Thermal Performance of Silica-Coated Wood Particles

Abstract

Wood is one of the most widely used sustainable lignocellulosic materials, with numerous applications in consumer goods and the construction sector. Despite its positive properties, such as a high strength-to-weight ratio, thermal insulation, and low density, wood’s natural thermal degradation can limit its potential applications. In composite applications like wood–plastic composites, the particle morphology and surface topography must be preserved to support intimate polymer–wood contact and mechanical interlocking. This study investigated the efficacy of a thin silica coating for thermal protection, which was applied via an in situ sol–gel method using the precursor tetraethoxysilane (TEOS). The wood particles and treatments were characterized using particle size analysis, physisorption, FTIR, SEM, XRD, and TGA analyses. After treatment, the specific and microporous surface area of wood particles increased by 118% and 97%, respectively, an effect of the porosity of silica itself. FTIR spectra of the silica-treated wood displayed peaks corresponding to Si stretching, and SEM micrographs confirmed a successful silica coating formation. TGA showed that the silica coating increased the temperatures needed to degrade the underlying hemicellulose and cellulose by 16 °C for all treatment levels. This particle-scale coating provided a promising method for producing thermally protected, functionalizable wood fillers for composites that maintain the filler geometry and potential mechanical interlocking, offering an attractive upcycling pathway for wood residues.

1. Introduction

The importance of forests is more than ever in the spotlight, as they play an immense role in critical and complex ecological systems that support the earth as well as humans. Within the European Union, this importance has been recognized by those in leadership placing the support of forest products (i.e., wood) and responsible forestry at the forefront of both the European Bauhaus [1] initiative and the European Green Deal [2]. To support the ambitious goals of these initiatives, manufacturing and building using locally available, natural materials is a great place to start. Wood is one such material that has been used in many applications to support human life, from simple combustion for heat to multi-storey, tall timber constructions. Wood is abundant in nature, renewable, aesthetically pleasing, non-toxic, carbon-neutral, and relatively low-cost [3]. Moreover, it is ideal for structural uses, with an excellent strength-to-weight ratio, thermal insulation, and low-impact processing [4]. These advantages have made wood and wood-based products popular in construction, as well as in other consumer products (e.g., paper, packaging, etc.). Engineered wood products have greatly increased the number of applications that wood products can be used for. By using engineered products and composites, we can utilize smaller-diameter trees, use lower-quality wood, and even use wood residuals from timber processing industries. These residuals are commonly used in low-value applications as fuel to generate energy, livestock bedding, landscaping materials, etc. On the other hand, there are valuable properties possessed by these residuals, such as their light weight, insulation, porosity, easy accessibility, and low cost, making them a potential raw material resource to utilize and obtain value-added products [5]. However, wood’s natural thermal degradation can limit its potential applications [6].

Different methods for reducing thermal degradation can be used to valorize these wood residues in more applications with heat restrictions. Thermal protection in wood is often focused on reducing the spread of flames on the wood surface, delaying the ignition time, and reducing the rate of heat release during combustion [7]. One strategy that these protective agents use is an insulating barrier, delaying the evaporation of pyrolysis products at high temperatures [5,6,7,8,9,10,11,12,13,14]. Silicates have been used to enhance mechanical properties of wood [15], improve its durability [16], and as coating compounds that are known to significantly improve the thermal performance of wood materials [17]. These compounds are commonly applied as coatings or directly mixed in composites rather than using impregnation processes, as they pose difficulties in impregnation and retention. When silicate compounds are applied as coatings, they act as a barrier on the sample surface, forming an insulating layer (ceramification), forming a char layer in the condensed phase, and trapping active radicals in the vapour phase [6,9,10]. This insulating layer is an interesting option for wood composites, which can be limited by thermal limits and can be damaged during manufacturing. Among existing procedures, the sol–gel technique is a relatively new process for imparting thermal stability to wood materials and is associated with a minimized environmental impact [18,19]. The sol–gel process utilizes the use of semi-metal alkoxides (such as tetramethoxysilane, tetraethoxysilane, titanium tetraisopropoxide) in hydrolysis and sequential condensation reactions to obtain a completely inorganic or hybrid organic–inorganic 3D network, depending on its structure [20,21]. Hribernik et al. [22] coated cellulose fibres with silica particles using a sol–gel method to improve the thermal performance of cellulose fibres. In their study, they used different percentages of the precursor tetraethoxysilane (TEOS). In the first step, a colloidal suspension of particles took place in a liquid medium. The particles then react with each other to form a 3D polymeric chain that turns into a cross-linked gel [23]. When the morphological and thermal properties of the treated fibres were examined, it was reported that silica particles formed a 300–400 nm thick surface layer on the surface of cellulose fibres and improved the thermal stability. It was also reported that the morphology of the silica layer, rather than the silica content, played the main role in imparting better thermal performance and that the sample group with 0.2% silica added exhibited the most effective thermal protective properties.

As hemicelluloses in wood can start to degrade at around 180 °C, processing temperatures of thermoplastic composites and wood-based panels can easily exceed this and can aesthetically damage products, harm their mechanical performance, and/or increase the need for processing aids and additives. Building from the success found in the Hribernik et al.’s [22] study, the goal of this study was to apply the sol–gel method for in situ silica application and analysis of the wood particles that can ultimately be scaled up for composite production. As much work using sol–gel silica focuses on continuous wood surfaces or fibrous materials, wood particles used in composite materials offer a unique challenge. As wood particles are heavily damaged and have rough surfaces compared to cellulose fibres, this study brings new insight into the efficacy of using this process with wood residuals and brings new application pathways for composites and upcycling of low-value materials.

An important bonding mechanism of particles within some particulate composites is that of mechanical interlocking. The particle morphology and surface topography must be preserved to support this bonding mechanism, which is why this study has looked at a thin coating that preserves the original particle morphology but that can simultaneously provide thermal protection and maintain good mechanical interlocking with a matrix. This is where this application departs from studies using silica for coatings on solid wood members or pre-formed silica fillers. It is therefore the objective of this study to apply a low-loading, in situ sol–gel silica coating to wood particles which preserves the particle morphology while providing a measure of thermal protection for higher-temperature processing conditions.

2. Materials and Methods

2.1. Wood Particles

Residual European beech (Fagus sylvatica) sawdust was donated by the Brst d.o.o sawmilling company (Črniče, Slovenia). A 20-mesh sieve (Cole Parmer, Vernon Hills, IL, USA) was used to remove any foreign objects or coarse particles. The material passing through the sieve was used for this study.

2.2. Silica Treatment of Wood Particles

Synthesis of the silica to be applied to the surface of the wood particles was based on the Stöber process [24]; tetraethyl orthosilicate as a Si precursor is hydrolysed in the presence of water, followed by condensation via silanol groups into a silica network. Wood particles were oven-dried at 103 °C for 24 h to remove any bound water (<1%). These particles were then introduced into a reaction mixture of tetraethyl orthosilicate (TEOS) (Sigma Aldrich, Saint Louis, MO, USA), ethanol (Merck, Rahway, NJ, USA), and deionized water. The preparation of silica was carried out at three different concentrations of precursor TEOS, 0.1, 0.2, and 0.3 wt%, which corresponded to water/TEOS molar ratios of 150, 75, and 50, respectively. The ethanol/water ratio was kept at 0.8. Then, 5 g of wood particles were impregnated with 30 mL of sol–gel mixture and kept in a closed container for 10 min to allow for almost complete absorption. Thoroughly wetted particles were then exposed to concentrated vapours of a 25% ammonia solution (Merck) for 1 h in a closed container, which catalysed the silica formation without directly contacting the wood particle/sol–gel reaction mixture.

2.3. Morphology and Characterization of Wood Particles

The size of the particles was monitored to see if treatment resulted in a noticeable increase in diameter. This was measured using a Horiba Scientific LA-960A2 particle analyser (HORIBA, Kyoto, Japan), using the LA-960 software version 9.00.

The porosity of the samples was also measured to measure the effectiveness of the treatment. Porosity was assessed by physisorption analysis (Autosorb iQ-XR-AG-AG, Anton Paar Quantachrome Instruments, Boynton Beach, FL, USA), using the ASiQwin software version 5.21. The specific surface area (SSA) and total pore volume (TPV) were determined using nitrogen gas according to the Brunauer–Emmett–Teller (BET) model, while the microporous (<2 nm) surface area (µSA) and micropore volume (µPV) were determined using CO₂ gas according to the density functional theory (DFT).

Attenuated Total Reflectance–Fourier Transform Infrared (ATR-FTIR) measurements were performed with a Bruker AlphaII instrument (Billerica, MA, USA). Measurements were carried out at ambient atmospheric conditions using an ATR crystal. Spectra were collected in the wave number range of 400–4000 cm⁻¹ with a resolution of 4 cm⁻¹ using OPUS V8.7.10 software. Ten replicates were tested for each sample, and average spectra were generated.

Scanning electron microscope (SEM) analysis was used to determine the morphology and surface topography of the samples. An FE-SEM SUPRA 35 VP (Carl Zeiss, Oberkochen, Germany) electron microscope was employed using the SmartSEM software version 6.08. Samples were prepared by immobilizing a thin layer of wood particles onto the sample holder with a double-sided adhesive carbon tape. Samples were sputtered with a layer of gold (Denton Vacuum LLC, Moorestown, NJ, USA). SEM analysis was performed at approximately 4.5 mm and an accelerating voltage of 1 kV.

2.4. Thermal Analysis

Thermogravimetric analysis (TGA) was carried out on a TGA 5500 (TA Waters Instruments, New Castle, DE, USA). Samples of 1–2 mg were placed in platinum pans. The samples were heated from 25 °C to 600 °C at a rate of 25 °C/min under a nitrogen gas flow of 25 mL/min, and the weight loss was recorded. The weight derivative curves (dTG) were determined using Trios software, version 5.3.0.48151.

2.5. Powder X-Ray Diffraction

The structural organization of samples was assessed with a Bruker D2 Phaser diffractometer (Bruker, Karlsruhe, Germany) (Cu–Kλ radiation; 1.5406 Å). The measurement software used was the DIFFRAC Measurement Center version 7.5.0 and evaluation software was the DIFFRAC.EVA version 6.1. Measurements were performed with a step size of 0.03° and a time/step of 0.5 s. Samples were analysed in the form of a pellet, pressed on a laboratory press: 0.05 g of each sample was used. XRD spectra were collected in the 2θ range from 5° to 70°. Scattering spectra were deconvoluted with PeakFit software version 4.12, using Gaussian function for the fitting of deconvoluted peaks and amorphous region. Values of integrated areas of deconvoluted peaks were used to calculate the crystallinity index as a ratio between integrated areas of crystalline peaks and integrated total area of the scattering spectra [25].

3. Results and Discussion

3.1. Morphology and Characterization

The particle size distribution data of silica-treated and non-treated beech particles used during this study are given in

Table 1. Particle size distribution of beech particles.

Treatment	D10 [µm]	D90 [µm]	Median [µm]
Reference	322.9	496.7	404.8
0.1% TEOS	322.4	510.2	407.1
0.2% TEOS	325.1	514.9	413.3
0.3% TEOS	323.5	513.6	410.5

Average of 3 repetitions.

.

D10 and D90, refer to the particle diameters at which 10% and 90% of the particles are smaller than that size. The median can also be expressed as D50. The median particle size was 405 µm in the reference wood particles, while the diameter of silica-treated particles was found to increase. This increase indicates that the silica was effectively applied and fixed to the particles. As this is an in situ synthesis, we can also have silica nucleating and growing within wood cell lumens or other anatomical features (i.e., vessels) of the particles.

The effect of this silica treatment on particle porosity is given in

Table 2. Porosity evaluation of beech wood particles.

Treatment	SSA, m2/g	µSA, m2/g	TPV, cc/g	µPV, cc/g
Reference	3.14	48.65	0.007	0.018
0.1% TEOS	3.40	92.32	0.009	0.035
0.2% TEOS	2.32	76.71	0.013	0.029
0.3% TEOS	6.86	96.06	0.011	0.035

. The specific surface area (SSA) of the reference wood particles was found to be less than the SSA of wood particles treated with silica particles. The SSA of reference particles was 3.14 m²/g, while for treated particles (0.1%, 0.2%, 0.3% TEOS), it was 3.40, 2.32, and 6.86 m²/g, respectively. The increases in SSA for samples produced with 0.1% and 0.3% TEOS indicate a successful treatment of the surface, supporting thermal protection of the wood. The 0.2% TEOS treatment had a lower SSA than the reference, which is likely due to the specific morphology of the wood particles used in those measurements. When looking at the microporous surface area (µSA) and pore volumes (TPV and µPV), we can see that even with the 0.2% TEOS treatment, there was an increase from the reference. This indicates that there was still silica added to the particles, as silica is reported to have a relatively high surface area and porosity [26]. Figure 1 shows a closeup view of silica that has formed on the wood particle surface during this study’s sol–gel process, highlighting the porous nature of the silica. Raabe et al. [27] also reported that the specific surface area, pore volume, and pore diameter of cellulose particles increased by up to 209% after incorporation of nanosilica particles. This increased surface area and microporosity can provide improvements in thermal stability by acting as a barrier layer, reducing heat transfer in a thin layer on particle surfaces. This is desirable for composite applications in which the initial decomposition of the wood material is the target and where the particle morphology is to be maintained, as opposed to flame retardancy applications, where higher silica mass can be more beneficial.

By using TEOS, silica particles were bonded together to form a strong silica network and coating. Figure 2 shows silica covering the surfaces of the wood particles, as well as entering anatomical features of the wood and voids created during processing. However, even with this study’s small increases in the amount of TEOS, it is seen that the silica coating on the surface provides good coverage even at 0.1% TEOS treatment, without causing significant clumping, as found by Yan et al. [28]. In the study by Yan et al., different ratios (0–5%) of silica particles were used to modify wood coatings. It was reported that the adhesion performance decreased when the silica content increased from 0% to 5%. This was due to increased silica particle agglomeration and uneven distribution, resulting in a decrease in adhesion. It must be noted that Yan et al. used pre-formed silica particles, and thus adhesion was the predominant way for SiO₂ immobilization to occur. In our study, bringing the wood particles into contact with sol–gel reaction mixture ensured that cellulose OH groups were involved in a condensation reaction with TEOS [29]. Here, a thick layer of silica was not a target, and we pursued a treatment method that can later be scaled to a higher volume solution and kept the silica content to a minimum by impregnating the wood particles with sol–gel mixture prior to catalysing nucleation and growth of SiO₂. This approach enabled controlled formation of a network that efficiently embeds the entire surface of wood particles in a thin layer of nano-sized SiO₂ particles. It can also be seen in Figure 2C that agglomerations of silica spheres have formed on top of the encrusted network of silica. These agglomerations are likely what Yan et al. experienced at higher loading ratios.

The FTIR spectra of non-treated and silica-treated beechwood particles are displayed in Figure 3. The spectra show some differences in peaks and intensities, suggesting that some chemical changes occurred after incorporation of silica.

The broad peak near 3400 cm⁻¹ can be attributed to the occurrence of O-H stretching corresponding to polar hydroxyl groups in wood. The peak at 2900 cm⁻¹ corresponding to aliphatic C-H stretching that is associated with cellulose polymers [30] was more visible in non-treated wood particles and started to decrease gradually with the incorporation of silica. The change in the band’s intensity can be attributed to the alteration of cellulose crystallinity during the chemical sol–gel process. The band at 1736 cm⁻¹ corresponds to unconjugated C=O stretch occurring in hemicellulose. The C=O band was present in the untreated wood spectrum and was reduced in the silica-treated samples. The peak at around 1600 cm⁻¹ likely corresponds to C=C stretching vibrations of aromatic rings in lignin polymers. Lazari and Elmi [31] reported the occurrence of C=C vibrations of lignin in FTIR spectra of a wood/chitosan/silica composite at 1613 and 1569 cm⁻¹. In the region between 1300 and 1500 cm⁻¹, several peaks were observed in the spectra of all samples. Prior research [32,33] assigned peaks at 1455, 1425, 1364, and 1335 cm⁻¹ to C-H vibrations in cellulose, hemicellulose, and lignin. Meanwhile, peaks around 1500 and 1310 cm⁻¹ were attributed to aromatic ring vibrations in lignin and CH₂ groups from cellulose and hemicellulose. The peak at 1200 cm⁻¹ attributed to O-H bending was observed in the spectrum of untreated wood, but it decreased progressively in the silica-treated samples. The peak between 1025 and 1050 cm⁻¹ was observed in all samples and assigned to the occurrence of C-O and C=C groups and C-C-O stretching from lignocellulosic biomass components [34]. The peaks around 450 and 800 cm⁻¹ were exclusively observed in silica-treated samples and absent in the reference. Moreover, the intensity of these peaks increased with the increasing silica loading from 0.1% to 0.3% TEOS; thus, they were attributed to the occurrence of Si groups. Indeed, previous studies [35,36] reported peaks at 810 cm⁻¹ and 820 cm⁻¹, corresponding to Si-O-Si and Si–CH3, and to Si-O bending, respectively. Additionally, the peak at 450 cm⁻¹ was attributed to Si-O rocking frequency. The detection of these peaks in the spectra of treated samples (Figure 3) presents a relatively fast and easy method of determining the efficacy of the treatment on wood particles.

3.2. Thermogravimetric Analysis

Thermogravimetric (TG) and derivative thermogravimetry (dTG) curves of non-treated and silica-treated wood particles are plotted in Figure 4 and Figure 5. Data from the thermal decomposition test are summarized in

Table 3. Thermogravimetric analysis of particles.

Treatment	Tonset, °C	Tmax, °C	Tendset, °C	T10%, °C	T50%, °C	Ash, wt%
Reference	300 ± 1	324 ± 1	363 ± 2	251.7 ± 9	319.0 ± 5	0.8 ± 6
0.1% TEOS	316 ± 3	327 ± 4	360 ± 1	255.9 ± 2	323.1 ± 6	1.2 ± 8
0.2% TEOS	316 ± 4	326 ± 3	363 ± 2	254.8 ± 1	322.0 ± 7	2.8 ± 2
0.3% TEOS	316 ± 6	326 ± 2	358 ± 7	255.5 ± 7	322.8 ± 4	4.3 ± 3

Average of 5 repetitions.

.

The TGA thermogram of the non-treated and treated wood samples display three weight loss phases (Figure 4). The first phase occurs at temperatures lower than 100 °C and was characterized with slight weight loss, attributed to dehydration and evaporation of moisture [37]. The second phase occurs between 240 °C and 360 °C and involves the basic constituents of wood [38]. This is of particular interest in this study, and a zoomed-in view of this temperature range in Figure 4 (lower plot) shows that the non-treated particles started degrading more rapidly than all treated particles. With increased silica, there was more protection of the particles to thermal degradation. It can also be seen that the wood particles treated with 0.1% TEOS experienced greater degradation around 350 °C, matching that of the non-treated wood, while the 0.2% and 0.3% treatments maintained their increased thermal protection until approximately 450 °C, when all material was degraded, leaving only ash. The dTG plot in Figure 5 provides another view of this thermal degradation. The increased thermal protection of wood particles treated with silica can be seen starting around 200 °C, when hemicellulose is degrading in the non-treated wood. The largest peak seen in this dTG was between 300 °C and 360 °C and resulted from primarily cellulose decomposition [37]. Here it can also be seen that the 0.1% treatment degraded more than the other treatments. Based on the onset data seen in Table 3, all treatments levels provided a 16 °C increase in thermal protection at this temperature. However, there is also to a lesser extent some protection occurring at lower temperatures around 200 °C. The third phase in the TG and dTG curves was between 360 °C and 480 °C, where additional mass loss was observed and was associated with the degradation of lignin and any remaining cellulose. These findings are in line with previous research [37,38,39,40,41,42].

The onset of thermal decomposition of the reference sample occurred at lower temperatures (300 °C) compared to the silica-treated samples (316 °C) (Table 3). Moreover, the maximum degradation temperatures of silica-containing samples were delayed by up to 33 °C (in the case of 0.2% TEOS-treated samples) compared to the reference, which can be observed from the peaks in dTG curves (Figure 5). These observations indicated that the incorporation of silica increased the resistance to thermal degradation of the beechwood particles. Similar findings were reported by Sheykhnazari et al. [43] when they investigated the thermal properties of silica-treated bacterial cellulose composites. They found that the inclusion of silica significantly enhanced the thermal stability of the bacterial cellulose and that the sample loaded with 7 wt.% silica showed the highest thermal resistance among other samples. The percentage of residual ash at the end of the thermal analysis (600 °C) increased with increasing silica contents, as this increased the amount of inorganic contents present. This increase in onset temperature can grant more flexibility in wood–polymer processing conditions and perhaps open up new polymer types with higher melting temperatures.

3.3. XRD Analysis

Wide-angle X-ray diffraction patterns of non-treated and silica-coated wood particles are presented in Figure 6. The diffractograms of all samples are dominated by a characteristic scattering pattern of cellulose, the component of wood particles with the highest content and structural organization: peaks at approximately 14.9° and 16.7° (not entirely resolved), and 22.7° and 34.9° correspond to Miller indices of 1–10, 110, 200, and 004, respectively, and are assigned to Iβ cellulose allomorph [44]. All samples also exhibit a certain degree of background, which is a result of amorphous scattering of present wood components with a poorly established structural arrangement.

Scattering patterns clearly differentiate in the intensities of the peaks, with non-treated wood particles exhibiting the highest intensity, which progressively decreases with an increase in silica content of coated particles. A higher loading of amorphous silica contributes to the decrease in the overall intensity of the coated samples’ spectra because of the introduction of a poorly scattering phase (i.e., silica coating); the employed procedure for particle modification alone is not expected to trigger changes in the crystalline arrangement of the wood particles. The influence of silica on scattering is clearly seen with the sample with 0.1% TEOS, which is almost identical to the reference sample (smallest amount of silica add-on), while the sample with 0.3% TEOS, with the highest silica content, clearly exhibits a lower scattering intensity, which is reflected in their crystallinity index values of 69.5% and 61.8%, respectively (

Table 4. Crystallinity index (CI) of silica-coated beech wood particles.

Treatment	CI %
Reference	69.5
0.1% TEOS	69.5
0.2% TEOS	67.1
0.3% TEOS	61.8

). Since the only effect of silica coatings on wood particles, as determined with XRD, is a decrease in scattering intensity due to a higher loading of amorphous material, and no other changes, e.g., peak shifts, are discernible, we can safely assume that the sol–gel treatment used here does not affect the crystalline structure of wood particles. Modification of particles’ surface primarily ensures, without interfering with its structural arrangement, that the intact mechanical properties of the wood filler can contribute to the final composites’ strength.

4. Conclusions

The thermal degradation of wood and wood-based products can be significantly improved by coating or impregnating with sol–gel for the production of a silica network. This treatment, in the amounts used in this study, provided a 16 °C increase in the onset of thermal degradation compared to reference wood particles, while maintaining the look of natural wood. This visibly imperceptible layer can effectively reduce degradation of the underlying wood particles.

While this study focused on confirming the efficacy of the sol–gel treatment used and methods for detection, future work will focus on making larger scale sol–gel treatments in larger reactors. This scaling up will generate enough particles that can then be used in wood–plastic composites. Here, the benefit of creating thin layers of particulate coatings that follow the contours of the wood particles closely without concealing their surface features is expected to extend even beyond thermal protection: the preserved surface area of wood particles can result in enhanced interfacial adhesion and mechanical interlocking between the filler and polymer matrix. In addition to mechanically securing sufficient contact points between composite constituents, silica-populated surfaces of wood particles also offer a potent platform for subsequent chemical functionalization, taking on the role of compatibilizer as well. While other researchers have found that increasing the amount of silica will eventually decrease adhesion in composites, the levels being used in this study are much lower than the 5% used in the study by Yan et al. [29]. This has added new insight that at very low silica precursor amounts, this work demonstrates (a) measurable increases in specific surface area and microporosity and (b) meaningful TGA shifts (≈+16 °C onset), showing that the morphology and placement of silica can deliver protection instead of simply adding silica mass. This particle-scale coating provides a promising route to producing thermally protected, functionalizable wood fillers for composites that maintain the filler geometry and potential mechanical interlocking, offering an attractive upcycling pathway for wood residues.