Extraction and Modification of Cellulose Microfibers Derived from Biomass of the Amazon Ochroma pyramidale Fruit

Abstract

Microfibers are important to several areas of human lifestyle, and the knowledge about their physicochemical characteristics allows for proposing new technological applications. The in natura microfiber of Ochroma pyramidale fruit (IN sample) and its extracted pulp (PU sample) were evaluated by X-ray Diffraction (XRD), Scanning Electron Microscopy (SEM), Fourier Transform Infrared Spectroscopy (FTIR) and Thermogravimetry and Differential Scanning Calorimetry (TG/dTG and DSC). Microfibers were composed mainly of (68 ± 1)% holocellulose, (35.8 ± 0.1)% cellulose, (32 ± 3)% lignin and (3.7 ± 0.3)% extractives. The XRD pattern of the PU sample revealed that the mercerization process resulted in the change of the cellulose crystal structure from Iα type (triclinic) to type II (monoclinic). The SEM technique showed that the IN sample presented regular cylindrical/hollow-shaped wire-like microfibers with diameters ranging from 5 µm to 25 µm. However, the mercerization process changed their natural morphology. A significant change in the FTIR spectra after the removal of hemicellulose and lignin components was observed: weak bands at 1739 cm⁻¹ (C=O stretching of lignin and hemicellulose fractions), 1463 cm⁻¹ (CH₃ of lignin) and 1246 cm⁻¹ (C-O of lignin) were still observed in the PU sample, indicating that the lignin was not completely removed due to the natural difficulty of isolating pure cellulose. The TG/dTG and DSC evaluation revealed a temperature increase of the second thermal event (starting at 235 °C) in the PU sample, which was assigned to the cellulose and residual hemicellulose degradation. Then, this work aimed to disseminate and characterize a microfiber with unusual characteristics still little explored by the scientific community, as well as its cellulosic pulp, providing information that may be useful in its application in different industries, enabling the positive development of new biocompatible, renewable and sustainable materials.

1. Introduction

The demand for sustainable and environmentally friendly products has been increasingly becoming a new global concern. In this context, lignocellulosic sources have been considered an alternative material to nondegradable fossil-fuel-based polymers due to its abundance and biodegradability, as well as renewability, large specific surface area, low cost and rich chemical functional groups [1,2]. Cellulose represents the major compound of the lignocellulosic biomass, as well as the most abundant natural biopolymer. The chemical structure of cellulose is constituted of β-1,4-linked glucopyranose units, which bear three hydroxyl groups. The search for friendly materials has made cellulose one of the most studied natural polymers, especially because it can be extracted from a variety of sources and crystal structures [3,4].

Ochroma pyramidale (Cav. ex Lamb.) Urban, known in Brazil as “pau-de-balsa”, is a Malvaceae specie widely distributed from southern Mexico to Bolivia and Brazilian Amazon rainforest, and is well known for its low-density wood. It can reach 25 m in height and 1.2 m in diameter. Its flowers appear in the rainy season (April–July) and the fruits in the dry season (July–October). Figure 1a shows the O. pyramidale fruits, which are loculicidal, almost cylindrical, ligneous and dehiscent. These fruits present a very soft yellow-brown inner fiber (Figure 1b), which promotes the dispersion of seeds by the wind [5]. Reports on O. pyramidale have usually considered its wood [6,7,8], seed germination [9], or use for reforestation due to its fast growth [6,10]. Furthermore, some works also have reported its use as filling for beds, pillows and lifeguards due to its softness and buoyancy [11]. On the other hand, another specie from the same family known as kapok (Ceiba pentandra L.) also presents a fruit with extremely similar morphological fiber. Reports are giving attention to its application as biofuels [12,13], oil absorbents [11,14,15], acoustic and thermal insulation [11,16], thermoplastic composites [17], paper and textile yield, polymeric matrix and even drug release [18], suggesting similar application potential of O. pyramidale [16].

Lignocellulosic fibers have been extensively evaluated as reinforcement in thermoplastic, polymeric composites and composites for building [1,19,20,21], oil adsorbents [22,23], bioplastics [24], biofilters [25], biofuel [26], acoustic and thermal insulators [27,28], chemical products and cosmetic [2] and biomedical application [3,29,30]. The study of new fibers represents an important topic of research due to the possibility of proposing new technological applications. In this context, producing new data about O. pyramidale fruit fibers proves to be a good proposal. For this reason, the aim of this paper was to perform a systematic physicochemical, structural, morphological and thermal investigation of the O. pyramidale fruit fibers: X-ray Diffraction technique (XRD) was applied in order to examine the long-range order achieved as a consequence of very short-range interactions. Fourier Transform Infrared Spectroscopy (FTIR) was applied for molecular structural characterization, as well as to confirm the success of the proposed method of cellulose extraction. Scanning Electron Microscopy (SEM) was useful to assess the fibers’ morphology. Finally, Thermogravimetric analysis (TG/dTG) and Differential Scanning Calorimetry (DSC) were conducted to evaluate the thermal stability of the fibers and pulp extracted from the Amazon Ochroma pyramidale fruit.

2. Experimental

2.1. Materials

Plant Material

O. pyramidale fruits (SISGEN n° A26CD5E) were collected at the Instituto Federal de Educação, Ciência e Tecnologia do Amazonas (IFAM)—Campus Zona Leste, Manaus/AM (3°4′51.8268 S and 59°56′2.9328 W). Fibers in natura were separated from seeds and maintained at 25 °C until further analysis (labeled as IN sample). Botanical identification was carried out at the Federal University of Amazonas (UFAM) following the established protocol.

2.2. Methods

2.2.1. Physicochemical Composition

The physicochemical composition of the IN sample was performed according to adapted TAPPI methodologies [31]. All measurements were performed in triplicate.

To measure the Moisture Content (M%), 2 g (m_i) of the IN sample were submitted to a thermal treatment in an oven at 105 °C for 3 h, according to TAPPI 412 om-16 [31]. M% was determined by Equation (1):

$$\left(\frac{m_i-m_f}{m_i}\right)\times 100=M\%$$

(1)

where m_i and m_f represent, respectively, the initial and final mass of the IN sample.

An amount of 0.3 g (m_i) of dried IN sample was carbonized in an industrial oven on asbestos leaf for 1 h. Then, the resulting material was placed in a muffle under a heating rate of 9.6 °C/min for 3 h up to 600 °C to measure the Ash Content (A%) [32]. A% was determined by Equation (2):

$$\left(\frac{m_a}{m_i}\right)\times 100=A\%$$

(2)

where m_a represents the mass of ashes.

For Cold-Water Solubility (CW%) measurement, 2 g of dried IN sample were added in water (300 mL) at 25 °C for 48 h under constant stirring and then washed with 200 mL of distilled water and filtered. For the Hot-Water Solubility (HW%) measurement, 2 g of dried IN sample were added in boiling water (200 mL) for 3 h, washed with 200 mL of boiling distilled water and filtered. Both water-treated fibers were placed in an oven at 103 °C until they reached constant weight [33]. Then, the CW% and HW% were determined by Equation (1).

To remove and evaluate the Extractives Content (E%) as a previous process, 4 g of dried IN sample were submitted to solvent extraction using a Soxhlet system using 180 mL of ethanol and acetone 2:1 at 70 °C for 4 h. The solvent with extractives was dried in an oven at 105 °C until constant weight. E% was determined according to Equation (2), but considering the mass of extractives, and the Extractives Content Corrected (E_cor%) was measured following Equation (3) [34]. Then, the treated and dried fibers were submitted to the extraction of holocellulose, cellulose and lignin.

$$\left(\frac{E\%\times (100-M\%)}{100}\right)=E_{cor}\%$$

(3)

To determine the Lignin Content (L%), 1 g (m_i) of treated fibers (without moisture and extractives) was added to H₂SO₄ (17 mL, 72% v/v) for 24 h at 25 °C. Then, the system was diluted to 4% using distilled water and agitated at 70 °C for 4 h. The resulting material was filtered using a sintered filter type 3 and maintained in an oven at 105 °C until constant weight [35]. L% was determined according to Equation (4), and Lignin Content Corrected (L_cor%) was measured following Equation (5):

$$\left(\frac{m_{lig}}{m_i}\times 100\right)-A\%=L\%$$

(4)

$$\left(\frac{L\%\times [100-\left(M\%+E_{cor}\%\right)]}{100}\right)=L_{cor}\%$$

(5)

where m_lig represents the mass of lignin.

To determine Holocellulose Corrected (Ho_cor%), Hemicellulose Corrected (He_cor%) and Cellulose Corrected (C_cor%) contents, 3 g of the treated fibers (without moisture or extractives) were submitted to a system containing NaClO₂ (5 g) and CH₃COOH (2 mL) in distilled water (240 mL) at 70 °C for 1 h for oxidation of lignin and reapplied for another hour aiming the obtainment of holocellulose. Then, 1 g of the obtained holocellulose (m_ho) was submitted to NaOH (15 mL, 17.5% m/m, 10 min) under maceration, and then distilled water (40 mL) was added. The final solution was filtered on a sintered filter type 2 to obtain the pulp material (labeled as PU sample) [36]. The obtained holocellulose and cellulose were dried in an oven at 105 °C until constant weight. Ho% and Ho_cor% were determined according to Equations (2) and (5), respectively. C%, C_cor% and He_cor% were measured using the following equations:

$$\left(\frac{m_c}{m_{ho}}\right)\times 100=C\%$$

(6)

$$\left(\frac{H{o}_{cor}\%\times C\%}{100}\right)=C_{cor}\%$$

(7)

$$H{o}_{cor}\%-C_{cor}\%=H{e}_{cor}\%$$

(8)

where m_ho, m_c and He_cor% represent, respectively, the mass of holocellulose, mass of cellulose and Hemicellulose Content Corrected.

2.2.2. X-ray Diffraction (XRD) and Percentage of Crystallinity

X-ray diffraction technique (XRD) was performed on a Panalytical Empyrean diffractometer, CuK_α, 40 kV and 40 mA. Measurements were performed using a Si single-crystal sample holder from 5° to 100° (2θ), step of 0.01313° and 5 s/step. A LaB₆ standard NIST (660 b) was used to consider the instrumental effects. The Rietveld method implemented in the GSAS-2 software package was used to refine structural parameters and line widths, following the recommendations of IUCr [37]. Crystal data of cellulose polymorphs I_α (triclinic; a = 10.400 Å; b = 6.717 Å; c = 5.962 Å; α = 80.370°; β = 118.080° and γ = 114.800°) [38] and type II (monoclinic; a = 8.100 Å; b = 9.030 Å; c = 10.310 Å; α = 90°; β = 90° and γ = 117.100°) [39] were considered as initial parameters. The percentage of crystallinity was measured using the deconvolution method [40,41].

2.2.3. Scanning Electron Microscopy (SEM)

SEM experiments were performed on a Carl Zeiss equipment model Supra 35, using 1.0 kV at 25 °C. Samples were placed on carbon tape and covered with a thin gold layer prior to analysis. The fiber diameter distribution was determined using the software ImageJ 1.53 t [42] and statistical program OriginPro 8^® [43] considering 500 units of IN sample.

2.2.4. Fourier Transform Infrared Spectroscopy (FTIR)

FTIR spectrum was recorded using a Shimadzu IR Prestige-21 Spectrometer TA Instruments^® from 4000 to 400 cm⁻¹, resolution of 1 cm⁻¹ and 64 scans.

2.2.5. Thermogravimetry/Derivative Thermogravimetry (TG/dTG) and Differential Scanning Calorimetry (DSC)

TG/dTG and DSC techniques were performed on a SDT Q600 TA Instruments at the Laboratório de Materiais da Amazônia e Compósitos (LaMAC-FT/UFAM). Measurements were carried out using approximately 5.5 mg of samples in alumina crucibles, N₂ atmosphere (flow of 30 mL/min) at a heating rate of 10 °C/min, from 20 °C to 600 °C.

2.2.6. Thermal Conductivity

The TCi equipment, developed by the C-Therm Company, Fredericton, NB, Canada, was utilized to perform thermal conductivity measurements. This equipment employs the modified transient plane source (MTPS) method to determine the thermal conductivity of the sample. The measurements were conducted at room temperature, and microfibers served as the reference material in all instances. The correlation factor R², which assesses the agreement between the experimental results and the reference material, approached 99%.

3. Results and Discussion

3.1. Physicochemical Characterization

Figure 1 shows the parts composing the O. pyramidale fruits, including their microfibers and seeds. Structural and nonstructural compounds were extracted from their microfibers and quantified.

Table 1 shows the main percentual compounds of the IN sample and other wood and nonwood values for comparation. The values shown are the corrected contents.

Holocellulose (Ho%) considers the sum of hemicellulose and cellulose. A value of (68 ± 1)% was obtained, which was similar to those found in E. binata (64%) [44], P. dactylifera (70%) [45] and H. sabdariffa (74%) [46]. The presence of hemicellulose in a bleached sample may be useful depending on the purpose and can be an alternative to the pure cellulose due to its easier obtention, lower cost and higher yield. Purnawati et al. 2018 [16] and Prachayawarakorn et al. 2013 [17] obtained Ho% values higher than 80% in balsa and kapok fruit fibers. On the other hand, the cellulose value of (35.8 ± 0.1)% represented one of the smallest values shown in Table 1, but similar to those found in C. pentandra (38.09%) [16], C. nucifera (35.62%) [23] and P. dactylifera (35–44%) [45]. The L% value was found around (32 ± 3)%, which was almost double the value found in balsa fruit [16] and one of the highest lignin values shown in Table 1. This characteristic assures its high hydrophobicity acting synergically with the extractives, as observed in Figure 1d. The difference between values from Table 1 can be explained, because these natural constituents are significantly related to the soil type, weather, plant age, harvest, maturation, presence of pests or diseases as well as methods of extraction, reagents and manipulation [47,48,49]. The A% value (2.03 ± 0.03)% agrees with that found for the residual mass obtained through TG/dTG analysis.

The O. pyramidale fruit fiber presented a surface constituted of extractives (E% = 3.7 ± 0.3) [16] based on secondary metabolites. Despite presenting a low concentration of extractives, their association to lignin allows for a hydrophobic–lipophilic fiber’s property [15,17] preventing their interaction to water [50]. Similar behavior was observed in kapok fibers [15]. This information is consistent with their role in nature: the high concentration of secondary metabolites is not necessary, because their function, as far as we know, is just to disperse their seeds by the wind.

Considering the use of nonlipophilic fibers for oleophilic purposes, a chemical surface modification might be applied [14]. However, in the case of O. pyramidale and C. pentandra fruit fibers, this modification may not be necessary, representing an important advantage when compared to other plant fibers.

Table 1. Main compounds of the O. pyramidale fibers (in natura) compared to other wood and nonwood sources.

Specie	M%	E%	He%	C%	L%	A%	Reference
Balsa fruit fiber (Ochroma pyramidale)	8.7 ± 0.2	3.7 ± 0.3	33 ± 1	35.8 ± 0.1	32 ± 3	2.04 ± 0.03	(Present work)
Balsa fruit fiber (Ochroma pyramidale)	11.45	2.29	37.35	44.62	16.6	0.94	[16]
Kapok fruit fiber (Ceiba pentandra)	11.23	2.34	45.64	38.09	14.1	1.05	[16]
Kapok bark (Ceiba pentandra)	7.46	0.38	17.53	60.9	23.5	1.05	[51]
Kapok fruit fiber (Ceiba pentandra)	–	23	23	64	19	–	[17]
Coconut fiber (Cocos nucifera)	–	–	8.51	35.62	37.59	0.97	[23]
Jute stem (Corchorus capsularis)	–	0.5	14–20	61–71	12–13	–	[51]
Sisal leaf (Agave sisalana)	–	2	12	65	9.9	–	[51]
Kenaf stem (Hibiscus cannabinus)	–	0.3	20.3	72	9	–	[51]
Oil palm leaves (Elaeis guineensis)	–	–	34	42.67	22.9	–	[51]
Umbrella thorn bark (Acacia tortilis)	6.47	17.43	–	61.89	21.26	4.33	[1]
Buriti leaf fiber (Mauritia flexuosa)	9	6	1	58	19	2	[19]
Date palm (Phoenix dactylifera)	5.4–15.6	–	9.75–26	35–44	11–29	3–12	[45]
Sabai grass (Eulaliopsis binata)	–	–	21.1	42.9	18.5	13.4	[44]
Portia tree bark (Thepesia populnea)	9.8–11.5	0.7–0.8	12–16	64–70	16–18	1.7–2.1	[52]
Roselle stems (Hibiscus sabdariffa)	–	–	16–20	58–64	6–10	–	[46]

Table 1. Main compounds of the O. pyramidale fibers (in natura) compared to other wood and nonwood sources.

Specie	M%	E%	He%	C%	L%	A%	Reference
Balsa fruit fiber (Ochroma pyramidale)	8.7 ± 0.2	3.7 ± 0.3	33 ± 1	35.8 ± 0.1	32 ± 3	2.04 ± 0.03	(Present work)
Balsa fruit fiber (Ochroma pyramidale)	11.45	2.29	37.35	44.62	16.6	0.94	[16]
Kapok fruit fiber (Ceiba pentandra)	11.23	2.34	45.64	38.09	14.1	1.05	[16]
Kapok bark (Ceiba pentandra)	7.46	0.38	17.53	60.9	23.5	1.05	[51]
Kapok fruit fiber (Ceiba pentandra)	–	23	23	64	19	–	[17]
Coconut fiber (Cocos nucifera)	–	–	8.51	35.62	37.59	0.97	[23]
Jute stem (Corchorus capsularis)	–	0.5	14–20	61–71	12–13	–	[51]
Sisal leaf (Agave sisalana)	–	2	12	65	9.9	–	[51]
Kenaf stem (Hibiscus cannabinus)	–	0.3	20.3	72	9	–	[51]
Oil palm leaves (Elaeis guineensis)	–	–	34	42.67	22.9	–	[51]
Umbrella thorn bark (Acacia tortilis)	6.47	17.43	–	61.89	21.26	4.33	[1]
Buriti leaf fiber (Mauritia flexuosa)	9	6	1	58	19	2	[19]
Date palm (Phoenix dactylifera)	5.4–15.6	–	9.75–26	35–44	11–29	3–12	[45]
Sabai grass (Eulaliopsis binata)	–	–	21.1	42.9	18.5	13.4	[44]
Portia tree bark (Thepesia populnea)	9.8–11.5	0.7–0.8	12–16	64–70	16–18	1.7–2.1	[52]
Roselle stems (Hibiscus sabdariffa)	–	–	16–20	58–64	6–10	–	[46]

Purnawati et al. 2018 [16] performed the fiber’s wettability analysis and obtained angles higher than 118°, confirming their high hydrophobicity. On the other hand, through CW% (7.5 ± 0.6%) and HW% (8.4 ± 0.2%) treatments performed in this present work, a slight solubility was suggested, but not identified in the immediate contact fibers water, requiring the use of a mechanic (CW%) or thermal action (HW%).

3.2. XRD Analysis

XRD analysis was carried out as a tool to investigate the crystal structure of both IN and PU samples. Figure 2a,b shows the overlapped diffraction patterns.

Cellulose polymorphism has been associated with the source and extraction method [53,54]. The cellulose polymorph in most plants is generally Iβ type (monoclinic) [52], but both Iα and Iβ polymorphs can coexist in this material, presenting diffraction peaks at similar angular positions. Therefore, depending on the chemical procedure, some extracted cellulose can also present the Iα type (triclinic) after refinement, as previously reported [53].

All XRD measurements were performed under the same experimental conditions. Thus, the differences between the intensities of the XRD patterns were attributed to different crystallinity levels. Despite presenting lignin, hemicellulose and partial cellulose as noncrystalline phases, the IN sample also presented the contribution of the crystalline cellulose phase. Naturally, bulk cellulose consists of highly ordered, crystalline regions disposed along disordered (amorphous/noncrystalline) regions in varying proportions, which depends on the cellulose source [55].

The XRD pattern of the IN sample was accurately represented by the Iα type cellulose (triclinic) [16]. The peaks located at 15.9°, 22.4° and 35.5° (2θ) were related to (001), (011) and (300) plans, respectively (Figure 2a). However, the XRD pattern of the PU sample (Figure 2b) presented a considerable phase changing. The peaks located at 12°, 20° and 22° corresponded, respectively, to (010), (110) and (020) plans. The mercerization process has been considered an irreversible method to extract cellulose able to change its crystal structure. In our case, the cellulose crystal structure Iα changed to type II (monoclinic) [24,54].

Table 2. Cell parameters obtained from Rietveld refinement.

	IN Sample [38]	IN Sample	PU Sample [39]	PU Sample
a (Å)	10.4	10.49 ± 0.06	8.10	7.87 ± 0.01
b (Å)	6.717	6.74 ± 0.03	9.03	9.01 ± 0.01
c (Å)	5.962	6.26 ± 0.02	10.31	10.19 ± 0.02
α (°)	80.37	76.3 ± 0.3	90	90
β (°)	118.08	112.5 ± 0.4	90	90
γ (°)	114.8	127.4 ± 0.2	117.1	117.77 ± 0.07
Volume (Å3)	-	325 ± 2	-	641 ± 2
χ2	-	1.104	-	1.205
Rwp (%)	-	0.0862	-	0.1091

shows the lattice parameters obtained by the Rietveld refinement.

The percentage of crystallinity of the PU sample was estimated using the deconvolution method. A value of 87% was estimated, which was similar to that found in a previous report [56], suggesting that the alkaline pretreatment considerably removed the noncrystalline contribution [57]. However, this method was not applied to estimate the crystallinity of the IN sample due to the air- scattering and their noncrystalline structural components, which may influence the XRD pattern.

3.3. Morphological Analysis

The SEM technique was performed to investigate the surface morphology of both IN and PU samples. Figure 3a,b shows that the IN sample presented a regular wire-like shape with a smooth surface. The large fiber’s length size did not accurately allow their measurements. However, [16] reported their range from 10.36 nm to 15.7 mm. The IN sample presented diameters ranging from 5 µm to 25 µm (with average diameter of 15 µm). Furthermore, Figure 3c,d shows that the IN sample presented a tube-like morphology, characterized by its cylindrical/hollow shape and slightly roughened inner surface. The high percentage of lignin may be responsible for the observed cylindrical/hollow morphology due to its role in the support of vegetable wall cell structure [47,58]. Figure 3e,f shows the fiber’s morphology after rupture, showing well-defined and interconnected microfibrils of cellulose composing the tubes. This morphology is similar to that of kapok fruit fibers [11,14,18], described previously as a hollow tube with porosity around 80% [15]. This characteristic can differentiate it from other vegetable fibers, which are generally massive but porous-like sponges.

The removal of lignin, hemicellulose and other components by bleaching treatment can be observed in the PU sample (Figure 4a,b). Fibers lost their original cylindrical/hollow morphology (as shown in Figure 4c,d), exposing their rough surface (Figure 4e,f) in a random-fold fashion. This characteristic is similar to that of cellulose reported previously [46,59]. Different methods of cellulose extraction have been proposed, showing their influence on the material surface [60] and physicochemical properties, as is also observed in the present work.

3.4. FTIR Analysis

Figure 5 shows the FTIR spectra of both IN and PU samples and the summarized peaks are at Table 3. The bands resulting from the chemical bonds of the IN sample were observed (considering all lignocellulosic components). A significant change in the spectral profile of the PU sample after the removal of hemicellulose and lignin components was observed.

Bands related to the –OH vibration were observed around 3500 cm⁻¹, which may result from the presence of water associated with the free hydroxyl groups in the main chain of cellulose [61]. The main region of interest in lignocellulosic materials ranged from 1800 cm⁻¹ to 800 cm⁻¹, where the main characteristic bands are usually located.

The bands at 1597 cm⁻¹ and 1503 cm⁻¹ (C=C of lignin) [46,50,55,60] presented in the IN sample completely disappeared in the PU sample. On the other hand, the bands at 1739 cm⁻¹ (C=O stretching of lignin and hemicellulose fractions) [55,59,61], 1463 cm⁻¹ (CH₃ of lignin) [19,62] and 1246 cm⁻¹ (C-O of lignin) [55,59,63] were still observed. However, their intensities were considerably reduced: the first two bands presented only a discrete shoulder in the PU sample spectrum, indicating that the lignin was not completely removed. This fact may be explained by the difficulty of isolating pure cellulose from vegetable sources due to the complex chemical composition of lignin. This point was explored in the TG/dTG and DSC analysis. The bands at 1337 cm⁻¹, 1327 cm⁻¹ and 1282 cm⁻¹ (CH₂ wagging and C–O aromatic ring of cellulose) [55,59,62,63] were more evident in the PU sample, as well as the bands at 1429 cm⁻¹ (CH₂ symmetrical band) and 897 cm⁻¹ (β-glycosidic linkages between glucose units of cellulose) [24,46,63].

Table 3. Vibrations modes from the FTIR spectra of IN and PU samples.

Wavenumber (cm−1)	Vibrational Modes	References
3350	O–H groups of cellulose or moisture	[19,62,64]
2913	C–H stretching	[47,64]
1739	C=O stretching of lignin and hemicellulose fractions	[55,59,61]
1644	Adsorbed water	[23,46,55]
1597/1503	C=C stretching of aromatic ring of lignin	[46,55,60]
1463	CH3 deformation of lignin	[60,62]
1429	CH2 symmetrical bending of cellulose	[24,46,47,63]
1375	C–H bending	[46,61]
1337/1327/1282	CH2 wagging vibration and C–O aromatic ring of cellulose	[55,59,62,63]
1246	C–O stretching of lignin	[47,55,63]
1202/1164/1113/ 1059/1034	Multiple peaks of C–O–C pyranose ring	[30,61,62]
897	β-glycosidic linkages between glucose units of cellulose	[24,46,63]

Table 3. Vibrations modes from the FTIR spectra of IN and PU samples.

Wavenumber (cm−1)	Vibrational Modes	References
3350	O–H groups of cellulose or moisture	[19,62,64]
2913	C–H stretching	[47,64]
1739	C=O stretching of lignin and hemicellulose fractions	[55,59,61]
1644	Adsorbed water	[23,46,55]
1597/1503	C=C stretching of aromatic ring of lignin	[46,55,60]
1463	CH3 deformation of lignin	[60,62]
1429	CH2 symmetrical bending of cellulose	[24,46,47,63]
1375	C–H bending	[46,61]
1337/1327/1282	CH2 wagging vibration and C–O aromatic ring of cellulose	[55,59,62,63]
1246	C–O stretching of lignin	[47,55,63]
1202/1164/1113/ 1059/1034	Multiple peaks of C–O–C pyranose ring	[30,61,62]
897	β-glycosidic linkages between glucose units of cellulose	[24,46,63]

3.5. TG/dTG Analysis

Lignocellulosic fibers present cellulose, hemicellulose and lignin as main components. For this reason, the thermal behavior depends on the concentration, individual characteristics and interaction of these components during pyrolysis [49,65]. Curves of TG/dTG of both IN and PU samples are shown in Figure 6a.

Considering the IN sample, the first thermal event ranged from 22 °C to 100 °C and was assigned to water release, resulting in 9% of mass loss [44,60,62]. The second thermal event was observed between 213 °C and 363 °C, and was attributed to the hemicellulose followed by cellulose decompositions, resulting in 54% of mass loss [66,67]. The hemicellulose degradation was observed before the cellulose decomposition, because the former is composed of multiple noncrystalline polysaccharides (such as xilans, arabinosis and galactans) presenting low-energy activation [49,59]. The third thermal event occurred between 363 °C and 526 °C and was attributed to the lignin decomposition and residual components, resulting in 35% of mass loss, a similar value found in the composition analysis shown in Table 1. The large temperature range of lignin decomposition is related to its molecular structure consisting of a complex network of cross-linked aromatic molecules resulting in high thermal stability [67]. The IN sample presented a remaining sample mass of 2%, as reported previously [19,52].

Considering the PU sample, the first thermal event was also related to the moisture release, occurring from 20 °C to 100 °C. The second thermal event was assigned to cellulose and residual hemicellulose degradation, which started at 235 °C, resulting in 63% of mass loss. This was the main observed thermal event, because cellulose represents the major component, resulting in the breaking of C–C and C–H bonds, saccharides carbonization [48], as well as releasing of CO, CO₂, CH₄ and H₂ [67]. The associated dTG curve presented the largest and narrowest peak with a maximum at 339 °C. The degradation range of cellulose is commonly reported between 280 °C and 360 °C [49]. Cellulose is more resistant to decomposition than hemicellulose due to its semicrystalline structure. However, after it started, it shows a high decomposition rate in a short temperature range, representing a typical pyrolysis behavior of linear polymers [65]. Finally, the third thermal event was initiated at 364 °C and assigned to residual lignin and other components degradation. The obtainment of pure cellulose continues to be a significant challenge due to the complex molecular structure of lignin, as observed in this work by FTIR and thermal analysis. The lignin degradation is reported in the range of 100 and 900 °C, but its pyrolysis occurs mainly around 380 °C or at higher temperatures [49,67].

3.6. DSC Analysis

The DSC curves (Figure 6b) allowed for verifying the heat flow as a function of sample mass (W/g). All events and results were related to TG/dTG events. Therefore, the first thermal event of both IN and PU samples was assigned to the moisture release, occurring between 20 °C and 100 °C. The endothermic minimum peaks were observed at 50 °C in the IN sample and 62 °C in the PU sample [67]. The second thermal event was exothermic and represented simultaneously the degradation of the main compounds, hemicellulose and cellulose. The presence of hemicellulose within a cellulosic sample extracted from lignocellulosic material is always subtle, and it may not be completely removed during the bleaching process. For this reason, the presence of hemicellulose was also observed in the DSC and FTIR results. The third thermal event was also exothermic, with the maximum peaks at 500 °C and 391 °C in the IN and PU samples, respectively, evidencing the degradation of lignin as well as residual components. The IN sample presented the most evident third thermal event, showing the necessity of higher energy for degradation due to the structural complexity and high amount of lignin. In the PU sample, the third event occurred in a temperature range prior to that observed in the IN sample due to the reduced amount of lignin, as well as the fact that its structure was already exposed as a consequence of the previous chemical treatment. This fact allowed for faster degradation. After the third event, other smaller peaks were observed and may be related to the decomposition of carbonization by-products [47], [68].

Table 4. Analysis of TG/dTG and DSC curves in the range of 20–600 °C obtained for IN and PU samples.

IN Sample
	1st Event	2nd Event	3rd Event
Temperature range (°C)	22–100	213–363	363–526
Mass loss (%)	9	54	35
Tmax dTG (°C)	47	314	490
DSC effect/temperature (°C)	Endo/50 °C	Exo/343 °C	Exo/500 °C
Remaining sample mass = 2.0%
PU sample
	1st Event	2nd Event	3rd Event
Temperature range (°C)	20–100	235–364	364–600
Mass loss (%)	8.8	63.0	26.0
Tmax dTG (°C)	52	339	529
DSC effect/temperature (°C)	Endo/62 °C	Exo/357 °C	Exo/391 °C
Remaining sample mass = 2.2%

summarizes the thermal events (TG/dTG and DSC curves) in the range of 20–600 °C obtained for IN and PU samples.

3.7. Thermal Conductivity

The thermal conductivity values are shown in

Table 5. Values of thermal conductivity of the IN sample, cotton, EPS and their comparison with referential values.

Material	Thermal Conductivity (W/m·K)	References
O. pyramidale fibers	0.036	Current work
Cotton	0.06	Current work
EPS	0.44	Current work
Cork	0.23–0.406	[71]
Cellulose	0.031
Bamboo	0.077–0.088	[51]
Corn	0.101–0.139
Hemp	0.039–0.123
Kenaf	0.026–0.044
Sunflower	0.038–0.05
Rice husk	0.048–0.08
Cotton	0.058–0.082
Pineapple	0.035–0.057
Wood fiber	0.038–0.05

. The analysis was realized comparing the IN sample to cotton and EPS, materials already known as traditional insulating, and it was observed that the IN sample presented a lower value than the comparatives, indicating that O. pyramidale microfibers have excellent thermal insulating properties to be mainly explored as substitutes to nonbiodegradable and petroleum-based ones [69,70]. It is interesting to point out that the result of the IN sample was comparable to cellulose, cork and wood fiber, but with the advantage of being a fiber harvested from the fruit, which makes it an even more sustainable material since it is not necessary to destroy the complete individual and does not require chemical treatment to obtain it.

The thermal conductivity depends on the mean temperature difference of material, moisture content, porosity and density, and the O. pyramidale microfibers have very low density, even lower than cotton, a fact that contributes significantly to the insulating property, preventing the air movement, which provides the thermal resistance; thus, less material is needed to produce the same [71].

4. Conclusions

The physicochemical characterization of the O. pyramidale fiber revealed that it is constituted mainly by lignocellulosic components and extractives. This study suggests that the pulp extraction using the conventional method is feasible; however, the α-cellulose isolation step causes changes in its original structure, and, depending on the application, it can be dispensed, like in textile, paper, pharmaceutical and cosmetic industries, where its atomic arrangement is not the focus. Therefore, it is suggested that new more ecological routes for the extraction of cellulosic pulp be explored, aiming at the quality and yield of pulp while reducing steps and the use of polluting reagents. On the other hand, it is possible to explore different treatments to obtain different crystalline structures. The tubular morphology of this microfiber is its great differential, as it naturally has a format and dimension that are much sought after in research, such as for carrying bioactive matter. Moreover, it was found that this same characteristic gives microfiber the property of being a thermal insulator as efficient as other materials already used, with the advantages of being biodegradable and from a renewable source. It is expected that this paper will provide a fuse of necessary data for the wide range of research possibilities with this microfiber.