Tropical mobola plum (Parinari Curatellifolia): a full characterization of wood and bark within the scope of biorefineries

The bark of P. curatellifolia is dark grey, rough, and corky (Fig. 1a), and the wood is brown to yellow-red. The average bark thickness was 1.2 (± 0.2) mm, including the rhytidome and the phloem. The phloem was 0.5 (± 0.1) mm thick and comprised the conducting phloem distinct from the nonconducting phloem by different colors due to the amount of sclerenchyma tissue. Growth increments were detected in the rhytidome and observable by the naked eye (Fig. 1b).

The general anatomy of P. curatellifolia bark had some similarities with another genus of the Chrysobalanaceae studied by Roth (1981), i.e., in some Licania sp. concerning the abundance of mechanical tissue, especially in the form of the nodules of sclereids, fibro-sclereids and reduced fibers, type of rays and dilatation tissue.

The rhytidome is moderately developed, appearing as small scales (Fig. 1b); a structural pattern in the rhytidome is recognized with a variable number of undulated periderm layers (2–4) intercalated with phloemic tissue, which includes compact nodules of sclerified tissues (Fig. 1b-e); in the periderm produced by the phellogen, the phellem cells (typical cork cells, Fig. 1c-g) with thin walls are relatively abundant and arranged regularly in radial rows, some with brown cell contents alternating with others, colorless and curving slightly, forming discontinuous arching layers in cross-section (Fig. 1c and e) and appearing somewhat radially stratified. The phelloderm is poorly developed, with 2–4 cells sometimes lignified with content and distinguished from cortical-like cells by their neat radial alignment (Fig. 1f-g); the observation of the phellogen is complex (Fig. 1g).

The phloem tissue produced by the vascular cambium is non-storied, comprised of the conducting and nonconducting phloem (Fig. 2a and c). The conducting phloem near the vascular cambium represents a tiny part of the phloem. It includes the sieve tube elements with companion cells (conducting cells), the axial parenchyma (storage tissue), rays (storage/transversal conduction), and the fibers (mechanical support) (Fig. 2a-b). The transition between conducting and nonconducting tissue is somewhat abrupt, marked by the collapse of sieve tube elements and the formation of prominent nodules of sclerenchyma cells (Fig. 2a and c).

In the transverse view, the sieve tube elements appear round to polygonal with thin and unlignified cell walls dispersed between parenchyma cells and fibers, solitary or in small groups (2–3 cells); they can distinguish from neighboring parenchyma cells by their somewhat larger diameter. The sieve tube elements presented a mean length of 423 μm (± 17), ranging between 127 and 670 μm and a tangential diameter of 25 μm (± 9), ranging between 6 and 42 μm. The companion cells are difficult to recognize in transverse and longitudinal sections. Sieve plates are scalariform compound oblique, with > 10 areas per plate equally spaced (Fig. 2b).

In the conducting phloem, the axial parenchyma is relatively scarce in small layers near the sieve tube elements; parenchyma cells appeared rectangular or polygonal in the transverse section and have thin unlignified walls, but in nonconducting phloem cells could enlarge or radially divide giving rise to a dilatation tissue between the fibers and nodules of cluster sclereids; in the outer phloem beneath the recently formed periderm, a considerable amount of cells like cortical-cells appeared (Fig. 1f); strands of crystal-bearing axial parenchyma of up to 10 cells were found along the margins of the fiber band (Fig. 3b) near the nodules of sclereids. These crystalliferous parenchyma strands accompanied by fibers were a conspicuous bark feature observed in other barks (Carlquist 2002; Junikka and Koek-Noorman 2007; Şen et al. 2011; Quilhó et al. 2013; Mota et al. 2021; Sousa et al. 2021).

The rays, when observed in a transverse section, followed a straight to an undulated direction in the initial phloem but soon distorted due to the collapse of the sieve tubes and early formation of nodules of clusters sclereids (Fig. 3c); rays are non-storied, mainly uniseriate (Fig. 3a) and homogeneous with procumbent cells (Fig. 3d); the cells composing the phloem rays had thickened walls and dark contents (Fig. 3a). Dilatation of rays is moderate toward the bark outside (Fig. 3c).

The sclerenchyma tissue was abundant and formed a high proportion of P. curatellifolia bark; sclerenchyma is primarily in the form of sclereids and fiber-sclereids, and comparatively few fibers developed. A structural pattern in the nonconducting tissue of this species is recognized, starting with the formation of round and conspicuous nodules of cluster sclereids arranged more or less parallel to the vascular cambium that progressively becomes radially aligned and diminish towards the periphery (Figs. 1b and 3c). According to Roth (1981) the abundance of mechanical tissue, especially in the form of sclereids, adds the hard and brittle consistency to the barks and constitutes a mechanical barrier against the collapse of living cells; this provides mechanical support of the tissue in the phloem of both tropical (Machado et al. 2005; Baptista et al. 2013) and temperate species (Şen et al. 2011; Cardoso et al. 2015; Vangeel et al. 2021).

Phenolic compounds were observed in rhytidome and nonconducting phloem by dark color staining in axial and ray parenchyma cells, cortical-like parenchyma, and sclereids (Fig. 1c and f). Crystals, presumably of calcium oxalate, mainly were present as prismatic crystals and found in parenchyma cells and sclereids (Figs. 3b and 4c). Starch grains were observed in our samples (Fig. 5a-b), as Halilu et al. (2008) also found in the bark of Parinari sp. Silica was also noticed (Fig. 3d), agreeing with records in the Flora Malesiana (1989) concerning the occurrence of silica in the phloem of Chrysobalanaceae. According to An and Xie (2022), few studies regarding phytoliths in bark exist. Still, the authors referred to studies in taxa from West Africa (Collura and Neumann 2017), where more than 90% of bark samples produced phytoliths, suggesting that silica production is concentrated in bark more than in wood.

Phenolic compounds and calcium oxalate crystals are common in the barks (Evert 2006) and, in conjunction with sclerenchyma, have evolved secondary defense functions (Franceschi et al. 2005).

The general wood structure of P. curatellifolia resembles the structure of other Parinari species described in InsideWood (2004), i.e., P. anamensis, P. campertis, P. congensis, P. excelsa, concerning the type of wood porosity, vessel pits, axial parenchyma, the composition of the rays and presence of silica grains.

Growth rings are poorly distinct and marked by the thick-walled latewood fibers (Fig. 6b). The xylem produced by the vascular cambium is non-storied. It includes the vascular elements (conducting cells), the axial parenchyma (storage tissue), rays (storage/transversal conduction), and the fibers (mechanical support) (Figs. 6b, 7a and 8a).

Vessels were diffuse-porous, of two classes of diameter (Fig. 6b-d), exclusively solitary (> 90%), and tended to be oval in outline with the oblique arrangement. The mean tangential diameter was 155 μm (± 16), ranging between 56 and 277 μm, which is within the range of the values reported in the literature, i.e., 100–200 μm or ≥ 200 μm for P. campestris (InsideWood 2004). Simple perforation plates in horizontal end walls were observed (Fig. 6c-d). Vessel-ray pits are oblique and vertical (Fig. 6c), and inter-vessel pits are minute and alternate (Fig. 6c). Helical thickenings are absent. The mean vessel element length was 678 μm (± 44), with values between 342 and 915 μm, following the range reported for Parinari species: 350–800 μm (InsideWood 2004). Numerous vasicentric tracheids with distinctly bordered pits were observed, most conspicuous in tangential sections (Fig. 7b).

Axial parenchyma is non-storied in narrow bands up to 3 cells wide (Fig. 6b) and in strands of over 5 cells (Fig. 7a-b). The parenchyma cells were rectangular and pitted (Fig. 7c); spiral thickenings were not observed in our samples, agreeing with observations of Ter Welle (1975); this author mentioned the presence of spiral thickenings in most representative genera of the Chrysobalanaceae, but highlight the lack of spiral in the genus Parinari. Crystals were also absent.

Libriform fibers were generally non-septate (Figs. 6d and 7c) and aligned in a radial row when observed in the transverse section. Fibers were very pitted with bordered pits in radial and tangential walls (Figs. 7b and 8b) and well defined in individualized fibers (Fig. 8c). Bordered pits are signaled by Feitosa et al. (2012) in fibers of this species.

The mean fiber length was 1 363 μm (± 19), ranging between 934 and 1 673 μm, which is in the range of values reported for various genus of Parinari: 900–1600 μm (InsideWood 2004) although shorter than those referred to by Massuque et al. (2021b): 1.80 mm. The fibers mean width was 29 μm (± 0.51), ranging from 18 to 42 μm; the fibers were thick-walled with a mean fiber wall thickness of 7 μm (± 0.15), ranging between 4 and 10 μm. These fiber width and wall thickness values follow the literature for this species, i.e., 25.16 μm and 7.5 μm, respectively (Massuque et al. 2021b).

Rays were non-storied exclusively uniseriate (Fig. 7a-b) with a mean ray height of 440 μm (± 22), ranging between 124 and 827 μm and 4–38 number of cells high. The rays were heterogeneous, with body ray cells procumbent with mostly 2–4 rows of upright and square marginal cells and sometimes rays with procumbent, square, and upright cells mixed throughout the ray (Fig. 8a-b). Radial canals were absent. Dark deposits were common in ray cells, but crystals were absent. Abundant silica was observed in ray cells and aggregations as irregularly shaped or globular bodies (Fig. 8b-c). Plants differ in their abilities to accumulate silica, varying significantly between species (Currie and Perry 2007). Silica bodies in ray cells were mentioned in various species of Chrysobalanaceae, i.e., Licania leptostachya, and as a diagnostic feature of species of different families, i.e., Lauraceae, Dipterocarpaceae (IAWA 1989). The role of phytoliths and microscopic amorphous silica structures as mechanical barriers against biotic and abiotic stress, and their use as taxonomic tools is highlighted by Nawaz et al. (2019).

Traumatic structures, i.e., pith fleck (Fig. 9a), presence of traumatic tyloses (Fig. 9b), and numerous traumatic thin vessels with tangential alignment (Fig. 9c), were detected, probably due in response to biotic or abiotic stress effects, i.e., drought, fire, insects, injury to the cambium that induces a disorganized formation of parenchymatous cells (IAWA 1964; Schweingruber 2007). According to Ceccantini (1996) and Mota et al. (2017), water deficits may cause the formation of pith flecks.

Wood and bark are heterogeneous materials due to their biological origin (Zobel and van Buijtenen 1989). Their formation involves production and growth of new cells by the vascular cambium and their subsequent differentiation, induced by hormonal signals, i.e., auxin, cytokinin, and gibberellin (Aloni 2021) and affected not only by internal aspects but also by external factors i.e. site and ecological (Zobel and van Buijtenen 1989; Meinzer et al. 2011; Benhura et al. 2013; Gricar et al. 2015). In addition, the dimensions of xylem and phloem cells have been shown to vary as a function of tree age, height, or tree diameter (Quilhó et al. 2000; Petit and Crivellaro 2014; Clerx et al. 2020), giving rise to variations in cell structure between trees in most species.

Table 1 summarizes the cell dimensions in the bark and wood of P. curatellifolia in studied trees. For bark, the height of the rays, the length and tangential diameter of the sieve-tube elements, and the length and lumen of the fibers were significantly different (p < 0. 05) between trees; also, for wood ANOVA revealed significant differences (p < 0. 05) concerning rays (height and nº cells), vessel elements (length and tangential diameter), and fiber (length, width, and lumen) among trees.

According to the results and when comparing the cell dimensions between bark and wood, the wood has larger cells than the bark (Table 1); i.e., bark fibers are shorter than those of the wood (on average, 1363.30 μm in wood and 1245.47 in bark) agreeing with Quilhó et al. (2013) in Quercus faginea, but in opposition to the results found in tropical trees by Parameswaran and Liese (1974) and in Eucalyptus globulus by Jorge et al. (2000).

There are few records concerning the correlation between xylem and phloem cells. Most were correlated with changes in conduit diameter, i.e., sieve tube diameter was strongly correlated with vessel diameter (Jacobsen et al. 2018). Still, our results also show that trees with large vessels have narrow sieve tube elements. Unfortunately, studies concerning bark and wood cell size variations between trees in Parinari sp. were not found in the literature.

As expected, the basic density of wood was higher than bark (580 vs. 505 kg m^− 3). P. curatellifolia, wood density values reported reached 593 kg.m^− 3 (Abbot et al. 1997) and 720 kg.m^− 3 (Orwa et al. 2009), which is in accordance with our results. However, no studies on the density of the bark were found. Density is crucial in determining a feedstock’s economics and fuel quality since it affects transport costs, combustion performance, and energy per volume.

Chemical summative analysis of wood and bark from P. curatellifolia is presented in Table 2.

The bark contains suberin in 5.4% (but no data about its chemical composition is available), a higher content in ash (3.2%), lignin (42.4%), and extractives (12.2%, where polar compounds make up 70.5% of the total); and a moderate content in polysaccharides (42.5%). Conversely, wood has a lower extractives content (10%), ash (1.6%), lignin (28.4%), and polysaccharides (52.8%), compared to bark and the values reported in the literature: 3.21% extractives, 2.55% ash, 29.59% lignin and 64.64% polysaccharides (Massuque et al. 2020). The differences could be attributed to geographical location, soil, and climate conditions, as Benhura et al. (2013) suggested. In general, P. curatellifolia presented a higher lignin content when compared with other miombo species, such as Julbernardia globiflora (23.8%), Brachystegia spiciformis (24.8%), Brachystegia boehmii (25.9%) and very similar to the species Uapaca kirkiana (29.4%), Pericopsis angolensis (29,2%) (Sangumbe et al. 2018; Massuque et al. 2021a).

The quantitative proportions of the chemical families identified by GC-MS in the DCM extracts of wood and bark are presented in Table 3. It was possible to identify about 92.2% (from wood) and 90.3% (from bark) of the total ion chromatogram area; there is a distinct difference in chemical composition between wood and bark. The wood revealed a high content of fatty acids (86.2%), where de hexadecanoic acid (C_16:0), 9,12-octadecadienoic acid (C_18:2), 9-octadecenoic acid (C_18:1) and octadecanoic acid (C_18:0) respectively with 28.4%, 17.2%, 30.8%, and 5.1% were the predominant compounds. In turn, the bark is rich in triterpenoids compounds (67.7%), where betulinic acid (32.2%), corosolic acid (10.1%), ursolic acid (8.8%), and oleanolic acid (6.5%) were identified as the main compounds. It is not easy to compare the results obtained since, to our knowledge, this is the first study reporting the lipophilic profile of P. curatellifolia. Various studies focus on the phytochemical screening of plant fractions from P. curatellifolia, such as leaves, barks, wood, seeds, and roots with different polar solvents. These studies have identified the presence of saponins, alkaloids, flavonoids, steroids, and tannins (Ogunbolude et al. 2009; Peni et al. 2010).

The yield of the extraction ethanol:water (80:20, v/v), phytochemical profile, and antioxidant activity are presented in Table 4. The extraction yield was similar between wood (8.57%) and bark (8.29%) and slightly lower than the content of the polar extractives determined by sequential solvent extraction (9.6% for wood and 11.3% for bark, Table 2). The Soxhlet apparatus can attribute this result. The sample is repeatedly brought into contact with fresh portions of solvent, and the system remains at a relatively high temperature, furthermore, no filtration is required after leaching, and sample yield can be increased by performing several simultaneous extractions (Luque de Castro and Priego-Capote 2010).

Bark presented the higher values for the phytochemical profile: total phenolics (443.8 mg GAE/g extract), flavonoids (461.0 mg CE/g extract), and condensed tannins (287.7 mg CE/g extract) in comparison with wood (102 mg GAE/g extract, 72.8 and 55.5 mg CE/g extract, Table 4). Generally, bark presents high contents of total phenolics (Ferreira et al. 2018).

Regarding the antioxidant activities, bark also presented higher FRAP values than wood (4.7 vs. 0.9 mM TEAC/g extract) and the same DPPH (1410 vs. 459 mg TEAC/g extract, Table 4). The antioxidant capacity expressed by AAI value was 8.8 in bark and 2.8 for wood, corresponding to a very strong antioxidant classification since AAI > 2 (Neiva et al. 2018). Overall, both tissues are interesting to be used as feedstock for antioxidant extraction and bio-based product development.

The results from analytical pyrolysis are presented in Table 5, where generally, there were no great differences between trees, so only the mean values will be presented and discussed. Bark presented higher content in lignin-derived compounds (33.0 vs. 31.8%) but lower carbohydrate derivatives when compared to wood (42.5 vs. 47.7%). In both tissues, the main compound identified was levoglucosan (peak 72, presented in Table SI 1), representing 7.6 (bark) and 8.8% (wood) of the total pyrogram area. Levoglucosan results from the depolymerization process of dehydrated cellulose (Dobele et al. 1999), and its yield is influenced by the cellulose characteristics, pyrolysis conditions, and inorganic compounds (Lourenço et al. 2019). The lignin in both bark and wood was mainly composed of guaiacyl and syringyl-units, reflected in the S/G ratio, whose values were 0.2 and 0.4, typical of hardwoods (Lourenço et al. 2019) and comparable to the values 0.35 attained by alkaline oxidation of Parinari curatellifolia wood (Massuque et al. 2021a), but lower compared to 0.89 and 1.72 attained by oxidation with copper (II) oxide of Brachystegia spiciformis and Pericopsis angolensis other Miombo species (Sangumbe et al. 2018). In bark, the H:G:S monomeric composition was 1:22:5, while in wood, it was 1:35:12, revealing a slightly higher content of guaiacyl units in bark (25.9 vs. 22.5%) and more syringyl units in wood (8.5 vs. 5.7%). Still, both had minor amounts of H-units (0.7 vs. 1.3 in wood and bark, respectively). The results of pyrolysis and wet chemistry analysis are different regarding lignin content. The difference was 5.2% for wood samples, similar to the ones reported by Costa et al. (2023). According to the previous author, this difference (6.1%) is mainly due to the formation of low molecular compounds during pyrolysis that can be produced either from carbohydrates or lignin but mainly accounts for the carbohydrate’s sum.

Table 6 presents the proximate, ultimate analysis, higher heating value (HHV), and mineral composition of P. curatellifolia samples. Generally, the bark had higher moisture content (9.0 vs. 8.0%), ash (3.3 vs. 1.6%), total volatiles (27.5 vs. 20.7%), and less fixed carbon (69.1 vs. 77.7%) compared to wood. The wood values are similar to the literature, except for ash content of 6.8%, the total volatiles of 21.5%, and fixed carbon of 71.7% (Massuque et al. 2021a). The higher ash content in the bark is a problem for biofuel production since they are responsible for the sintered molten deposit in the biomass boilers (Labbé et al. 2020).

The ultimate analysis presented more similar values between bark and wood, e.g., the ratio between hydrogen and carbon in bark and wood (0.12 vs. 0.13) and the ratio between oxygen and carbon (0.93 vs. 1.05). Bark presented slightly higher carbon content (48.7%) than wood (45.7%), but wood presented more oxygen than bark (48.1 vs. 45.2%). Interestingly, the hydrogen content was similar in both tissues (5.8 vs. 6.1%), while bark presented more than double nitrogen content (0.3%). Some of the data obtained here is similar to the literature: both bark and wood presented similar hydrogen 5.58%, but differ slightly in carbon (48.7 vs. 47.4%), nitrogen (0.3 vs. 0.9%), and oxygen (45.2 vs. 45.8%) (Massuque et al. 2020).

Regarding the HHV, there is not much difference between bark and wood (20.9 vs. 19.1 MJ/kg), but for wood of P. curatellifolia, there was reported a lower value of 18.5 MJ/kg (Massuque et al. 2020), slightly lower than 20.0 MJ/kg from wood of B. spiciformis and B. utilis, the favorite species of rural communities (Menéndez and Curt 2013). For the mineral composition, the most important minerals were potassium (2560.3 and 2387.8 mg/kg), calcium (6509.2 and 912.3 mg/kg), and magnesium (374.9 and 874.6 mg/kg), respectively, in bark and wood. The mineral composition is similar in fruits and seeds (Benhura et al. 2013; Lesten et al. 2018).