Among-family variations of chemical and Kraft-pulp properties in the heartwood of the third-generation Acacia mangium in Indonesia

The Acacia mangium sampled trees were 10-year-old and planted in Alas Kethu Special Purpose Forest Area in Wonogiri, Central Java, Indonesia (7°47′S, 110°56′E), using seedling-seed orchard from the third generation. The third-generation A. mangium is an open-pollinated family consisting of 40 families from two provenances and four sub-lines (Nirsatmanto et al. 2015; Masendra et al. 2023). At the age of 10 years, the progeny trial was thinned to keep the seedling-seed orchard. A total of 99 thinned individuals were used as the study materials, while the remaining trees formed the seedling-seed orchard. The selected trees represented 20 families (five families from each sub-line) with approximately five trees per family. Table 1 shows the mean and standard deviation of stem diameter at 1.3 m above the ground, tree height, and basic density for all selected trees. A log with a length of 30 cm was collected from the stem at a height of 1.0–1.3 m above the ground in each tree. The logs were then sawn into radial boards with a thickness of 30 mm and air-dried for two months at room temperature without an air-conditioner. The air-dried board was cut in half at the pith position. Due to different concentrations of extractives between inner heartwood (near the pith) and outer heartwood (near the boundary between heartwood and sapwood) in hardwood species (Hillis 1987), the heartwood was cut and separated into inner and outer heartwood. The differences in the two parts were investigated visually on the radial surface of the heartwood: the inner heartwood has a light color, and the outer heartwood has a dark color, as shown in Fig. 1 (Masendra et al. 2024).

Basic density was measured using the water displacement method (Panshin and de Zeeuw 1980). The strip specimens with a thickness of 1 cm from the pith to bark were collected at 1.3 m above the ground from 99 trees. The strip specimens were cut into small block specimens at 1 cm intervals from the pith toward the bark side. As a result, four to 14 block specimens (depending on the stem diameter) were obtained. The mean value of all block specimens from a tree was calculated, and then the calculated value was regarded as the representative value of the tree.

Wood powder (42–82 mesh) was prepared from the inner and outer heartwood sections of the radial boards using a rotary speed mill (P-14, Fritsch, Idar-Oberstein, Germany) to assess the chemical properties. Ash content was measured by heating the wood sample (1.0 g) in a muffle furnace (FO 100, Yamato, Tokyo, Japan) at 600 °C for 1 h, according to the method described by Kuroda (2000). Hot-water extract was determined by boiling 1.0 g of the sample in distilled water in a 200 mL Erlenmeyer flask connected with a Liebig condenser for 3 h. The 1% NaOH extract content was measured by heating a 1.0 g sample in 100 mL of 1% NaOH (Kanto Chemical Co., Tokyo, Japan) aqueous solution for 1 h in a water bath at 80 °C. To determine the content of organic-solvent extracts, a 5.0 g sample was extracted with a 120 mL mixture of 95% ethanol (EtOH) (Kanto Chemical Co., Japan) and toluene (Kanto Chemical Co., Japan) (1 : 2, v : v) for 6 h using a Soxhlet extractor. After extraction, the solvent was evaporated using a rotary evaporator (N-1100, EYELA, Tokyo, Japan). The ash and extractive contents were measured gravimetrically.

The content of Klason lignin was determined according to the modified method described by Kuroda (2000). The organic solvent extract-free wood powder (1.0 g) was reacted with 20 mL of 72% H₂SO₄ (FUJIILM Wako Pure Chemical Corporation, Osaka, Japan) for 4 h at room temperature to hydrolyze and solubilize the carbohydrates. The suspension was then diluted with 765 mL of distilled water to reduce the H₂SO₄ concentration to 3% and further boiled for 2 h. Then, the suspension was allowed to settle before filtration. The residue was washed with 500 mL of hot-distilled water until reaching a neutral pH. The obtained insoluble residue was determined as the Klason-lignin content.

Holocellulose content was determined according to the method described by Kuroda (2000). The organic solvent extract-free sample (2.5 g) and distilled water (150 mL) were put in a 300 mL Erlenmeyer flask. A 0.2 mL CH₃COOH (FUJIFILM Wako Pure Chemical Corporation, Osaka, Japan) and 1.0 g NaClO₂ (Kanto Chemical Co., Tokyo, Japan) were added to the flask and heated in a water bath at 70 °C. CH₃COOH and NaClO₂ were added to the flask every hour for 3 h. The mixture was cooled down, and the residue was filtered and washed with 500 mL of distilled water and 50 mL of acetone (FUJIFILM Wako Pure Chemical Corporation, Osaka, Japan). The residue was dried in the oven at 105 ± 3 °C and weighed to determine holocellulose content.

The content of α-cellulose was determined using a holocellulose sample. The 25 mL of 17.5% NaOH-aqueous solution was added to a 1.0 g holocellulose sample in a 50 mL beaker, the sample was left for 4 min, and then homogenized with a glass stick for 5 min. After 21 min, 25 mL of distilled water was added to the suspension and then left for 5 min. The content in a beaker was filtered within 5 min, the residue was washed first with 200 mL of distilled water, and the solution was collected to determine β- and γ-cellulose contents. Then, the residue was washed with 40 mL of 10% CH₃COOH and left for 5 min. After that, the residue was rewashed with 1000 mL of distilled water and dried in the oven at 105 ± 3 °C. The content of α-cellulose was determined gravimetrically (Kuroda 2000).

To determine β-cellulose, the sample solution after α-cellulose determination was diluted to a volume of 800 mL. The 40 mL of 30% CH₃COOH was added to the solution and then heated until the color changed. The solution was cooled for 2 h, and filtered through a paper filter (No. 2, Whatman, Toyo Roshi Kaisha Limited, Tokyo, Japan), and dried in the oven at 105 ± 3 °C. The content of β-cellulose was determined gravimetrically. The γ-cellulose proportion was calculated by subtracting α- and β-cellulose from holocellulose (Kuroda 2000).

Wood sticks (approximately 1 [T] × 1 [R] × 20 [L] mm) were cut from the inner and outer heartwood sections of the radial boards to assess Kraft-pulp properties. A 5 g sample of wood sticks was used for Kraft pulping under the following conditions: an active alkali charge of 18% (as NaOH), a sulfidity index of 25%, a liquor-to-wood ratio of 4:1, and a cooking time of 90 min at 170 °C. After cooking, the cooked pulp was defibrillated with a pestle and a mortar. The cooked pulp was filtered through a glass filter (1G1) and then washed with 2 L of distilled water. The washed pulp was dried at 105 °C and then weighed as Kraft-pulp yield (Istikowati et al. 2016).

The kappa number was determined according to the Japanese Industrial Standards (JIS P8211: 2011), applicable for measurements within the range of 5–100. Briefly, 50 mL of sulfuric acid was mixed with 50 mL of 0.02 M potassium permanganate solution (FUJIFILM Wako Pure Chemical Corporation, Osaka, Japan) in a 1 L conical beaker. To the mixture, 1 g of oven-dried Kraft-pulp was added. The sample solution was then diluted to 500 mL with distilled water. The suspension was then mixed with a magnetic stirrer for 10 min. Ten mL of 1 M potassium iodide solution (Kanto Chemical Co., Tokyo, Japan) was added to the sample mixture. The mixture was then titrated using 0.2 M sodium thiosulfate solution (FUJIFILM Wako Chemicals, Osaka, Japan) until the color changed. The kappa number was calculated by the following equations (Eqs. 1 and 2):

where KN is kappa number, p is the volume of 0.02 M potassium permanganate solution (mL) consumed by pulp sample, w is oven-dried Kraft-pulp weight (g), b is the volume of 0.2 M sodium-thiosulfate solution consumed in the blank determination, a is the volume of 0.2 M sodium-thiosulfate solution (mL) consumed by kraft-pulp sample, C is the molar concentration (M) of sodium-thiosulfate solution, and f is a correction factor to 50% (mass/mass) potassium permanganate consumption that depending on the value of p.

The R software version 4.3.2 (R Core Team 2023) was used for statistical analysis. The phenotypic variations among families in the measured traits were assessed using a linear mixed-effects model. The variance components of the measured traits were estimated using the “lmer” function in the “lme4” package (Bates et al. 2015), with family as a random effect (3).

where Y_ij is the trait at the ith tree of the jth family, µ is the grand mean, Family_j is the random effect of the jth family, and e_ij is the residual. Furthermore, to categorize the 20 families for optimal wood utilization, principal component analysis (PCA) and cluster analysis were applied. Phenotypic values (the family mean) in each measured trait were used for PCA. The cluster analysis was applied using the Ward hierarchical clustering algorithm, which uses the 1 st and 2nd principal component scores from PCA as variables.

To evaluate the relationships between the traits, phenotypic correlation coefficients (r) were calculated using Pearson’s correlation coefficient based on the values of both the inner and outer heartwood samples. Moreover, a t-test was performed to differentiate between the traits in the inner and outer heartwood samples.

Table 2 shows the extractive and ash contents in the 10-year-old third-generation A. mangium. The mean values of inner heartwood in each sub-line ranged from 5.9 to 7.7% for hot-water extracts, 14.4–16.5% for alkali extract contents, 5.1–6.3% for organic-solvent extracts, and 0.5–0.7% for ash content. Meanwhile, the outer heartwood ranged from 7.0 to 7.8% for hot-water extracts, 16.2–17.0% for alkali extracts, 7.9 to 8.7% for organic-solvent extracts, and 0.5–0.7% for ash content. Compared to the previous studies, the third-generation A. mangium showed higher hot-water, alkali, and organic-solvent extract contents than those reported for A. mangium 3-year-old (Susanto et al. 2025) and Acacia aulocarpa 27-year-old (Lukmandaru et al. 2025), respectively.

The results of the t-test showed significant differences between the inner and outer heartwood for hot-water and alkali extracts in sub-lines B and D, and organic-solvent extracts for all sub-lines (p < 0.05, Table 2). In woody species, the hot-water and alkali extracts contain nonstructural components of low molecular weight, such as simple sugars, degraded hemicellulose (xylans and mannans), polyphenols, fatty acids, and ash-related components, while organic-solvent extracts (EtOH-toluene extracts in the present study) represent fatty acids, steroids, and polyphenols (Hillis 1987). Those components are responsible for wood coloration in the heartwood formation to distinguish the difference between sapwood, outer heartwood, and inner heartwood (Hillis 1987). Previously, it was reported that extractive contents in the outer heartwood were higher than those in the inner heartwood in some hardwood species, such as Tectona grandis wood (Haupt et al. 2003; Lukmandaru and Takahashi 2008). Therefore, the higher values of extracts at the outer heartwood in A. mangium (Table 2) indicate that those components were deposited as newly formed heartwood substances during heartwood formation. In addition, the lower values of these components in the inner heartwood are also related to polymerization and autoxidation that occur in these substances within the inner heartwood (Masendra et al. 2024). On the other hand, extractives and ash contents are undesirable wood chemical components in the pulping process. These components are not part of wood cell walls that are not converted into cellulosic pulp (Sjöström and Alén 1999). Furthermore, they hinder the penetration of reagents, reduce pulping yield, increase bleaching costs, cause equipment impregnation, and reduce pulp quality (dos Santos et al. 2024; Neiva et al. 2024; Lukmandaru et al. 2025). As the fatty acid in A. mangium wood is the main component responsible for pitch deposition in cellulosic pulp (Lange and Hashim 2001; Pieterinen et al. 2004), considering the organic-solvent extracts from the 20 families may be important to understand pulp and paper quality from this species.

The mean values of inner heartwood in each sub-line ranged from 24.9 to 25.9% for Klason lignin, 77.9–79.7% for holocellulose, 43.8 to 46.1% for α-cellulose, 5.7–6.5% for β-cellulose, and 27.1 to 28.4% for γ-cellulose, respectively (Table 3). The outer heartwood values ranged from 26.0 to 27.9% for Klason lignin, 77.6–78.9% for holocellulose, 46.7 to 48.7% for α-cellulose, 6.1 to 6.7% for β-cellulose, and 22.7 to 26.1% for γ-cellulose, respectively. The Klason lignin content in the present study was lower than that of 4-year-old A. crassicarpa (Martins et al. 2020) and 27-year-old A. aulocarpa (Lukmandaru et al. 2025), but higher than that of 3-year-old A. mangium (Susanto et al. 2025). Furthermore, the α-cellulose content in this study was comparable to that reported for other Acacia species (Martins et al. 2020; Lukmandaru et al. 2025) but lower than that of A. mangium 3-year-old reported by Susanto et al. (2025).

The significant differences (p < 0.05) in Klason-lignin content between inner and outer heartwood samples were found in sub-lines C and D (Table 3). The holocellulose content differed in the inner and outer heartwood of sub-lines A and D. The α- and γ-celluloses were different between the inner and outer heartwood in all sub-lines, except for α-cellulose in sub-line D (p = 0.170). Hardwood lignin is composed of both guaiacyl and syringil units derived from trans-coniferyl and trans-sinapyl alcohols (Sakakibara and Sano 2001). The radial distribution of lignin in the hardwood species has been reported in the previous studies (Yokoi et al. 1999; Lhate et al. 2010; Feng et al. 2025): lignin content varies among species and either increases from the inner to the outer heartwood or follows the opposite trend. For example, in T. grandis, the decrease in lignin in the inner heartwood may be due to a higher syringyl-to-guaiacyl ratio (S/G ratio) (Feng et al. 2025). In E. camaldulensis, the S/G ratio was also higher on the pith side and decreased gradually towards the bark (Yokoi et al. 1999). In addition, the S/G ratio was strongly positively correlated with acid-soluble lignin content but weakly negatively correlated with Klason lignin in E. globulus (Alves and Rodrigues 2022). These findings suggest that syringyl units contribute to the solubilization of lignin under acidic conditions, resulting in lower Klason lignin content. Therefore, the lower Klason lignin content in the inner heartwood than that in the outer heartwood for all sub-lines (Table 3) might be due to its higher S/G ratio, which enhances lignin solubilization under acidic conditions. Furthermore, the lower lignin content of inner heartwood corresponded with the lower kappa number (Table 4), as described in Kraft-pulp properties section. On the other hand, α-cellulose generally indicates undegraded, high-molecular-weight cellulose; β- and γ-cellulose represent degraded cellulose and hemicelluloses, respectively (Baeza and Freer 2001). The distribution of polysaccharide content has been reported in broad-leaved species (Lhate et al. 2010; Lukmandaru 2018; Feng et al. 2025). For example, the contents of hemicellulose and cellulose decreased from the outer heartwood to the middle and outer heartwood in 18-year-old T. grandis (Feng et al. 2025). Eucalyptus pellita showed a higher cellulose content in the inner heartwood than in the outer heartwood, while hemicellulose had the opposite result (Lukmandaru 2018). However, in Icuria dunensis and Pterocarpus angolensis, the cellulose of outer heartwood was higher than inner heartwood, while the hemicellulose was conversely (Lhate et al. 2010). Thus, the present results corresponded to those of the broad-leaved species reported by Lhate et al. (2010). In addition, the degree of polymerization and crystallinity of cellulose increased with tree age in the species of Trema orientalis (Jahan and Mun 2005). Therefore, due to the outer heartwood produced lastly, the higher content of cellulose was accompanied by lower hemicellulose contents in the outer heartwood of the third-generation A. mangium of the present study (Table 3), indicating that the degree of polymerization and crystallinity of cellulose in this outer heartwood part is high. Lower lignin contents are desirable for pulping because they must be removed to increase cellulosic pulp yield (dos Santos et al. 2024). In contrast, since cellulose is the primary source of pulpwood, especially the dissolving pulp component, the α-cellulose content is crucial for pulp quantity. The higher the cellulose content in the pulp, the better the yield and quality of the final product (Lourenço et al. 2008; Martins et al. 2020). In addition, as γ-cellulose is similar to hemicellulose, the presence of this component may be responsible for the high consumption of alkali during the Kraft-cooking process. Furthermore, producing hexenuronic acid from glucuronoxylan (hemicellulose) during cooking is also undesirable in the pulp-bleaching process (Haque et al. 2019; Neiva et al. 2024).

The results of Kraft-pulp yield and kappa number are shown in Table 4. The mean values of inner heartwood in each sub-line ranged from 51.3 to 52.7% for Kraft-pulp yield and 18.0 to 20.1 for kappa number, respectively. The outer heartwood ranged from 49.6 to 51.3% for Kraft-pulp yield and 21.9–24.7 for kappa number. In comparison, the pulp yield in the present study was within the range reported for Acacia and Eucalyptus species in the previous studies (Stackpole et al. 2011; Martins et al. 2020), indicating that the third-generation A. mangium exhibits comparable pulping performance and fiber quality to other major pulpwood species.

The t-test results found significant differences between the inner and outer heartwood in the Kraft-pulp yield and kappa number for all sub-lines (p < 0.05, Table 4). According to the results in Table 4, inner heartwood generally might provide a higher Kraft-pulp yield, possibly due to its lower lignin content and higher cellulose fraction. Lower kappa numbers in the inner heartwood suggest better delignification efficiency during the pulping process, requiring less bleaching effort and chemical usage. In contrast, the higher kappa number and lower Kraft-pulp yield of outer heartwood imply better lignin retention, which could lead to increased chemical and energy consumption during Kraft pulping. However, the outer heartwood also contains a higher α-cellulose fraction, making it more suitable for dissolving pulp applications than Kraft-pulp (dos Santos et al. 2024). These differences highlight a challenge for the pulp industry, as maintaining uniform chip quality is critical for efficient and consistent pulping operations, yet they also provide an opportunity for product diversification.

Regarding pulp bleaching, the preferred kappa number for bleachable pulp of Acacia species is 15 to 25 (Malinen et al. 2006; Haque et al. 2019). Therefore, the present results indicate that the Kraft-pulp yield of the third-generation A. mangium is bleachable pulp, with a kappa number ranging from 18.0 to 24.7 (Table 4). Furthermore, due to the high concentration of organic-solvent extracts and γ-cellulose (Tables 2 and 3) in the heartwood of these families, besides bleaching, dispersing and washing pulp also might be needed to remove the fatty acids and hexenuronic acid compounds that cause low-quality for cellulosic pulp (Pieterinen et al. 2004; Haque et al. 2019; Lukmandaru et al. 2025).

It was reported that fast-growing tree species, such as E. globulus, E. camaldulensis, and A. crassicarpa, have high potential for pulp and paper production (Stackpole et al. 2011; Martins et al. 2020, 2024; Mboumba et al. 2022). In comparison, the present study is the first to examine the chemical and Kraft-pulp properties of inner and outer heartwood third-generation A. mangium, highlighting their potential for promoting product diversity and improved utilization of pulp and paper using this species in Indonesia.

Table 5 represents the phenotypic correlations (r). Significant positive correlations were found between stem diameter and tree height (r = 0.306) or basic density (r = 0.371). In the inner heartwood, negative correlations were observed between organic solvent extracts and stem diameter (r = − 0.423) or tree height (r = − 0.205). In the outer heartwood, α-cellulose content was positively correlated with stem diameter (r = 0.250) or basic density (r = 0.245). Positive correlations were found between Kraft-pulp yield and holocellulose or α-cellulose contents of 0.263 and 0.358 in the inner heartwood. Meanwhile, the correlation coefficients in the outer heartwood were 0.347 and 0.315. In addition, a negative correlation was found between Klason lignin and holocellulose contents in the inner heartwood (r = − 0.319) and outer heartwood (r = − 0.352), respectively. Furthermore, a negative correlation between Kraft-pulp yield and organic-solvent extracts was only found in the inner heartwood sample (r = − 0.441).

Relationships between growth and wood traits have been investigated in previous works on hardwood species (Quang et al. 2010; Istikowati et al. 2016; Sunarti et al. 2022). For example, a significant positive correlation was found between stem diameter and basic density in 4.5-year-old A. mangium second-generation planted in Indonesia (r = 0.400, p < 0.05) (Sunarti et al. 2022) or cellulose content in 10-year-old Eucalyptus europhylla planted in Vietnam (Quang et al. 2010). Basic density was positively correlated with α-cellulose content (r = 0.702, p < 0.01), but it was negatively correlated with kappa number in fast-growing tree species, namely, Artocarpus elasticus, Neolitsea latifolia, and Alphitonia excelsa from the secondary forest in Indonesia (Istikowati et al. 2016). In the present study, significant phenotypic correlations were found between growth and wood traits (Table 5), suggesting that wood traits related to pulp and paper quantity depend on growth traits. This positive phenotypic association in A. mangium suggests that trees with superior growth could also enhance pulp production.

Under nearly the same pulp cooking conditions as those of the present study, correlations between chemical components and Kraft-pulp properties have been reported in previous studies (Ona et al. 1995; Istikowati et al. 2016; Mboumba et al. 2022). Significant positive correlations of Kraft-pulp yield and holocellulose were found in E. camaldulensis (r = 0.820, p < 0.001) (Ona et al. 1995), and in A. elasticus, N. latifolia, and A. excelsa (r = 0.667, p < 0.05) (Istikowati et al. 2016). The Kraft-pulp yield also was positively correlated with α-cellulose content (r = 0.799, p < 0.001), but negatively correlated with extractive contents in E. camaldulensis (r = − 0.896, p < 0.001) (Ona et al. 1995), or Klason lignin in A. elasticus, N. latifolia, and A. excelsa (r = − 0.657, p < 0.001) (Istikowati et al. 2016). In E. camaldulensis 4-year-old planted in Thailand, the correlations between chemical components and Kraft-pulp properties were found, but not significant (Mboumba et al. 2022). In the present study, the relationships between chemical components and Kraft pulp properties corresponded to those of previous studies (Ona et al. 1995; Istikowati et al. 2016). These findings suggest that families with higher Kraft-pulp yield or α-cellulose content might result in a reduction in Klason lignin and organic solvent extracts, which are favorable responses for improving pulp and paper quality.

Table 6 summarizes the variance components of the family in measured properties based on mixed-effects model with the random effect of family. Organic solvent extracts of the outer heartwood and γ-cellulose of the inner heartwood showed nearly zero family variance, indicating negligible differences among families. For the remaining traits, the family variance component ratios ranged from 1.7% (Klason lignin of the inner heartwood) to 23.5% (Kraft-pulp yield of the outer heartwood). Relatively large variance component ratios were observed in ash (inner heartwood, 22.5%), holocellulose (inner heartwood, 22.4%), β-cellulose (inner and outer heartwood, 22.3 and 22.2%), hot-water extracts (inner heartwood, 21.2%), α-cellulose (outer heartwood, 20.1%), and stem diameter (19.2%). By contrast, Klason lignin exhibited relatively smaller family variance component ratios (1.7–8.8%), and the ratio in the organic-solvent extracts of inner heartwood was also low (4.9%). All fixed-effect parameters were statistically significant (p < 0.001). These results indicate that phenotypic differences among families were more pronounced for traits associated with polysaccharides and pulp yield, possibly reflecting differences in seed origin.

Figure 2 shows the PCA scores of breeding values for measured traits, a dendrogram based on the principal component (PC) scores, and loading values. The cumulative proportion of variance in PC1 and PC2 was 59.8% (Fig. 2). PC1 had high loading values for stem diameter, tree height, Klason lignin, α-cellulose, and Kraft-pulp yield, whereas basic density and organic-solvent extracts showed high loading values on PC2 (Fig. 2). The cluster analysis using obtained loading values revealed that 20 families were divided into two groups (Fig. 2). The eight families in group II showed superior radial growth, high Kraft-pulp yield, and low Klason lignin contents. The opposite trend was found in the 12 families in group I (Fig. 2). Therefore, the wood from the families of group II might be suitable for pulp and paper production. In addition, group II was dominated by almost all the families from sub-lines A and B, originating from PNG provenance. The growth characteristics and wood properties related to pulp and paper qualities may be due to differences in seed origin. In conclusion, families with superior growth tended to show higher pulp yield and favorable chemical composition, suggesting potential for improving pulp properties in the third-generation A. mangium.