study aims to evaluate the feasibility of producing single-layer particleboards from grape stalk particles (Vitis vinifera L.) bonded with a urea–formaldehyde resin, with a focus on their suitability for interior non-structural applications. Particleboards were manufactured at three target densities (550, 650, and 750 kg/m³) and assessed mainly for their physical and mechanical performance in relation to the requirements of EN 312. Internal bond strength showed a clear dependence on board density: panels produced at 650 and 750 kg/m³ met the minimum threshold for P2-type particleboards, whereas those produced at 550 kg/m³ did not comply with the standard. Thickness swelling decreased with increasing density, with only the highest-density boards fulfilling the reference criterion. Overall, the results indicate that grape stalk residues can be effectively converted into particleboards with adequate mechanical performance when manufactured at densities of at least 650 kg/m³. The study highlights the potential of this agro-industrial residue as a low-impact raw material for particleboard production, supporting circular bioeconomy strategies and development in wine-producing regions.

1. Introduction

The accelerating depletion of fossil-based and forest-derived raw materials, combined with intensifying regulatory pressure to reduce the environmental footprint of the construction sector, has catalyzed growing scientific and industrial interest in lignocellulosic agricultural residues as alternative feedstocks for engineered composite materials [1,2]. Within this context, the circular bioeconomy framework provides a conceptual and operational foundation for the systematic reintegration of biomass waste streams into productive value chains, thereby decoupling economic output from primary resource extraction [3,4,5].

Lignocellulosic residues generated by agro-industrial processes comprising cellulose, hemicellulose, and lignin as primary structural constituents represent particularly suitable candidates for the manufacture of wood-based composite panels, including particleboards, medium-density fiberboards, and oriented strand boards [3,6]. A wide range of non-wood biomass sources has been investigated for this purpose, including wheat straw, rice husk, sugarcane bagasse, sunflower stalks, hemp shives, and grapevine prunings [7,8,9]. Among these, grapevine-derived residues have attracted increasing attention due to their global abundance the worldwide vineyard area exceeded 7.3 million hectares in 2022 according to the International Organization of Vine and Wine and their consistently documented lignocellulosic composition [10,11].

Particleboards manufactured from lignocellulosic agricultural residues offer demonstrated potential to simultaneously reduce pressure on forest-based wood supplies, lower greenhouse gas emissions associated with biomass incineration, and advance material circularity in the construction supply chain [10,12]. For interior non-structural applications including dry-condition furniture substrates, ceiling panels, lightweight partition systems, and decorative cladding these bio-based panels can be technically adequate while requiring less stringent mechanical performance than load-bearing structural elements [9,13,14,15]. Fire resistance behavior represents an increasingly decisive criterion in this context, particularly in Chile, where the Ordenanza General de Urbanismo y Construcción (OGUC) mandates compliance with NCh1974 for the assessment of flame retardancy in wood-based building products [16].

Grape stalks (escobajos) are the woody vascular rachis remaining following mechanical grape destemming, representing approximately 3–7% of total grape harvest weight and generated at an estimated volume of 500,000 tons per year in Chile alone [12]. Currently, these residues are predominantly managed by open-field burning, direct soil incorporation, or low-value composting, with only marginal quantities directed toward higher-value biorefinery applications [11,12]. Chemically, grape stalk rachis is distinguished from conventional woody raw materials by its parenchyma-dominated anatomy, elevated acetone-soluble extractives fraction (12–20% w/w), high inorganic ash content (4–9% w/w), and comparatively reduced cellulose content relative to hardwood species [10,11]. These compositional characteristics impose specific constraints on adhesive bond formation, moisture resistance, and ignitability in composite panel manufacturing, and must be explicitly addressed in materials design to achieve acceptable performance across multiple property domains.

These compositional characteristics impose specific constraints on adhesive bond formation, moisture resistance, and ignitability in composite panel manufacturing, and must be explicitly addressed in materials design to achieve acceptable performance across multiple property domains. For interior non-structural applications, the EN 312 [17] defines internal bond strength (IB) and thickness swelling (TS) as the critical performance thresholds, reflecting the predominance of cohesion and dimensional stability requirements over flexural performance in dry-condition end uses [18].

2. Materials and Methods

The overall experimental workflow encompassing raw material collection and preparation, chemical characterization, board manufacturing, and multi-property performance evaluation is schematically depicted in Figure 1.

2.1. Raw Material Sourcing and Preparation

Grape stalk biomass was sourced from a commercial wine production facility in the Maule and Biobío regions of central-south Chile, following the 2021 autumn harvest campaign. The material comprised rachis residues from four Vitis vinifera L. cultivars Cinsault, Carménère, País, and Muscatel collected immediately after mechanical destemming. Stalks were comminuted using an industrial drum chipper (Gardenwood NWT-15-220; Segener S.A., Santiago, Chile) and air-dried to a moisture content of 10–12% (w/w, wet basis) prior to chemical analysis and board manufacture.

2.2. Chemical Characterization of Grape Stalk Cell Wall Components

Cell wall chemical composition was determined on an oven-dry mass basis following standardized analytical procedures. Acetone-soluble extractives were quantified by Soxhlet extraction (16 h, 90% v/v acetone (Sigma-Aldrich, St. Louis, MO, USA)). Structural carbohydrates and Klason lignin were determined by a two-stage acid hydrolysis protocol: ground samples (300 mg, 45–60 mesh) were initially hydrolyzed with 72% (w/w) H₂SO₄ (Sigma-Aldrich, 95–98%) at 30 °C for 60 min, followed by dilution to 4% (w/w) acid concentration and secondary hydrolysis by autoclave at 121 °C (1 atm) for 60 min. Acid-insoluble (Klason) lignin was determined gravimetrically from the hydrolysis residue dried at 105 °C to constant mass; acid-soluble lignin was quantified spectrophotometrically at 205 nm using a molar absorptivity of 110 L/g·cm. Monosaccharide monomers (glucose, xylose) and acetic acid were resolved and quantified by HPLC (Merck Hitachi, Tokyo, Japan) equipped with a refractive index detector and an Aminex HPX-87H column (Bio-Rad, Hercules, CA, USA) operated at 45 °C with 5 mM H₂SO₄ as the mobile phase at a flow rate of 0.6 mL/min. Stoichiometric correction factors of 0.900 (glucose → glucan) and 0.880 (xylose → xylan) were applied to account for polysaccharide hydration during hydrolysis. Crude protein content was estimated by the Kjeldahl nitrogen method with a nitrogen-to-protein conversion factor of 6.25, following Sluiter et al. [19]. Ash content was determined by combustion at 575 °C for 3 h in a muffle furnace.

2.3. Morphoanatomical Characterization

Morphoanatomical parameters were determined using an automated L&W Fiber Tester quality analyzer (Lorentzen & Wettre, Stockolm, Sweden), recording: arithmetic mean fiber length (mm), fiber width (µm), fines content (% w/w; particles < 0.2 mm), and coarseness (µg/m). A minimum of 1000 individual fiber elements were measured per sample.

2.4. Adhesive System

A moisture-resistant urea-formaldehyde (RH-UF) resin specifically formulated for particleboard and MDF manufacturing (Adelite 3859; Oxiquim S.A., Concepción, Chile) was used as binder. The resin presented a viscosity of 150–300 cP, a gel time of 15–25 min, and a pH of 8.8–9.4. The resin was applied at a dosage of 8%, 10%, and 12% (w/w on oven-dry particle mass basis) for the 550, 650, and 750 kg/m³ target density formulations, respectively. Ammonium sulphate ((NH₄)₂SO₄) was incorporated as an acid hardener at a loading of 3% (w/w relative to resin solids content).

2.5. Particleboard Manufacture and Experimental Design

Particle size distribution analysis of the ground grape stalk rachis revealed a bimodal pattern with a fine fraction (10%) ranging from 20 to 50 mm and a predominant coarse fraction (90%) between 80 and 120 mm (Figure 2).

Single-layer flat-pressed particleboards (500 × 500 mm nominal dimensions, target thickness 15 mm) were manufactured at three target density levels: 550, 650, and 750 kg/m³. Hot-pressing was conducted at 190 °C using a hydraulic press (ZL–3022B, Dongguan Zhongli Instrument Technology Co., Ltd., Dongguan, China) at a nominal clamping force of 0.40 MN, corresponding to a specific pressure of 1.6 MPa over the panel surface area (0.25 m²). Press times were 120 s for panels at 550 and 650 kg/m³, and 150 s for panels at 750 kg/m³. Following pressing, boards were equilibrated at 20 ± 2 °C and 65 ± 5% relative humidity for ≥7 days prior to specimen preparation. A minimum of five boards were manufactured per density level. Single-layer flat-pressed particleboards with nominal dimensions of 500 × 500 mm and a target thickness of 15 mm were manufactured at three pressing density levels 550, 650, and 750 kg/m³. Blending of particles with resin and hardener was performed in a laboratory drum blender, with atomized resin application to ensure uniform distribution. Mat formation was conducted by hand felting into a forming box, and hot-pressing was executed using a single-opening hydraulic press at a platen temperature of 190 °C and a maximum specific pressure of 14–15 MPa. Pressing times were set at 120 s for the 550 and 650 kg/m³ boards and 150 s for the 750 kg/m³ boards, reflecting the higher mat thickness and density required for the latter formulation. Following pressing, all boards were equilibrated at 20 ± 2 °C and 65 ± 5% relative humidity for a minimum of 7 days prior to specimen preparation and mechanical testing. A minimum of five boards were manufactured per density level.

2.6. Mechanical Properties: Internal Bond Strength

Internal bond (IB) strength defined as the tensile strength perpendicular to the board plane, reflecting interlaminar particle cohesion was determined in accordance with EN 319:2010. Square test specimens (50 ± 1 mm side) were cut from each board, and metal loading blocks were adhesively bonded to both faces. Specimens were subjected to monotonic tensile loading at a constant crosshead displacement rate of 4 mm/min until failure. IB strength was calculated as (Equation (1)):

IB = F_max/A (N/mm²),

(1)

where F_max is the maximum load at failure (N) and A is the nominal bonded area (mm²). A minimum of ten replicate specimens per board density were tested.

2.7. Physical Properties: Thickness Swelling and Water Absorption

Thickness swelling (TS) and water absorption (WA) were determined after 24 h immersion in accordance with EN 317:1993. Square test specimens (50 × 50 mm) were cut, initial thickness measured at the geometric centre using a calibrated dial gauge (resolution ± 0.01 mm), and initial mass recorded to ±0.01 g. Specimens were fully immersed in distilled water at 20 ± 1 °C for 24 h, after which final thickness and mass were recorded immediately following surface blotting (Equation (2)).

TS (%) = (E₂ − E₁)/E₁ × 100,

(2)

where E₁ and E₂ are initial and final thickness (mm), respectively.

WA (%) = (m₂ − m₁)/m₁ × 100,

(3)

where m₁ and m₂ are initial and final mass (g), respectively.

2.8. Reaction-to-Fire Assessment Under NCh1974

The reaction-to-fire behavior of the manufactured boards was evaluated according to NCh1974 Of.86, the Chilean standard for the determination of flame retardancy in coatings and wood-based substrates, which is methodologically equivalent to ASTM D1360-90a (Standard Test Method for Fire Retardancy of Paints Cabinet Method) [16]. Prior to testing, all specimens were oven-dried at 50 °C until reaching constant mass (variation < 0.002 g over 1 h) to eliminate moisture-related confounding effects on combustion behavior.

Combustion experiments were conducted within a standardized enclosed cabinet (internal dimensions: 400 × 700 × 810 mm) maintained at a continuous, laminar air flux of 0.2 m/s. Individual specimens were secured at a 45° inclination relative to the horizontal plane, positioned at the geometric centre of the combustion chamber. Ignition was affected by the controlled combustion of a precisely volumetrically measured quantity of absolute ethanol in a stainless-steel crucible positioned directly beneath the specimen. The test was considered complete upon total consumption of the fuel charge and complete extinction of any residual flaming on the specimen surface. After a controlled cooling period to ambient temperature, final specimen mass was recorded to ±0.001 g.

Two quantitative fire performance metrics were determined. Mass loss percentage (WL%) was calculated as (Equation (4)):

WL (%) = [(m₁ − m₂)/m₁] × 100,

(4)

where m₁ and m₂ denote specimen mass before and after combustion (g), respectively.

The carbonization index (CI), representing the three-dimensional extent of thermally degraded material, was determined by measuring the maximum carbonization depth along the three principal specimen axes following ash removal (Equation (5)):

CI = L₁ × L₂ × L₃ (cm³),

(5)

where L₁, L₂, and L₃ correspond to the maximum carbonization dimensions in the length, width, and depth directions, respectively.

Three replicate specimens per density level were evaluated. Benchmark reference values for commercial particleboard (PB: WL = 3.04 ± 0.2 g; CI = 5.16 ± 0.3 cm³) and medium-density fiberboard (MDF: WL = 8.0 ± 1.0 g; CI = 19.46 ± 5.4 cm³), both evaluated under NCh1974, were sourced from Garay [20].

2.9. Statistical Analysis

All experimental data are presented as arithmetic mean ± one standard deviation (SD). The effect of board pressing density on each measured response variable was assessed by one-way analysis of variance (ANOVA). Where the ANOVA F-test indicated statistically significant differences (p < 0.05), pairwise post-hoc comparisons were performed using Tukey’s honest significant difference (HSD) test at a family-wise significance level of α = 0.05.

3. Results and Discussion

3.1. Chemical Composition of Grape Stalk Rachis

The cell wall chemical composition of grape stalk particles is presented in Table 1. The glucan (cellulose) content of 31.6 ± 1.6% is consistent with values reported in the literature for V. vinifera rachis tissue from multiple cultivars (30–38%) [15,16], and is notably lower than typical hardwood values (ca. 45%), reflecting the prevalence of parenchymatous, thin-walled cells in the rachis vascular anatomy relative to the thick-walled libriform fibre-dominated structure of wood. According to Spigno et al. [21], who characterized six distinct V. vinifera cultivars using identical analytical protocols, the value obtained in the present study corresponds to the lower portion of the reported cultivar range, consistent with known genotype-dependent variability in cell wall biogenesis.

The most analytically distinctive feature of the chemical profile was the acetone-soluble extractives fraction (15.9 ± 0.3% w/w), which is approximately 1.6–2.0 times higher than values typically reported for temperate hardwoods, and substantially exceeds values documented for most annual crop residues employed in particleboard research. This elevated extractive content is characteristic of grapevine rachis tissue and has been ascribed to the accumulation of low-molecular-weight phenolic compounds including stilbenes, flavonoids, and condensed tannins together with fatty acids, waxes, and pectic substances within the intercellular spaces and epidermal layers of the rachis [21,22]. The presence of these compounds is known to interfere competitively with adhesive penetration into lignocellulosic substrates and to occupy hydroxyl-accessible surface sites, thereby reducing the effective contact area for UF resin bond formation [23]. Concurrently, the elevated extractives content is expected to reduce the activation energy for thermal degradation, increasing surface ignitability under direct flame exposure [24,25].

Total Klason lignin content comprising acid-insoluble (16.2 ± 1.0%) and acid-soluble (1.2 ± 1.1%) fractions was 17.4% w/w, consistent with the range reported by Lorenzo et al. [26] for rachis-derived biomass but notably lower than values documented by Ping et al. [10] (ca. 22–27%), a discrepancy attributable to inter-cultivar variability and differences in vine maturity stage at harvest. The hemicellulose fraction (19.1 ± 0.7%), estimated from xylose, arabinose, and acetyl group quantification, fell within the broad range reported across the literature (13.9–35.3%), with extreme values associated with marked differences in extractive removal efficiency and analytical methodology [10,11]. The ash content (9.3 ± 0.7%) substantially exceeded values reported in several comparative studies (3.9–8.6%) [27,28,29], consistent with the relatively high soil mineral uptake characteristic of perennial vine root systems and is potentially advantageous for fire resistance through the formation of a thermally insulating inorganic char layer during combustion [25].

Table 1. Cell wall chemical composition of grape stalk rachis particles and comparative wood reference values [30]. All values expressed as percentage of oven-dry sample mass (% w/w, mean ± standard deviation).

Component (% w/w)	Grape Stalk Rachis	Woods [30]
Glucans (cellulose)	31.6 ± 1.6	40–45
Hemicellulose (total)	19.1 ± 0.7	25–35
Lignin—acid-soluble	1.2 ± 1.1	—
Lignin—acid-insoluble (Klason)	16.2 ± 1.0	—
Total lignin	17.4 ± 1.5	20–30
Acetone-soluble extractives	15.9 ± 0.3	2–5
Ash (inorganic fraction)	9.3 ± 0.7	0.1–1
Crude protein (Kjeldahl × 6.25)	6.1 ± 0.1	<0.5

—: not reported.

Table 1. Cell wall chemical composition of grape stalk rachis particles and comparative wood reference values [30]. All values expressed as percentage of oven-dry sample mass (% w/w, mean ± standard deviation).

Component (% w/w)	Grape Stalk Rachis	Woods [30]
Glucans (cellulose)	31.6 ± 1.6	40–45
Hemicellulose (total)	19.1 ± 0.7	25–35
Lignin—acid-soluble	1.2 ± 1.1	—
Lignin—acid-insoluble (Klason)	16.2 ± 1.0	—
Total lignin	17.4 ± 1.5	20–30
Acetone-soluble extractives	15.9 ± 0.3	2–5
Ash (inorganic fraction)	9.3 ± 0.7	0.1–1
Crude protein (Kjeldahl × 6.25)	6.1 ± 0.1	<0.5

—: not reported.

3.2. Fiber Morphoanatomical Characterization

Fiber morphoanatomical characterization revealed a mean fiber length of 0.5 ± 0.1 mm, fiber width of 20.2 ± 0.8 µm, fines content of 25 ± 0.8% (w/w), and coarseness of 122 ± 3.0 µg/m. The short mean fiber length and elevated fines fraction (particles with projected length < 0.2 mm) are consistent with the parenchyma-dominated cellular architecture of the rachis, in which thin-walled, isodiametric parenchyma cells constitute the predominant tissue type, with only a limited complement of vascular bundle-associated fibers [31,32]. Approximately 60% of the particle population fell within the 0.2–0.4 mm length class, with an additional 20% in the 0.4–0.6 mm range. The high fines fraction is expected to reduce mat air permeability during hot-pressing, potentially affecting the uniformity of the through-thickness density profile, and to increase the specific surface area available for resin coating, thereby influencing the overall binder efficiency and interparticle bond development [33].

3.3. Physical and Mechanical Properties of Grape Stalk Particleboards

The physical and mechanical properties of the manufactured particleboards are compiled in Table 2. Internal bond strength exhibited a pronounced positive response to pressing density, increasing from 0.246 N/mm² at 550 kg/m³ to 0.517 N/mm² at 650 kg/m³ and 1.008 N/mm² at 750 kg/m³. This monotonic increase is mechanistically consistent with the density-dependent enhancement of interparticle contact area, adhesive film continuity, and specific bonding interface development during hot-pressing [33]. Boards manufactured at 650 and 750 kg/m³ satisfied the minimum IB requirement for general-purpose interior particleboards in dry conditions (EN 312 P2: IB ≥ 0.35 N/mm²), while the 550 kg/m³ board remained below this threshold, indicating insufficient interparticle cohesion at the lower compaction level.

The IB values obtained are consistent with findings reported for particleboards produced from other V. vinifera-derived lignocellulosic fractions. Auriga et al. [17] reported that vine pruning waste boards at 650 kg/m³ with 50% agricultural residue content met EN 312 IB requirements, while 100% vine pruning boards at 550 kg/m³ did not, a density composition interaction pattern analogous to that observed in the present study. Ntalos and Grigoriou [34] attributed reduced IB in vine pruning-based boards to competitive interference of high extractives and pith-derived parenchyma fractions with UF resin cross-linking at particle bonding interfaces. The comparatively high IB of 1.008 N/mm² at 750 kg/m³ substantially exceeding values reported for 100% vine-derived boards in the literature [18,34] suggests that the combination of elevated pressing density and the humidity-resistant UF formulation (Adelite 3859) effectively compensated for the adhesion-limiting effect of the elevated extractives content (15.9% w/w), consistent with findings by Hoseinzadeh et al. [35] for hybrid grapevine-wood particleboards bonded with 12% UF resin.

Measured board densities (579, 664, and 819 kg/m³) systematically exceeded nominal target values across all formulations, a well-documented outcome in laboratory-scale particleboard production attributable to elastic springback of compressed particles following press opening and moisture-induced thickness changes during post-press conditioning [33]. The 650 kg/m³ board (measured density: 664 kg/m³) fell within the industrial target density range of 620–670 kg/m³ cited by Wong et al. [36] for general-purpose interior particleboards and is therefore the formulation most directly comparable to commercial product standards.

Thickness swelling following 24 h water immersion decreased monotonically with pressing density (18.45% → 16.69% → 12.71%), with only the 750 kg/m³ board satisfying the ≤14% reference criterion [20]. The relatively elevated TS values at 550 and 650 kg/m³ may be attributed to the combined effect of: (i) the hygroscopic character of the parenchyma-rich grape stalk tissue, in which abundant thin cell walls and intercellular spaces facilitate rapid moisture sorption; (ii) the elevated extractives content (15.9% w/w), which may partially impair UF resin cure by interfering with isocyanate group reactivity; and (iii) the intrinsic susceptibility of UF polymer networks to hydrolytic chain cleavage under prolonged aqueous exposure [23]. These TS values are nevertheless considerably lower than those reported for 100% vine pruning boards by Martínez-García et al. [36] (TS: 26.0–31.6%), which were produced without water-repellent additives. Water absorption values (49.67–75.47%) substantially exceeded the ≤40% commercial reference [20] across all density levels, indicating that the current formulation while suitable for dry condition applications requires supplementary moisture resistance intervention (e.g., wax emulsion addition, MUF or pMDI resin substitution, or surface impregnation) for deployment in humid service conditions [37].

Table 2. Physical and mechanical characterization of grape stalk particleboards bonded with humidity-resistant UF resin (Adelite 3859) at three pressing density levels. Values are arithmetic mean ± standard deviation.

Property	550 kg/m3	650 kg/m3	750 kg/m3	Standard Requirement
Internal bond, IB (N/mm2)	0.246	0.517 *	1.008 *	≥0.35 [EN 312 P2]
Board density (kg/m3)	579	664 *	819	620–670 [36]
Thickness swelling, TS (%/24 h)	18.45	16.69	12.71 *	≤14 [20]
Water absorption, WA (%/24 h)	63.71	75.47	49.67	≤40 [20]

* Meets the respective standard threshold. EN 312 P2: general-purpose interior particleboards for use in dry conditions.

Table 2. Physical and mechanical characterization of grape stalk particleboards bonded with humidity-resistant UF resin (Adelite 3859) at three pressing density levels. Values are arithmetic mean ± standard deviation.

Property	550 kg/m3	650 kg/m3	750 kg/m3	Standard Requirement
Internal bond, IB (N/mm2)	0.246	0.517 *	1.008 *	≥0.35 [EN 312 P2]
Board density (kg/m3)	579	664 *	819	620–670 [36]
Thickness swelling, TS (%/24 h)	18.45	16.69	12.71 *	≤14 [20]
Water absorption, WA (%/24 h)	63.71	75.47	49.67	≤40 [20]

* Meets the respective standard threshold. EN 312 P2: general-purpose interior particleboards for use in dry conditions.

The physical and mechanical performance of the grape stalk particleboards was compared with typical values reported for commercial particleboards used in interior applications. In terms of internal bond (IB), the panels manufactured at 650 kg/m³ (0.517 N/mm²) and 750 kg/m³ (1.008 N/mm²) exceeded the minimum requirement established by EN 312 for P2-grade particleboards (IB ≥ 0.35 N/mm²), indicating that their internal cohesion is within the range of commercially acceptable materials. In particular, the IB value obtained at 750 kg/m³ falls within, and even surpasses, the upper range commonly reported for conventional industrial particleboards (typically 0.4–0.8 N/mm²), suggesting that sufficient bonding can be achieved despite the high extractives content of grape stalk biomass. Measured board densities (664 kg/m³ for the nominal 650 kg/m³ formulation) were consistent with the typical industrial density range for general-purpose particleboards (620–670 kg/m³), supporting the comparability of the results with commercial products. However, differences become more pronounced when considering dimensional stability. Thickness swelling (TS) after 24 h immersion reached 12.71% at 750 kg/m³, thereby meeting the commonly accepted commercial threshold (≤14%). In contrast, boards at 550 and 650 kg/m³ exceeded this limit, indicating reduced dimensional stability relative to standard commercial panels. Water absorption (WA) values (49.67–75.47%) were substantially higher than typical values reported for commercial particleboards (generally ≤30–40%), reflecting the hygroscopic nature of the raw material and the absence of water-repellent additives in the formulation. This suggests that, although internal bonding performance is comparable to commercial materials, moisture resistance remains a limiting factor. Overall, the results indicate that grape stalk-based particleboards can achieve mechanical performance comparable to commercial P2-grade panels, particularly at densities ≥650 kg/m³. Nevertheless, their physical performance, especially in terms of water resistance, is inferior to that of conventional industrial particleboards, highlighting the need for adhesive system optimization or hydrophobic additives to reach full commercial equivalence.

3.4. Reaction-to-Fire Performance Under NCh1974

The macroscopic combustion behavior of the boards during the NCh1974 flame test is illustrated in Figure 3. A visually progressive and consistent reduction in the geometrical extent of the carbonized surface zone is apparent with increasing board density, providing qualitative evidence of the protective effect of higher compaction on limiting heat conduction and char front propagation within the board matrix [24].

Quantitative fire performance data are presented in

Table 3. Mass loss (g) and carbonization index (cm³) of grape stalk particleboards evaluated under NCh1974 [16] (mean ± SD, n = 3). Commercial reference values for particleboard (PB) and MDF from Garay [20].

Board Density (kg/m3)	Initial Mass (g)	Post-Combustion Mass (g)	Mass Loss (g)	Carbonization Index (cm3)
550	460.4 ± 8.6	450.0 ± 8.9	10.9 ± 0.6	48.8 ± 8.4
650	579.5 ± 23.4	569.3 ± 23.0	10.1 ± 0.9	31.4 ± 4.9
750	567.2 ± 14.8	557.1 ± 14.8	9.5 ± 0.95	18.3 ± 3.7
Reference values under NCh1974 [20]
PB commercial	—	—	3.04 ± 0.2	5.16 ± 0.3
MDF commercial	—	—	8.0 ± 1.0	19.46 ± 5.4

—: not applicable. PB: commercial particleboard; MDF: medium-density fiberboard.

. Both mass loss and carbonization index exhibited statistically significant (p < 0.05) monotonic decreases with increasing board density. Mass loss declined from 10.9 ± 0.6 g at 550 kg/m³ to 10.1 ± 0.9 g at 650 kg/m³ and 9.5 ± 0.95 g at 750 kg/m³, representing a net reduction of 12.8% across the density range. The carbonization index exhibited a substantially more pronounced density dependence, decreasing from 48.8 ± 8.4 cm³ at 550 kg/m³ to 31.4 ± 4.9 cm³ at 650 kg/m³ and 18.3 ± 3.7 cm³ at 750 kg/m³, yielding a 62.5% overall reduction. The markedly asymmetric magnitude of the density effect on CI versus mass loss 62.5% versus 12.8% carries mechanistic significance: it implies that board compaction primarily constrains the lateral and vertical propagation of the carbonization front by reducing interparticle void volume and the associated oxygen-diffusion pathways that sustain oxidative char combustion, rather than substantially reducing the total volatile mass fraction released during the initial pyrolytic decomposition phase [24,38]. The 750 kg/m³ board attained a CI of 18.3 ± 3.7 cm³, which is statistically indistinguishable from the commercial MDF reference value (19.46 ± 5.4 cm³) [20], and is therefore the formulation most proximate to MDF-equivalent fire performance in absolute terms.

When fire performance metrics are normalized to initial specimen mass thereby enabling direct comparison across density levels independently of specimen weight, and alignment with the percentage-based thresholds established by Garay & Henríquez [39] a more nuanced pattern emerges (Figure 4). Percentage mass loss values of 2.37% (550 kg/m³), 1.74% (650 kg/m³), and 1.67% (750 kg/m³) were obtained, all of which fall below the 3.4% reference threshold for commercial PB [39], indicating that the mass-specific fire load generated per unit board weight is equivalent to or lower than that of commercial panels across the full density range investigated. For the normalized carbonization index, the 550 kg/m³ board (10.60%) exceeded the 8% commercial PB reference, while boards at 650 kg/m³ (5.42%) and 750 kg/m³ (3.23%) both satisfied this criterion. Consequently, boards manufactured at ≥650 kg/m³ fulfil both percentage-based fire performance thresholds for commercial particleboard under NCh1974 [39].

The high ash content (9.3% w/w) and lignin fraction (17.4% w/w) contribute to a thermally resistant fraction that does not participate in combustion, reducing the mass-specific fire load per unit board weight [25,38]. These findings are corroborated by results for other Vitis vinifera-derived panels [40,41].

4. Conclusions

This study evaluated the feasibility of producing single-layer particleboards from grape stalk rachis residues derived from the Chilean wine industry, with emphasis on their physical, mechanical, and reaction-to-fire performance.

(1)

Grape stalk rachis exhibits a distinctive lignocellulosic composition, characterized by moderate glucan content (31.6 ± 1.6%), elevated acetone-soluble extractives (15.9 ± 0.3%), high inorganic ash content (9.3 ± 0.7%), and parenchyma-dominated morphoanatomy. These features influence both interparticle bonding behavior and thermal response during combustion.

(2)

Pressing density was identified as a key processing parameter controlling panel performance. Boards manufactured at 650 and 750 kg/m³ achieved internal bond strength values of 0.517 and 1.008 N/mm², respectively, meeting the EN 312 requirement for P2-type particleboards in terms of internal cohesion. However, thickness swelling met the reference criterion only at 750 kg/m³, and water absorption remained above recommended values for all formulations, indicating that moisture resistance requires further optimization.

(3)

Reaction-to-fire performance improved with increasing board density. A substantial reduction in carbonization index (62.5%) was observed from 550 to 750 kg/m³, while mass loss showed a smaller decrease (12.8%), suggesting that densification primarily limits char propagation rather than volatile release. When normalized to specimen mass, boards at ≥650 kg/m³ met the reference thresholds reported for commercial particleboards under NCh1974.

(4)

Based on the properties evaluated in this study, grape stalk residues show potential as a raw material for particleboard production. However, the absence of bending properties (MOR and MOE) limits a comprehensive assessment of their suitability for interior applications according to EN 312 requirements. Therefore, conclusions regarding end-use performance should be considered preliminary.

(5)

Although the valorization of grape stalk residues represents a promising pathway for the utilization of agro-industrial biomass, broader sustainability claims cannot be fully supported within the scope of this study. Additional assessments such as formaldehyde emissions, long-term durability, and life-cycle analysis are required to evaluate the environmental performance of panels produced with urea–formaldehyde resin systems.

Future work should include full mechanical characterization (MOR/MOE), optimization of adhesive systems to improve moisture resistance, and evaluation of environmental performance indicators, including emissions and life-cycle impacts, to support a more comprehensive assessment of application potential.