Physicomechanical Properties of Recycled Gypsum Composites with Polyvinyl Acetate Emulsion and Treated Short Green Coconut Fibers

Abstract

The reintegration of waste into the production chain represents a sustainable method of reducing environmental impact while promoting economic growth. This also aligns with social and environmental demands. In this study, composites were produced from commercial and recycled gypsum, polyvinyl acetate (PVA) emulsions, and chemically treated short green coconut fibers, and characterized by physical and mechanical analyses. The addition of PVA improved paste workability, extended setting time, and reduced porosity, while fiber pretreatment enhanced adhesion and tensile performance. XRD, FTIR, and TGA-DTA confirmed modifications in crystallinity, bonding, and thermal stability due to the combined action of PVA and fibers. Compared with the recycled gypsum reference (RG), the optimized composite (R50C50P5F10) exhibited a 69.1% reduction in sorptivity (from 5440 × 10⁻⁴ to 1680 × 10⁻⁴ kg/m²·s^0.5), a 27.9% increase in flexural tensile strength (from 2.65 to 3.39 MPa), and a 15.1% increase in compressive strength (from 6.18 to 7.12 MPa). Surface hardness values remained statistically equivalent to RG but complied with normative requirements, maintaining all formulations within the moderate hardness category (55–80 Shore C). The results demonstrate the technical feasibility of incorporating recycled gypsum and agro-industrial fibers into gypsum composites, providing a sustainable route for developing more durable construction materials.

1. Introduction

Gypsum products are widely employed in interior construction for partition walls, lining systems, molded elements, and ceiling panels due to their low density, fire resistance, acoustic benefits, and ease of finishing. Nonetheless, gypsum’s susceptibility to water uptake and its brittle fracture limit broader use without modification. At the same time, construction and demolition activities generate significant plasterboard waste, making gypsum recycling both an environmental necessity and an opportunity for circularity in the built environment. In Brazil, national drywall sales in 2020 exceeded 100 million square meters [1].

Recent environmental assessments and sectoral analyses indicate that, under appropriate logistics and processing, recycled gypsum (RG) can be technically, environmentally, and economically viable compared with natural quarry or FGD gypsum, reinforcing its potential as a secondary raw material for binders and prefabricated components [2,3,4]. In Brazil, according to Cavalcanti and Póvoas [5], approximately 10% of the total volume of gypsum boards produced annually is lost due to waste generated during assembly, production, and demolition. Therefore, the recycling of waste gypsum boards could provide a viable and sustainable resource.

A broad body of research shows that fiber-reinforcement is an effective strategy to mitigate gypsum’s brittleness and to improve crack control, energy absorption, and impact resistance. Recent reviews and studies document gains in flexural/compressive performance and functional properties (e.g., thermal/acoustic behavior, fire response) when gypsum is reinforced with natural or synthetic fibers and/or polymer modifiers. Notably, short natural fibers (e.g., hemp, sisal, jute, wood, coir) increase toughness and can raise flexural strength by ~10–30% depending on fiber type, volume, and fiber–matrix compatibility; some systems even achieve pseudo-strain-hardening with ultra-high ductility when interfacial tailoring and fiber dispersion are optimized [6,7,8].

In particular, coir (coconut) fiber—abundant agro-residues—have been shown to enhance flexural and compressive strengths and lower thermal conductivity in gypsum matrices, with the best performance typically observed around ~10% fiber content by volume or mass (depending on the study’s basis and geometry) [8].

Alongside fibers, polyvinyl acetate (PVA) emulsions/films are frequently adopted as processing aids and microstructural modifiers. PVA can adsorb on hemihydrate/dihydrate surfaces, interfering with crystal growth (notably along the c-axis), delaying nucleation/setting, and producing more compact crystal networks that improve cohesion and, in several reports, compressive strength [9,10]. When recycled gypsum is the base binder, both fibers and PVA can simultaneously address increased water demand, altered setting kinetics, and variability introduced by impurities or particle-size differences typical of reclaimed streams, thereby helping close the performance gap to commercial gypsum [2].

By improving fracture toughness, indentation/impact resistance, and moisture management (via reduced sorptivity), fibrous, polymer-modified RG composites can expand gypsum’s use beyond conventional decorative casts toward higher-demand interior, non-load-bearing systems: (i) impact/abuse-resistant partitions and linings for schools, hospitals, corridors, and garages; (ii) prefabricated gypsum blocks/tiles with upgraded surface hardness; (iii) lightweight panels targeting thermal/acoustic retrofits; and (iv) repair mortars and molded elements requiring greater flexural capacity. These application classes are consistent with current commercial specifications for abuse/impact-resistant gypsum panels and moisture-resistant interior products. Scope-wise, the present work remains focused on interior, non-structural uses, where improved toughness and lower capillary uptake are most relevant.

To avoid ambiguity in comparison with standard values, this study benchmarks fresh, physical, and mechanical results against the following explicit standards: (i) ABNT NBR 13207:2017 [11] and NBR 13207-3:2023 [12] for construction gypsum requirements and mechanical test methods; (ii) UNE-EN 13279-1:2009 [13] for minimum flexural strength of cast gypsum; (iii) ASTM C28/C28M-10 [14] for compressive strength of gypsum plasters; (iv) ISO 15148:2002 + A1:2016 [15] for capillary water absorption (sorptivity); and (v) ABNT NBR 16494:2017 [16] for surface hardness classification of gypsum blocks. These references guided the selection of test methods and the interpretation of compliance thresholds throughout the paper.

Building on the above, this study introduces a hybrid reinforcement strategy that combines recycled gypsum with alkali-treated green coconut fibers and polyvinyl acetate (PVA) emulsion. Unlike previous works that evaluate fibers or polymers separately, the present approach demonstrates a synergistic effect that improves setting behavior, reduces sorptivity, and enhances flexural, compressive, and surface hardness properties. By coupling waste valorization with chemical fiber pretreatment and polymer modification, and confirming the mechanisms through XRD, TGA, and FTIR analyses, this research establishes a technically innovative and sustainable pathway for producing high-performance gypsum composites. The optimized formulation not only recovers the performance losses typically associated with recycled gypsum but also broadens the potential applications of gypsum-based materials in moisture-resistant, impact-resistant, and prefabricated construction systems.

2. Materials and Methods

2.1. Materials

This study employed a commercial gypsum that is commonly used for producing construction components and is defined by the Brazilian standard NBR 13207 [11] as “fast-setting gypsum”. Recycled gypsum (Figure 1) was obtained from waste generated during the cutting and assembly of drywall panels, supplied by a local company in Ilhéus, BA, Brazil. The preparation and characterization of recycled gypsum is detailed in the following section.

The PVA emulsion used in this study was Cascorez Extra white glue. It is a white liquid with a density of 1.02–1.07 g/cm³ at 25 °C and a solid content ranging from 45% to 47%. Short green coconut fibers with lengths varying between 20 and 30 mm were extracted from the mesocarp of discarded green coconut waste collected in the coastal region of Una municipality in the state of Bahia, northeastern Brazil. The average fiber diameter was 0.275 ± 0.065 mm, measured by profilometry (Rtec Instruments, UP-2000 series, coupled to the Universal Tribometer MFT-5000, San Jose, CA, USA) and optical microscopy on 105 specimens, with multiple measurements per fiber to estimate the cross-sectional area. Single-fiber tensile tests, performed in accordance with ASTM C1557-20 [17] and adapted from ASTM D3822-96 [18], were conducted using 50 mm templates with a 30 mm gauge length, cyanoacrylate-bonded ends, and sandpaper reinforcement to ensure proper gripping. The results indicated a maximum load (4.31 ± 2.49 N), tensile strength (71.80 ± 25.42 MPa), maximum deformation (6.54 ± 1.85%), and elastic modulus (2.57 ± 0.24 GPa). At least 10 valid specimens were tested per condition, ensuring statistical robustness of the results. The tests were performed on an electromechanical testing machine (Instron, EMIC 23-10, São José dos Pinhais, PR, Brazil) equipped with a 2 kN load cell, under displacement control at a rate of 2 mm/min.

The determination of the percentage of PVA emulsion and coconut fiber to be incorporated was based on the results obtained by Brandão [19], who investigated different compositions and modifications of gypsum with PVA emulsion, ranging from 2.5% to 30% by mass of gypsum, in increments of 2.5%. Additionally, the study evaluated the addition of treated green coconut fiber in proportions ranging from 1% to 10% by volume.

2.2. Preparation and Characterization of Recycled Gypsum

The recycled gypsum was reclaimed from waste drywall panels by manually removing the adhered cardboard and metallic elements (which were subsequently sent for reuse and recycling), followed by manual and electromechanical crushing and grinding of the gypsum waste. The ground gypsum was sieved through a mesh with 0.290 mm openings and calcined in a forced-air circulation oven at 150 °C for 10 h. After temperature stabilization, the recycled gypsum was stored in 1 kg packages. These stages are illustrated in Figure 2.

Table 1. Physicochemical parameters of gypsum powder.

Physical Parameters
Properties (Unit)	Gypsum Type	Result (CV) 2	p-Value 3	Normative Limit	Standard
Granulometry (%) 1	Commercial	96 (2.1)	2.8930 × 10−3	≥90	NBR 13207 [11]
	Recycled	91 (3.2)
Unit mass (kg/m3)	Commercial	830 (3.0)	1.2796 × 10−12	≥600
	Recycled	636 (3.4)
Specific mass (kg/m3)	Commercial	3096.54 (6.2)	4.2841 × 10−17	N/A 4
	Recycled	2442.66 (9.8)
Specific surface area (m2/kg)	Commercial	540.74 (2.9)	4.9864 × 10−12	200–800	NBR 16372/2015 [20]
	Recycled	346.54 (3.7)
Chemical Parameters
Properties (unit)	Gypsum Type	Result (CV) 1	p-Value	Normative Limit	Standard
Free water content (%)	Commercial	1.23 (8.1)	1.0143 × 10−2	≤1.3	NBR 13207 [11]
	Recycled	1.07 (9.3)
Water of crystallization (%)	Commercial	6.1 (3.3)	2.2101 × 10−3	4.2–6.2
	Recycled	4.9 (8.2)
Calcium oxide content (%)	Commercial	41.4 (5.6)	7.0134 × 10−11	>38
	Recycled	32.7 (5.4)
Sulfur trioxide content (%)	Commercial	51.6 (7.5)	7.6898 × 10−10	>53
	Recycled	40.3 (9.1)

¹ Percentage that passes through a sieve with 0.290 mm openings; ² CV: Statistical coefficient of variation; ³ p-value: Probability level obtained from the statistical test (ANOVA followed by Tukey’s post hoc test) used to evaluate whether differences between mixtures are statistically significant (values < 0.05 indicate significance at the 95% confidence level); ⁴ N/A: No applicable standard.

presents the results of the physicochemical characterization of the commercial and recycled gypsum powders (hemihydrate). The values in parentheses are the coefficients of variation (%). All obtained data showed a normal distribution within each set of samples and statistically significant differences when comparing the results of commercial and recycled gypsum. The lower CaO and SO₃ in RG are consistent with dilution by residual paper/cardboard fibers and other additives common in drywall (e.g., perlite, alumina sources), and, in minority, due to incomplete calcination/rehydration, as corroborated by XRD (Section 3.6).

All physical parameters of the commercial and recycled gypsum powders complied with the minimum normative criteria. The free water content and water of crystallization of both samples also met the minimum requirements. However, the calcium oxide content was satisfactory only for the commercial gypsum; that of the recycled gypsum was 13.9% below the normative minimum. Neither sample met the minimum requirement for sulfur trioxide content.

The results were statistically analyzed using the Shapiro–Wilk test. Each test was conducted on three to six specimens of each composite with K variables. Statistical tests were performed using the open-source software Past (version 4.03). Significance was assessed using t-tests between the measured parameters and hypothetical values. The hypothetical values adopted served as a reference, and the explanatory variables were considered to have no significant effect on the response variable at p > 0.05.

Considering that some datasets have different units, the statistical coefficient of variation was employed to indicate the variability of the results within the same dataset. This approach expresses the variability of the data while excluding the influence of the magnitude of the variable.

2.3. Pretreatment of Coconut Fibers

The green coconut fibers were donated by the Polymer and Systems Laboratory of the State University of Santa Cruz, Brazil, after manual peeling and defibrillation. Subsequently, the fibers were sent to the Technology Laboratory of the Federal Institute of Bahia, Ilhéus campus, Brazil, for manual carding to remove visible impurities resulting from defibrillation and storage.

After manual cleaning, the green coconut fibers were subjected to alkaline pretreatment following a methodology adapted from [9]. The choice of chemical pre-treatment was based on the results obtained in [19], where the authors compared the effects of different processes on green coconut fibers, including thermal treatment, wetting–drying cycles, and alkaline treatment. Their findings showed that sodium hydroxide treatment was the most effective in removing amorphous components such as hemicellulose and lignin, increasing cellulose crystallinity, and improving fiber–matrix adhesion, which justifies its adoption in the present study.

The fibers were immersed in an aqueous solution consisting of 5% (w/w) sodium hydroxide (NaOH) in distilled water for 1 h (Figure 3a). Next, the fibers were thoroughly rinsed with distilled water until a neutral pH was achieved. The pH was monitored using pH indicator strips (Figure 3b). Once neutralization was complete, the fibers were dried in an oven at 60 °C for 10 h. Finally, the samples were transferred to a laboratory bench and allowed to dry naturally for 24 h (Figure 3c).

2.4. Characterization of Gypsum Pastes in the Fresh State

The properties of the gypsum pastes in the fresh state were analyzed following the Brazilian standard NBR 12128 [21]. The evaluated parameters included the normal consistency and setting times (initial and final). These parameters are essential for evaluating the mixing behavior and strength of gypsum to ensure technical compliance for practical applications.

The normal consistency of the gypsum pastes was evaluated using a Vicat apparatus specifically adapted for gypsum, as shown in Figure 4a. After establishing the water/gypsum ratio, the setting times were measured using a standard Vicat apparatus (Figure 4b). The initial setting time was when the needle stabilized 1 mm from the base, and the final setting time was when the needle no longer penetrated the paste, leaving only a superficial mark.

Visible differences in the consistency of the mixtures were observed during the preparation of the reference pastes (without additives). The commercial gypsum sample exhibited greater fluidity (Figure 5a), whereas the recycled gypsum (Figure 5b) and fibrous gypsum samples (Figure 5c) had a more paste-like texture. The water/gypsum ratios required to achieve normal consistency were 0.6, 0.7, and 0.65 for the commercial, recycled, and fibrous gypsum mixtures, respectively.

2.5. Production of Gypsum-Based Composites

Table 2. Composition of the experimental matrices.

Sample	Gypsum		PVA Emulsion (%)	Fiber Content (%)	Water/Gypsum Ratio
	Recycled (%)	Commercial (%)
Reference—no additives
Commercial gypsum (CG)	0	100	0	0	0.6
Recycled gypsum (RG)	100	0	0	0	0.7
Composite group 1
R100P5F1	100	0	5	1.0	0.65
R100P5F2.5	100	0	5	2.5	0.65
R100P5F5	100	0	5	5.0	0.65
R100P5F7.5	100	0	5	7.5	0.65
R100P5F10	100	0	5	10.0	0.65
Composite group 2
R70C30P5F1	70	30	5	1.0	0.65
R70C30P5F2.5	70	30	5	2.5	0.65
R70C30P5F5	70	30	5	5.0	0.65
R70C30P5F7.5	70	30	5	7.5	0.65
R70C30P5F10	70	30	5	10.0	0.65
Composite group 3
R50C50P5F1	50	50	5	1.0	0.65
R50C50P5F2.5	50	50	5	2.5	0.65
R50C50P5F5	50	50	5	5.0	0.65
R50C50P5F7.5	50	50	5	7.5	0.65
R50C50P5F10	50	50	5	10.0	0.65

presents the compositions of the matrices used in this study. All samples are denoted based on the type and content of raw materials used: R, recycled gypsum; C, commercial gypsum; P, PVA emulsion; and F, treated green coconut fibers. The numbers following each letter represent the percentage of that raw material in the composite. The percentage of gypsum and PVA emulsion is expressed by weight, while the percentage of coconut fiber is measured by volume.

To prepare the test specimens (Figure 6), water and PVA emulsions were initially mixed to form a homogeneous liquid, which was then transferred to a waterproof container. The dry materials (gypsum and fibers) were first homogenized to avoid fiber entanglement and promote even distribution. The coconut fibers were manually carded and dispersed within the gypsum powder to separate individual strands and improve coating. This dry mixture was then gradually sprinkled over the PVA–water solution while stirring for 1 min, followed by a 2 min resting period to allow initial wetting, and a final 1 min of mixing. This procedure ensured that the fibers were uniformly coated and dispersed within the matrix, preventing clumping or preferential orientation. The resulting mixture was immediately transferred to a mold.

After 24 h, the specimens were demolded, placed into a sealed recipient, and cured at room temperature for 28 d at 24 ± 3 °C before further testing. Prismatic molds (40 mm × 40 mm × 160 mm) were used to prepare samples for flexural strength and capillary absorption tests, whereas cubic molds (50 mm × 50 mm × 50 mm) were employed to prepare samples for hardness and compression tests.

2.6. Physicomechanical Characterization of Hardned Composites

2.6.1. Sorptivity Test

Water absorption by capillarity (sorptivity) tests followed the procedure described in the European standard EN ISO 15148:2002 + A1:2016 [15]. Three prismatic specimens of each mixture were molded. After curing for 28 days, each specimen was cut in half, resulting in six specimens with dimensions of 40 mm × 40 mm × 80 mm for each composite.

Before the test, the lateral faces of each specimen were sealed with a water- and vapor-impermeable material, as shown in Figure 7a, to allow only unidirectional water flow. The samples were placed in a tray with a 5–10 mm layer of water, as illustrated in Figure 7b. The cut faces were placed downward, and the samples were supported by a holder to ensure the base was immersed in the water and to prevent air bubbles from being trapped beneath the samples.

2.6.2. Flexural Tensile Strength Test

Flexural tensile strength tests were conducted according to the Brazilian standard NBR 13279 [22]. A universal testing machine (AGX-Plus, Shimadzu do Brasil, Barueri, SP, Brazil) with a 100 kN load cell was used (Figure 8). Five specimens were tested for each composite, and the load was applied at a rate of 50 ± 10 N/s until failure. The flexural tensile strength was calculated using Equation (1), where R_f is the flexural tensile strength (MPa), F_t is the load applied vertically at the center of the prism (N), and L is the distance between the supports in millimeters, mm.

$$R_f={\displaystyle \frac{1.5\times F_t\times L}{{40}^3}}$$

(1)

2.6.3. Axial Compressive Strength Test

The compressive strength was evaluated according to the Brazilian standard NBR 13207-3 [12] using a servo-hydraulic press (PCS200, EMIC, Instron Brasil, São José dos Pinhais, PR, Brazil). Five specimens were tested for each composite, utilizing the samples from the hardness tests on unused and unsanded faces.

To ensure the load was applied uniformly, metal plates with dimensions of (50.0 ± 0.1) mm (length) × (50.0 ± 0.1) mm (width) × 15 mm (thickness) were placed on the top and bottom faces of the specimens, ensuring a minimum application area of 2500 mm², as illustrated in Figure 9.

The load was applied continuously at rates ranging from 250 to 750 N/s until the specimens ruptured, and the compressive strength was calculated using Equation (2), where R_c is the compressive strength (MPa), F_c is the load at rupture (N), and 2500 (mm²) is the area of the square section of the loading device (50 mm × 50 mm).

$$R_c={\displaystyle \frac{F_c}{2500}}$$

(2)

2.6.4. Surface Hardness Test

The surface hardness was determined following the recommendations of the Brazilian standard NBR 13207-3 [12] using a calibrated portable Shore C durometer (Figure 10a,b). Five cubic specimens were tested for each composite (Figure 10c), and the readings were taken directly from the device.

Results of flexural tensile strength, axial compressive strength, and surface hardness were complemented by statistical analysis. The recycled gypsum (RG) sample was used as the reference group, and pairwise comparisons were conducted between RG and each composite. An independent two-sample t-test (Student’s t, unequal variances assumed) was applied to compare the mean values of each mixture with the reference, adopting a 95% confidence level (p < 0.05). For each comparison, the mean, standard deviation, percentage variation relative to RG, t-value, and p-value were determined. Differences were considered statistically significant when p < 0.05.

2.6.5. X-Ray Diffraction Analysis

The mineralogical compositions of the composites were analyzed using X-ray diffraction (XRD) with a Bruker D2 Phaser diffractometer equipped with a copper target tube (λ = 0.15406 nm) operating at 30 kV and 10 mA, without a secondary monochromator. The XRD patterns were obtained in the 2θ range of 5–70° in continuous mode at a rate of 0.3°/s. The crystalline phases in the samples were identified using DIFFRAC plus EVA software, version 6, with the Crystallography Open Database.

From the XRD results, the crystallinity index of the PVA emulsion and the untreated and treated fibers were calculated using Equation (3), based on the empirical method proposed by Segal et al. [23], where I_c is the crystallinity index (%), I₁ is the intensity of diffraction peaks corresponding to the crystalline portion, and I₂ is the intensity of diffraction peaks corresponding to the amorphous portion.

$$I_c={\displaystyle \frac{I_1-I_2}{I_1}}\times 100$$

(3)

Based on the obtained XRD patterns, the crystallite size was calculated using the Scherrer equation, where t is the crystallite size (nm), K is a constant that depends on the particle shape (equal to 0.9 for spherical particles), λ is the X-ray wavelength (nm), β is the full width at half maximum of the diffraction peaks, and θ is angle of the maximum peak intensity (rad).

$$t={\displaystyle \frac{K\lambda }{\beta coscos\left(\theta \right)}}$$

(4)

Additionally, the interplanar spacing of the crystal lamellae was calculated based on Bragg’s law, using the (220) plane as a reference, using Equation (5), where d is the distance between the inter-reticular planes (nm), θ is Bragg’s angle (rad), n is an integer corresponding to the number of wavelengths, and λ is the X-ray wavelength (nm).

$$2dsin\left(\theta \right)=n\times \lambda$$

(5)

2.6.6. Thermogravimetric Analysis

Thermogravimetric analyses, including thermogravimetry (TG) and differential thermogravimetry (DTG), were performed on powdered composite samples using a Shimadzu TGA-50 (Shimadzu Corporation: Kyoto, Japan) thermogravimetric analyzer under a nitrogen atmosphere. Powdered samples (10 ± 1 mg) were placed in platinum pans and tested under N₂ (flow: 50 mL/min), heating rate 10 °C/min to 1100 °C.

2.6.7. FTIR Spectroscopy

Fourier-transform infrared (FTIR) spectroscopy was conducted at the Advanced Materials Research and Innovation Laboratory of the State University of Santa Cruz, Brazil, using a Shimadzu spectrophotometer (IR Prestige 21, Shimadzu Corporation: Kyoto, Japan) in transmission mode with 32 scans. The samples were analyzed in KBr mode within the spectral range of 4000 to 500 cm⁻¹, using 16 scans and a resolution of 4 cm⁻¹.

3. Results and Discussion

3.1. Setting Times

The recycled gypsum paste exhibited shorter initial and final setting times than the commercial gypsum paste, with reductions of 34.7% and 33.0%, respectively, as shown in Figure 11. Although the recycled gypsum presented a lower specific surface area and coarser particle size than the commercial counterpart (Table 1), a reduction in setting time was nevertheless observed. Similar findings were reported by Santana et al. [24], who investigated the feasibility of producing blocks with recycled gypsum powder (RGP) obtained from plasterboard waste calcined at 150 °C for 24 h. Despite the fact that commercial gypsum powder (CGP) exhibited significantly higher fineness, the recycled material still showed a slightly shorter initial setting time.

This behavior can be attributed to structural and chemical factors introduced during the recycling process. The calcination of waste gypsum generates particles with a higher density of lattice defects and amorphous regions, which dissolve more readily than the well-crystallized phases of commercial gypsum. Moreover, the recycled material is typically free from the chemical retarders commonly added to commercial gypsum to extend workability. The absence of these additives allows the intrinsic reactivity of hemihydrate and residual soluble anhydrite to dominate the hydration process. As a result, despite the lower surface area, the enhanced dissolution kinetics accelerated supersaturation of the pore solution, leading to earlier nucleation of calcium sulfate dihydrate and a shorter setting time.

According to Seffrin et al. [25], the properties of gypsum depend on the hydration and curing conditions, with the morphology of calcium sulfate dihydrate crystals being directly related to the formation conditions and impurities present. This relationship helps to explain the observed differences between the commercial and recycled gypsum pastes.

Concerning the effect of PVA, it was observed that the addition of the emulsion increased the initial and final setting times, as well as the working time (i.e., the interval between the initial and final setting), for all mixtures compared with the pure recycled gypsum paste. This delay can be attributed to the interaction between PVA and gypsum constituents, possibly through the formation of a surface layer or microstructural modifications that hinder the crystallization of calcium sulfate dihydrate (CaSO₄·2H₂O). The extension of the working time indicates that PVA acts as a modulating agent of hydration kinetics, which is advantageous for applications requiring longer malleability and adjustment periods before complete solidification.

The influence of PVA on gypsum setting, however, remains controversial in the literature. Portella et al. [26] investigated the drying kinetics of commercial gypsum over 45 min at constant temperature and reported that PVA extended the setting time by up to 80%, an effect associated with interference in crystallization that slowed nucleation and promoted gradual nucleus formation. In contrast, Zhou et al. [27] evaluated PVA mainly at 0.2 wt.% addition and found a slightly accelerated early hydration of β-hemihydrate, although the overall setting time was not significantly affected. Other works support the retarding influence of PVA: Abbood et al. [28] tested additions of approximately 4 wt.% at water/gypsum ratios of 0.30 and 0.40 and observed only moderate effects, whereas Abo Dhaheer et al. [29] demonstrated that increasing PVA concentrations from 1% to 6 wt.% progressively extended the setting time in both ordinary and high-purity gypsum pastes. Collectively, these findings suggest that the effect of PVA on gypsum hydration and setting is strongly concentration-dependent and may also vary with the type of gypsum, with low dosages producing negligible or slightly accelerating effects, while higher dosages typically result in moderate to pronounced retardation.

3.2. Sorptivity

The water absorption by capillarity curves (Figure 12) revealed distinct differences in the behaviors of the commercial, recycled, and composite gypsum pastes. The pure commercial and recycled gypsum pastes exhibited accelerated absorption in the initial stage; these stabilized at approximately 480 min (170 s^0.5) and were classified as Type B. For these samples, water reached the upper surface of the specimen before the test concluded.

The analyzed composites exhibited low and gradual water absorption throughout the test, and were classified as Type A. Notably, the R50C50P5F7.5 and R50C50P5F10 samples showed the lowest capillary water absorption rates, as presented in

Table 3. Sorptivity (S) of the studied gypsum and composite samples, with the statistical coefficient of variation shown in parenthesis.

References	S [×10−4 kg/(m2·s0.5)]	Composites
		Group 1	S [×10−4 kg/(m2·s0.5)]	Group 2	S [×10−4 kg/(m2·s0.5)]	Group 3	S [×10−4 kg/(m2·s0.5)]
CG	4280 (1.4)	R100P5F1	2210 (10.8)	R70C30P5F1	2110 (7.8)	R50C50P5F1	1730 (9.6)
RG	5440 (1.9)	R100P5F2.5	2180 (9.8)	R70C30P5F2.5	2080 (10.2)	R50C50P5F2.5	1720 (8.7)
		R100P5F5	2180 (8.8)	R70C30P5F5	2070 (11.2)	R50C50P5F5	1700 (8.9)
		R100P5F7.5	2170 (9.8)	R70C30P5F7.5	2060 (10.1)	R50C50P5F7.5	1690 (9.5)
		R100P5F10	2160 (8.8)	R70C30P5F10	2040 (9.3)	R50C50P5F10	1680 (8.0)

.

According to EN ISO 15148:2002 + A1:2016 [15], construction materials are classified by their capillary water absorption coefficient (sorptivity, S) as quick suction when S > 333.0 × 10⁻⁴ kg/(m²·s^0.5), water-resistant when 83.3 < S ≤ 333.0 × 10⁻⁴ kg/(m²·s^0.5), almost impermeable when 0.167 < S ≤ 83.3 × 10⁻⁴ kg/(m²·s^0.5), and impermeable when S ≤ 0.167 × 10⁻⁴ kg/(m²·s^0.5). Based on this classification, all samples in the present study were categorized as quick suction materials. Nevertheless, The R50C50P5F10 sample stood out by exhibiting a 69.1% reduction in its absorption coefficient compared to the RG sample. This improvement is attributed to the alkaline treatment of the coconut fibers, which removed amorphous components such as hemicellulose and lignin, increased surface roughness, and reduced fiber water absorption. These modifications enhanced fiber–gypsum adhesion and minimized voids at the fiber–matrix interface, thereby reducing the connectivity of pores and resulting in a denser microstructure. Additionally, PVA emulsion acted as a sealing agent, decreasing porosity and hindering water penetration. The use of commercial gypsum also contributed to a less permeable matrix due to its higher density and lower contaminant content. These combined factors improved the physical properties of the composite.

Supporting these findings, Munhoz and Renofio [30] observed that adding 40% of a PVA-based additive to gypsum resulted in an 80% reduction in capillary water absorption after 3 h, compared to gypsum without additives. After 6 h of water exposure, this difference decreased to 15%, remaining constant over the following 72 h.

Corroborating these findings, Lima et al. [31] observed that the addition of 1.0% of a commercial water-repellent additive to gypsum resulted in a significant reduction in capillary water absorption, classifying the mixture as water-resistant. Additionally, Rodrigues [32] evaluated the durability of gypsum with the incorporation of a water-repellent agent and found that, even after accelerated aging cycles, there was a considerable reduction in capillary water absorption in the specimens containing the additive compared to the reference specimens. These results reinforce the effectiveness of additives in improving water resistance properties in gypsum-based composites.

These studies demonstrate that the incorporation of water-repellent additives can significantly enhance capillary water absorption properties in gypsum composites, making them more suitable for applications requiring higher moisture resistance.

3.3. Flexural Tensile Strength

Figure 13 shows the three-point flexural tensile strength results for the recycled gypsum (RG) reference and the studied composites. All samples exceeded the minimum requirement of 2.0 MPa defined by the European standard UNE-EN 13279-1 [13]. The incorporation of fibers, proportional to their content, generally increased flexural strength, and this effect was further enhanced by the addition of commercial gypsum (C).

Table 4. Relative flexural tensile strength of modified compositions compared to the reference gypsum (RG). Percentage variation (%Δ mean vs. RG) is calculated from the mean values. Statistical analysis was performed using Welch’s t-test; results include t- and p-values, with significance determined at p < 0.05.

Comparison	%Δ Mean vs. RG	t-Value	p-Value	Significant?
RG vs. R100P5F1	+1.7	0.312	0.7664	No
RG vs. R100P5F2.5	+4.4	0.977	0.3645	No
RG vs. R100P5F5	+6.6	1.786	0.115	No
RG vs. R100P5F7.5	+8.5	1.864	0.1097	No
RG vs. R100P5F10	+10.0	2.361	0.0523	No
RG vs. R70C30P5F1	+9.3	2.228	0.0627	No
RG vs. R70C30P5F2.5	+12.3	3.055	0.0187	Yes
RG vs. R70C30P5F5	+14.5	3.184	0.018	Yes
RG vs. R70C30P5F7.5	+17.7	4.232	0.0043	Yes
RG vs. R70C30P5F10	+20.2	5.252	0.0011	Yes
RG vs. R50C50P5F1	+14.9	4.129	0.0038	Yes
RG vs. R50C50P5F2.5	+17.9	4.27	0.0041	Yes
RG vs. R50C50P5F5	+20.7	3.591	0.014	Yes
RG vs. R50C50P5F7.5	+23.9	4.922	0.0027	Yes
RG vs. R50C50P5F10	+27.7	6.635	0.0004	Yes

presents the statistical analysis with pairwise comparisons against the RG reference. While fiber addition alone only slightly improved the mean flexural tensile strength, statistically significant gains were observed for composites containing 30% commercial gypsum with at least 2.5% fibers, as well as for all mixtures with 50% commercial gypsum regardless of fiber content. Among these, the R50C50P5F10 composite exhibited the best performance, achieving a flexural strength approximately 27.7% higher than RG and 69.5% above the normative minimum.

The literature reports results on the flexural strength of coconut fiber-reinforced gypsum composites, mostly using untreated fibers and commercial gypsum.

Aramwit et al. [33] investigated the use of untreated coconut and rice husk fibers as reinforcements for gypsum composites, focusing on their flexural behavior. The composites were prepared with commercially available gypsum supplied by local vendors and a constant water-to-gypsum ratio of 0.67. The reference gypsum samples exhibited a flexural strength of approximately 2.3 MPa. The incorporation of coconut fibers led to a steady and significant improvement, with strength increasing as fiber content rose, reaching about 5.6 MPa at 30% coconut. This enhancement, corresponding to more than a twofold increase compared to neat gypsum, was attributed to the continuous fibrous structure, high elongation capacity, and lignin-rich composition of coconut, which promoted effective stress transfer and crack-bridging. In contrast, the addition of rice husk resulted in a reduction in flexural performance, with a maximum strength of only 1.5 MPa at 10% rice husk. The authors associated this decline with the shell-like morphology and poor bonding of rice husk within the gypsum matrix, which induced brittle behavior and limited reinforcement efficiency.

Rodríguez-Robalino et al. [34] also developed gypsum composites reinforced with untreated coconut fibers. The composites were prepared using commercial gypsum plaster with a water-to-binder ratio of 0.7, and coconut fibers were incorporated in increasing volumes, ranging from 2.5% up to 17.5% by volume of the binder. The reference gypsum (without fibers) exhibited a flexural strength of approximately 3.61 MPa. With fiber addition, strength progressively improved, reaching an optimum at 15% coconut fiber, where the flexural strength peaked at 4.35 MPa, corresponding to a 20.5% improvement over the control. However, further increasing the fiber content to 17.5% led to a reduction in flexural strength to 3.25 MPa, mainly due to fiber agglomeration and increased porosity that impaired load transfer. SEM analysis confirmed strong fiber–matrix adhesion, with the formation of calcium sulfate dihydrate crystals around the fibers, which anchored the reinforcement and promoted crack-bridging mechanisms.

Guna et al. [35] investigated the reinforcement of gypsum ceiling tiles with untreated coconut (coir) and sheep wool fibers, both individually and in hybrid combinations. The composites were produced with commercial gypsum and a fixed water-to-gypsum ratio of 0.67, while fiber content was varied between 10% and 40% by weight of gypsum. The reference gypsum exhibited a flexural strength of approximately 1.8 MPa. When fibers were added separately, the inclusion of coconut fibers produced a progressive increase in strength, peaking at 3.4 MPa with 30% coir, corresponding to an ~89% improvement over the control. Wool fibers showed a similar trend, though with lower values, reaching 3.0 MPa at 30% addition. However, when fiber dosage was increased to 40% for either reinforcement, flexural strength dropped sharply, to 0.9 MPa for coir and 1.1 MPa for wool, due to fiber agglomeration and insufficient gypsum matrix to ensure effective bonding. These results demonstrated that the optimum fiber content was around 30%, with coconut fibers outperforming wool in terms of flexural strength enhancement.

Among vegetal fibers explored for reinforcing gypsum composites, coconut (coir) fibers have received particular attention due to their high elongation capacity (15–47%), lignin-rich composition, and effective crack-bridging ability. Studies reviewed by Jia et al. [36] indicate that when incorporated at contents up to 30% by weight of gypsum and with typical fiber lengths of 20–40 mm, coconut fibers can nearly double the flexural strength of plain gypsum, in addition to markedly improving ductility and reducing brittleness. However, the review also highlights that excessive fiber addition may lead to agglomeration and porosity, which compromise performance. In comparison, sisal and hemp fibers showed strong reinforcing capacity at lower dosages (1–3% by volume) and with shorter lengths (1–4 cm), with hemp in particular benefiting from biological pretreatments that improved fiber–matrix adhesion and flexural resistance, in some cases surpassing glass fibers. Abaca fibers, typically used at 1–2%, also enhanced flexural strength, though to a lesser degree. Overall, while sisal and hemp are efficient reinforcements at low fiber volumes, coconut fibers stand out for their ability to provide substantial flexural strength gains and toughness when used at higher dosages, making them especially attractive for the development of sustainable gypsum composites.

In this work, the fiber content was limited to 10% by volume to ensure processability and specimen homogeneity. While higher fiber fractions could be accommodated with commercial gypsum—by slightly increasing the water-to-gypsum ratio—such adjustment was impractical for recycled gypsum, which required disproportionately high water contents that impaired molding and strength.

Figure 14 illustrates the fracture of a gypsum matrix containing coconut fibers, demonstrating that even after cracking, complete rupture is delayed, indicating good adhesion between the fibers and the matrix. This behavior is consistent with the findings of Freitas [37] and Jia et al. [36], who highlighted that the pretreatment of green coconut fibers—removing amorphous components such as hemicellulose and lignin—can promote stronger cellulose–cellulose hydrogen bonding, increase surface roughness, and improve adhesion. Such structural modifications make the fibers more resistant, less absorbent, and require greater energy for rupture, thereby enhancing fiber anchorage within the gypsum matrix.

It is worth noting that the increase in fiber content up to 10% did not result in aggregation or fiber clumping. This behavior can be attributed to the dispersing effect of the PVA emulsion, which promotes fiber wetting and coating, ensuring uniform distribution throughout the matrix. Furthermore, the gradual mixing procedure adopted (Section 2.6) prevented the formation of fiber clusters, improving fiber–matrix packing and filling microstructural gaps. These factors explain the stable performance at higher fiber contents.

3.4. Axial Compressive Strength

Figure 15 presents the axial compressive strength results for the recycled gypsum (RG) reference and the studied composites. The Brazilian standard NBR 13207-3 [12] does not establish a minimum requirement for compressive strength; therefore, the American standard ASTM C28/C28M-10 [14], which specifies a minimum compressive strength of 5.20 MPa after 28 days, was adopted as a reference. All samples exceeded both this normative value and the strength observed in the RG sample. The overall behavior was consistent with the flexural strength results, showing that compressive strength tended to increase with fiber content. However, based on statistical analysis with pairwise comparisons (see

Table 5. Relative compressive strength of modified compositions compared to the reference gypsum (RG). Percentage variation (%Δ mean vs. RG) is calculated from the mean values. Statistical analysis was performed using Welch’s t-test; results include t- and p-values, with significance determined at p < 0.05.

Comparison	%Δ Mean vs. RG	t-Value	p-Value	Significant?
RG vs. R100P5F1	+2.2	0.458	0.6602	No
RG vs. R100P5F2.5	+2.7	0.547	0.6002	No
RG vs. R100P5F5	+3.4	0.694	0.5083	No
RG vs. R100P5F7.5	+4.3	0.964	0.3636	No
RG vs. R100P5F10	+5.0	1.192	0.2676	No
RG vs. R70C30P5F1	+4.9	1.071	0.3160	No
RG vs. R70C30P5F2.5	+5.6	1.345	0.2157	No
RG vs. R70C30P5F5	+6.4	1.609	0.1469	No
RG vs. R70C30P5F7.5	+7.4	1.559	0.1591	No
RG vs. R70C30P5F10	+8.3	1.582	0.1566	No
RG vs. R50C50P5F1	+8.4	1.793	0.1118	No
RG vs. R50C50P5F2.5	+9.9	2.297	0.0507	No
RG vs. R50C50P5F5	+12.2	2.743	0.0255	Yes
RG vs. R50C50P5F7.5	+13.4	2.897	0.0205	Yes
RG vs. R50C50P5F10	+15.1	3.315	0.0109	Yes

), significant improvements in compressive strength were observed only for the R50C50 blends starting from 5% fiber addition. The best performance was achieved by the R50C50 composite with 10% fibers, which exhibited an increase of approximately 15.1% compared with RG and 58.3% above the normative minimum.

Similar findings were descried by Rodríguez-Robalino et al. [34]. The authors evaluated the compressive strength of coconut fiber-reinforced (commercial) gypsum composites and reported a progressive increase with fiber incorporation. The highest compressive strength, 8.77 MPa, was obtained for the composite with 12.5% fiber content, representing an increase of 24.2% compared with the reference, which exhibited a compressive strength of 7.06 MPa. However, when the fiber dosage exceeded the optimum level, compressive strength decreased; for instance, the specimen with 17.5% fiber content reached 5.58 MPa, corresponding to an 18.3% reduction relative to the reference. This reduction was attributed to increased porosity and the non-homogeneous distribution of fibers at higher dosages, which compromised the load-bearing capacity of the composites. While the optimum fiber content for compressive strength was identified as 15%, the best flexural strength performance was achieved at 12.5%, as previously described in the preceding section.

While coconut fiber contributed to a more uniform stress distribution, reducing localized failures and enhancing compressive performance, the addition of PVA emulsion reduced the internal porosity of the matrix and improved compaction. This effect can be observed in Figure 16, which presents cross-sections of recycled gypsum; recycled gypsum with 5% PVA; 70% recycled gypsum + 30% commercial gypsum + 5% PVA; and 50% recycled gypsum + 50% commercial gypsum + 5% PVA. The recycled gypsum sample displays a matrix with higher porosity, whereas the incorporation of PVA and commercial gypsum results in a visibly denser structure.

These findings are consistent with those reported by Gonçalves [34], who observed greater density, improved compaction, and a more homogeneous microstructure in samples containing commercial gypsum and PVA emulsion, characterized by a smoother surface texture and a more refined esthetic finish.

3.5. Surface Hardness

Surface hardness is an important property for gypsum-based components, as it directly influences their resistance to wear, scratching, and localized damage during service. It also reflects the microstructural quality of the material, since higher hardness values are typically associated with reduced porosity and improved matrix compaction. Thus, the incorporation of PVA emulsion can contribute to increased surface hardness by enhancing particle bonding and promoting matrix densification, as previously observed in Figure 15. Figure 17 presents the surface hardness results of the analyzed samples, while

Table 6. Relative surface hardness of composites compared to the reference gypsum (RG). Percentage variation (%Δ mean vs. RG) is calculated from the mean values. Statistical analysis was performed using Welch’s t-test; results include t- and p-values, with significance determined at p < 0.05.

Comparison	%Δ Mean vs. RG	t-Value	p-Value	Significant?
RG vs. R100P5F1	+0.3	0.016	0.9880	No
RG vs. R100P5F2.5	+0.3	0.017	0.9865	No
RG vs. R100P5F5	+0.8	0.042	0.9676	No
RG vs. R100P5F7.5	+1.2	0.058	0.9548	No
RG vs. R100P5F10	+1.3	0.068	0.9478	No
RG vs. R70C30P5F1	+2.0	0.099	0.9236	No
RG vs. R70C30P5F2.5	+2.3	0.115	0.9112	No
RG vs. R70C30P5F5	+3.0	0.144	0.8889	No
RG vs. R70C30P5F7.5	+3.1	0.156	0.8800	No
RG vs. R70C30P5F10	+3.5	0.173	0.8671	No
RG vs. R50C50P5F1	+4.0	0.194	0.8511	No
RG vs. R50C50P5F2.5	+4.5	0.221	0.8304	No
RG vs. R50C50P5F5	+5.0	0.246	0.8121	No
RG vs. R50C50P5F7.5	+5.7	0.276	0.7896	No
RG vs. R50C50P5F10	+5.9	0.293	0.7769	No

summarizes the statistical analysis. Although a slight tendency toward increased hardness with higher fiber content can be noted, the variations are small and fall within the experimental variability. In fact, the statistical analysis presented confirms the absence of significant differences between the samples. According to ABNT NBR 16494:2017 [16], gypsum blocks are classified as high hardness (D ≥ 80, Shore C), medium hardness (55 ≤ D < 80, Shore C), and low hardness (40 ≤ D < 55, Shore C). The differences observed among the studied mixtures were relatively small, with all mean values falling within the medium hardness range (55–80, Shore C) established by the standard. More importantly, all formulations exceeded the minimum normative requirement, confirming that the composites meet the classification criteria for gypsum blocks in construction applications. Similar findings were reported by Álvarez et al. [38], who investigated the performance of gypsum composites incorporating 1% by volume of glass, basalt, polypropylene, and wood fibers in a lightened gypsum-based matrix. The reinforced composites exhibited surface hardness values ranging from 60 to 70 Shore C, which are classified within the moderate hardness range (55–80 Shore C). Although a slight tendency toward increased hardness was observed with fiber addition, statistical analysis confirmed that the variations were not significant. Nevertheless, all formulations surpassed the normative minimum, indicating that the incorporation of fibers effectively preserved the required hardness classification and ensured the applicability of the lightened gypsum composites in prefabricated blocks and other building elements.

3.6. Phase Formation

The XRD patterns of the PVA emulsion and coconut fibers are shown in Figure 18 and Figure 19, respectively. The PVA emulsion and coconut fibers are both semicrystalline materials.

Table 7. Peak intensity and crystallinity index values (in percent) for PVA emulsion and coconut fibers.

Material	I1 1	I2 2	Ic (%) 3
PVA emulsion	759.6	1393.90	45.49
Coconut fibers untreated	3636.40	5454.55	33.3
Coconut fibers pretreated with 5% NaOH (w/w)	4090.91	6363.64	35.71

¹ I₁: Intensity of diffraction peaks related to the crystalline portion; ² I₂: intensity of diffraction peaks related to the amorphous portion; ³ I_c: crystallinity index as a percentage.

lists the peaks corresponding to the crystalline and amorphous fractions, as well as the crystallinity indices, which were calculated according to Equation (3).

The diffractogram of the PVA emulsion, presented in Figure 18, exhibits characteristics of material amorphization, as evidenced by the absence of well-defined diffraction peaks and the presence of broad bands around 22.5° (2θ), indicating partial organization within the polymer chains, a feature typical of a semicrystalline arrangement. Meanwhile, the peak around 42° reflects a higher degree of structural disorder in the amorphous regions. These findings are corroborated by the crystallinity index, which indicates that 45.49% of the structure is crystalline or semicrystalline, as shown in Table 7. The absence of crystalline peaks suggests that the PVA emulsion does not introduce a significant crystalline phase into the composite and may act as a binding agent between the other components.

The diffractograms of the untreated and treated fibers (Figure 19) exhibited diffraction peaks at similar 2θ positions, located at 16° (plane 101) and 22° (plane 002), which are characteristic of crystalline cellulose present in lignocellulosic fibers. According to Araújo et al. [39], the diffraction peak at 22° is associated with the glycosidic rings of type I cellulose. The higher intensity of this peak in the treated fiber, related to the crystallographic plane (002), indicates an increase in structural density and, consequently, in fiber crystallinity. This increase in crystallinity may enhance the adhesion between the fiber and the matrix, as well as contribute to improvements in the composite’s mechanical strength, a result also evidenced by Pinheiro et al. [40].

Although fibers with a high cellulose content generally exhibit two diffraction peaks near 2θ = 16°, only one was observed in Figure 19. This may be due to the presence of amorphous material that obscured one of the peaks, as noted by Pereira et al. [40]. Alkaline pretreatment increased the degree of crystallinity by 6.75% compared with that of the untreated fibers. This is because pretreatment removes surface substances and allows better packing of the cellulose microfibrils.

The commercial gypsum sample mainly comprised gypsum hemihydrate, with peaks of bassanite (CaSO₄·0.5H₂O), anhydrite (CaSO₄), and unconverted dihydrate (CaSO₄·2H₂O) (Figure 20). In addition to these peaks, the XRD pattern of recycled gypsum also showed low-intensity peaks associated with cellulose, perlite, and alumina. The presence of cellulose is attributed to the paper from the drywall sheets, whereas the other phases are related to the additives used in the production of plasterboard. Compared with the commercial sample, the recycled gypsum thus exhibited a more heterogeneous composition, characterized by residual dihydrate and additional impurity phases not present in the commercial material.

The crystallite sizes and lamellar spacings in the (220) plane, presented in

Table 8. Crystallite size (t) and lamellar spacing between lamellae in the (220) plane of the studied samples.

Sample	t (nm)	Spacing (Å)
Hemihydrate gypsum
Commercial gypsum (CG)	54	4.3288
Recycled gypsum (RG)	43	4.3462
Dihydrate gypsum
RG + CG + PVA emulsion	70	2.3732
RG + CG + treated coconut fiber	37	2.3696
RG + CG + PVA emulsion + treated coconut fiber	52	2.3989

, show that recycled gypsum had smaller crystallites, possibly due to internal stresses and structural defects introduced during the recycling process. The larger lamellar spacing of recycled gypsum suggests distortions in the crystal lattice caused by contaminants, which may affect properties such as strength and reactivity [40].

The XRD patterns of the composite samples (Figure 21) revealed characteristic peaks at 2θ angles of 11.8° and 20.8°, which are associated with the (220) and (112) planes of dihydrate gypsum, respectively. The crystallite size was larger in the sample with the PVA emulsion, indicating well-defined crystals owing to the nucleation effect promoted by the additive. By contrast, the addition of coconut fibers reduced the crystallite size, suggesting that the alkaline-pretreated coconut fibers inhibited nucleation or generated lattice stress. The combination of PVA and coconut fibers resulted in a balanced effect, with PVA promoting growth and the fibers mitigating this process.

The lamellar spacings in the (220) plane, related to density and mechanical strength, were similar for each composite, with a slight expansion observed for the sample comprising gypsum, PVA emulsion, and coconut fibers, suggesting subtle structural changes due to the interaction of the additives. The more intense diffraction peaks in the samples with additives, compared with those of the standard dihydrate, indicated favorable gypsum crystal growth, particularly due to the stabilizing interaction of the PVA emulsion (5 wt%), which promoted crystal formation.

3.7. Results of Thermogravimetric Analysis

Figure 22 presents the TGA/DTG and DrTGA curves of the composites. The thermal behavior is dominated by a single major mass-loss event (Region A), while subsequent events (Regions B and C) contribute only marginally to the overall degradation.

✓

Region A (≈ 80–180 °C): This stage accounts for the principal mass loss of the composites (~18–20%). It corresponds to the dehydration of gypsum dihydrate (CaSO₄·2H₂O) into hemihydrate (CaSO₄·½H₂O), with a theoretical mass reduction of ~15.7%. The slightly higher experimental value arises from overlapping processes: (i) removal of physically adsorbed moisture and emulsion water, and (ii) the early onset of the hemihydrate → anhydrite transition (CaSO₄·½H₂O → CaSO₄). The combination of these mechanisms under dry purge conditions explains why the entire ~20% water release appears concentrated in this first event.

✓

Region B (≈ 220–400 °C): Only a very small additional weight loss is detected in this interval. This may be attributed to the degradation of organic components—PVA chain scission or hemicellulose/cellulose decomposition in coconut fibers—as well as residual contributions from hemihydrate conversion to anhydrite. However, the magnitude of the loss is negligible compared with Region A.

✓

Region C (≈ 500–750 °C): A weak and broad signal is observed at high temperatures, especially in fiber-containing mixtures. This can be assigned to the slow decomposition of lignin and oxidation of residual char. Nevertheless, the associated mass loss remains minimal and does not significantly affect the total residue.

At 800 °C, all systems preserve a stable residue corresponding to anhydrite (CaSO₄) and the inorganic fraction of the fibers. Thus, the thermal profile is governed almost entirely by the dehydration of gypsum, with organic additives contributing only minor secondary effects.

3.8. Results of FTIR Spectroscopy

Figure 23 presents the FTIR spectra of the composites, highlighting information on functional groups and chemical bonds. Similar absorption bands were identified in all samples. In the range of 3700 to 3300 cm⁻¹, O–H bonds were observed, indicating the presence of chemically bound water (hydrated gypsum) or hydroxyl groups from PVA and coconut fiber.

The range of 1700 to 1400 cm⁻¹ was associated with carbonate (CO₃²⁻) vibrations, possibly related to new functional groups exposed by the alkaline treatment of the coconut fiber, which increases its chemical interaction with the gypsum matrix. In the range of 1200 to 1000 cm⁻¹, strong bands indicated the integrity of sulfates in the gypsum and the presence of methoxyl groups in the fiber, while in the range of 800–500 cm⁻¹, vibrations confirmed the structural interaction between SO₄²⁻, the fiber, and PVA.

The treated fiber promoted hydrogen bonding with calcium sulfate dihydrate (CaSO₄·2H₂O), densifying the matrix and reducing porosity, which provided greater thermal and mechanical stability. Meanwhile, the PVA emulsion, as a hydrophilic polymer, physically interacted with gypsum mainly through adsorption and hydrogen bonding of its hydroxyl groups with the dihydrate crystals and retained water. These physicochemical interactions promoted preferential precipitation of gypsum on the PVA-coated surfaces, leading to a compact interfacial transition zone (ITZ) and a denser microstructure. This interpretation is consistent with Zhu et al. [41], who observed that PVA fibers in gypsum composites accelerated hydration and produced a remarkably compact ITZ compared with polypropylene fibers.

These interactions, reflected in the bands from 1200 to 1000 cm⁻¹, contributed to a more uniform and cohesive matrix, improving properties such as mechanical strength and water absorption. Similar results were observed by Brandão [6], who analyzed wood–plastic composites with coconut fiber and gypsum waste molded by injection. The FTIR spectra indicated that the presence of PVA significantly interferes with band intensities, highlighting the bands belonging to PVA. Furthermore, XRD results suggested that chitosan and PVA chains were inserted into the spatial layers of clays, forming exfoliated and intercalated structures.

Similarly, Ferreira [42] developed and evaluated nanocomposites of poly (ethylene-co-vinyl acetate) (EVA) and clay minerals as polymeric additives to improve the flow properties of paraffinic and petroleum systems. The nanocomposites were able to reduce the pour point, whereas pure EVA was not effective, highlighting the importance of interactions between the composite components in enhancing the final properties of the material.

4. Conclusions

This study evaluated gypsum composites prepared with commercial and recycled gypsum, polyvinyl acetate (PVA) emulsion, and alkaline-treated green coconut fibers. The investigation combined XRD, TGA, FTIR, and mechanical testing to understand how these constituents influence hydration, crystallization, and performance. The main conclusions are as follows:

✓

Role of PVA and fibers: Alkaline treatment of coconut fibers improved fiber–matrix interaction by removing hemicellulose and lignin, exposing hydroxyl groups that favor adhesion and reduce premature degradation. The PVA emulsion physically interacted with gypsum crystals, mainly by coating their surfaces, forming hydrogen bonds, and adsorbing water. This delayed rehydration, refined the microstructure, and reduced pore connectivity.

✓

Microstructural modifications: XRD and TGA analyses confirmed that PVA influenced gypsum crystallization by promoting nucleation and modifying crystallite size and lamellar spacing, while fibers provided reinforcement through interfacial bonding. The combination of both additives produced intermediate properties that optimized performance compared with single-additive systems.

✓

Spectroscopic evidence: FTIR spectra revealed changes in absorption bands associated with gypsum hydrates, sulfate groups, and organic functional groups from PVA and fibers. These shifts confirmed that the incorporation of PVA and treated fibers alters the structural organization of the composites through physical interactions.

✓

Best-performing hybrid composite: The mixture containing 50% commercial gypsum, 50% recycled gypsum, 5% PVA emulsion, and 10% treated fibers achieved the most favorable balance of properties, with notable increases in flexural tensile strength (+27.92%), compressive strength (+15.10%), and surface hardness (+5.76%), as well as a 69.1% reduction in sorptivity. The superior behavior of this blend is attributed to three factors: (i) the higher reactivity and purity of commercial gypsum compensating for impurities in recycled gypsum; (ii) the PVA coating effect, which reduced pore size and improved crystal interlocking; and (iii) the reinforcing role of treated fibers, which bridged microcracks and enhanced the interfacial transition zone. Together, these mechanisms explain the synergistic improvement in both strength and durability.

Future work can evaluate wet–dry and humidity-cycle durability, dimensional stability/creep, screw-pull-out and impact resistance, fire response in the presence of fibers/PVA, and recyclability/LCA of the hybrid composites.