*Article* **Chemical-Anatomical Characterization of Stems of Asparagaceae Species with Potential Use for Lignocellulosic Fibers and Biofuels**

**Agustín Maceda <sup>1</sup> , Marcos Soto-Hernández <sup>2</sup> and Teresa Terrazas 1,\***


**\*** Correspondence: tterrazas@ib.unam.mx; Tel.: +52-55-5622-9116

**Abstract:** During the last decades, the possibility of using species resistant to droughts and extreme temperatures has been analyzed for use in the production of lignocellulosic materials and biofuels. Succulent species are considered to identify their potential use; however, little is known about Asparagaceae species. Therefore, this work aimed to characterize chemically-anatomically the stems of Asparagaceae species. Stems of 10 representative species of Asparagaceae were collected, and samples were divided into two. One part was processed to analyze the chemical composition, and the second to perform anatomical observations. The percentage of extractives and lignocellulose were quantified, and crystalline cellulose and syringyl/guaiacyl lignin were quantified by Fourier transform infrared spectroscopy. Anatomy was observed with epifluorescence microscopy. The results show that there were significant differences between the various species (*p* < 0.05) in the percentages of extractives and lignocellulosic compounds. In addition, there were anatomical differences in fluorescence emission that correlated with the composition of the vascular tissue. Finally, through the characterization of cellulose fibers together with the proportion of syringyl and guaiacyl, it was obtained that various species of the Asparagaceae family have the potential for use in the production of lignocellulosic materials and the production of biofuels.

**Keywords:** Asparagaceae; lignocellulose; crystalline cellulose; syringyl/guaiacyl; anatomy

## **1. Introduction**

Asparagaceae is one of the most important families in Mexico due to its biological, economic, and cultural importance [1]. Several species are used in the production of fibers, intoxicating drinks, food preparation, and the consumption of plant parts (flower, stem, and leaf) [2]. Asparagaceae has several subfamilies, including Agavoideae with *Agave*, *Furcraea*, *Manfreda*, and *Polianthes*; Yuccoideae including *Yucca*, *Hesperaloae*, and *Hesperoyucca* [3], and the subfamily Nolinoideae that includes the genera *Beaucarnea* and *Nolina* [4]. Most of the species of Agavoideae and Nolinoideae are distributed in arid and semiarid, and warm temperate regions of North and Central America [3] and in other parts of the world naturally or introduced. Several species of *Agave* have been used and studied the most because ethanol is produced in the form of intoxicating beverages such as *mezcal* and *tequila* [5].

In recent years, mainly the fibers and bagasse waste of several agave species, mainly *A*. *tequilana* [6,7], *A. angustifolia* [8], and *A. salmiana* [9] have been studied because a large amount of waste is produced annually from the production of *tequila* and *mezcal*. In addition, the subfamilies Agavoideae and Nolinoideae present acid metabolism of the crassulacean (CAM), which is considered raw material for the production of biofuel [10,11]. Furthermore, these species are part of the second generation of plants focused on biofuels. They are not part of the plants essential for human consumption [12], and they tolerate drought conditions and high temperatures [13].

**Citation:** Maceda, A.; Soto-Hernández, M.; Terrazas, T. Chemical-Anatomical Characterization of Stems of Asparagaceae Species with Potential Use for Lignocellulosic Fibers and Biofuels. *Forests* **2022**, *13*, 1853. https://doi.org/10.3390/f13111853

Academic Editors: Vicelina Sousa, Helena Pereira, Teresa Quilhó and Isabel Miranda

Received: 3 October 2022 Accepted: 3 November 2022 Published: 6 November 2022

**Publisher's Note:** MDPI stays neutral with regard to jurisdictional claims in published maps and institutional affiliations.

**Copyright:** © 2022 by the authors. Licensee MDPI, Basel, Switzerland. This article is an open access article distributed under the terms and conditions of the Creative Commons Attribution (CC BY) license (https:// creativecommons.org/licenses/by/ 4.0/).

Other species within the group of the second generation are cacti, such as *Opuntia* spp. [14], which also withstand extreme drought conditions and have CAM metabolism [15]. However, except for some genera such as *Opuntia* spp. and *Selenicereus* spp. [16,17], the other cacti species have slow vegetative development or very small sizes [18], so they could not be profitable in their use as biofuels or production of paper, while in species of the genus *Agave* growth and yield are higher [19].

However, even though many species of Asparagaceae exist in Mexico, there is not much information on the composition of the main lignocellulosic structural components, the anatomical distribution, or the potential use for farmers to cultivate and protect the plants in their natural environment [1]. Therefore, the objectives were to characterize the different Asparagaceae species with the extractives and lignocellulosic percentages, obtain crystallinity indexes, syringyl/guaiacyl (S/G) lignin ratio, and the anatomical distribution, with which it will be possible to identify the potential use of the different species as biofuels or in the paper industry, in addition to the possible biological implications.

#### **2. Materials and Methods**

#### *2.1. Plant Materials and Extractives*

Healthy adult plants were donated from the *Universidad Nacional Autónoma de México* (UNAM) Botanical Garden collection (Table 1), located at 19◦1804400 N, 99◦1104600 O and 2320 m a. s. l. The climate of the area is temperate, with rain in summer, with the rainy season from June to October, and the dry season from November to May. It has an average annual temperature of 15.6 ◦C and a rainfall of 833 mm. The plants were collected in their natural populations and grew in the garden. The leaves and stems of the ten species were cut, and only the stem was selected for the study. Stem samples were cut into pieces and dried in an oven for two weeks at 70 ◦C. Subsequently, the samples were ground (40–60 mesh size, Cyclone Sample Mill, (UDY Corporation, Fort Collins, CO, USA) until a particle size of 0.4 mm was obtained.

**Table 1.** Asparagaceae Species.


The samples were analyzed in triplicate based on the TAPPI T-222 om-02 standard and based on the method proposed by Maceda et al. [20,21]. From each ground sample, 2 g were taken, which were placed in filter paper cartridges to carry out successive extractions for six hours in a Soxhlet with ethanol-benzene (1:2 *v*/*v*) and subsequently in ethanol (96%). After each extraction, the cartridges were allowed to dry for 24 h at 70 ◦C to record their constant weight.

Subsequently, the cartridges were discarded, and the samples were kept in a reflux system for 1 h in water at 90 ◦C. The samples were filtered through a medium pore Büchner filter and dried at 70 ◦C for 24 h to record constant dry weight. The formula used was the following:

Total extractives (%) = [(*A* + *B* + *C*)/*W*0] × 100

where *A* is the weight lost (g) after extraction with ethanol:benzene, *B* is the weight lost (g) after extraction with ethanol (96%), *C* is the weight lost (g) after extraction with water at

90 ◦C, and *W*<sup>0</sup> is the initial weight of each sample. The percentage of extractive-free lignocellulose was obtained by subtracting from 100% the initial weight of the total percentage of extractives.

#### *2.2. Lignocellulosic Purification*

Klason lignin. From the extractive-free lignocellulose of each species, 0.2 g were taken, and 15 mL of concentrated sulfuric acid (72%) was added at a temperature of 2 ◦C. The mixture was kept under constant stirring and at room temperature (18 ◦C) for 2 h. Then, 560 mL of distilled water was added, and the mixture was refluxed and boiled for 4 h. The samples were filtered through a fine-pore Büchner filter and dried at 105 ◦C for 24 h to record constant dry weight. Lignin was quantified as follows:

$$\text{Klason lignin (\%)} = (\mathcal{W}\_L / \mathcal{W}\_W) \times 100$$

where *W<sup>L</sup>* is the obtained weight of lignin (g), and *W<sup>W</sup>* is the extractives-free lignocellulose (g).

Cellulose. From the extractive-free lignocellulose, 0.2 g were taken to purify the cellulose using the Kûshner-Höffer method [21]. Twenty-five mL of HNO3/ethanol (1:4 *v/v*) were added to each sample and kept in a reflux system, and boiled for one hour. The sample was allowed to decant to discard the HNO3/ethanol solution, and another 25 mL was added again. This cycle was repeated three more times, and in the last cycle, 25 mL of an aqueous solution of KOH at 1% was added and kept for 30 min at reflux and boiling to finally filter the sample through a fine-pore Büchner filter. The sample was left to dry at 70 ◦C for 12 h to record the constant dry weight and obtain the percentage of cellulose based on the following formula:

$$\text{Cellulose } (\%) = (W\_{\mathbb{C}} / W\_{W}) \times 100$$

where *W<sup>C</sup>* is the obtained weight of cellulose(g), and /*W<sup>W</sup>* is the extractives-free lignocellulose (g).

Hemicellulose. The purification was carried out based on the methodology proposed by Li. et al. [22]. From the extractive-free lignocellulose, 0.5 g were taken and placed in a reflux system with 10 mL of water for 3 h (solid-to-liquid ratio 1:20 g/mL). The system was cooled to room temperature and filtered. The filtrate was concentrated at 1.25 mL and purified into 3.75 mL of ethanol (95%) with stirring. The mixture was held for 1 h without stirring, and the hemicellulose precipitated. In order to obtain the dry weight (*H*0), the sample was centrifuged at 4500× *g* for 4 min, and then lyophilized. The residue insoluble in water was dried at 60 ◦C for 16 h, then successive extractions were performed with different concentrations of KOH (0.6, 1.0, 1.5, 2.0, and 2.5%) in a ratio of 1:20 (g/mL) at 75 ◦C for 3 h in each extraction. In the last concentration of 2.5% KOH, ethanol (99.7%) was added in a ratio of 2:3. The five mixtures were filtered and acidified to pH 5.5 with glacial acetic acid and concentrated to 1.25 mL. The mixtures were poured into 3.75 mL of ethanol (95%) with constant stirring. The mixtures were kept for 1 h and finally were centrifuged (4500× *g* for 4 min) and lyophilized. The constant dry weight was recorded in each extraction stage (*H*0.6, *H*1.0, *H*1.5, *H*2.0, *H*2.5), and the percentage of cellulose was obtained with the following formula:

## Hemicellulose (%) = (*WH/WW*) × 100

where *W<sup>H</sup>* is the sum of *H*0.6 to *H*2.5 and *W<sup>W</sup>* is the extractive-free lignocellulose (g).

#### *2.3. Fourier Transform Infrared Spectroscopy Analysis*

Lignin analysis. The ratio of syringyl/guaiacyl monomers (S/G) was obtained by Fourier transform infrared spectroscopy (FTIR) analysis. Klason lignin samples were kept dry until analyzed by FTIR (30 scans with a resolution of 4 cm−<sup>1</sup> , 15 s per repeat). Three FTIR readings (Agilent Cary 630 FTIR) were made from each sample, and then the

baseline correction was performed to separate the peaks of the fingerprints (wavelength of 800–1800 cm−<sup>1</sup> ) [23] in the MicroLab PC program (Agilent Technologies). The peaks of 1269 to 1272 cm−<sup>1</sup> and 1328 to 1330 cm−<sup>1</sup> were used to quantify the proportion of guaiacyl (G) and syringyl (S), respectively [24]. The value of each peak was obtained by drawing a line connecting the lowest values and a similar line for the highest values of each peak. A vertical line was drawn from the base of the *X*−axis to the highest part of the peak. The portion of the line between the top and the base is the value of each peak, so the S/G ratio was calculated by dividing the values of each peak [24].

Cellulose analysis. The proportion of crystalline cellulose was obtained by analyzing the dried samples with FTIR [21]. From each sample and doing the analyzes in triplicate, a small portion of the sample was placed in the FTIR Spectrometer (Agilent Cary 630 FTIR), and the spectrum was obtained in a range of 400–650 cm−<sup>1</sup> (30 scans with a resolution of 4 cm−<sup>1</sup> , 15 s per repeat). Samples were converted from transmittance to absorbance, and spectra were averaged using the Resolution Pro FTIR Software program (Agilent Technologies, Santa Clara, CA, United States).

The crystallinity indices used were: Total crystallinity index (TCI) proposed by Nelson and O'Connor [25] or also called the crystallinity ratio [26,27]. The lateral order index (LOI) [25,27] or second proportion of crystallinity [26]; and hydrogen bonding intensity (HBI) [28]. TCI was calculated with the ratio between the absorption intensity of the peaks 1370 cm−<sup>1</sup> and 2900 cm−<sup>1</sup> [27], LOI was calculated from the ratio between the absorption intensity of the peaks 1430 cm−<sup>1</sup> and 893 cm−<sup>1</sup> [26], while HBI was calculated with the ratio between 3350 cm−<sup>1</sup> and 1315 cm−<sup>1</sup> [28].

#### *2.4. Statistical Analysis*

The data obtained from the percentages of extractives and lignocellulosic components were analyzed with the non-parametric Kruskal-Wallis test and Dunn's post hoc analysis since the values did not present normality based on the results of Kolmogorov-Smirnov and Shapiro−Wilk, even when they were transformed with the square root of the arc sine. In addition, a multivariate principal component analysis was performed to separate the groups based on the values of the structural components.

#### *2.5. Lignocellulosic Anatomical Distribution*

Stem fragments were saved from each sample and were fixed, embedded, and cut based on the procedures of Arias and Terrazas [29] for succulent hardwood species. The transverse sections were stained with acridine orange and calcofluor [30] to observe the distribution of cellulose and lignin in the stem, in addition to comparing the anatomical results with the chemical ones.

#### **3. Results**

#### *3.1. Extractives and Lignocellulosic Structural Compounds*

The Asparagaceae species had significant differences between species (Table 2) in the variables of extractives and lignocellulosic components. The percentages of extractives were heterogeneous between the species of the same genus; however, the differences occurred mainly between *Agave* and *Yucca*-*Nolina*. In Tables 3 and 4, the means and standard deviation of the extractives are presented, and the different superscript capital letters show the species that are significantly different. In the ethanol extractives, *A. striata* had the lowest percentage and was significantly different from *Y. gigantea,* which had the highest percentage (Table 3). In hot water extractives, a similar situation was shown; *A. convallis* had the highest percentage, while *N. excelsa* and *Y. periculosa* had the lowest percentage. In the ethanol:benzene extractives, the statistical differences were presented between *N. excelsa*, with the lowest percentage, and *F. longaeva*, with the highest percentage. The species with the highest content of extractives and the lowest content of lignified tissue were *A. convallis* and *Y. gigantea*. On the contrary, the species with the lowest content of extractives and the highest amount of lignified tissue were *N. excelsa* and *Y. periculosa*.


**Table 2.** Kruskal-Wallis analysis for the lignocellulosic and extractives variables.

**Table 3.** Extractives percentage from the 10 Asparagaceae species.


Different letters in each column indicate significant differences (*p* < 0.05). Mean ± standard deviation (SD).

**Table 4.** Lignin, cellulose, and hemicellulose percentage of dry biomass of Asparagaceae species.


Different letters in each column indicate significant differences (*p* < 0.05). Mean ± standard deviation (SD).

The percentages of lignocellulosic components had significant differences in the percentages of lignin between *Y. filifera* and *F. longaeva* (Table 4). In cellulose, the species that had the lowest percentage were *A. celsii* and *A. convallis,* and *N. excelsa* was the species with the highest percentage. Finally, in the hemicelluloses, *Y. gigantea* presented the least quantity and *Y. filifera* the largest.

In the Principal Component (PC) analysis, the first two PCs had eigenvalues above 1, while PC3 was less than 1; however, PC3 was considered in the analysis due that it explained 11% of the variance, so the three PCs explained 84.5% of the total variation (Table 5). In each PC, the highest negative or positive values were those that influenced the separation of each species, as shown in Figure 1. For PC1, the variables that determined the separation of the different groups were hot water extractives and extractive-free lignocellulose. In PC2, the determinant variables of the variation were hemicelluloses and ethanol extractives, while in PC3 was the cellulose percentage (Table 5).


**Table 5.** Vectors, eigenvalues, and cumulative proportion of the variation are explained by each variable. **Variables PC1 PC2 PC3** 

**Table 5.** Vectors, eigenvalues, and cumulative proportion of the variation are explained by each vari-

ethanol extractives, while in PC3 was the cellulose percentage (Table 5).

In the Principal Component (PC) analysis, the first two PCs had eigenvalues above 1, while PC3 was less than 1; however, PC3 was considered in the analysis due that it explained 11% of the variance, so the three PCs explained 84.5% of the total variation (Table 5). In each PC, the highest negative or positive values were those that influenced the separation of each species, as shown in Figure 1. For PC1, the variables that determined the separation of the different groups were hot water extractives and extractive−free lignocellulose. In PC2, the determinant variables of the variation were hemicelluloses and

\* The highest values are in bold on each PC. \* The highest values are in bold on each PC.

(SD).

able.

*Forests* **2022**, *13*, x FOR PEER REVIEW 21 of 21

*Nolina excelsa* 24.5 ± 3.2 AB 52.2 ± 2.1 C 12.0 ± 4.1 AB *Yucca filifera* 11.4 ± 1.1 A 38.9 ± 1.2 ABC 30.7 ± 0.9 C *Yucca gigantea* 24.9 ± 1.7 AB 45.3 ± 1.8 BC 5.7 ± 1.5 A *Yucca periculosa* 24.5 ± 0.5 AB 41.6 ± 1.8 ABC 21.0 ± 2.3 ABC Different letters in each column indicate significant differences (*p* < 0.05). Mean ± standard deviation

**Figure 1.** The three−dimension plot of the principal components from Asparagaceae stems. Blue points: group 1, green point: group 2, pink point: group 3, brown point: group 4. **Figure 1.** The three-dimension plot of the principal components from Asparagaceae stems. Blue points: group 1, green point: group 2, pink point: group 3, brown point: group 4.

When plotting the species based on the first three PCs and the variables that influenced each PC (Figure 1), the species were separated into four groups. The first group (blue points) included the species of *N. excelsa* and *Y. periculosa,* which were the species with the lowest content of hot water extractives and higher extractive−free lignocellulose content. The second group (green point), represented by *Y. filifera* had the highest percentage of hemicelluloses. The third group (pink point), conformed by *Y. gigantea* was separated from the other Yuccas and the other species because they had lower percentages of hemicelluloses, but it was one of the species with the highest percentage of cellulose and percentage of total extractives. In the fourth group (orange points), the remaining species that belong to *Agave*, *F. longaeva,* and *B. gracilis* were clustered (Figure 1).

#### *3.2. Cellulose Crystallinity*

In the cellulose spectra (Figure 2) and Table 6, the main cellulose peaks are observed. In order to determine the purity of the cellulose, the absence of xylans and hemicellulose was obtained by not detecting the peak at 1735 cm−<sup>1</sup> . Lignin was not detected with the peaks 1595 cm−<sup>1</sup> , 1512 cm−<sup>1</sup> , and 1463 cm−<sup>1</sup> . Only weak lignin signals were observed in *A.*

*3.2. Cellulose Crystallinity* 

When plotting the species based on the first three PCs and the variables that influenced each PC (Figure 1), the species were separated into four groups. The first group (blue points) included the species of *N. excelsa* and *Y. periculosa,* which were the species with the lowest content of hot water extractives and higher extractive−free lignocellulose content. The second group (green point), represented by *Y. filifera* had the highest percentage of hemicelluloses. The third group (pink point), conformed by *Y. gigantea* was separated from the other Yuccas and the other species because they had lower percentages of hemicelluloses, but it was one of the species with the highest percentage of cellulose and percentage of total extractives. In the fourth group (orange points), the remaining species

In the cellulose spectra (Figure 2) and Table 6, the main cellulose peaks are observed.

was obtained by not detecting the peak at 1735 cm−1. Lignin was not detected with the peaks 1595 cm−1, 1512 cm−1, and 1463 cm−1. Only weak lignin signals were observed in *A.* 

*striata* and *Y. periculosa*. In addition, the absence of hemicellulose was observed without the presence of the 1269 cm−<sup>1</sup> peak, except for *A. celsii,* which had a weak peak. *striata* and *Y. periculosa*. In addition, the absence of hemicellulose was observed without the presence of the 1269 cm−1 peak, except for *A. celsii,* which had a weak peak.

**Figure 2.** Cellulose FTIR spectra of Asparagaceae species. **Figure 2.** Cellulose FTIR spectra of Asparagaceae species.


*Forests* **2022**, *13*, x FOR PEER REVIEW 21 of 21

that belong to *Agave*, *F. longaeva,* and *B. gracilis* were clustered (Figure 1).


In the crystallinity indexes (Table 7), the TCI values showed that most of the species had values above one because they had a higher percentage of crystalline cellulose, except *A. striata, B. gracilis,* and *Y. periculosa*, which presented a higher percentage of amorphous In the crystallinity indexes (Table 7), the TCI values showed that most of the species had values above one because they had a higher percentage of crystalline cellulose, except *A. striata, B. gracilis,* and *Y. periculosa*, which presented a higher percentage of amorphous cellulose. In LOI, the species that had the highest value was *A. striata*, while the species that had the lowest value was *F. longaeva*, which had the highest value in TCI. In HBI, similarly, *A. striata* had the lowest value, while *B. gracilis* had the highest value.

**Table 7.** Crystallinity indexes of Asparagaceae species.

670 C−O−H out−of−plane stretching


Mean ± standard deviation (SD).

#### *3.3. Lignin S/G Ratio* vibration of C−O and the glucopyranose cycle of guaiacyl and syringyl, respectively. The

*3.3. Lignin S/G Ratio* 

Mean ± standard deviation (SD).

Figure 3 shows the FTIR spectra for the 10 species of Asparagaceae. Representative peaks of lignin were 1501 cm−<sup>1</sup> which showed the C=C aromatic ring vibration of syringyl and guaiacyl monomers. The 1325 cm−<sup>1</sup> peak reflected the breathing of the ring of syringyl in addition to C-O stretching. The 1271 cm−<sup>1</sup> and 1225 cm−<sup>1</sup> peaks reflected the symmetric vibration of C-O and the glucopyranose cycle of guaiacyl and syringyl, respectively. The 1030 cm−<sup>1</sup> peak reflected the C-H in-plane deformation of guaiacyl and C-O deformation in primary alcohol; finally, the 913 cm−<sup>1</sup> peaks showed the =CH out-of-plane deformation in the aromatic ring of syringyl and guaiacyl monomers (Figure 3). 1030 cm−1 peak reflected the C−H in−plane deformation of guaiacyl and C−O deformation in primary alcohol; finally, the 913 cm−1 peaks showed the =CH out−of−plane deformation in the aromatic ring of syringyl and guaiacyl monomers (Figure 3). Whit the peaks 1325 and 1271 cm−1 the S/G ratio was calculated (Table 8). The species with the lowest proportion of S/G was *A. attenuata* because it had a higher percentage of guaiacyl in its structure, while in the genus *Yucca*, the three species presented high values of syringyl, for which the proportion of S/G was in the range of 2.8 to 3.9. The other species had similar proportions, so the syringyl monomer prevailed except in *A. convallis*.

Figure 3 shows the FTIR spectra for the 10 species of Asparagaceae. Representative peaks of lignin were 1501 cm−1 which showed the C=C aromatic ring vibration of syringyl and guaiacyl monomers. The 1325 cm−1 peak reflected the breathing of the ring of syringyl in addition to C−O stretching. The 1271 cm−1 and 1225 cm−1 peaks reflected the symmetric

cellulose. In LOI, the species that had the highest value was *A. striata*, while the species that had the lowest value was *F. longaeva*, which had the highest value in TCI. In HBI,

**Species TCI (A1370/A2900) LOI (A1430/A893) HBI (A3400/A1315)** 

similarly, *A. striata* had the lowest value, while *B. gracilis* had the highest value.

*Agave attenuata* 1.12 ± 0.01 0.49 ± 0.01 1.16 ± 0.02 *Agave celsii* 1.15 ± 0.05 0.49 ± 0.03 1.18 ± 0.09 *Agave convallis* 1.27 ± 0.03 0.56 ± 0.28 1.17 ± 0.08 *Agave striata* 0.85 ± 0.17 0.64 ± 0.03 0.92 ± 0.12 *Beaucarnea gracilis* 0.9 ± 0.01 0.59 ± 0.01 1.21 ± 0.02 *Furcraea longaeva* 1.18 ± 0.05 0.47 ± 0.02 1.09 ± 0.04 *Nolina excelsa* 1.09 ± 0.1 0.54 ± 0.02 1.09 ± 0.01 *Yucca filifera* 1.16 ± 0.02 0.52 ± 0.01 1.11 ± 0.02 *Yucca gigantea* 1.15 ± 0.03 0.51 ± 0.01 1.11 ± 0.02 *Yucca periculosa* 0.78 ± 0.05 0.59 ± 0.02 1.03 ± 0.05

*Forests* **2022**, *13*, x FOR PEER REVIEW 21 of 21

**Table 7.** Crystallinity indexes of Asparagaceae species.

**Figure 3.** FTIR spectra of the lignin fingerprint of the Asparagaceae species. **Figure 3.** FTIR spectra of the lignin fingerprint of the Asparagaceae species.

**Table 8.** Percentages of syringyl and guaiacyl, and S/G ratio of Asparagaceae species. Whit the peaks 1325 and 1271 cm−<sup>1</sup> the S/G ratio was calculated (Table 8). The species with the lowest proportion of S/G was *A. attenuata* because it had a higher percentage of guaiacyl in its structure, while in the genus *Yucca*, the three species presented high values of syringyl, for which the proportion of S/G was in the range of 2.8 to 3.9. The other species had similar proportions, so the syringyl monomer prevailed except in *A. convallis*.


**Table 8.** Percentages of syringyl and guaiacyl, and S/G ratio of Asparagaceae species.

#### *3.4. Anatomical Distribution*

The species of Asparagaceae studied showed the same type of vascular tissue, with closed vascular bundles forming isolated patches or two or more patches joined through lignified parenchyma. In the vascular bundle, the presence of tracheary elements with mainly reticular type of secondary wall thickenings was observed, with completely lignified fibers and non-lignified phloem with contents inside with fluorescence emission in bluish

tones (Figure 4). In most species, patches of non-lignified parenchyma were observed on the edges of the stem, while in the center, the parenchyma was completely lignified. In the species *A. attenuata* and *A. convallis*, the fluorescence tones of the tracheary elements and fibers predominated in yellow-green tones, while in the *Yucca* species, the fluorescence emission was observed in green to bluish-green tones. In all species, the presence of crystals was observed, mainly raphides and prisms. Cellulose fluoresced in bluish tones but differed from lignin due to the intensity of fluorescence emission since the cellulose had lower intensity compared to lignin. *Forests* **2022**, *13*, x FOR PEER REVIEW 21 of 21

**Figure 4.** Fluorescence images from representative Asparagaceae species. (**a**) *A. attenuata.* (**b**) *A. convallis*. (**c**) *B. gracilis*. (**d**) *N. excelsa*. (**e**) *Y. gigantea*. (**f**) *Y. periculosa*. f: fiber, v: vessel, p: parenchyma, ph: phloem, cr: crystal. components that allow its grouping based on the values of extractive compounds, hemi- **Figure 4.** Fluorescence images from representative Asparagaceae species. (**a**) *A. attenuata.* (**b**) *A. convallis*. (**c**) *B. gracilis*. (**d**) *N. excelsa*. (**e**) *Y. gigantea*. (**f**) *Y. periculosa*. f: fiber, v: vessel, p: parenchyma, ph: phloem, cr: crystal.

The Asparagaceae family presented heterogeneity in the percentages of the structural

**4. Discussion** 

#### **4. Discussion**

The Asparagaceae family presented heterogeneity in the percentages of the structural components that allow its grouping based on the values of extractive compounds, hemicellulose, and cellulose. In addition, the presence of high percentages of cellulose, a majority proportion of crystalline cellulose, and the presence of S/G ratios greater than one make it possible to use in biofuels and cellulosic compounds.

#### *4.1. Extractives and Lignocellulose*

Asparagaceae species had similar percentages of extractives when extracted with ethanol: benzene and ethanol. However, hot water extractives were low except for *A. convallis* (Table 3). Compared with the literature on other *Agave* species (Table 9), the extractive percentages in water were similar to the values reported here. *A. tequilana, A. angustifolia*, and *A. salmiana* have values between 4.4%–6.0% (Table 9). On the contrary, the values reported in *Agave* for the percentages of extractives in ethanol: benzene and ethanol (Table 9) were low (1.5%–4.0% in ethanol: benzene and 1.3%–5.0% in ethanol) compared to the values reported here (Table 3). For the genera *Beaucarnea, Furcraea,* and *Nolina*, there are no reports of extractive percentages in their stems or leaves. In the *Yucca* genus, the literature report that *Y. gloriosa* presents a percentage of total extractives of 1.1 [31], which is low compared to the percentages obtained here (Table 3).

Although extractives in ethanol: benzene and in ethanol are hardly reported in the literature for the species of the genera analyzed here, they have been reported for other succulent species such as cacti [20,21], which had lower percentages of ethanol: benzene (2.5%–4.2%), similar in ethanol (1.0%–9.1%), higher in hot water (8.2%–44.5%) and total extractives (16.7–49.2 %).

In the percentages of lignocellulosic compounds, Table 4 shows that there was heterogeneity in the results within the same genus. In the species that are considered arborescent, it was obtained that the percentages of lignin were low. However, the percentage of cellulose was higher (37.3%–52.2%) and in herbaceous species was lower (31.6%–37.1%). In the hemicelluloses, there is homogeneity between the agaves (20.9%–27.1%), while in the *Yucca* genus, notable differences are observed between *Y. gigantea* (5.7%) concerning the other two species of *Yucca* (21.0% and 30.7%). The differences possibly were due to the type of environment in which they live; *Y. gigantea* is distributed in regions with higher humidity [32], while *Y. filifera* [33,34] and *Y. periculosa* [35] grow in arid zones and semi-desert, respectively. In celluloses, the percentages reported in the literature (Table 9) were high compared to those obtained here (Table 4), possibly due to the differences in the species or the conditions in which they have been developed [36,37].

When comparing the results obtained with succulent species such as cacti, it was observed that the percentages of lignocellulose, in general, are higher in agave species than in non-fibrous wood cacti species [20], while in fibrous species, the percentages were similar [38].


**Table 9.** Percentages of extractives and lignocellulosic compounds of *Agave*, *Furcraea*, and *Yucca.*


**Table 9.** *Cont.*

S: sample; Ce: cellulose, He: hemicellulose, Li: lignin, TE: total extractives, W: hot water extractives, E−B: ethanol−benzene extractives, Et: ethanol, Ref: references. \* B: Bagasse, F: Fiber, L: Leaf, St: Stem.

The differences in the extractive and lignocellulosic compounds between species allowed us to identify four groups through the principal component analysis. The first group with *N. excelsa* and *Y. periculosa* presented a lower amount of hot water and total extractives. This could have implications for their use because various authors consider the hot water and total extractives to represent the nonstructural sugars that are used in fermentation processes for the production of ethanol [8,48,51]. However, in the case of *N. excelsa*, the percentage of cellulose presented in the stem would allow it to be used to obtain cellulosic products (Table 4). The characteristic that distinguishes *Y. filifera* from the other species studied was the presence of a large amount of hemicellulose. This hemicellulose can be transformed into usable sugars either through enzymatic hydrolysis processes or with temperature [9], in addition to the fact that *Y. filifera* presented lower percentages of lignin so that the purification of cellulose would be more efficient [64]. The main characteristic of *Y. gigantea* was the lower amount of hemicelluloses; however, it had a high percentage of cellulose that could be used to obtain cellulosic fibers. The three species of *Yucca* studied here revealed the higher differences in the extractives and lignocellulosic compounds between them, and other species of *Yucca* should be studied in the future to support these findings. All agave species plus *Beaucarnea* and *Furcraea* were grouped together by similar values in the extractive compounds (Table 4). In the lignocellulosic components, the agaves had low percentages of cellulose and lignin; however, they had a higher percentage of hemicelluloses. Finally, chromatographic analyzes can be carried out on all the extractives compounds to identify components with potential use, such as flavonoids and triterpenes, which have the antioxidant capacity and can be used mainly in the food and cosmetic industries [65]. In addition, lignocellulose can be treated in different

ways to obtain various derivatives, as mentioned by Palomo-Briones et al. [66] for *Agave tequilana*: through the elimination of lignin by alkaline, organosolv or enzymatic methods, lignin can be solubilized and fermented to obtain ethanolic derivatives. Hemicelluloses and celluloses can be degraded by acids or enzymes to obtain insoluble fractions and use microcrystalline and nanocrystalline cellulose, while the solubilized carbohydrates can be used in the fermentation to obtain ethanol. Solubilized carbohydrates can also be fermented by anaerobic digestion or dark fermentation to obtain biogas and H<sup>2</sup> which can be used to obtain energy [66]. Therefore, the species analyzed in this work could be used in different ways, both energetically and in the production of paper or cellulose components [21].

#### *4.2. Cellulose Crystallinity*

The crystallinity indexes allowed identifying the species that had the highest proportion of crystalline cellulose. The purity of the cellulose could affect these indexes by presenting hemicellulose or lignin residues that alter the peaks [67]. The FTIR spectra can determine the purity of the samples and calculate the proportion of crystalline cellulose [68,69] by the absence of peaks belonging to hemicelluloses [70] and lignin [71].

The most commonly used indexes in conjunction with FTIR are TCI, LOI, and HBI [68,72]. The TCI index provides information about the amount of crystalline or amorphous cellulose in a sample. The peak indicating the presence of crystalline cellulose is 1370 cm−<sup>1</sup> , while the peak of 2900 cm−<sup>1</sup> [27] indicates the presence of amorphous cellulose. Therefore, if the value of the ratio is greater than one, there will be a greater amount of crystalline cellulose [28]. The LOI index is related to the order of crystalline cellulose, in addition to the fact that the 1430 cm−<sup>1</sup> peak shows the presence of crystalline cellulose of type I and the peak 893 cm−<sup>1</sup> the presence of type II crystalline cellulose and cellulose amorphous [73]. The order of the crystalline cellulose and LOI peaks can be altered by the type of chemical extraction, and the type of purification used [74]. The HBI index reflects the crystallinity of the sample and its water absorption. Low values reflect a greater amount of crystalline cellulose, while high values indicate the presence of cellulose II or amorphous [27], however, TCI and HBI values are related in terms of cellulose structure and stability, so if they are similar, the cellulose structure has greater stability [28].

Generally, the samples had a higher proportion of crystalline cellulose; however, the type of extraction modifies the order of the cellulose, so the low LOI values reflect the presence of type II crystalline cellulose [75], in addition to possibly the presence of NaOH during purification and the temperature used would alter the order of the crystalline [73]. Furthermore, LOI is correlated with the overall degree of cellulose order, while TCI is directly proportional to the percentage of crystalline cellulose [28]. The similar values of HBI with TCI confirm that the highest proportion of crystalline cellulose predominates among the Asparagaceae species, except the species *A. striata, B. gracilis,* and *Y. periculosa*, which had lower crystalline cellulose, and this is reflected in the LOI values that were also the highest [76].

The presence of a greater amount of crystalline cellulose (between 50 and 56% so that the proportion of TCI is 1 to 1.27) in most of the Asparagaceae species agrees with that reported for stems (Table 10). Furthermore, the percentages were also similar for other structures such as fibers in *Agave* (50.07 [46]), *Furcraea* (52.6% [62]), and *Yucca* spp. (55–56% [77]). The other structures have a higher percentage, possibly due to the species analyzed, the type of sample, such as bagasse that no longer has extractives, and some structural sugars, such as hemicelluloses. However, in future studies, X−ray diffraction analysis [78] could be used to confirm the proportion of crystalline cellulose in samples of Asparagaceae species.


**Table 10.** Crystalline index *Agave*, *Furcraea*, and *Yucca.*

#### *4.3. Lignin S/G Ratio*

The proportion of S/G obtained for the 10 species reflected that most of the species presented high percentages of syringyl monomers, except for *A. attenuata* (36.9%) and *A. convallis* (48%). The species with the highest percentage was of the *Yucca* genus since it had percentages of 73.9 to 79.5%. The presence of higher percentages of syringyl in species with stem potential allows the purification of cellulose and the degradation of lignin by hydrolytic processes to be more efficient [80] by presenting more bonds of the β-O-4 type and being less condensed [81]. In species with higher percentages of guaiacyl, the stem tissue is hard, and the lignin is difficult to degrade, which is why it is called recalcitrant lignin [82].

Therefore, in the *Yucca* species and generally in the other species of *Agave, Beaucarnea, Furcraea*, and *Nolina*, by presenting a higher percentage of syringyl, the hydrolyzation process with Kraft would be more efficient [83]. In the literature, there are few reports on the proportion of S/G for *A. fourcroydes*, proportions are reported in fibers (1.05) and spines (1.2) [84], in leaves of *A. sisalana* (2.0) [85], and leaf fibers (3.0–3.5) [86]. In stems, S/G values have been reported for *A. americana* (1.27), *A. angustifolia* (1.29), *A. fourcroydes* (1.40), *A. salmiana* (1.33) y *A. tequilana* (1.57) in untreated samples [11], while for *A. tequilana* S/G values of 4.3 have also been reported in untreated samples [87] while in *A. sisalana* the proportion of S/G in stem fibers is 3.6 [88]. In the other genera analyzed here, there are no reports on the composition of lignin. In general, it is observed that the values obtained are similar to those reported by the aforementioned authors.

The presence of high percentages of syringyl in lignin has also been reported in various fibrous species where it provides resistance to and cellular support [89–91]. However, in succulent species such as cacti, the presence of syringyl monomers is more associated with tissues with non−lignified parenchyma and not fibers [38,92,93], so the presence of syringyl could be associated with a defense mechanism against pathogens [93] as it has been reported for bryophytes, conifers and angiosperms [94–98].

#### *4.4. Cellulose and Lignin Anatomical Distribution*

Analyzing the anatomical distribution of lignin and cellulose in the vascular bundles of Asparagaceae species was useful in identifying the location of the main structural components, in addition to explaining the results obtained both in the percentages of lignocellulosic components and in the proportion of S/G [49,99].

In Figure 4, the fluorescence emitted by the tracheary elements, the fibers, and the non−lignified and lignified parenchyma showed that lignin had different emissions in its fluorescence since it ranged from bluish tones to yellow tones. In the species of *A. attenuata* and *A. convallis*, the presence of yellow tones was observed, which would be related to the type of lignin present in the lignified walls. As proposed by Maceda et al. [93], based on a tone scale, yellow to green tones would reflect the presence of guaiacyl-type lignin, while syringyl−type lignin would have shades of lime-green to blue. The difference in the

tones in the emission of fluorescence has already been reported for various species with similar tones [100]. The presence of blue tones in fibers in species with S/G ratios above two [93], as for *Ferocactus hamatacanthus* and *F. pilosus*, whose S/G ratio is 11.7 and 3.5, respectively [38]. Therefore, in the species of Yuccas, whose proportion ranges from 2.8 to 3.9, the presence of fibers and tissue in blue tones responds to the presence of lignin of the syringyl type.

The presence of lignified parenchyma in the stems of the *Yucca* species (Figure 4c–f) agrees with the lignin percentages shown in Table 4 since these species, except for *B. gracilis*, had high lignin values, in contrast to the species of *A. convallis* and *A. attenuata* (Figure 4a,b). Therefore, the presence of lignin is a structural support factor for species of tall size and arborescent shape. However, these species also had high percentages of cellulose (Table 4), with a predominance of crystalline cellulose (Table 7). Cellulose and lignin not only provide structural rigidity but also the proportion of S/G, and the presence of crystalline cellulose could improve water conduction in tracheary elements [101] in addition to protecting against pathogens [91,102]. The species analyzed in this study are naturally distributed in Mexico and *Y. guatemalensis* in Mexico and Guatemala, so it would be interesting to expand the number of genera to determine if there is heterogeneity in a greater number of species of the same genus or if they present homogeneity in the composition of cellulose and lignin.

#### **5. Conclusions**

The presence of high percentages of cellulose, the predominance of syringyl−type lignin, percentages above 10% of total extractive components, and the high percentages of holocellulose (structural sugars) show that Asparagaceae species have potential in the use of both productions as biofuels as in the production of paper. In addition, the chemical composition that Asparagaceae species present would be related to biological implications such as conduction, support, and protection against pathogens.

**Author Contributions:** Conceptualization, A.M. and T.T.; methodology, A.M., M.S.-H. and T.T.; validation, A.M. and T.T.; investigation, A.M. and T.T.; resources, T.T.; writing—review and editing, A.M., T.T. and M.S.-H. All authors have read and agreed to the published version of the manuscript.

**Funding:** Funding was provided by DGAPA−UNAM postdoctoral fellowship (document number: CJ IC/CTIC I5OO7I2O2I) to AM.

**Data Availability Statement:** Raw data and FTIR spectra are available in the Figshare repository: https: //doi.org/10.6084/m9.figshare.21259230.v1 https://figshare.com/articles/dataset/Chemical-anatomical\_ characterization\_of\_stems\_of\_Asparagaceae\_species\_with\_potential\_use\_for\_lignocellulosic\_fibers\_and\_ biofu-els/21259230 (accessed on 3 November 2022)

**Acknowledgments:** The authors thank Abisaí Josué García for providing the plants from the Botanic Garden, UNAM; thanks to Rubén San Miguel−Chávez for allowing us to use the FTIR in COLPOS. Thanks to Elizabeth Navarro Cerón for allowing us to use the laboratory LANISAF, thanks to Pedro Mercado Ruaro from Laboratorio de Morfo−Anatomía y Citogenética (LANABIO, UNAM), and thanks to Steffany Aguilar Moreno for the support and the laboratory glassware provided.

**Conflicts of Interest:** The authors declare no conflict of interest.

#### **References**


del sur de México. *Acta Bot. Mex.* **2019**, *126*, e1461. [CrossRef]


**Pingping Guo <sup>1</sup> , Xiping Zhao 1,\*, Qi Feng <sup>1</sup> and Yongqiang Yang <sup>2</sup>**


**Abstract:** *Tilia amurensis* Rupr. and *Tilia mandshurica* Rupr. and Maxim. are two essential commercial species, though there is surprisingly little concern about whether their branches can be used in the current situation of a wood shortage in China. In this study, tissue proportions and fiber morphology, physical and mechanical properties, and chemical composition of the branchwood were studied and compared with stemwood to evaluate the potential for papermaking. The branchwood and stemwood showed similar cell arrangement but different tissue proportions and fiber morphology. The branchwood had more than 40% fiber proportion, 90%–97% below 0.9 mm in length, 75%–90% less than 33 in slenderness ratio, and 80% less than 1 in Runkel ratio. The branchwood was as light and soft as stemwood with a density of 0.32–0.36 g/cm<sup>3</sup> and a compressive strength of about 30 MPa. The branchwood had 6% water extractives, 66% holocellulose, and 22% lignin for *T*. *amurensis*, 58% holocellulose and 30% lignin for *T. mandshurica*. The results suggest the branchwood is favorable for mechanical chipping, has the potential to obtain high pulp yield and its fibers can be mixed with wide, long and thick fibers from other tree species to produce specific paper products. In contrast, *T. mandshurica* branchwood may not be suitable for chemical pulping.

**Keywords:** *Tilia amurensis* Rupr.; *Tilia mandshurica* Rupr. & Maxim.; branch; wood properties; papermaking material

### **1. Introduction**

Branches are an essential part of a tree, the proportion of branchwood to the volume of the whole tree is about 20%, which varies with species, tree height, and stand condition [1]. Unfortunately, many branches are discarded when harvesting operations concentrate on stems. Using branchwood for pulping and papermaking can be an additional measure to ease the conflict between timber supply and demand [2,3]. Branchwood has been used in pulp and paper making in China since the mid-1960s [4]. With improvements in harvesting methods [5] and technological advances in wood pulping [6], the prospects for increased use of branchwood for papermaking materials are promising.

The qualities of pulp and paper products highly depend on the properties of raw materials, e.g., wood density, fiber dimensions, and chemical composition. Higher pulp yield is consistent with higher cellulose and hemicellulose content and lower lignin and extractive content [7,8]. Species with low wood density are indicated to produce printing and writing sheets, while high-density wood is favorable to make tissue paper [9]. Long fibers are beneficial for improving the fracture toughness of paper, and short fibers give superior tear resistance [10,11]. However, wood properties closely related to pulping and papermaking are lacking for branch, especially for broadleaves grown in China. Branches and stems of the same tree may differ in wood properties. For example, branchwood has a smaller cell size than stemwood, because some types of cells are more or less abundant in branchwood than stemwood [2,12–14]. These differences are important factors affecting the utilization of wood as a papermaking material [15–17].

**Citation:** Guo, P.; Zhao, X.; Feng, Q.; Yang, Y. Branchwood Properties of Two *Tilia* Species Collected from Natural Secondary Forests in Northeastern China. *Forests* **2023**, *14*, 760 . https://doi.org/10.3390/ f14040760

Academic Editors: Vicelina Sousa, Helena Pereira, Teresa Quilhó and Isabel Miranda

Received: 17 February 2023 Revised: 31 March 2023 Accepted: 1 April 2023 Published: 7 April 2023

**Copyright:** © 2023 by the authors. Licensee MDPI, Basel, Switzerland. This article is an open access article distributed under the terms and conditions of the Creative Commons Attribution (CC BY) license (https:// creativecommons.org/licenses/by/ 4.0/).

*Tilia amurensis* Rupr. and *Tilia mandshurica* Rupr Maxim., two native linden trees in the temperate zone of the northern hemisphere, are not easily visually distinguishable from each other [18,19]. They are of commercial importance for honey production and as essential timber trees in China [20]. The wood of the two linden trees is diffuse-porous in its gross structure and exhibits well-differentiated growth rings, but there are no color differences between heartwood and sapwood [21,22]. Linden wood has a minor use in construction and is commercialized mainly for the manufacture of furniture, craft items, and plywood [23–25]. A survey shows that most pencil boards are made of linden wood in China [26].

The work described in this paper will examine tissue proportions and fiber morphology, physical and mechanical properties, and chemical composition of the branchwood of the two *Tilia* species. We expect that the branches will have similar wood properties to the stem and could meet the basic requirements of pulping and papermaking.

#### **2. Materials and Methods**

#### *2.1. Materials*

For each species, three healthy straight trees were randomly selected from the Maoershan Forest Ecosystem Research Station in Heilongjiang Province, northeastern China (127◦300–340 E, 45◦200–250 N, 400 m elevation). The site has a continental monsoon climate with a warm summer, cold, dry winter, and temperate species-rich deciduous broadleaf forests [27]. A live branch was sampled from each tree's upper, middle, and lower canopy. The characteristics of the trees and branches sampled were measured (Table 1). A segment (about 1 m long) was cut from each sample branch. It should be stressed that efforts to ensure sustainable management in the station are crucial; only one 5 mm increment core (from pith to bark) was collected at 1.3 m stem height (b.h., breast height) as non-destructively as possible to the stem.

**Table 1.** Characteristics of the trees and branches sampled <sup>1</sup> .


<sup>1</sup> The data is represented as Means <sup>±</sup> SD.

#### *2.2. Xylem Anatomy*

A strip (located in the middle of the branches) with 20 (radial) × 10 (tangential) × 10 (longitudinal) mm and a chip with 20 (radial) × 2 (tangential) × 10 (longitudinal) mm were cut from the lateral wood (mid-way between the tension wood and the opposite wood) of each branch segment. Two 20 mm long segments (including 3–5 growth rings) were cut from the middle of each tree core, one for slicing and the other for macerating. A 15-µm-thick transverse section was cut from each strip and core segment (embedded in paraffin blocks) using a rotary microtome (RM 2235, Leica Microsystems, Wetzlar, Germany) equipped with a knife holder, stained with safranin, and then fixed on microscopic slides [28]. The wood chip and core segment were macerated using the chromic acid-nitric acid method described by Jeffrey [29]. The macerated material was rinsed and placed on microscopic slides. All microscopic slides were photographed using a digital light microscopy (Mshot-MD50; Microshot Technology Limited, Guangzhou, China) to measure tissue proportions with an image analysis system (TDY5.2; Beijing Tian Di Yu Technology Co., Ltd., Beijing, China) [30]. Fiber dimensions were measured manually. At least 60 measurements were done per parameter. Two derived indices were calculated using fiber dimension: slenderness ratio as fiber length/fiber diameter and Runkel ratio as two times fiber cell wall thickness/lumen diameter [31].

#### *2.3. Physical and Mechanical Properties*

Five defect-free cubic test pieces with 20 (radial) × 20 (tangential) × 20 (longitudinal) mm and prismatic test pieces with 20 (radial) × 20 (tangential) × 30 (longitudinal) mm were cut from each branch segment to test basic wood density (hereafter referred to as "wood density") and compressive strength parallel to grain (hereafter referred to as "compressive strength") by the Chinese Standard GB/T1933-2009 [32] and GB/T 1935–2009 [33], respectively. The green volume and absolute weight of each cubic test piece were measured using the immersion and electronic balance weighing methods, respectively. Wood density was calculated based on weight and volume values. The prismatic test pieces were adjusted to a moisture level of 12%, and then the compressive strength was tested using a universal testing machine (WDW-50B, Jinan Shengfeng Testing Machine Co., Ltd., Jinan, China) with a loading capacity of 50 kN and a testing speed of 25,000 ± 5000 N/min.

#### *2.4. Chemical Composition*

After mechanical tests, all damaged columns for each branch were ground to 40 to 60 mesh wood powder.

The extractives were determined according to the Chinese Standard GB/T 35816– 2018 [34]. Cold water-soluble extraction was carried out at room temperature (23 ± 2 ◦C) for 48 h. Hot-water-soluble and 1% sodium hydroxide (NaOH) extraction were carried out by boiling water bath for 3 and 1 h, respectively. The extractive content of the wood powder was calculated by weight lost when extracted.

The holocellulose, cellulose, and lignin were determined by glacial acetic acid-sodium chlorite, nitric acid-ethanol, and sulphuric acid hydrolysis methods, respectively, according to the Chinese Standard GB/T 35818–2018 [35]. The contents were calculated based on the final residue. The hemicellulose content was calculated as the subtraction of cellulose content from holocellulose content.

#### *2.5. Data Analysis*

Data processing and analysis were performed using IBM SPSS Statistics software (Version 24.0, International Business Machines Corporation, Armonk, NY, USA). Interspecific and differences between branch and stem, and the two species in wood properties were evaluated by variance analyses with alpha significance less than 0.05 and 0.01. The fitting curves of the fiber length and derived parameter distributions were developed using a normal distribution function, and skewness and kurtosis were used to check the normality of the data sets.

#### **3. Results and Discussion**

#### *3.1. Wood Tissue Proportions*

The stemwood of the *T. amurensis* and *T. mandshurica* was diffuse-porous, with vessels arranged as clusters, multiples, or solitary (Figure 1a,b). Growth rings were distinct. There were many, but relatively narrow, rays. Axial parenchyma was sparce, axial parenchyma apotracheal short tangential lines, (terminal) and in marginal or in seemingly marginal lines, as reported in previous studies [22], *T. mandshurica* showed a more frequent diagonal pattern in vessels arrangement than the *T. amurensis*. Compared to stemwood, branchwood showed diffuse to echelon arrangement in vessel groups; latewood vessels were about half the diameter of earlywood vessels (Figure 1c,d).

The mean fiber proportion of stemwood was over 50% and above the value reported by Fang et al. [36]. Fang et al. also measured the proportions of vessels (26.6%), rays (13.6%) and axial parenchyma (13%) that sightly differed from the present results (Table 2). In the study by Fang et al., the linden trees were sampled from southwest China, which is about 3300 km from the sampling site (northeast China). Northeast China is a sub-humid region in the cold temperate zone, the mean annual precipitation is 629 mm, and January and July air temperature are −18 and 22 ◦C, respectively [27]. Southwest is a humid region in the subtropical zone, the mean annual precipitation is 800–1600 mm, and January and July air

temperature are 5–12 and 20–30 ◦C, respectively [37]. Thus, variation in provenance may be the main reason for the difference [38]. Interspecific differences in tissue proportions were significant for stemwood (*p* < 0.05, except ray proportion) but not for branchwood. These differences in tissue proportions can lead to differences in porosity, shrinkage, and treatment capacity. January and July air temperature are 5–12 and 20–30 °C, respectively [37]. Thus, variation in provenance may be the main reason for the difference [38]. Interspecific differences in tissue proportions were significant for stemwood (*p* < 0.05, except ray proportion) but not for branchwood. These differences in tissue proportions can lead to differences in porosity, shrinkage, and treatment capacity.

is about 3300 km from the sampling site (northeast China). Northeast China is a sub-humid region in the cold temperate zone, the mean annual precipitation is 629 mm, and January and July air temperature are −18 and 22 °C, respectively [27]. Southwest is a humid region in the subtropical zone, the mean annual precipitation is 800–1600 mm, and

*Forests* **2023**, *14*, x FOR PEER REVIEW 4 of 11

**Figure 1.** Cross-sections of the stemwood of *T. amurensis* (**a**) and *T. mandshurica* (**b**), and of the branchwood of *T. amurensis* (**c**) and *T. mandshurica* (**d**): axial parenchyma (P), ray parenchyma (R), vessels (V) and fibers (F). The scale bar represents 200 µm. **Figure 1.** Cross-sections of the stemwood of *T. amurensis* (**a**) and *T. mandshurica* (**b**), and of the branchwood of *T. amurensis* (**c**) and *T. mandshurica* (**d**): axial parenchyma (P), ray parenchyma (R), vessels (V) and fibers (F). The scale bar represents 200 µm.


**Table 2.** Comparison of tissue proportions in the branchwood and stemwood from two *Tilia* species 1. **Table 2.** Comparison of tissue proportions in the branchwood and stemwood from two *Tilia* species <sup>1</sup> .

cies at *p* = 0.01. \* and \*\* indicate significant differences between branchwood and stemwood at *p* = 0.05 and 0.01, respectively. <sup>1</sup> The data is represented as Means <sup>±</sup> SE. <sup>2</sup> Data in bold indicate significant differences between species at *<sup>p</sup>* = 0.01. \* and \*\* indicate significant differences between branchwood and stemwood at *p* = 0.05 and 0.01, respectively.

The proportion of fiber in branchwood was lower than the stemwood, but it was more than 40%, indicating the potential of branchwood to obtain a high pulp yield [39]. The proportion of fiber in branchwood was lower than the stemwood, but it was more than 40%, indicating the potential of branchwood to obtain a high pulp yield [39]. Specified proportions of parenchyma are good for press-dried paper, increasing the bond strength in the paper [40]. However, redundant vessel elements in the sheets reduce mechanical properties and cause linting problems [16].

#### *3.2. Fiber Morphology*

Interspecific differences in fiber dimensions were significant (*p* < 0.01) for the stemwood of the two species (Table 3). The average fiber length for stemwood was similar (0.7–1.6 mm) to most of the hardwood species [41], and the mean length for *T. amurensis* stemwood was similar to the medium-length fibers (0.9–1.6 mm) according to the IAWA [42]. The stemwood fibers were wide (average 31–41 µm) and similar to *Paulownia fortune* and *Alniphyllum fortunei* [43], and were up to a particular wide grade (>30 µm) proposed by Cheng et al. [44]. The average lumen diameter and double wall thickness of the fibers were about 20 µm for *T. amurensis* stemwood, and 15 µm for *T. mandshurica*. The average fiber lumen diameter of the linden stemwood was similar to *Betulaplaty phylla* [45] and *Alnus sibirica* [31], but the fiber wall was too thick. For example, a thicker cell wall would make the fiber more flexible but lead to a massive void in the paper produced in *Prunus domestica* [46].

**Table 3.** Comparison of fiber dimensions and their derived indices of branchwood and stemwood from two *Tilia* species <sup>1</sup> .


<sup>1</sup> The data is represented as Means <sup>±</sup> SE. <sup>2</sup> Data in bold indicate significant differences between species at *<sup>p</sup>* = 0.01. \*\* indicate significant differences between branchwood and stemwood at *p* = 0.01, respectively. n.s. indicate no significant differences.

Average fiber dimensions (except lumen diameter) for branchwood were significantly lower (i.e., shorter and narrower fibres) compared to stemwood (*p* < 0.01). Similar results were found in many hardwoods [14,31]. These may be ascribed to cambial age and distance from apical meristem [28,47]. Short and narrow fibers could have resulted from a faster growth rate during wood formation in the branches and the extent of invasive growth of the tip of the fibers during their differentiation [48]. The average fiber dimensions were significantly lower in branchwood of *T. amurensis* than those of *T. mandshurica* (*p* < 0.01), except for cell wall thickness. Compared to long and wide fibers, short and narrow fibers make it challenging in specific usage of lignocellulosic materials [38,49]. For example, longer fibers are preferred to shorter fibers due to their capacity to produce paper with greater tensile strength and toughness [10]. However, shorter fibers could give superior tear resistance at higher levels of sheet density [11].

The average fiber length for *T. amurensis* branchwood was 0.59 mm, which does not fall in the range values (0.7–1.6 mm) for most of the hardwood species [41]. Even the average fibers in *T. mandshurica* branchwood that were 0.72 mm long only met the IAWA category for the shorter fibers [42]. Figure 2a–d shows the distribution of fiber length. Table 4 shows that the skewness and kurtosis of distributions appeared to meet the normality assumption (|skewness| < 2.1 and |kurtosis| < 7.1) set by West et al. [50]. However, the fact was that all distributions were not standard but slightly skewed. About 50% and 30% of the fibers had a length greater than 0.9 mm in *T. amurensis* and *T. mandshurica* stemwood, respectively. Only 10% of the fibers in *T. mandshurica* branchwood with a length of more than 0.9 mm, and less *T. amurensis*, only 3%. About 80% of the fibers showed 0.4–0.8 mm and 0.5–0.9 mm in length in *T. amurensis* and *T. mandshurica* branchwood, respectively. These results showed that the short fibers of the branchwood would be a factor limiting its application in pulp and papermaking. stemwood, respectively. Only 10% of the fibers in *T. mandshurica* branchwood with a length of more than 0.9 mm, and less *T. amurensis*, only 3%. About 80% of the fibers showed 0.4–0.8 mm and 0.5–0.9 mm in length in *T. amurensis* and *T. mandshurica* branchwood, respectively. These results showed that the short fibers of the branchwood would be a factor limiting its application in pulp and papermaking.

Table 4 shows that the skewness and kurtosis of distributions appeared to meet the normality assumption (|skewness| < 2.1 and |kurtosis| < 7.1) set by West et al. [50]. However, the fact was that all distributions were not standard but slightly skewed. About 50% and 30% of the fibers had a length greater than 0.9 mm in *T. amurensis* and *T. mandshurica*

*Forests* **2023**, *14*, x FOR PEER REVIEW 6 of 11

**Figure 2.** Relative frequency (histogram), cumulative frequency (dotted line), and fitted normal distribution (solid line) of fiber length, slenderness ratio and Runkel ratio. (**a**) Fiber length of *T. amurensis* branchwood. (**b**) Fiber length of *T. amurensis* stemwood. (**c**) Fiber length of *T. mandshurica* branchwood. (**d**) Fiber length of *T. mandshurica* stemwood. (**e**) Slenderness ratio of fibers in *T. amurensis* branchwood. (**f**) Slenderness ratio of fibers in *T. amurensis* stemwood. (**g**) Slenderness ratio of fibers in *T. mandshurica* branchwood. (**h**) Slenderness ratio of fibers in *T. mandshurica* stemwood. (**i**) Runkel ratio of fibers in *T. amurensis* branchwood. (**j**) Runkel ratio of fibers in *T. amurensis* stemwood. (**k**) Runkel ratio of fibers in *T. mandshurica* branchwood. (**l**) Runkel ratio of fibers in *T. mandshurica* stemwood. **Figure 2.** Relative frequency (histogram), cumulative frequency (dotted line), and fitted normal distribution (solid line) of fiber length, slenderness ratio and Runkel ratio. (**a**) Fiber length of *T. amurensis* branchwood. (**b**) Fiber length of *T. amurensis* stemwood. (**c**) Fiber length of *T. mandshurica* branchwood. (**d**) Fiber length of *T. mandshurica* stemwood. (**e**) Slenderness ratio of fibers in *T. amurensis* branchwood. (**f**) Slenderness ratio of fibers in *T. amurensis* stemwood. (**g**) Slenderness ratio of fibers in *T. mandshurica* branchwood. (**h**) Slenderness ratio of fibers in *T. mandshurica* stemwood. (**i**) Runkel ratio of fibers in *T. amurensis* branchwood. (**j**) Runkel ratio of fibers in *T. amurensis* stemwood. (**k**) Runkel ratio of fibers in *T. mandshurica* branchwood. (**l**) Runkel ratio of fibers in *T. mandshurica* stemwood.

**Table 4.** Skewness and kurtosis (data in brackets) of normal distribution curves for fiber length, slenderness ratio and Runkel ratio. **Table 4.** Skewness and kurtosis (data in brackets) of normal distribution curves for fiber length, slenderness ratio and Runkel ratio.


The average slenderness ratio of the fibers for both branchwood and stemwood was smaller than the acceptable value (33) in papermaking [51]. About 10% of the fibers with the slenderness ratio were above 33 for *T. mandshurica*, and 25% for *T. amurensis* (Figure The average slenderness ratio of the fibers for both branchwood and stemwood was smaller than the acceptable value (33) in papermaking [51]. About 10% of the fibers with the slenderness ratio were above 33 for *T. mandshurica*, and 25% for *T. amurensis* (Figure 2e–h). Therefore, it is difficult for both species to produce high-quality pulp. In fact, many paper products on the market have slenderness ratio of less than 33 due to a severe shortage of

wood resources, as witnessed, for example, by study of paper and paper egg trays used in Southwestern Nigeria by Amoo et al. [52].

The average Runkel ratio of fibers in the branchwood was less than 1, which implied that the fibers would collapse and provide a large surface area for bonding during papermaking [46]. The distribution of the Runkel ratio were skewed to the right (Figure 2i–l), indicating that most of the data were below its average value. The distribution of the Runkel ratio in *T. mandshurica* branchwood was steep, with kurtosis up to 2.02 (Table 4), indicating that the data were relatively centralized. More than 55% of the fibers in stemwood had a Runkel ratio less than 1, while up to 80 % in branchwood. Therefore, the short fibers of branchwood can be mixed with some long fibers (for example, softwood fibers) in different proportions to produce particular products, such as newsprint, packaging, and hygienic tissue products [48,53].

#### *3.3. Physical and Mechanical Properties*

Wood density is generally believed to reflect fiber wall thickness [54]. However, these results showed that linden trees with thick-walled fibers (Table 3) did not appear to produce high wood density (Table 5). Despite the low density of wood, the values still met the density requirement (0.3 to 0.5 g/cm<sup>3</sup> ) of papermaking raw material [55]. The correlation between wood density and pulp yield is still controversial, but the correlation between wood density and pulp and paper quality is recognized [7,56]. Colodette et al. [57] suggest that species with low density wood should be directed towards manufacturing refined paper (printing and writing grades). Referring to the research of Kennedy et al. [58], pulp produced from linden wood is suitable for manufacturing fine paper. Wood density was similar between stemwood and branchwood. Similar wood density creates a favorable condition for mixing stemwood and branchwood in cooking to optimize the production process and pulp quality [56].


**Table 5.** Comparison of physical and mechanical properties of the branchwood and stemwood from two *Tilia* species <sup>1</sup> .

<sup>1</sup> The data of branchwood is represented as Means <sup>±</sup> SE.

Generally, the mechanical strength is weak in species with low wood density [60]. The compressive strength of linden wood is below 45 MPa, and even below the values of fast-growing *Populus deltoides* (36.46 MPa) reported by Feng [61], belonging to the low strength range according to the classification of wood mechanical properties [62]. Although having similar densities, the branchwood has a slightly lower compressive strength than the stemwood's value [59]. The low density and strength of linden wood confirm that it is rarely used in construction (see Section 1). Still, it also implies less energy consumption in mechanical chipping [63] and an enormous potential to produce printing and writing sheets.

#### *3.4. Chemical Properties*

Table 6 shows that the branchwood of *T. amurensis* had more cold and hot water extractives than the stemwood reported by Lu [64], suggesting that branchwood contained more water-soluble material, such as starch and soluble sugar. This suggestion has already been demonstrated in previous studies [27,65]. The cellulose content in *T. amurensis* branchwood was lower (36.73%) and hemicellulose content was higher (30.93%) than in stemwood, and the contents exceeded the reference 45%–50% for cellulose content and 20%–25% for hemicellulose content [66]. Similar results were found in other tree species [45,67]. A high proportion of sapwood in the branches may be the main reason [45]. Still, branchwood

contained a large amount of holocellulose (total cellulose and hemicellulose) and a small amount of lignin (22.34%), which was beneficial for pulping [8,68]. The branchwood of *T. mandshurica* showing cold and hot water extractives were not much different from those of *T. amurensis*. However, NaOH extractive in the branchwood of *T. mandshurica* was about 6% less than that of *T. amurensis*. This may be because the branchwood of *T. mandshurica* had less hemicellulose (20.98%) than that of *T. amurensis*. Some studies [69,70] have shown that although hemicellulose is a structural carbohydrate, it is not as stable as cellulose. Wood treated with 1% NaOH can dissolve a portion of hemicellulose in addition to water-soluble substances. Thus, low hemicellulose content in *T. mandshurica* branchwood leads to low 1% NaOH extractives. The cellulose content in *T. mandshurica* branchwood was close to that of *T. amurensis* branchwood. However, *T. mandshurica* branchwood had a high lignin content (30.48%), and the content was almost close to the critical reference 20%–30% [66], which was not beneficial for pulping [68].


**Table 6.** Comparison of chemical composition of branchwood and stemwood from two *Tilia* species <sup>1</sup> .

<sup>1</sup> The data is represented as Means <sup>±</sup> SE. <sup>2</sup> Data in bold indicate significant differences between species at *<sup>p</sup>* = 0.01. — indicate no data for reference.

#### **4. Conclusions**

The branchwood and stemwood showed significant differences in their tissue proportions and fiber dimensions despite showing similar cell arrangements (*p* < 0.05). The branchwood have more than 40% fiber proportion, indicating there is the potential to obtain high pulp yield. The branchwood have a short, narrow, and thin fiber dimensions and small slenderness ratio, but a suitable Runkel ratio to papermaking. Branchwood fibers can be mixed with large fibers from other tree species to produce specific paper products. The density and compressive strength of branchwood and the reported stemwood values are similar and relatively weak, which would be beneficial for mechanical chipping. Compared to stemwood, *T. amurensis* branchwood has less cellulose but more holocellulose (about 67%) and less lignin (22%), which was beneficial for pulping. *T. mandshurica* branchwood may not be suitable for chemical pulping due to its high lignin content (up to 30%).

**Author Contributions:** Conceptualization, P.G., X.Z. and Q.F.; methodology, P.G. and X.Z.; software, P.G., Q.F. and Y.Y.; validation, P.G., X.Z. and Q.F.; formal analysis, P.G. and X.Z.; investigation, Q.F. and Y.Y.; resources, X.Z.; data curation, P.G. and Y.Y.; writing—original draft preparation, P.G.; writing review and editing, X.Z., Q.F. and Y.Y.; visualization, P.G.; supervision, X.Z.; project administration, P.G. and X.Z.; funding acquisition, X.Z. All authors have read and agreed to the published version of the manuscript.

**Funding:** This research was funded by the National Science Foundation of China, grant number 32171701.

**Data Availability Statement:** The data used in the study are published in this paper.

**Acknowledgments:** The authors would like to thank Xingchang Wang of the Maoershan Forest Ecosystem Research Station for collecting samples. We thank a thoughtful reviewer and editor for helping to improve this work.

**Conflicts of Interest:** The authors declare no conflict of interest.

#### **References**


**Disclaimer/Publisher's Note:** The statements, opinions and data contained in all publications are solely those of the individual author(s) and contributor(s) and not of MDPI and/or the editor(s). MDPI and/or the editor(s) disclaim responsibility for any injury to people or property resulting from any ideas, methods, instructions or products referred to in the content.
