3.1. Raw Materials
Natural fibers are composed of cellulose, lignin, and hemicellulose, in addition to other minor components, such as water, proteins, peptides, and inorganic compounds [
59,
60].
Table 2 presents the results of the chemical characterization of henequen and palm fiber. The results indicate that the primary component of each fiber is cellulose, with a percentage composition of 33.64% for henequen and 46.43% for palm fiber. The difference in results is related to the intrinsic characteristics of each species, the part of the plant being analyzed, as well as environmental aspects [
61]. For example, the lignin content is higher in the palm leaf fiber, corresponding to 36.04%. This can be associated with a higher fiber stiffness, since the henequen filaments are extracted from the pith of the plant and therefore from the less stiff zone [
14,
39]. Furthermore, non-volatile compounds extractable with organic solvents were identified in 7.59% of the henequen filaments and 13.33% of the palm fibers.
Table 3 provides an overview of the most used fibers that have been subjected to chemical characterization. The cellulose content of henequen fiber is comparable to that reported for hemp [
62,
63], oats [
30], and coconut [
64]. In contrast, palm fiber has a percentage similar to that reported for flax [
65,
66], sunflower [
63,
65], wheat straw [
63], and sisal [
30,
64].
The intrinsic thermal conductivity and density of both raw materials were measured to ascertain their suitability for use in the development of a thermal insulation material. As shown in
Figure 3, the density of palm fiber was found to be 0.8289 ± 0.0046 g/cm
3, while its thermal conductivity was determined to be 0.0409 ± 0.0071 W/mK. These values are comparable to those obtained by sunflower fiber [
70,
71] and coconut fiber [
51,
64,
72,
73,
74]. In contrast, henequen fiber exhibited a density of 1.2890 ± 0.0581 g/cm
3 and a thermal conductivity of 0.0255 ± 0.0011 W/mK, rendering it comparable to flax fiber [
27,
51,
70,
73,
74,
75]. The values observed in both fibers are suitable for their application in thermal insulation, as they are comparable with those reported for other fibers that have been investigated for similar applications, including sugarcane bagasse [
13], rubber, recycled paper [
73,
74], reed [
76], and sargassum [
77]. In general terms, both values fall within the characteristic range of lignocellulosic fibers. However, the density is considerably higher than that of conventional insulating materials such as expanded polystyrene, which has a density range of 10 to 30 kg/m
3. It is therefore necessary to conduct further research to determine whether this has any effect on thermal inertia [
5].
Lignocellulosic fiber structures exhibit three distinct types of ordered and stable molecular conformations, with cellulose being the most prevalent [
61].
Figure 4 illustrates the X-ray diffraction curves obtained for the two raw materials under study. Broad signals were identified which are associated with the cellulose crystals that are characteristic of natural vegetable fibers [
52,
78,
79].
The graphs of henequen and palm leaf fiber demonstrate that the maximum peaks recorded are 196,880 u.a. and 198,430 u.a., respectively. The intensities at 2θ = 18° are 144,444 and 124,988, resulting in a percentage of crystallinity of 26.6355% for henequen and 37.0156% for palm leaf fiber [
78]. The calculated values are relatively low in comparison with other fibers, such as hemp (87.9%) or sisal (75%), and can be attributed to the fact that they correspond to non-structural parts of the plant. However, for palm fiber, the result is consistent with that reported in the literature for a similar fiber (Borassus fiber with a percentage of crystallinity of 38.4%) [
23,
26]. The mineralogical composition indicates that both fibers exhibit a notable presence of sulfur trioxide (SO
3), potassium oxide (K
2O), calcium oxide (CaO), and iron oxide (Fe
2O
3). The values are presented in
Table 4.
The primary components of plant fiber are α-cellulose, hemicellulose, and lignin, which were identified as being characteristic of plant fibers in the spectrum between 400 and 4000 cm
−1 of the FTIR test, as illustrated in
Figure 5.
The graphs obtained for both fibers are characteristic of lignocellulosic fibers; however, the magnitudes differ. The band recorded at 1027 cm
−1 is indicative of C-O-C stretching in the xylan of hemicellulose [
80]. The absorption band at 1700 cm
−1 refers to a stretching of a C=O double bond, which is associated with acid-like carboxyl groups that are generally present in hemicellulose. The signal observed at approximately 2920 cm
−1 corresponds to the stretching vibration of the C–H group [
81]. Furthermore, a robust and extensive absorption band is observed between 3650 and 3000 cm
−1, with a peak at 3350 cm
−1. Additionally, a moderate peak at 1410 cm
−1 corresponds to O–H stretching vibrations, which are indicative of the presence of hydroxyl groups in compounds such as cellulose, hemicellulose, and lignin [
26,
80].
The results of the TGA are presented in
Figure 6, which depicts three principal mass losses associated with distinct processes as the temperature increases. The test indicates that the initial mass loss process occurs between room temperature and approximately 125 °C, with a loss of 6.1% for both fibers. This process is associated with dehydration and the loss of volatile substances [
80]. In this temperature range, the fiber exhibits thermal stability associated with its chemical characterization (mainly lignin content) and its crystallinity index. Consequently, this value is lower than that reported for other fibers, such as blue agave, Brazilian banana, and sugarcane bagasse. However, it allows for the safe use of fiber in thermal insulation applications in residential buildings [
52]. Subsequently, the most significant degradation occurs between 125 and 375 °C, which is associated with the elimination of biopolymers and the decarbonization of important substances, beginning with hemicellulose and continuing with α-cellulose. Finally, a gradual degradation occurs up to 900 °C, corresponding to the loss of fixed carbon, which is mainly associated with lignin [
39,
82].
Figure 7a–d corresponds to micrographs of the henequen fiber. As shown in
Figure 7a,b, the cross-section reveals a highly compact cell wall structure with an approximate fiber diameter of 480 to 500 μm. Consequently, the low bulk density of the resulting composite is attributed to the interweaving of the strands, which allows for the formation of air cavities. The surface of the fibers (
Figure 7c,d) exhibits a consistent degree of roughness, which would facilitate adhesion in its use as reinforcement in the manufacture of composites with a binder. It is therefore recommended to inquire into this type of fiber for multiple applications, and it also displays minor defects that are typically associated with the extraction and handling process [
52,
53,
82,
83].
Figure 7e–h depicts micrographs of palm leaf fiber, which exhibits a markedly distinct morphology, with a thickness ranging from 200 to 225 μm. The cross-section (
Figure 7e,f) reveals the presence of minute cavities with diameters between 10 and 20 μm, which could serve as reservoirs for air bubbles, enhancing the fiber’s insulating capacity. In contrast,
Figure 7g,h of the longitudinal plane (fiber surface) demonstrate a regular structure with minimal roughness and the presence of small impurities, which may be calcium oxalates, a common feature of natural vegetable fibers [
53,
82,
83].
3.2. Experimental Design
The sample manufacturing procedure involves boiling the fiber in a sodium hydroxide solution at varying concentrations, which results in a modification of the manner in which the hydrogen bond is removed from the network, as described by Equation (4) [
35,
84].
The consequences of exposing the fiber to the alkaline solution are the removal of surface impurities such as waxes and other non-cellulosic materials, which generates an external, rougher structure. Moreover, the fiber bundles are fragmented into smaller components in a process referred to as defibrillation [
84]. The chemical composition of the fiber has been reported to undergo a general reduction in the percentage contents of hemicellulose and lignin according to the duration of the treatment. Conversely, the cellulose content is observed to increase [
85,
86].
A Fourier transform infrared (FTIR) analysis was conducted on both fibers prior to and following their exposure to the alkaline solution (
Figure 8). In their original state, both fibers exhibited an O-H stretching band at 3350 cm
−1, which is indicative of O-H bonding. This bonding is primarily associated with cellulose vibrations and hydrogen bonds of the hydroxyl groups, as previously reported in the literature [
39]. The band at 2920 cm
−1 indicated the presence of C-H bonds of methyl and methylene groups. The peak at 1700 cm
−1 indicates the presence of C=O stretching in the acetyl groups of hemicelluloses, which is subsequently absent in alkali-treated fibers. This observation is consistent with the removal of hemicelluloses, as mentioned [
87]. A similar phenomenon occurs with the peak visible at 1244 cm
−1, which is associated with C-O stretching in hemicelluloses. This peak was subsequently not visible, confirming the lack of hemicelluloses. The band at 1027 cm
−1 exhibited a gradual decrease with alkaline treatment, indicative of a reduction in lignin content [
61,
83,
84,
86].
The removal of surface compounds from the fiber allows for the exposure of the cellulose fibrils, which can interact with each other. The alkaline solution depolymerizes the cellulose molecules, breaking intramolecular hydrogen bonds and facilitating the modification of these bonds. This modification allows for the formation of non-covalent interactions, which are susceptible to the presence of moisture [
60]. This reaction allows for the agglomeration of the fiber without the incorporation of additional external substances and confers rigidity to the developed material. The degree of rigidity is dependent on the NaOH concentration, the boiling time, and the blending time.
3.2.1. Orthogonal Array
The study, conducted using the Taguchi L9 matrix, revealed that, for each raw material, the optimal prototype corresponded to the third arrangement, as detailed in
Table 5.
The results demonstrated that, in regard to palm-based material density, only fiber length was a significant contributor, with 1.54% (p < 0.05; see
Table 6a). For thermal conductivity for palm fiber, the fiber length contributed to 24.65% (p < 0.001), followed by blending time with 22.73% (p < 0.001), NaOH concentration with 13.42% (p < 0.001), and boiling time with 8.5% of contribution (p < 0.01; see
Table 6b). In the case of henequen-based material density, the boiling time was the most significant contributor with 24.21% (p < 0.001), NaOH concentration 19.6% (p < 0.001), blending time 18.58% (p < 0.001), and fiber length 14.32% (p < 0.001; see
Table 6c). Concerning the thermal conductivity of the henequen fiber, it can be observed that there was no significant contribution (see
Table 6d).
Another crucial aspect pertains to the yield of the fiber introduced into the process in comparison to that which results from the manufacturing of the prototype. In the case of the optimal prototype based on henequen fiber, its yield is 84.47%. In contrast, the yield of the prototype based on palm fiber is 52.57%. The recorded loss is associated with the manufacturing process. On the one hand, there is the boiling phase in sodium hydroxide where hemicellulose is hydrolyzed, impurities are extracted, and lignin is solubilized from the fiber. On the other hand, material losses are associated with the manufacturing process [
39].
3.2.2. Density, Porosity, and Thermal Conductivity
For both optimal prototypes, the bulk density was calculated to be 0.0698 ± 0.006 g/cm
3 for the henequen and 0.04506 ± 0.00616 g/cm
3 for the palm fiber prototype. The discrepancy in the values can be attributed to the inherent characteristics of each raw material, as well as the efficacy of the pulping process, which facilitates the alignment of the fibers to form the prototype. The results of the fiber density and the bulk density of the prototypes were employed to ascertain the porosity of each optimal sample, which impacts the insulating capacity of the material. In the case of the henequen-based insulating, the porosity was determined to be 94.5849%, while for the palm-fiber-based material, the porosity achieved was 94.5639%. These values, although relatively high, are advantageous in dissipating heat transfer due to the ability to contain airtight air, which has a thermal conductivity of 0.028 W/mK [
88], resulting in a low weighted conductivity. This phenomenon can be attributed to the mechanisms of heat transfer in porous insulating materials. These mechanisms include heat conduction from the solid network (fiber) and the gas phase (air), convection of heat from moving air in the pores, and radiation between the fibers [
89]. Consequently, the high porosity of the samples enables the attainment of thermal conductivity of the optimized sisal and palm prototypes of 0.0353 ± 0.0019 W/mK and 0.0367 ± 0.0012, respectively. The outcomes permit the establishment of a positive linear correlation between the density and thermal conductivity of the insulating materials, as illustrated in
Figure 9. This is because, as the solid phase fraction increases with decreasing porosity, the contribution of solid conduction increases with increasing density [
89]. The values obtained for both developed prototypes are comparable to conventional thermal insulation materials such as expanded polystyrene (0.0361 W/mK) and even lower than mineral wool (0.045 W/mK) [
14,
70,
88]. On the other hand, when compared with other lignocellulose-based insulation materials, it can be observed that the thermal conductivity is similar to that of the insulator based on hemp, flax, and sheep wool. These materials have been widely used for this type of application and have lower thermal conductivity than insulators based on straw, sugar cane bagasse, and wood fibers. This suggests that these raw materials have great potential for use in construction in regard achieving aspects of sustainability and energy efficiency [
5,
27,
51,
70,
73,
74,
75].
3.2.3. Flexural Strength
Figure 10 presents the outcomes of the bending test conducted on the optimized insulating material prototypes. The results indicate that the henequen prototype exhibits a value of 0.0129 ± 0.0006 MPa at 5% deformation, with a modulus of elasticity of 0.2486 MPa. The maximum stress observed for the palm fiber prototype was 0.01827 ± 0.0009 MPa, with a modulus of elasticity of 0.4059 MPa. These results are compared with those of a prototype of expanded polystyrene and a prototype of a eucalyptus-bark-based insulator, both of which have been previously reported in the literature [
39]. Furthermore, it can be observed that the behavior of the natural insulators is quite similar, with a maximum stress value much lower than that reported for EPS (0.20 ± 0.01 MPa). Therefore, it is advisable to conduct a more detailed study to limit their applications in construction elements according to their intrinsic properties. Although the flexural properties are reduced, the material could be employed similarly to that of non-rigid insulators such as sheep wool or mineral wool.