Waste Nutshell Particulate Biocomposites with Geopolymer Matrix

Abstract

The objective of this study was to explore the potential of creating advanced insulating biocomposites using waste almond and hazelnut shells as particulate fillers, combined with a geopolymer binder, to develop sustainable materials with minimal environmental impact. Optimal conditions for the preparation of biocomposites were determined by measuring the compressive strengths. The aforementioned optimal conditions included a geopolymer to waste nutshell mass ratio of 2, room-temperature curing, and the use of metakaolin geopolymers activated with potassium solutions. Notably, the highest compressive strengths of 4.1 MPa for hazelnut shells biocomposite and 6.4 MPa for almond shells biocomposite were obtained with milk of lime pretreatment at 80 °C for 1 h. Scanning electron microscopy with energy dispersive X-ray spectroscopy (SEM/EDX) and Fourier transform infrared spectroscopy (FTIR) analyses revealed better adhesion, as well as improved geopolymer gel polymerization. Furthermore, thermal conductivity and diffusivity measurements demonstrated values characteristic of insulating materials, reinforcing their potential for eco-friendly construction applications.

1. Introduction

The pursuit and development of sustainable and eco-friendly materials is one of the main goals of contemporary material research. Innovative solutions in this field of science reach out to waste materials and by-products of different industries as a resource with vast potential for transformation into value-added materials [1]. Some of these materials are by-products of industrial agriculture, such as waste nutshells (WN), which are usually utilized as fuel due to their favorable calorific value. However, the application of these waste materials in different technologies and engineering applications offers a promising solution due to their abundance, low cost, and unique lignocellulosic composition [2,3]. As two of the most commonly cultivated nuts around the globe, almond and hazelnut production result in high amounts of WN due to 50% of the produced mass being the hard nutshell, which could find use in biocomposite materials [4]. Furthermore, WN are lignocellulosic materials (LCM) with interesting composition and microstructure, whose composition is influenced by the climate and growth conditions of the parent plant and its fruit. For hazelnut nutshell (HN), the average composition is 35% hemicellulose, 32% lignin, and 30% cellulose, while almond nutshells (AN) are composed of approximately 50% cellulose, 28% hemicellulose, and 20% lignin. The remainder of WN consists of minor fractions of extractive organic compounds and inorganic phases [5,6].

Variability in the composition and size of different LCMs, as well as the number of compounds present on the surface of LCMs and their hydrophilic nature, hinders their simple exploitation in biocomposite materials. As a consequence, a large number of pretreatment methods are being developed and investigated [7]. They fall into three main categories: chemical, physical, and biological, each playing a distinct role in modifying the composition, morphology, and surface properties of LCMs [8,9]. Physical methods primarily focus on altering the shape and morphology of LCMs, with the main group being mechanical processes that include milling, grinding, and extrusion. Other physical methods include steam explosion, microwave-assisted processing, cold plasma treatment, and ultrasonication. These methods are mostly employed to facilitate the chemical treatments, although they can themselves influence the composition of LCMs and positively influence properties of their biocomposites [10,11,12,13]. Chemical pretreatments are one of the most effective strategies for improving the compatibility of LCMs with both organic and inorganic matrices. In addition to changes in surface properties, chemical pretreatments are able to remove unwanted compounds from LCMs, such as lignin, hemicellulose, pectin, waxes, and sugars. Besides impeding the setting and hardening processes of inorganic matrices, these compounds are hydrophilic, which influences the workability of inorganic pastes [14,15]. Alkali treatment, specifically mercerization using sodium hydroxide (NaOH), is a widely applied method of pretreatment. This process enhances fiber surface roughness, cleans the surface from different extractive compounds, improves crystallinity, and increases the number of surface hydroxyl groups, leading to better adhesion in biocomposites with inorganic matrices [16,17,18]. Silane treatment, acetylation, and acid hydrolysis are alternative chemical methods that modify surface polarity and promote better bonding with matrices, especially organic ones, contributing to improved composite strength and water resistance [19,20]. Milk of lime pretreatment is a chemical method that employs a calcium hydroxide (Ca(OH)₂) water suspension. It is explored for selective lignin removal and enhancement of interfacial properties in biocomposites. During this process, calcium hydroxide is precipitated on the surface of LCMs and can then react with inorganic matrices and form numerous calcium phases, which usually contribute to the mechanical properties of such biocomposites [21,22]. Biological pretreatments employ enzymatic or microbial processes to selectively degrade lignin and hemicellulose while preserving cellulose integrity. These methods are advantageous due to their eco-friendliness and specificity, though they often require longer processing times compared to chemical and physical techniques [8,9]. Natural pretreatment methods are alternative approaches for LCM pretreatment, utilizing naturally occurring agents. These methods typically employ plant extracts for chemical pretreatment or microorganisms for biological pretreatment. For instance, tannic acid is used to modify various natural fibers [23]. This type of pretreatment enhances adhesion between the hydrophobic organic matrix and hydrophilic natural fillers by altering the surface properties of LCM, making it hydrophobic. Additionally, the polyphenol composition of tannic acid imparts biocidal properties to such biocomposites [24]. However, these pretreatment methods face challenges when applied to biocomposites with an inorganic matrix. The hydrophilic nature of these matrices, combined with the retardation effect of phenolic compounds on hydration reactions, can lead to the deterioration of hardened concretes through interactions with phenols [25].

Geopolymers (GP) are a class of inorganic binder materials that have emerged as a feasible alternative to conventional Portland cement [26]. They are ceramic amorphous materials, which form through the reactions of geopolymerization that occur in the geopolymer pastes after mixing the aluminosilicate powder precursors with alkaline or acidic activation solutions. This process can be divided into three steps. Firstly, the aluminosilicate precursors are dissolved by the strong alkaline or acidic solutions, where the main building blocks of geopolymers, silicate and aluminate tetrahedrons, are released. Secondly, during the initial curing and hardening time, the released tetrahedrons condense into monomeric and oligomeric structures. Finally, the monolithic geopolymer stone is achieved in the third step, where the monomeric and oligomeric units polymerize into an aluminosilicate amorphous 3D network [27,28,29]. As eco-friendly materials, GPs offer significant environmental advantages due to their notably lower carbon footprint compared to traditional cement. They can incorporate industrial by-products like fly ash, ground granulated blast-furnace slags, red mud, as well as different natural aluminosilicate materials, including calcined clays such as metakaolin, shale, laterite, and natural pozzolans [30,31].

As inorganic monolithic materials, hardened geopolymers exhibit limitations such as low flexural strength and low toughness. The incorporation of different materials as fillers into geopolymer matrices creates composite materials with enhanced properties and expanded application potential [32]. Fiber reinforcements are the most prevalent, offering improved mechanical properties through stress redistribution and crack inhibition under external loads. Natural fibers, particularly lignocellulosic ones, are increasingly favored in biocomposite preparation due to their sustainable, cost-effective, and ecological qualities, especially for composites with GP matrix [33,34]. Even though the incorporation of natural fibers enhances some of the biocomposite properties, the optimal addition varies for different types of these fibers, while their content does not exceed 10 wt.%. Contents greater than the aforementioned upper limit result in lower GP paste workability and lower mechanical properties of the resulting biocomposites [35]. Among innovative developments are biocomposites, which integrate fungal mycelium or hyphae from various fungi species as fillers. These types of composites are being studied as potential substitutes for polymeric foams and insulating materials (thermal and acoustic) used in the construction industry, as well as for applications in air purification, packaging, and more [36]. Additionally, particulate LCM fillers are commonly added to GP matrices as fine powders to improve their mechanical or thermal properties, as well as creating novel materials. Despite numerous investigations of the application of powdered LCM particulate fillers, biocomposites utilizing non-powdered lignocellulosic materials remain largely unexplored. One notable exception is the study by Roper et al. [37], which investigated the properties of geopolymer biocomposites incorporating cork particles with an average particle size of 2 mm. To the best of our knowledge, this remains the only research specifically focused on such a type of biocomposite.

This work continues our investigation into the influence of two alkaline pretreatments—mercerization and lime milk treatment—on hazelnut and almond waste nutshells (WN) for use in particulate biocomposites with a geopolymer matrix. Our prior research suggested that both treatments can enhance the properties of WNs, with mercerization using 6 wt.% NaOH at 80 °C for 2.5 h showing particular promise based on adhesion parameter predictions. Mercerization appeared more effective at removing undesired lignocellulosic constituents (e.g., lignin, hemicellulose, and waxes), while lime milk pretreatment was associated with high concentrations of precipitated calcium phases, potentially beneficial for mechanical performance through interactions with the geopolymer matrix [38].

Given the limited research on such biocomposite systems, this study aims to further explore their mechanical and thermal behavior, focusing on their potential application as insulating materials in civil engineering. Specifically, we investigate how different pretreatments influence WN–geopolymer adhesion and composite performance, with particular emphasis on validating parameter-based adhesion predictions. Compressive strength was selected as the primary performance criterion, while thermal conductivity and thermal diffusivity were also evaluated to provide a more comprehensive assessment. Scanning electron microscopy with energy dispersive X-ray spectroscopy (SEM/EDS) and Fourier transform infrared spectroscopy (FTIR) were used to evaluate differences in adhesion quality and phase development within the geopolymer matrix of differently pretreated WN biocomposites. Statistical analysis was performed using the Welch’s t-test to determine the statistical significance and to reject/accept the null hypothesis.

2. Materials and Methods

2.1. Materials

Geopolymer matrices were prepared using fly ash from the Plomin Power Station (Plomin, Croatia) and metakaolin, i.e., calcined commercial kaolin clay (technical grade, VWR Chemicals, Paris, France) as solid precursors. These powders were activated with either sodium (technical grade, VWR Chemicals, Paris, France) or potassium (technical grade, Ivero, Zagreb, Croatia) water glass, combined with a 12 M solution of NaOH or KOH (p.a., Gram-mol, Zagreb, Croatia). Waste almond (AN) and hazelnut (HN) nutshells were sourced from small farmers in the Šibenik-Knin and Zagreb County, Croatia. Sodium hydroxide (NaOH) solutions were prepared by dissolving NaOH microgranules (p.a., Gram-mol, Zagreb, Croatia) in demineralized water, while milk of lime was obtained by mixing freshly calcined calcium oxide (CaO), derived from calcium carbonate (CaCO₃, p.a., Merck, Darmstadt, Germany), with demineralized water in a 1:7 ratio.

2.2. Sample Preparation

Waste nutshells were milled and hand-sieved on the 800 μm mesh sieve to remove smaller particles and dust, which could hinder the geopolymerization reactions. Subsequently, the sieved WNs were pretreated with two alkaline solutions. Mercerization was carried out using 6 and 9 wt.% NaOH solutions for 1 and 2.5 h at 80 °C. Lime milk pretreatment with a 1:7 ratio of CaO and demineralized water was used by mixing WN and milk of lime in a 1:8 ratio. The WN were kept at 80 °C for 1 and 2.5 h, while another batch of treated WN was prepared at room temperature for 24 h.

As a first step of biocomposite preparation, fresh geopolymer pastes were made by mixing fly ash or metakaolin with a 12 M NaOH or KOH solution mixed with sodium or potassium water glass to achieve a water-to-solid ratio of 0.66 and a molar ratio Al:Na,K = 1:1. After 5 min of mixing, hazelnut and almond nutshells were directly added to fresh pastes, while mixing continued for 2 min. Freshly mixed biocomposites were pressed into disposable syringes, which were utilized as cylindrical plastic molds. They were filled up to a predetermined point to form samples with a height of 37 mm and a diameter of 16 mm. Afterwards, the molds were sealed and the samples were cured in controlled conditions for 24 h, after which the samples were kept in the molds at room conditions to harden. In addition to examining the effect of pretreatment, the impact of the mass ratio between the geopolymer paste and waste shells was studied using mass ratios of 1:1, 1.5:1, and 2:1. Furthermore, the influence of the nature of the solid precursor and activation solution was investigated, and the effect of temperature was examined by curing the biocomposites at room temperature and 40 °C. Compressive strengths, as the primary criterion of the influence of the aforementioned variables in biocomposite preparation, were tested on cylindrical samples, with a diameter of 16 mm and a height of 37 mm. Further in the text, the prepared biocomposites will be named with abbreviations that indicate their composition, type of shell, pretreatment method, and geopolymer used in their formulation. The abbreviations will consist of the following components: CPHN-SPT, which are elucidated in

Table 1. Prepared biocomposite sample abbreviation description.

Abbreviation	Description	Details
Waste nutshell
C	Concentration of sodium hydroxide used for pretreatment	6 or 9 for 6 and 9 wt.% solutions
P	Pretreatment method	N for mercerization, C for pretreatment with milk of lime
H	Duration of pretreatment	1 and 2 for 1 and 2.5 h at 80 °C, 24 for 24 h at room temperature
N	Type of nutshell in the biocomposite	H for hazelnut shell, A for almond shell, untreated designated as H or A
Geopolymers and curing
S	Type of activation solution used	N for sodium activation solution, K for potassium activation solution
P	Solid precursor used	P for fly ash, M for metakaolin
T	Curing temperature	No designation for room temperature, 40 for curing at 40 °C

. The first part of the abbreviation refers to the pretreatment and type of nutshell in the biocomposite, while the second part refers to the type of geopolymer and preparation method.

For example, 9N1H-KM40 indicates that the biocomposite sample was prepared from hazelnut shells treated with a 9 wt.% NaOH solution for 1 h at 80 °C, combined with metakaolin activated using a potassium activation solution, and cured at 40 °C.

2.3. Biocomposite Characterization

Compressive fracture forces were measured employing the WEB Thüringer Industriewerk Rauenstein (TIRA GmbH, Schalkau, Germany) universal testing machine, which has a maximum force capacity of 4800 N and a fixed support span of 4 cm. The fracture forces necessary for strength calculations were determined as the average value of measurements from at least three specimens and were measured after 1, 7, and 28 days of biocomposite hardening.

Compressive strengths were calculated from the measured fracture forces of cylindrical biocomposite specimens with a diameter of 16 mm and a height of 37 mm (Figure 1) using equation

$${\sigma }_C={\displaystyle \frac{F}{A}},$$

(1)

where σ_C is the compressive strength, F is the fracture force of the biocomposite (N), and A is the surface area of the applied force (m²).

For conducting the hypothesis test and to determine whether there is a significant difference between the strength values obtained for biocomposites prepared from different precursors and pretreated nutshells using various methods, the Welch’s t-test was conducted. The t-test was calculated using the following equations:

$$t={\displaystyle \frac{{\overline{X}}_1-{\overline{X}}_2}{\sqrt{\frac{{\sigma }_1^2}{n_1}}-\frac{{\sigma }_2^2}{n_2}}},$$

(2)

$$df={\displaystyle \frac{{\left(\frac{{\sigma }_1^2}{n_1}+\frac{{\sigma }_2^2}{n_2}\right)}^2}{\frac{{\left(\frac{{\sigma }_1^2}{n_1}\right)}^2}{n_1-1}+\frac{{\left(\frac{{\sigma }_2^2}{n_2}\right)}^2}{n_2-1}}},$$

(3)

where X₁ and X₂ are the mean values of the variables for the two systems, σ₁ and σ₂ are their standard deviations, and n₁ and n₂ represent their respective sample sizes. The df value provides the degrees of freedom necessary in combination with a p-value of 0.05 to obtain the critical value (t_crit), which is compared with the calculated t value from Equation (3). If the condition t > t_crit is met, the null hypothesis is rejected, indicating that the mechanical properties values of the two different samples are not equal when considering their standard deviations.

Adhesion between the geopolymer matrix and differently pretreated waste nutshells in biocomposites was examined using a Tescan Vega 3 scanning electron microscope (Tescan, Brno, Czech Republic) operating at 10 kV. Samples were fixed on specimen holders with double-sided carbon conductive tape and gold-coated using a Quorum SC 7620 sputter coater (Quorum Technologies, Laughton, UK). The EDS Bruker Quantax Compact detector (Bruker, Billerica, MA, USA) for energy-dispersive X-ray spectrometry was used to employ linear EDX mapping to elucidate the difference in composition of geopolymer matrix and nutshell filler and to corroborate the incorporation of precipitated calcium into the geopolymer matrix. FTIR analysis, which was carried out using a Bruker Vertex 70 spectrometer (Bruker Optics, Karlsruhe, Germany) in attenuated total reflectance mode (ATR) on samples pressed on a diamond, and the spectra measured between 400 and 4000 cm⁻¹, with a spectral resolution of 2 cm⁻¹ and an average of 32 scans, was utilized to discern the difference in the degree of geopolymer matrix interconnectivity or possibility of the development of calcium phases. Thermal conductivity and diffusivity of biocomposites were determined using the Linseis THB-100 commercial device (Linseis Messgeraete GmbH, Selb, Germany) on samples with dimensions of 50 × 30 × 10 mm, where the sensor was positioned between two identical samples in a standard holder. Thermal conductivity values were obtained from the average of 10 measurements, conducted at a constant current of 50 mA for 15 s for the almond shell biocomposite and 5 s for the hazelnut shell biocomposite.

3. Results and Discussion

The compressive strength testing of composites is often used to assess the efficiency of their preparation methods [39]. As already mentioned in the text above, in this study, compressive strength was used as a key property for selecting optimal preparation and pretreatment conditions for nutshell and geopolymer biocomposites. Furthermore, to determine the statistical significance of the obtained results, Welch’s t-test was applied with the null hypothesis stating that the mean compressive strengths of two different biocomposite samples are equal, considering the variation and standard deviation. The Welch’s t-test is suitable for cases with a small sample size [40,41]. Rejected null hypotheses will be designated as “+” while accepted null hypotheses will be designated as “−” in the further text. Figure 2 presents the measured compressive strengths of the samples and their variations for biocomposites prepared with different ratios of the matrix (geopolymer) and filler (nutshells). The tested geopolymer-to-nutshell ratios were 1, 1.5, and 2. However, data for samples with a 1:1 ratio were not included as these composites shattered during demolding. Their mechanical properties could not be determined even after 28 days of hardening. Almond shell biocomposites showed 35% higher compressive strength at a ratio of 2 compared to 1.5, reaching 4.8 MPa after 28 days. Hazelnut shell biocomposites exhibited a similar difference of 31%, with a maximum compressive strength of 2.5 MPa.

Using Equation (4), the degrees of freedom (df) were calculated, and the critical value (t_crit) of the variable t was read from the table for a two-tailed t-test with a significance level of 0.05 [42].

Table 2. Welch’s t-test parameters for determining the significance of the geopolymer-to-waste shell ratio.

	df			tcrit			$\left|\mathit{t}\right|$			Null Hypothesis Rejected
Days	1	7	28	1	7	28	1	7	28	1	7	28
A-1.5KM × A-2KM	3.08	4.26	4.56	3.15	2.72	2.66	6.60	3.97	10.38	+	+	+
H-1.5KM × H-2KM	3.72	3.61	5.29	2.89	2.93	2.53	2.53	5.86	6.66	−	+	+

contains the data for this system, based on which the null hypothesis was either confirmed or rejected. It is evident that in almost all cases, the calculated t-values exceed the critical values, allowing the null hypothesis to be rejected. The only exception where the null hypothesis cannot be rejected is the compressive strength of the H-1.5KM samples after one day of curing. However, since this measurement represents a time point where these samples exhibit the lowest strengths, and given the fact that the null hypothesis for the same sample was rejected after 7 and 28 days of curing, it can be concluded that a significant difference exists in the compressive strengths of biocomposites with different mass ratios of waste hazelnut shells and geopolymeric binder. For the further preparation of composites, the 2:1 ratio was selected, as it demonstrated the highest compressive strengths.

The next step involved the preparation of biocomposites with the previously determined ratio at different curing temperatures to determine the optimal curing conditions. The A-KM and H-KM composites were cured at room temperature and 40 °C, and their strength values are presented in Figure 3.

Even though elevated curing temperatures typically enhance geopolymer strength, this was not noticeable in this system. The primary reason for increased strength at higher temperatures is the kinetics of geopolymerization reactions, where higher temperatures facilitate faster reactions that lead to earlier development of higher strength. However, in systems that are not well isolated, water loss due to evaporation during curing may inhibit geopolymerization. Water present in geopolymer pastes serves as the medium for these reactions. Additionally, excessive reaction speeds can cause rapid formation of a product that encapsulates particles, preventing their further dissolution. Lastly, porous structures form at higher curing temperatures, potentially leading to weaker mechanical properties [43,44]. Alsina et al. investigated water absorption in various biocomposites and their lignocellulosic fillers as a function of temperature. While there was no significant difference in maximum absorbed water after 450 h of soaking, results from the first 50 h indicated higher absorption at elevated temperatures [45]. The compressive strength results of H-KM and A-KM biocomposites show a negative influence of increased curing temperature. A-KM biocomposites cured at room temperature exhibited 29% higher strength, while the difference for H-KM samples was 79%.

Statistical analysis of biocomposites prepared with varied curing temperatures is given in

Table 3. Welch’s t-test parameters for determining the significance of the curing temperature of biocomposites.

	df			tcrit			$\left|\mathit{t}\right|$			Null Hypothesis Rejected
Days	1	7	28	1	7	28	1	7	28	1	7	28
A-KM × A-KM40	3.14	4.12	4.96	3.13	2.75	2.58	6.45	4.37	7.35	+	+	+
H-KM × H-KM40	/	3.29	4.32	/	3.26	2.71	/	15.76	16.52	/	+	+

. It is evident that all of the null hypotheses are rejected, which confirms the room temperature curing conditions as optimal for the preparation of such biocomposites.

Using the selected ratio and curing temperature, biocomposites were prepared using untreated nutshells and geopolymers with two different solid precursors activated with potassium and sodium solutions. The purpose of this step was to identify the optimal solid precursor for further biocomposite preparation. Figure 4 presents the compressive strengths, while

Table 4. Welch’s t-test parameters for determining the significance of the effect of the type of solid precursor on hazelnut and almond shell biocomposites.

	df			tcrit			$\left|\mathit{t}\right|$			Null Hypothesis Rejected
Days	1	7	28	1	7	28	1	7	28	1	7	28
H-NM × H-NP	/	3.67	5.92	/	2.91	2.46	/	2.33	3.29	/	−	+
H-KM × H-KP	/	4.29	4.06	/	2.72	2.76	/	12.89	14.50	/	+	+
A-NM × A-NP	/	5.77	4.87	/	2.47	2.59	/	19.53	12.33	/	+	+
A-KM × A-KP	2.18	4.68	4.72	4.10	2.64	2.63	23.04	9.31	18.71	+	+	+

displays the Welch t-test parameters for biocomposite samples made from raw hazelnut and almond shells using different activation solutions and solid precursors. During the initial production of biocomposites, preparing reliable samples with fly ash proved challenging, as they frequently fractured upon removal from molds. To minimize the consumption of treated shells, the influence of the solid precursor was tested only on biocomposites made with untreated nutshells. From the graphical presentation in Figure 4, it is evident that biocomposites synthesized using metakaolin activation exhibit significantly superior mechanical properties compared to those made with fly ash, regardless of the type of nutshell used.

The final compressive strengths of metakaolin-based biocomposites containing HN and activated with sodium solutions are 28% higher than those of biocomposites made with fly ash. Additionally, activation with potassium solutions results in an even greater strength increase, 73% higher than the fly ash-based counterparts. For AN biocomposites, similar trends can be observed, where metakaolin-based geopolymers activated with sodium and potassium solutions achieve 56% and 62% higher strengths, respectively, compared to those using fly ash as a precursor. The highest compressive strength recorded for HN biocomposites was 2.53 MPa for the H-KM sample, whereas for AN biocomposites, the A-KM sample reached 4.79 MPa. Furthermore, biocomposites prepared with fly ash exhibited such low early strengths that they could not be measured after just one day of curing, except for the A-KP sample. This trend is commonly observed in fly ash-based geopolymers, as they contain a large proportion of highly crystalline phases that are poorly reactive, contributing little to strength development. As a result, the final geopolymer structure is often highly porous, which negatively impacts its mechanical performance [46]. Statistical analysis using the Welch’s t-test confirms the rejection of the null hypothesis in all cases, except when comparing metakaolin and fly ash-based geopolymers activated with sodium solutions, measured after seven days of curing. Based on the results shown in Figure 4 and the statistical evaluation in Table 4, it was concluded that metakaolin should be used as the solid precursor for biocomposite preparation, regardless of the nutshell type.

As the final step in evaluating the preparation conditions, the influence of the type of activation solution on the compressive strength of the biocomposites was examined. Alongside composites made from untreated nutshells, additional biocomposites were prepared using nutshells treated with a 6 wt.% NaOH solution, incorporating metakaolin activated with sodium and potassium solutions. It is well known that sodium-based activation solutions facilitate a more efficient first stage of geopolymerization, ensuring better dissolution of solid precursors and the formation of a less porous microstructure during curing. The key difference between geopolymerization processes involving potassium and sodium ions is that potassium leads to a higher degree of condensation and increased porosity. While sodium-based geopolymers often exhibit higher compressive strength, their final performance is not solely dependent on the activation solution, and each system requires optimization to achieve desirable mechanical properties and other key characteristics. Additionally, geopolymer pastes activated with potassium solutions tend to be easier to handle, with faster setting times. Additionally, their resulting geopolymers have lower thermal conductivity, attributed to a more porous structure [47,48,49].

Figure 5 presents the compressive strengths for both types of nutshells with varying activation solutions. It can be observed that metakaolin activated with potassium solutions exhibits higher compressive strengths, which is further confirmed by the statistical analysis provided in

Table 5. Welch’s t-test parameters for determining the significance of the effect of the activation solution on hazelnut shell biocomposites.

	df			tcrit			$\left|\mathit{t}\right|$			Null Hypothesis Rejected
Days	1	7	28	1	7	28	1	7	28	1	7	28
H-NM × H-KM	2.22	5.88	5.53	4.05	2.46	2.51	7.55	9.58	11.01	+	+	+
6N1H-NM × 6N1H-KM	5.96	5.98	3.95	2.44	2.45	2.77	2.84	1.42	2.85	+	−	+
6N2H-NM × 6N2H-KM	3.48	4.50	4.20	2.98	2.67	2.74	3.31	0.97	3.23	+	−	+

and

Table 6. Welch’s t-test parameters for determining the significance of the effect of the activation solution on almond shell biocomposites.

	df			tcrit			$\left|\mathit{t}\right|$			Null Hypothesis Rejected
Days	1	7	28	1	7	28	1	7	28	1	7	28
A-NM × A-KM	3.99	3.63	6	2.78	2.936	2.45	3.82	2.74	6.39	+	−	+
6N1A-NM × 6N1A-KM	4.63	4.51	4.74	2.65	2.67	2.62	1.28	2.69	3.24	−	+	+
6N2A-NM × 6N2A-KM	5.97	3.60	4.83	2.45	2.93	2.61	1.81	4.4	9.00	−	+	+

. Additionally, since potassium-activated geopolymers have lower thermal conductivity [49] and the goal of this study is to develop a biocomposite insulation material, potassium activation solution was selected for biocomposite preparation and for the investigation of the effects of nutshell pretreatment on its properties.

Since this study focuses exclusively on raw and pretreated nutshell samples mixed with metakaolin activated with alkaline potassium solutions, the part of the abbreviations indicating the type of geopolymer in the biocomposite material will be omitted. As previously mentioned, this work is a continuation of our research on the effects of pretreatment on the properties of waste nutshells. In the previous study, adhesion parameters of WN and different GPs were calculated. The obtained results indicated that the optimal pretreatment for both types of nutshells is treatment with a 6% NaOH solution at 80 °C for 2.5 h. These pretreatment conditions should result in the highest mechanical properties of biocomposites prepared with these kinds of nutshells. However, lime milk treatment, despite being less effective in removing undesirable components, contributed to an increase in compressive strength, which can be seen in Figure 6 and Figure 7 [38]. This observation supports the hypothesis that the deposition of calcium compounds on the nutshell surface enhances material properties, as they form different calcium cementing phases through reactions with the geopolymer paste during curing. To avoid unnecessary complexity in statistical analysis, results from Welch’s t-test will be presented only for the compressive strength values measured after 28 days of geopolymer curing. Additionally, only selected sample combinations will be shown to examine claims related to improvements in compressive strength.

The highest compressive strength of hazelnut shell biocomposites, reaching 4.15 MPa after 28 days of curing, is exhibited by the C1H samples. These samples show a 23% increase in strength compared to 6N2H biocomposites, which were initially predicted, based on adhesion parameters, to form the best composite materials. Additionally, C1H samples display 39% higher compressive strength than biocomposites made from untreated hazelnut shells. At first glance, Figure 6 suggests that C24H shells should yield the best mechanical properties, based on strength measurements taken after 1 and 7 days of curing. However, due to the significant deviations in measured strengths across all samples, it cannot be definitively stated that C24H samples demonstrate the highest early-day strengths or that the observed decrease in strength at 28 days accurately reflects the material’s actual performance. The results of the statistical analysis, presented in

Table 7. Welch’s t-test parameters for determining the significance of the effect of hazelnut shell pretreatment on the compressive strengths of prepared biocomposites.

Sample	Hx6N2H	Hx9N2H	HxC1H	6N2Hx9N2H	C1HxC2H	6N2HxC1H
df	4.48	5.67	4.92	4.79	5.66	4.87
tcrit	2.67	2.48	2.59	2.61	2.49	2.60
$\left|t\right|$	4.36	2.49	8.26	2.62	1.08	4.66
Null hypothesis rejected	+	+	+	+	−	+

, confirm these observations. The only case in which the null hypothesis is not rejected is the comparison between C1H and C2H samples, indicating that the duration of pretreatment with lime milk at 80 °C does not significantly affect the compressive strength of hazelnut shell biocomposites.

The compressive strengths of almond shell biocomposites are presented in Figure 7, demonstrating that, similar to hazelnut shells, shells pretreated with lime milk exhibit the highest values. The C1A composites achieve an average compressive strength of 6.41 MPa, marking an increase of 25% compared to untreated shells and 17% compared to 6N2A-treated nutshells. The C24A samples show similar properties to the C24H biocomposites, while the statistical analysis provided in

Table 8. Welch’s t-test parameters for determining the significance of the effect of almond shell pretreatment on the compressive strengths of prepared biocomposites.

Sample	Ax6N2A	Ax9N2A	AxC1A	6N2Ax9N2A	C1AxC2A	6N2AxC1A
df	4.15	5.54	5.95	3.68	5.81	3.99
tcrit	2.74	2.50	2.45	2.90	2.47	2.78
$\left|t\right|$	3.54	3.47	7.80	6.83	2.55	6.44
Null hypothesis rejected	+	+	+	+	+	+

supports the rejection of all null hypotheses. This confirms that C1A samples outperform C2A samples in mechanical properties. A potential explanation for the decline in compressive strength in C2H/A and C24H/A samples is the excessive deposition of calcium phases on the shell surface. This excess may negatively impact the development of strength during geopolymer curing, making these biocomposites structurally weaker.

To confirm the hypothesis that the properties of geopolymer improve through the formation of calcium phases from precipitated surface calcium carbonate, the biocomposite samples A-KM, 6N2A-KM, and C1A-KM were analyzed using SEM-EDS and FTIR techniques. The results obtained through SEM analysis (Figure 8) indicated poorer adhesion of untreated nutshells compared to 6N2A and C1A nutshells in relation to the geopolymer. Although improved adhesion was observed for the 6N2A samples, it was significantly more conspicuous in the C1A sample. In Figure 8b, the geopolymer phase appears to be in small, irregular fragments that only partially cover the surface of the shell, whereas in the C1A nutshells separated from the biocomposite (Figure 8c), a much better coverage of the surface with the geopolymer phase is visible. In the C1A sample, a matrix covers the surface with the plate-like structures, whereas for the biocomposite with untreated nutshells (Figure 8a), there are only a few large pieces of geopolymer matrix. Larger quantities of the leftover surface geopolymer phase could indicate better interface and adhesion in such biocomposites.

Through SEM-EDS mapping of the shell surface, an elemental analysis was conducted, showing that calcium is evenly distributed across the entire surface of the samples, regardless of whether it is the shell or the geopolymer area. Additionally, it was not possible to determine the calcium mass fraction in the A-KM and 6N2A-KM samples, despite its presence in metakaolin, whose composition was outlined earlier in this chapter. However, in the C1A sample, a calcium mass fraction of approximately 0.6 mas. % was obtained. Furthermore, a linear EDS profile scan, shown in Figure 9, revealed that the calcium content is slightly higher in the geopolymer phase. The characteristic radiation intensities of calcium along the line in Figure 9a were magnified tenfold to make the response visible in the graphical representation of the element distribution along the line (Figure 9b). The accuracy of the EDS analysis is influenced by the sample’s morphology, and since these samples have highly irregular, relief-like surfaces, it is challenging to determine the actual calcium content and its distribution.

To confirm the additional incorporation of calcium into the geopolymer structure, the formation of cementing calcium phases, and the consequent increase in strength, FTIR analysis was conducted on geopolymer matrix samples extracted from the biocomposites. The FTIR spectra (Figure 10) display standard bands characteristic of this type of material. Spectra of geopolymers extracted from the A-KM, 6N2A-KM, and C1A-KM biocomposite samples are presented. Broad bands in the range of 3600 to 3000 cm⁻¹, along with a smaller band at 1650 cm⁻¹, are attributed to structural and adsorbed water within the geopolymer sample. The next notable band, with significant intensity in the A-KM sample but weaker in others, appears at ~1570 cm⁻¹ and corresponds to bond vibrations in residual or incompletely removed extractives and lignin. Carbonate bands, resulting from reactions of excess alkalis and precipitated calcium with atmospheric CO₂, appear in the 1450–1300 cm⁻¹ range. A small band at 1260 cm⁻¹, visible in the C1A-KM sample spectrum, corresponds to C-O bond vibrations present in organic phases, which may stem from an incomplete removal of undesired components during lime milk treatment. The primary feature of these spectra is a broad, high-intensity band centered around 1000 cm⁻¹, which arises from Si-O bond vibrations within the geopolymer matrix. The position of this band can indicate the degree of network formation and the extent of the geopolymerization reaction. A shift to lower wavenumbers suggests a higher presence of non-bridging oxygens, i.e., terminated Si-O bonds that imply weaker geopolymer gel networking. A shift to higher wavenumbers signals better networking, though it may also be influenced by chemical bonds in cellulose potentially retained from the biocomposite. In the A-KM and C1A-KM spectra, a shoulder appears in the ~1195 to 1080 cm⁻¹ range (C-O bond vibrations), further indicating residual extractives and components retained from the biocomposite. Although these findings do not definitively confirm the formation of cementing calcium phases, the shift of the main band in the C1A-KM sample spectrum toward higher wavenumbers suggests an increased geopolymer gel polymerization. Since this band remains at similar positions in the A-KM and 6N2A-KM samples, it can be presumed that C-O bonds from shell components, which respond in this wavenumber range, do not contribute to the observed shift in C1A-KM. Additionally, the primary contributor to this band in lignocellulosic materials is cellulose, which is difficult to extract, making its presence in the geopolymer gel unlikely. Further confirmation of this assumption comes from the C1A-KM sample band at 795 cm⁻¹, which corresponds to a convoluted effect of Si-O bond vibrations in the amorphous structure and Al-O bonds in tetrahedral coordination.

From this, it can be concluded that C1A-KM shell-based biocomposites exhibit a higher proportion of amorphous structures, a more developed interconnected Si-O network, and an increased presence of aluminum in tetrahedral coordination. These characteristics, combined with improved adhesion observed in micrographs, explain the enhanced mechanical strength of biocomposites prepared from shells pretreated with lime milk [50,51,52,53,54].

According to literature sources, the pretreatment of lignocellulosic material is not expected to influence thermal conductivity and diffusivity, and if any impact does exist, it is not statistically significant [55,56]. Therefore, thermal conductivity and diffusivity were determined exclusively for the L-KM and B-KM samples. Their thermal conductivity values of 0.04 and 0.037 W m⁻¹ K⁻¹, along with diffusivity values of 2.68 and 5.31 mm² s⁻¹, are in agreement with characteristics typical of insulating materials, confirming the potential application of this material as a thermal insulator. In comparison, commercial wood wool cement boards like Drvolit D, which are ordinarily used as insulating material in civil engineering, have a higher thermal conductivity value of approximately 0.074 W m⁻¹ K⁻¹, which indicates better thermal insulation of prepared biocomposites. Furthermore, the compressive strengths of C1L-KM and C1B-KM samples, at 4.15 and 6.41 MPa, respectively, are significantly higher than the compressive strength of Drvolit D, which is around 0.15 MPa [57].

4. Conclusions

This study successfully demonstrated the potential of waste almond and hazelnut nutshells as particulate fillers in geopolymer-based biocomposites. Through systematic analysis, the optimal conditions for biocomposite preparation were identified, including a geopolymer-to-nutshell mass ratio of 2, room-temperature curing, and the use of metakaolin activated with potassium solutions. The application of alkaline pretreatments, particularly lime milk treatment, enhanced the adhesion between the nutshells and the geopolymer matrix, contributing to improved mechanical properties. The highest compressive strength values of 4.15 MPa for hazelnut shell biocomposites and 6.41 MPa for almond shell biocomposites were achieved with lime milk pretreatment at 80 °C for 1 h. FTIR and SEM-EDS analyses confirmed the beneficial effects of calcium incorporation, suggesting that precipitated calcium phases positively influenced geopolymer gel polymerization and matrix connectivity. Thermal conductivity and diffusivity values were consistent with those of insulating materials, emphasizing the usability of these biocomposites for thermal insulation applications. Overall, this research highlights the feasibility of repurposing waste nutshells into value-added materials, promoting sustainability and eco-friendly alternatives in construction and engineering. Future studies may focus on optimizing or combining pretreatment methods and further exploring the impact of biocomposite type and shape on its properties.