Cold-Pressed Insulation Boards from Recycled Cotton Fibers Using a Water-Borne PVAc-Starch Binder: Processing, Structure and Properties

1. Introduction

The built environment is responsible for approximately 37% of global annual CO₂ emissions, with more than half of this originating from heat losses due to inadequately insulated building envelopes [1].

From a materials design perspective, new sustainable materials should combine low thermal conductivity, mechanical integrity, and low-toxicity chemistry. In 2025, European Union legislation mandates the separate collection and recycling of textile waste under Directive 2008/98/EC, as amended by Directive 2018/851/EC. This regulatory framework demands the development of new functional materials for the cascading use of textile fibers into sustainable products with added value, for example new thermal insulation materials. Recycled lignocellulosic or textile fibers contribute to the circular economy by diverting post-consumer waste from landfills, while requiring minimal additional energy for processing [2].

Among such resources, post-consumer cotton exhibits an intrinsic solid-phase thermal conductivity of approximately 0.04 W·m⁻¹·K⁻¹ [3]. The primary challenge lies in transforming these hydrophilic, crimped fibers into cohesive, self-supporting boards without relying on formaldehyde-based or highly cross-linked synthetic resins.

Recent epoxy/flax-hemp laminates filled with waste glass dust achieved an effective thermal conductivity λ of approximately 0.30 W·m⁻¹·K⁻¹, but required a virgin epoxy matrix and curing at 120 °C [4]. Cellulose nanofiber aerogels crosslinked with melamine-urea-formaldehyde exhibited a λ of less than 0.025 W·m⁻¹·K⁻¹, though they were brittle and released free formaldehyde [5]. Cotton filter waste mats, hot-pressed with thermoplastic corn starch, reduced λ to 0.046 W·m⁻¹·K⁻¹, but their compressive strength remained under 20 kPa [6]. Epoxy syntactic foams containing hollow glass microspheres demonstrated hydrostatic pressure resistance and a λ of approximately 0.10 W·m⁻¹·K⁻¹, although at densities exceeding 500 kg·m⁻³ [7]. Hollow glass bead-filled carbon fiber/PEEK composites reduced λ by 45%, but their high cost and embodied energy limit their adoption in construction [8]. Nanocellulose-modified polyurethane foams exhibit a refined microstructure and a lower λ, yet they still rely on petrochemical-based materials [9]. A recent review highlights the rapid progress in natural fiber composite boards and bio-based adhesives, while also emphasizing the scarcity of panels that are both structurally sound and have low conductivity [10].

At the same time, regulatory and market signals are driving materials toward carbon transparency and circularity. The European Union's 2024 recast of the Energy Performance of Buildings Directive (EPBD) introduces phased requirements for whole-life carbon assessment and disclosure for new buildings, furthering the demand for low-impact insulation products [11]. Comparative life-cycle assessments (LCAs) indicate that plant-fiber and recycled-fiber materials can achieve lower cradle-to-grave impacts than mineral wool or petrochemical foams, particularly when functional performance is normalized and both biogenic carbon storage and end-of-life are properly accounted for [12]. These findings reinforce the rationale for utilizing recycled textile feedstocks, which not only diverts waste but also reduces processing energy.

From a materials and structural perspective, heat transfer in porous fiber-binder boards is governed by the solid fiber/polymer skeleton, gaseous conduction across the pore spectrum, and a radiative term that increases with cell size [13,14,15,16]. Thus, binder chemistry, volume fraction, and spatial distribution co-tune λ through interfacial contact resistance and the pore-size distribution established during consolidation [17,18,19]. For cellulosic (cotton) fibers, water-based PVAc offers low-toxicity processing and room-temperature film formation. However, moisture absorption can plasticize PVAc and weaken fiber-matrix load transfer on hydroxylated surfaces unless carefully controlled in the formulation [20,21,22]. In contrast, starch binders can be mildly cross-linked to enhance cohesive strength and reduce water uptake [23,24]. PVAc-starch hybrids thus improve wet resistance and interfacial adhesion while maintaining a water-based processing window [25,26].

In this work, a PVAc-starch hybrid binder is employed to enhance mechanical properties, maintain a fine pore structure that limits gaseous and radiative heat transfer, and reduce moisture-induced softening through mild cross-linking, enabling the design of bio-based fiber composites based on recycled cotton fibers for building insulation applications.

2. Materials and Methods

2.1. Recycled Cotton Fibers

The basic material used in this study was recycled cotton fibers EN-TEX (Enroll CZ, Nová Ves, Czech Republic), originating from mechanically processed post-consumer and post-industrial textile waste. Shredding and fiberizing refer to a purely mechanical size-reduction procedure in which textile residues are progressively torn, cut, and disintegrated into individual fibers, without inducing any chemical modification or thermal exposure. To establish the baseline characteristics of the recycled cotton feedstock, the fibers were analyzed in terms of their dimensions, morphology, and chemical composition. The size distribution analysis revealed a broad variability in both diameter and length. Fiber diameters mostly ranged between 8 and 15 µm, while fiber lengths were typically 400–900 µm. This heterogeneity reflects the random breakage of fibers during recycling.

Figure 1. Histograms of fiber diameter (a) and fiber length (b) for recycled cotton feedstock.

Figure 1. Histograms of fiber diameter (a) and fiber length (b) for recycled cotton feedstock.

In addition to the recycled fibers, the material contains integrated flame retardants based on aluminium hydroxide (Al(OH)₃) (ATH) in weight ration 2–3 % weight, applied in the form of a fine powder to the surface of recycled cotton fibers, ensuring classification in reaction-to-fire class E according to the European Technical Assessment ETA-19/0457 [27]. Further additives include antifungal agents compounds Sodium Bicarbonate (NaHCO3) in weight ration 1–2 % weight, and Zinc Pyrithione (C10H8N2O2S2Zn) in weight ration 1–2 % weight, that resist pests and rodents, which are commonly found in commercial cellulose-fiber-based insulation products. Both applied in the form of a fine powder to the surface of recycled cotton fibers; The fibers are supplied without synthetic binders or formaldehyde.

2.2. Binder Systems

One of the adhesives used in the composite production was a polyvinyl acetate PVAc-based water dispersion PROFIBOND (Profil Print Technology s.r.o., Všemyslice, Czech Republic). It is a one-component adhesive supplied as a white liquid dispersion with a density of 1.08 ± 0.01 g·cm^-³, a solids content of 50 ± 1%, a viscosity of 9500 ± 1500 mPa·s, and a pH of 4.75 ± 0.25. The adhesive provides water resistance corresponding to class D3 and does not contain formaldehyde.

Another adhesive used in the production of the composite boards was corn starch (Herold, Rakovník, Czech Republic), supplied as a fine white powder. The starch used in this study is a polysaccharide composed primarily of amylopectin (≈approximately 72–75%) and amylose (≈approximately 25–28%). Its granules are typically 5–25 µm in diameter, with a bulk density of 0.5–0.6 g·cm^-³, and a moisture content of less than 13%. The gelatinization temperature is approximately 62–72 °C, at which point the starch swells in water and forms a viscous gel.

2.3. Microstructural Characterization

Recycled cotton fibers were characterized to establish their baseline morphology and chemistry prior to board fabrication. Their heterogeneous origin from post-consumer and post-industrial waste necessitated an evaluation of potential synthetic residues and inorganic additives, which may impact binder compatibility and composite performance. Morphology was examined using a scanning electron microscope MIRA3 (Tescan Orsay Holding, Brno, Czech Republic) equipped with an energy-dispersive spectroscopy Xplore 15 (Oxford Instruments plc, High Wycombe, Great Britain) system. Images were acquired at 200×, 1000×, 2000×, and 10,000× magnification and analyzed in ImageJ (ImageJ 1.53f, National Institutes of Health, Bethesda, MD, USA), with a total of 250 fibers measured after scale calibration. Elemental composition was determined using the ZAF quantification method. The chemical structure was assessed using FTIR (Nicolet iS20, Waltham, MA, USA; spectra collected in the range of 4000–600 cm⁻¹) to identify cellulose functional groups and possible chemical modifications, and Raman spectroscopy WITec alpha 300 (WITec Wissenschaftliche Instrumente und Technologie GmbH, Ulm, Germany) to evaluate cellulose crystallinity and detect synthetic residues.

2.4. Preparation of Thermal Insulation Boards

The boards were prepared with target densities of 300-340 kg·m-³, the range was set due to a balance between thermal resistance and mechanical integrity in fiber-based materials [28,29,30]. The density of the material depends on the weight fraction of the components cotton fibers, PVAc and cornstarch. Different weight ratios of the three components affect the properties of the composite. After weighing the required proportions of components, the given ratio of PVAc and cornstarch was applied while constantly stirring with a mechanical stirrer LGB (Imalpal S.r.l., San Damaso, Italy) at the constant temperature (20 ± 2 °C) and humidity (65 ± 5 %) environment. Test specimens with dimensions of 500 (L) × 300 (W) × 80 (T) mm were produced from the homogenized fiber-binder mixture, placed into molds, and pressed in a universal testing machine TT 2850 (TIRA, GmbH, Schalkau, Germany). The molds were lined with standardized support plates to ensure uniform pressure transfer during the process, and the same support configuration was used for all specimens, regardless of their target density or the specific PVAc-to-starch binder ratio applied in each formulation. The pressing of the mixture of gelatinized starch (65-70 °C) and PVAc adhesive occurred in a cold press, with the entire process completed in a single 2-minute cycle. The applied force varied with target density, ranging from 7.5 kN for the lowest-density boards (300 kg·m⁻³) to 25 kN for the highest-density boards (340 kg·m⁻³). The specimens were kept under load for approximately 1 hour, then technologically dried in drying chamber Binder FDL (Binder GmbH, Tuttlingen, Germany) and then air-conditioned in a climate chamber Memmert HPP 750 (Memmert GmbH, Schwabach, Germany) at a temperature of 20 ± 2 °C and a relative humidity of 65 ± 5 % for 14 days to reach an equilibrium moisture content.

The complete formulations and pressing parameters are summarized in Table 1.

2.5. Bulk Density Determination of Insulation Boards

The bulk density of the insulation boards was determined after conditioning at 20 ± 2 °C and 65 ± 5% relative humidity. Density was calculated as the ratio of the specimen mass to its geometrical volume. The mass of each specimen was measured using a laboratory balance, while the length, width, and thickness were determined using a digital caliper.

The bulk density $\rho$ (kg·m⁻³) was calculated according to Equation (1):

$$\rho ={\displaystyle \frac{m}{V}}$$

(1)

where m is the mass of the conditioned specimen (kg) and V is the specimen volume (m³), calculated from the measured length, width, and thickness.

2.6. Thermal Parameters Characterization

Thermal conductivity $\lambda$ and the thermal diffusivity $a$ were measured using a thermal conductivity device ISOMET 2114 (Applied Precision Ltd., Bratislava, Slovakia) equipped with a needle probe sensor, in accordance with EN ISO 8301 [31]. Measurements were carried out on conditioned specimens under laboratory conditions (20 ± 2 °C, 65 ± 5% relative humidity). Based on the measured values, the volumetric heat capacity $\rho ·c$ and thermal diffusivity $a$ were calculated according to Equation (2). For each density level, three independent board specimens were prepared, and thermal conductivity was measured at three different points on each specimen, resulting in nine measurements per variant. The reported values are presented as mean ± standard deviation. The thermal parameters were calculated using the following relationship:

$$a={\displaystyle \frac{\lambda }{\rho ·c}}$$

(2)

where $\lambda$ is the thermal conductivity (W·m⁻¹·K⁻¹), $c$is the specific heat capacity (J·kg⁻¹·K⁻¹), $\rho$ is the bulk density (kg·m⁻³), $\rho ·c$ is the volumetric heat capacity (J·m⁻³·K⁻¹), and $a$ is the thermal diffusivity (m²·s⁻¹).

2.7. Moisture-Related Behavior Characterization

Moisture-related behavior of the boards was evaluated in accordance with EN ISO 29767 [32]. Specimens with dimensions of 200 mm (L) × 200 mm (W) × 80 mm (T) were prepared from conditioned boards. Water absorption and thickness swelling were determined by specimens weighed in the conditioned state and reweighed after 24 hours of immersion in tap water at a temperature of 20 ± 2 °C. After immersion, excess surface water was removed, and dimensional changes were recorded simultaneously. Measurements were performed under laboratory conditions (20 ± 2 °C, 65 ± 5% relative air humidity). For each density level, five specimens were tested, and the results are reported as mean ± standard deviation. Water absorption (WA) and thickness swelling (TS) were calculated using the following relations:

$$WA\left(\mathrm{\%}\right)={\displaystyle \frac{m_2-m_1}{m_1}}\times 100$$

(3)

$$TS\left(\mathrm{\%}\right)={\displaystyle \frac{t_2-t_1}{t_1}}\times 100$$

(4)

where $m_1$​ is the specimen mass after conditioning (g), $m_2$ is the specimen mass after immersion (g), $t_1$​ is the initial specimen thickness (mm), and $t_2$​ is the thickness after immersion (mm).

2.8. Mechanical Properties Characterization

Compressive behavior of the thermal insulation boards was determined in accordance with EN ISO 29469 [33] using a universal testing machine TT 2850 (TIRA GmbH, Schalkau, Germany) (Figure 2a). Stress-strain curves were recorded up to 10% relative deformation of the specimen thickness. The load was applied continuously at a crosshead speed corresponding to a strain rate of 10 % min⁻¹ (≈8 mm·min⁻¹ for 80 mm thick samples). For each density level, ten conditioned specimens with dimensions of 50 mm (L) × 50 mm (W) × 80 mm (T) were tested, and the results are reported as mean ± standard deviation. Compressive strength (σ₁₀) was calculated as follows:

$${\sigma }_{10}={\displaystyle \frac{F}{A}}$$

(5)

where ${\sigma }_{10}$ is the compressive stress at 10 % deformation (Pa), $F$ is the applied load at 10% deformation (N), and $A$ is the loaded cross-sectional area of the specimen (m²).

For the assessment of board cohesion and resistance to tensile loading perpendicular to the board plane, internal bond strength was determined in accordance with EN ISO 29766 [34] using a universal testing machine TT 2850 (TIRA GmbH, Schalkau, Germany) (Figure 2b). Specimens with dimensions of 50 mm (L) × 50 mm (W) × 80 mm (T) were prepared in the conditioned state. The load was applied continuously at a rate of 2 mm·min^-1 until failure. For each density level, ten specimens were tested, and the results are reported as mean ± standard deviation. Internal bond strength was calculated as follows:

$${\sigma }_{ib}={\displaystyle \frac{F}{A}}$$

(6)

where ${\sigma }_{ib}$ is the internal bond strength (Pa), $F$ is the maximum load at failure (N), and $A$ is the loaded cross-sectional area of the specimen (m²).

The experimental data obtained in this study were processed and evaluated using a combination of graphical and tabular methods. Statistical analyses were performed using STATISTICA (TIBCO Software Inc., Palo Alto, CA, USA), while selected graphs were prepared in Microsoft Excel (Microsoft, Redmond, USA). Basic descriptive statistics were applied to summarize the measured thermal, mechanical, and moisture-related properties of the insulation boards.