BASIC WOOD DENSITY, FIBER DIMENSIONS, AND WOOD CHEMICAL COMPOSITION OF FOUR Eucalyptus SPECIES PLANTED IN SOUTHERN BRAZIL

ABSTRACT In the Brazilian planted forest sector, most of the species used are from the genus Eucalyptus. Even though Southern Brazil has a suitable climate for species of the genus Pinus, the planting of frost-resistant Eucalyptus species has been increasing annually. The objective of this study was to evaluate the basic density, fiber dimensions, and chemical composition of Eucalyptus benthamii, Eucalyptus dunnii, Eucalyptus saligna, and Eucalyptus cloeziana woods. The trees used were from a six-year-old experimental plantation located in Canoinhas, state of Santa Catarina. For each species, three trees were selected, and discs were removed from each tree at 0%, 25%, 50%, 75%, and 100% of the commercial stem height. To evaluate the quality of the wood, the basic wood density, fiber dimensions, and chemical composition of the wood were determined by comparing the values between species and between the heights in the stem. The highest basic density was that of the wood of E. cloeziana; this species also had the greatest length and width of fibers. E. dunnii had the lowest levels of lignin and the highest levels of holocellulose. The highest lignin content was found in the wood of E. benthamii, E. saligna, and E. cloeziana. The relationship between wood properties and stem height was not relevant. The results highlight the tendency for higher extractives to be found at taller heights (100%), and holocellulose and lignin content are similar at all heights.


INTRODUCTION
In southern Brazil, the edaphoclimatic conditions are favorable for Pinus spp. plantations, with the Santa Catarina state being the second largest holder of forests of this genus. However, plantations of Eucalyptus spp. in southern Brazil have been increasing, and between 2009 and 2020, the area of Eucalyptus spp. in Santa Catarina more than doubled, from 100,140 hectares to 273,116 hectares, representing more than a third of all planted forests in that state (Indústria Brasileira de Árvores -IBÁ, 2021).
Eucalyptus spp. were introduced in Santa Catarina through the exploration of frost-resistant species, led mainly by companies in the pulp and paper sector. However, the wood of these species is also versatile and can be used to produce sawn wood in long cycles of rotations, energy, and wood panels (Oliveira and Pinto Júnior, 2021). Furthermore, wood from Eucalyptus spp. is harvested in short cycles of rotations of 6 to 7 years, while a similar volume of Pinus spp. takes 10 to 12 years to achieve (Lengowski et al., 2020).
Eucalyptus plantations in Southern Brazil show rapid growth and high productivity due to abundant rainfall, but many Eucalyptus species have low frost tolerance, a common climate stress in this region. The most planted tree in Brazil is the hybrid Eucalyptus grandis × Eucalyptus urophylla (Lengowski et al., 2020). However, this hybrid is not frost-resistant (Kleinpaul et al., 2010) and is not used in the expansion of Eucalyptus plantations in southern Brazil, for which other species are being tested.
Knowledge of wood properties, such as basic density, chemical composition, and fi ber dimensions, is of great importance in determining its proper application (Vivian et al., 2015) and diversifying the timber market and silvicultural expansion. This information is also useful in targeting forest genetic breeding, which not only seeks gain in volumetric productivity, but also an improvement of wood characteristics and phytosanitary resistance (Grattapaglia, 2021).
The basic density of wood is one of the most relevant properties because it is related to most other wood properties (Gonçalez et al., 2014), and its importance is reinforced by its ease of determination (Dias et al., 2018), whereas the dimensions of the fi bers are important because they are directly correlated with the properties of the paper produced (Pirralho et al., 2014). Finally, the chemical characteristics are useful for predicting and understanding the performance of wood in chemical processing, for energy production and in papermaking (Fearon et al., 2020).
Considering the expansion of Eucalyptus plantations in southern Brazil is recent and there is a lack of informations about wood from these plantations, the objective of this research was to evaluate the basic density, fi ber dimensions, and chemical composition of the wood of Eucalyptus benthamii Maiden & Cambage, Eucalyptus dunnii Maiden, Eucalyptus saligna Sm., and Eucalyptus cloeziana F. Muell., planted in Canoinhas in the state of Santa Catarina.

Wood origins and sampling
The wood specimens were obtained from the rural district of Marcílio Dias in Canoinhas, state of Santa Catarina, Brazil. The plantation is located at latitude 26°07′37′′ S, longitude 050°23′41′′ W, and at an altitude of 839 m above sea level (DATUM: SIRGAS2000). According to Köppen's classifi cation, the local climate is of the Cfb type, without a defi ned dry season, with cool summers and frequent frosts in winter (Alves et al., 2013).
The forest stand was a mixed experimental plantation in which the trees were planted in a 3 × 3 m spacing level in 2011. The species E. benthamii, E. dunnii, and E. saligna are suitable for the city of Canoinhas because they are frost-tolerant (Flores et al., 2016;Bonfatti Júnior and Lengowski, 2017), whereas E. cloeziana does not have satisfactory growth in the municipality (Bonfatti Júnior and Lengowski, 2017).
The sample collection was undertaken in 2017 when the plantation was six years old. From each species, three trees were randomly selected, and discs were removed at 0%, 25%, 50%, 75%, and 100% of the commercial stem height (minimum usable diameter: 6cm). Each wooden disc was divided into six wedges; a pair of opposing wedges was used to determine the basic wood density, another pair was ground to determine the chemical composition of wood, and the last pair was reduced to small fragments for the measurement of fi ber dimensions.

Wood properties
The basic wood density was determined by mass weighing and indirect volume measurement method according to NBR 11941:2003(Associação Brasileira de Normas Técnicas -ABNT, 2003 standard in two samples per disc, totalling six replicates per stem height per species. For the fi ber dimensions, small wood fragments were macerated. From the macerated tissue, the length, width, and wall fraction of 100 fi bers per stem height were measured using an optical microscope following the guidelines of the International Association of Wood Anatomists -IAWA (1989), totalling 300 replicates per stem height per species. Wood samples for chemical analysis were prepared according to TAPPI T 257 sp-14 (Technical Association of Pulp and Paper Industry -TAPPI, 2014) in fi ve samples per species. The total extractive content according to TAPPI T 204 cm-17 (TAPPI, 2017), acid-insoluble lignin content according to TAPPI T 222 om-15 (TAPPI, 2015), acid-soluble lignin content according to Goldschimid (1971), and ash content according to TAPPI T 211 om-16 (TAPPI, 2016) were determined. The total lignin content was determined by the sum of the acid insoluble and acid soluble lignin contents, and the holocellulose content was calculated by subtracting the total content of extractives and the total lignin content from 100%. Three replicates were determined, totalling nine replicates per stem height per species.

Statistical procedures
To check for residual normality and variance homoscedasticity, the data were subjected to the Shapiro-Wilk test and Bartlett test, respectively. How these assumptions were achieved, the data were subjected to analysis of variance (ANOVA), and when signifi cant diff erences were detected between species, the Tukey test was used at 5% of signifi cance. To assess the variation in wood properties in the axial direction of the stem (bottom to top), the data were plotted in histograms. All the statistical procedures were performed using the R language, version 4.1.0 (R Core Team, 2021).

Comparison between species
The results of the basic wood density and fi ber dimensions of the four evaluated Eucalyptus species, considering the diff erences between species, are shown in Table 1. E. cloeziana had a higher density at all heights, whereas E. benthamii and E. dunnii had lower densities E. benthamii had the highest content of extractives, E. dunnii had the lowest levels of lignin and the highest levels of holocellulose and ash, and E. cloeziana had the lowest ash content. The highest lignin content was found in E. benthamii, E. saligna, and E. cloeziana, and did not diff er statistically at all heights (Table 2).

Variation of wood properties in the axial direction of the stem
The four species showed diff erent patterns in the basic density profi le of wood. E. benthammii had the highest basic wood density at 0% and declined followed by stabilization at other heights. The highest basic wood density of E. dunnii was found at heights 0 and 50%, and the lowest, with no statistical diff erences between them, were found at heights, 25, 50, and 75%. Basic wood density of E. saligna showed a tendency to increase with increasing height, while E. cloeziana had a constant basic wood density along the stem (Figure 1).
The eff ect of height on the fi ber dimensions is shown in Figure 2, in which Figure 2A shows the fi ber length results, Figure 2B shows the fi ber width results, and Figure 2C shows the wall fraction results.
The fi ber length did not vary in the wood samples of E. benthamii and E. cloeziana, whereas in E. dunnii and E. saligna, the smallest lengths were found at 100% height. E. dunnii had longer fi bers at heights 0.50 and 75 %, and E. saligna had longer fi bers at 75%   height. Fiber width tended to decrease with increasing height in E. benthamii, E. dunnii, and E. saligna, whereas the fi ber widths of E. cloeziana did not vary in relation to the axial position of the stem. The wall fraction of E. dunnii was maintained at all heights, whereas in E. saligna, the wall fraction tended to increase with increasing height. E. benthamii and E. cloeziana exhibited contrasting results; in the former species, the smallest wall fraction was found at the base of the stem, while in the latter, the largest wall fraction was found at the stem base. Figure 3 shows the results of the chemical composition of the wood, Figure 3A shows the total extractive content, Figure 3B shows the total lignin content, Figure 3C shows the holocellulose content, and the ash content is shown in Figure 4D.
The highest total content of extractives was found at 100% height; however, in E. saligna, this value was statistically diff erent from the value found in the base, whereas in the other three species they did not vary. The total lignin content did not vary with height in the wood samples of E. benthamii and E. saligna, whereas in E. dunnii and E. cloeziana, the lowest levels of total lignin were found at the tallest heights. The holocellulose content did not vary in E. saligna, but in E. benthamii and E. cloeziana, the highest values were found at intermediate heights of 25 and 50%, respectively. In E. dunnii, the highest content was found at 25 and 100% heights. E. saligna and E. cloeziana showed a tendency to have higher ash content with increasing stem height. E. dunnii had the highest levels at heights 0 and 100%, and E. benthamii did not show a clear trend of variation in the ash content, with the lowest content found at the height of 25% and the highest at the height of 100%. The ash contents at heights 0.50 and 75% heights were statistically equal to the highest values found for that species.

Comparison between species
The basic density values found were within the typical values of wood of Eucalyptus species cultivated in Brazil for pulp and paper production (0.43 and 0.55 g cm −3 ) (Gomide et al., 2005). Fast-growing trees are expected to have low basic density, as was the case of the value found (0.38 g cm −3 ) for fi ve-year-old E.
grandis wood from the state of Rio Grande do Sul, Brazil, by Cremonez et al. (2019). The four species in the present study presented satisfactory values for wood applications in which higher basic wood densities are favourable, such as to produce cellulosic pulp and energy (Brand, 2010;Smook, 2016).
Considering the fi ber dimensions, all wood samples were suitable for producing paper. The diff erences found in length are not relevant to the paper properties, as the cellulosic pulp produced from hardwoods is considered as short-fi ber pulp (Smook, 2016). However, the wood of E. benthamii is more suitable to produce printing and writing paper as it has a lower wall fraction (Santos and Sansígolo, 2007), while the wood of other species is more suitable for producing tissue papers for absorbent purposes, as fi bers with a higher wall fraction absorb more liquid (Santos and Sansígolo, 2007).
To produce pulp, the lignin is removed by chemicals for fi ber individualization; therefore, the lignin content directly aff ects the process because this compound makes pulping diffi cult (Lengowski et al., 2020). However, the presence of lignin and total extractives is favorable for energy production, whereas ash is undesirable for any use (Wastowski, 2018). Despite having the highest ash content, wood of E. dunnii is the most suitable for producing cellulosic pulp because it has a higher holocellulose content and a lower lignin content. The highest levels of lignin were found in E. benthamii, E. saligna, and E. cloeziana, among which the most suitable for energy production was E. cloeziana, which had the lowest ash content.

Variation of wood properties in the axial direction of the stem
The idiosyncrasy of a decrease/stabilization at 0-25% height, followed by an increase at 25-75% height, and a decrease/stabilization in the fi nal portion of the commercial height constitutes a common model. This pattern of longitudinal variation of the basic wood density of trees of the Eucalyptus genus has been frequently reported in the literature (Alzate et al., 2005;Sette Junior et al., 2012;Gonçalez et al., 2014). In the present study, only E. dunnii fi t into this model, contrasting the fi ndings of Lopes et al. (2011) that there were no signifi cant diff erences in basic wood density at diff erent heights on the stems of E. dunnii trees.
The basic density of E. benthamii was higher at the base than at the other heights, which is in line with the fi ndings of Benin et al. (2017). E. saligna, on the other hand, showed an increase in basic density up to the apex, a pattern similar to that described by Alzate et al. (2005), whereas E. cloeziana did not show diff erences between heights. This homogeneity of the basic density along the stem of E. cloeziana was earlier reported by Sturion et al. (1987). It has been suggested that higher density is associated with the mechanical requirements to support the stem and canopy of the trees (Sette Junior et al., 2012), and that the four species studied have diff erent responses according to their mechanical requirements.
Earlier studies on the eff ect of stem height on the length of wood fi bers from species of the genus Eucalyptus have found varied patterns; the length increased at greater tree heights (Gonçalez et al., 2014), decreased with increasing height (Valente et al., 1992;Rocha et al., 2004;Jorge et al., 2000), as well as did not vary with tree height (Taylor, 1973). Therefore, it is diffi cult to characterize the pattern of fi ber length along the stem in Eucalyptus spp. This was the case for fi ber width and wall fraction as well; no standard pattern was found for these variables because anatomical variations along the stem for Eucalyptus species are not consistent (Wilkes, 1988).
The higher concentration of total extractives at the tallest height is related to the accelerated physiological activity of the tree in that portion due to the proximity to the leaves that are responsible for photosynthesis (Taiz et al., 2014). The higher content of extractives found at the base of the stem may be related to the conversion of sapwood into heartwood, in which extractives are formed and deposited in the wood (Silva and Trugilho, 2007).
It is expected that the lignin content will increase with the age of the tree and the holocellulose content will decrease until they stabilize (Valente et al., 1992;Silva et al., 2005). It can be considered that E. saligna trees had already reached this stabilization and trees of the other species were close to this, as there were few discs that diff ered signifi cantly, both in terms of the lignin content and the holocellulose content.
The variation in ash content is not commonly a research topic of wood chemistry studies, which is more commonly used to evaluate the general presence of ash and its eff ects on wood properties. In the present study, the highest ash content was found at the tallest height, being statistically equal to the 0% height position in E. benthamii and E. dunnii. In general, the variation in ash content was not similar to that of any of the other wood chemical constituents studied.

CONCLUSIONS
The wood with the highest basic density was found in E. cloeziana, and those with the lowest density was found in E. benthamii and E. dunnii.
The length of the fi bers of E. cloeziana was the largest, being the only one that exceeded 1 mm, and fi bers of greater width were also found in E. cloeziana; however only fi bers at 50% and 70% tree heights were in fact signifi cantly superior. The fi bers with the highest wall fraction were those of E. dunnii.
Chemical composition analysis showed that E. dunnii wood was more suitable for pulping with the lowest lignin content and the highest holocellulose content. E. benthamii, E. saligna, and E. cloeziana had higher lignin contents.
The relationship between Eucalyptus wood properties and stem height was considered weak. In terms of basic wood density, each species exhibited a diff erent pattern.
Considering the fi ber length, E. benthamii and E. cloeziana showed no variation in this variable along the stem; despite the diff erent variations, E. dunnii and E. saligna had the smallest lengths at the tallest height. It was not possible to determine any consistent behavior for the fi ber width and wall fraction.
For extractives, it can be stated there is a tendency for a higher concentration of them at the tallest height of the tree (100%), and the content of total lignin and holocellulose tend to be similar at any height of the tree.

AUTHOR CONTRIBUTIONS
Eraldo Antonio Bonfatti Júnior designed the research Project, wrote the paper obtained and discussed the statistical, experimental, and estimated data. Elaine Cristina Lengowski wrote the paper and supported the data analysis. Bruna Mulinari Cabral, Thiago Wendling Gonçalves de Oliveira, Jeinna Michelly Rodrigues de Barros, and Rudson Silva Oliveira supported the data analysis and the obtaining of experimental results. Alan Sulato de Andrade, Umberto Klock, and Dimas Agostinho da Silva obtained the experimental results and supported the writing of the paper and statistical procedures.