Introduction

Bamboo has gained significant attention in the construction industry due to its satisfactory mechanical properties, versatility, and positive impact on climate change through carbon dioxide sequestration during its cultivation. It also shows a lower global warming potential after processing compared to products made from traditional materials such as plastic, concrete, and steel [1]. However, despite this growing interest, bamboo is still not widely adopted in construction compared to conventional materials, mainly due to challenges related to durability, geometric anisotropy, and machining.

To improve bamboo’s physical and mechanical properties, it can undergo chemical and physical modifications such as the thermo-hydro-mechanical (THM) process [2]. This process increases the density, strength, and hardness of readily available and fast-growing bamboo by reducing pores and the spaces between fiber bundles [3, 4].

For example, Kadivar et al. [5] reported an 21% increase in apparent density (0.92 g/cm3) and 53% increase in flexural modulus of rupture (MOR) (127.7 MPa) for Phyllostachys edulis bamboo, and a 46% increase in density (1.11 g/cm3), and 42% increase in MOR for Dendrocalamus asper bamboo, using one-layer densified panels pressed at 140 °C with 2 MPa for 5 min for flattening, followed by 140 °C with 10 MPa for 20 min for densification. Similarly, Azadeh et al. [6] obtained an increase of 16% in apparent density (0.99 g/cm3), and 42% in MOR (241.3 MPa) for D. asper without outer skin and pith ring, pressed at 140 °C, with 4.34 MPa for 15 min. Notably, one-layer densified panels exhibited less variation in MOR when tested in flexural tension in both upward and downward positions (inner wall in tension or compression, respectively). This effect was attributed to the reduced gradient distribution of fibers across the wall thickness, in contrast with undensified samples where the inner layer contained fewer fibers than the outer layer. In another study, Li et al. [7] pretreated Phyllostachys bambusoides bamboo by partial delignification with an alkali solution before densification (150 °C, 5 MPa for 24 h), resulting in an 81% increase in apparent density (1.45 g/cm3) and 120% increase in MOR (327 MPa).

To produce multilayer bamboo panels or composites, however, more than one engineered bamboo panel must be connected with adhesives. Nkeuwa et al. [8] provided an extensive review on bamboo bonding and noted that many factors influence bonding quality, including bamboo characteristics (permeability, wettability, porosity of the natural material, presence of hydrophobic extractives on the outer skin, surface preparation), adhesive properties (viscosity, density, molecular weight), and gluing parameters (amount of adhesive, pressing time, pressing strength, temperature).

Zheng et al. [9] investigated the influence of assembling patterns without outer skin and pith ring (inner-inner, inner-outer, and outer-outer) on the bonding strength of glued P. edulis bamboo with phenolic resin. Panels glued through the inner layer (inner-inner) achieved higher compressive and shear strength (18.4 MPa), with failure occurring in the bamboo rather than the adhesive. The inner-outer and outer-outer samples averaged 13.8 MPa and 14.1 MPa, respectively.

Chen et al. [10] compared the gluing performance of bamboo composites with and without the pith ring and found that phenolic resin could penetrate the pith ring, achieving shear strengths comparable to laminates without inner skin (8.12 MPa and 9.71 MPa, respectively). Chen et al. [11] evaluated gluing curved bamboo laminates with PF resin after slightly treating the outer skin and the pith ring by sanding. This pretreatment improved surface wettability and permeability and removed the thin waxy outer skin. The study also examined gluing parameters (temperature, pressure, and time) and found temperature to have the greatest effect on shear strength, followed by pressure and time.

Polyurethane castor oil (PU) has recently gained attention as a promising resin for bamboo [12,13,14], offering a sustainable alternative to fossil-derived adhesives containing formaldehyde, a compound recognized as hazardous and toxic [15, 16].

Despite numerous studies on engineered bamboo glued materials, little research has addressed the gluing of densified panels, where the reduction in porosity and changes in microstructure directly affect adhesive penetration and bond formation. Furthermore, while PU has recently emerged as a sustainable alternative to fossil-based adhesives, its performance in densified bamboo composites remains largely unexplored. Considering the uneven fiber distribution across the wall thickness, the influence of the skin and pith ring on bonding effectiveness, and the effect of densification on the cross-section, this study investigates the physical and mechanical properties of two-layer flattened–densified D. asper bamboo panels glued with PU resin. Specifically, it evaluates how different configurations—fiber alignment (parallel or perpendicular) and layer orientation at the glue line (inner or outer layers with skin and pith ring)—affect panel performance.

Materials and methods

Materials and samples preparation

Dendrocalamus asper Backer ex K. Heyne bamboo was chosen due to its consistent fiber distribution in the radial direction, relatively large diameter, and thicker wall compared with other bamboo species. This makes it suitable for producing either densified or undensified bamboo laminates [17]. Three- to five-year-old bamboo culms were harvested at the University of São Paulo (USP-FZEA) (21°57′39′′S, 47°28′12′′W), with an average diameter of approximately 150 mm and a wall thickness of 14 mm.

The culms were treated using an 8% disodium octaborate tetrahydrate (DOT) aqueous solution in a pilot-scale immersion tank, as described by Gauss et al. [18], and then stored at room temperature in a protected environment for four months until reaching an equilibrium moisture content of 9–12%. From these treated culms, half-splits with a length of 300 mm were prepared, with outer curved widths ranging from 170 to 200 mm and thickness between 10 and 15 mm.

The panels (two-layered flattened-densified bamboo panels) were produced using PU resin supplied by the IMPERVEG company (Brazil). The PU resin is sustainable and ecological, 100% solid, biodegradable, compostable, and free of toxic solvents. The PU used in this study consists of two components: a prepolymer (component A) and a polyol (component B).

Flattening and densification

The manufacturing process of the glued flattened-densified (GFD) panels is shown in Fig. 1. Before pressing, the residual bamboo powder, dust, and surface contaminants were manually removed (Fig. 1 B). The half-split bamboo culms were soaked in water for 15 min (outer curved width of 170–200 mm, a length of 300 mm, and a thickness of 10–15 mm). The soaked bamboo culms were then heated for 20 min at 170 °C on the thermo-hydraulic press (HIDRAL-MAC, PHH, 10 ton).

Fig. 1
Fig. 1
Full size image

Bamboo panels manufacturing process. A half split B cleaned manually C soaked in water D driven in the hot press E flattened and densified F flattened and densified panel G adhesive H Flattened-densified panels Impregnated with the PU I adhesive application J Before hot and cold pressing, glue the panels as directed. K glued panels in hot press for 20 min L cold press for 24 h (M) glued flattened-densified panels

For the flattening–densification process, the half splits were pressed at 20 MPa pressure for 20 min at 170 °C. Following this process, the samples were kept under pressure at the same press for 24 h, without heating, to reduce internal stress within the densified panel and, consequently, minimize the spring-back effect [19].

PU adhesive was applied at a spread ratio of 200 g/m2 (similarly to Shah et al. [20]). The panels were hot-pressed at 110 °C for 20 min under 17 MPa pressure to ensure the activation and initial curing of the adhesive. Following this stage, the panels underwent cold pressing for 24 h under applied pressure to ensure proper curing and consolidation of the PU resin.

Sample arrangements

Figure 2 illustrates the sample arrangements for the gluing phase. The samples were divided into two main groups: panels glued in parallel (P) or perpendicular (PER). Each group was further subdivided into three categories based on layer orientation. It is important to note that the pith ring and the outer skin were not removed from the bamboo samples. For simplicity, orientations are referred to as inner and outer.

  • Parallel

    • P IN-IN: Inner surfaces facing each other.

    • P OUT-OUT: Outer surfaces facing each other.

    • P OUT-IN: One outer surface facing one inner surface.

  • Perpendicular

    • PER IN-IN: Inner surfaces facing each other.

    • PER OUT-OUT: Outer surfaces facing each other.

    • PER OUT-IN: One outer surface facing one inner surface.

Fig. 2
Fig. 2
Full size image

- Schematic of GFD panel groups with their orientations

Methods

Determination of chemical composition

The main chemical composition of the ash in the outer skin (cortex) and pith ring (inner layer) of bamboo was determined using an X-ray fluorescence (XRF) spectrometer. Pressed pellet samples were analyzed under STD-1 calibration with a Malvern Panalytical Zetium X-ray fluorescence spectrometer, covering chemical elements from fluorine to uranium. The results were normalized to 100%. Loss on ignition (LOI) was carried out at 1020 °C for 2 h.

Physical characterization

Apparent dry density, water absorption, and thickness swelling were determined for the panels. Measurements were taken using a digital balance (0.001 g precision) and a digital caliper (0.01 mm precision). For each of the six groups, nine samples were evaluated (54 samples in total).

Specimens with dimensions of (50 × 50 × t) mm were dried in an oven for 336 h at 60 °C until weight stabilization (mass difference < 0.5% between successive measurements). They were then immersed in water for 576 h (24 days) to evaluate water absorption and thickness swelling at short-term intervals (1, 3, 8, and 24 h) and long-term intervals (every 24 h until 576 h). Apparent dry density was determined after oven-drying.

Water absorption (WA) was calculated according to Eq. 1. Each sample was removed from water and surface-dried with a tissue before measuring.

$$WA \left(\%\right)=\left[\frac{{w}_{i}-{w}_{0}}{{w}_{0}}\right] \times 100$$
(1)

where wi is the weight after immersion (g) at each interval (1, 3, 8, 24, and each 24 h until 576 h), and w0 is the oven-dry weight before immersion.

Thickness swelling (TS) was determined by measuring oven-dry and wet specimen dimensions in the radial direction (Eq. 2):

$$TS \left(\%\right)=\left[\frac{{t}_{i}-{t}_{0}}{{t}_{0}}\right]\times 100$$
(2)

where ti is the thickness after immersion (mm) at each interval (1, 3, 8, 24, and each 24 h until 576 h), and t0 is the oven-dry thickness (mm). Three measurements were taken on each side of the samples, and the mean values were used for analysis (Fig. 3).

Fig. 3
Fig. 3
Full size image

Bamboo sample for physical characterization where the red lines were used to maintain consistency throughout measurements

Mechanical characterization

Specimens were prepared for three-point bending and shear tests. Dimensions were obtained using a digital caliper with precision of 0.01 mm. Samples were weighed prior to testing and then dried at (100 ± 2) °C for at least 48 h to establish moisture content at the time of the test.

The three-point bending test was carried out in a universal testing machine EMIC 23–300 INSTRON (Brazil) equipped with a 30 kN load cell (Fig. 4). Samples measured 300 mm in length, 48 mm in width, and 15 mm in thickness, with a span length of 240 mm (span-to-thickness ratio = 16:1). The test followed ASTM D7264:15 [21]. Displacement at midspan was measured with a deflectometer, and digital image correlation (DIC) was used as a complementary test to analyze strain distribution [22]. In parallel groups, fiber alignment matched the testing span direction, while in perpendicular groups, the lower layer fibers aligned with the span.

Fig. 4
Fig. 4
Full size image

Flexural test setup A three-point bending test B digital image correlation (DIC) system

The modulus of rupture (MOR), modulus of elasticity (MOE), and limit of proportionality (LOP) were calculated according to Eqs. 3, 4, and 5. The LOP corresponds to the maximum stress within the elastic range.

$$MOR=3{P}_{\text{max}}L/2b{t}^{2}$$
(3)
$$MOE=P{L}^{3}/4\Delta b{t}^{3}$$
(4)
$$LOP=3{P}_{\text{prop}}L/2b{t}^{2}$$
(5)

where Pmax is the maximum load at midspan; L is the span length; b and t are specimen width and thickness, respectively; Pprop is the maximum load within the elastic range; and Δ is midspan displacement.

Shear tests were performed with a mechanical servo-hydraulic universal testing machine (MTS model 370.02, Eden Prairie) following ASTM D3163 [23]. A total of 6 samples were tested per group. Specimen dimensions were: total length (L) = 215 mm, overlap length (L0) = 15 mm, grip-to-overlap length (Lg) = 63 mm, layer thickness (t) = 5 mm, and width (b) = 20 mm (Fig. 5).

Fig. 5
Fig. 5
Full size image

- Schematic of samples and the preparation for the shear test adapted from Shah et al. [20]

Scanning electron microscope (SEM)

Glue line sections were examined using a HITACHI TM-3000 scanning electron microscope with a 15 kV accelerating voltage and backscattered electron mode to observe bonding conditions in different panel orientations. Samples were polished using a TegraPol-11 Struers system, first with P320–P2000 grit sandpapers, followed by 6, 3, and 1 µm cloths with diamond solution.

Statistical analysis

All statistical analyses were performed in R software. One-way analysis of variance (ANOVA) was used to evaluate significant differences in mechanical properties, thickness swelling, and water absorption. When significant differences were found, Tukey’s post hoc test was applied, with a 95% confidence level (p < 0.05).

Results and discussion

Ash chemical composition in outer skin and pith ring

The chemical analysis (Table 1) shows that SiO2 represents 58.2% of the outer skin (cortex) of bamboo, compared to only 13.2% in the pith ring (inner layer). In the outer layer, silica deposits, in combination with cuticular waxes [24], create a hydrophobic and chemically inert surface that reduces wettability and adhesive penetration, particularly for polar or water-based adhesives. This barrier effect limits both chemical adhesion — by reducing the availability of hydroxyl groups — and mechanical interlocking, by blocking cell lumens and restricting resin infiltration into the cell wall. These findings are consistent with Yin et al. [25] who described a dense silicon-rich layer in the outer cortex that gradually decreases toward the inner surface. On the other hand, the pith ring exhibits lower SiO2 content and higher levels of CaO, MgO and P2O5, compositions that could enhance wettability and bonding due to its polarity.

Table 1 Chemical compositions of bamboo outer skin (cortex) and pith ring
Full size table

Physical properties

Table 2 displays the average apparent dry density for each of the 6 groups. The ANOVA statistical analysis indicated that there were no significant differences between them. The density of the GFD two-layered panels in this study reflects both the complete bamboo section and the glue line, with an overall average of 1.20 g/cm3. In comparison, Azadeh et al. [6] reported an apparent density of 0.99 g/cm3 for one-layer densified D. asper with skin, processed at 140 °C, 4.34 MPa for 15 min. Similarly, Kadivar et al. [5] found an oven-dry density of 1.11 g/cm3 for one-layered D. asper with outer skin and inner pith ring pressed at 10 MPa at 140 °C for 20 min. In the present study, half-splits were densified at 170 °C and 20 MPa for 20 min, followed by gluing at 110 °C to cure the PU resin. Thickness reduction was expected to occur only during densification. Given the similar pressing times across studies, the higher density obtained on the present investigation suggests that temperature and pressure are key parameters governing the degree of densification. This corroborates the findings of Kadivar et al. [19], which reported that higher temperatures result in greater densification.

Table 2 The average apparent density of GFD panels
Full size table

Figure 6 and Table 3 present the detailed analysis of water absorption (WA) over time. Statistical analyses are shown for 1, 24 and 576 h, as they are the most critical points. The parallel panels glued with the inner layers facing each other (P IN-IN) demonstrated significantly lower water absorption after 24 h (4.75%) and 576 h (36.22%). In contrast, parallel panels glued with the outer skin facing each other (P OUT-OUT) presented the highest WA values, reaching 45.86% after 576 h.

Fig. 6
Fig. 6
Full size image

a Short-term water absorption for all six groups b. long-term water absorption for all six groups

Table 3 The results of water absorption for different groups
Full size table

Differences in WA are associated with the layer exposed outward. The siliceous wax layer naturally present on the bamboo surface prevents liquid penetration, reducing both chemical treatment efficiency and bond strength [26].

For P IN-IN panels, the waxy outer layer had a higher contact area with water, and the high percentage of hydrophobic components such as silica (Table 1) reduced water uptake. Removal of these waxy layers by potassium hydroxide and sodium dodecyl sulfate has been shown to significantly increase absorption and swelling in Moso bamboo [26]. Even without the cortex, the outer layer still absorbs less water than the inner layer [27]. For the PER OUT-OUT group, the opposite occurs, with the inner layers exposed outward and absorbing more water. Interestingly, the pith ring was not removed in this study. Although its permeability (by airflow) has been reported to be lower than that of the outer skin [28], its lower contact angle compared to the outer cortex indicates higher wettability [11], which aligns with the higher WA observed.

Perpendicular groups followed a similar pattern during the first 24 h of immersion: PER OUT-OUT (inner surface exposed outward) absorbed more water (13.37%) than PER IN-IN (outer surface outward, 11.43%). However, the difference between perpendicular groups was smaller than that observed for parallel groups. After 576 h, both PER OUT-OUT and PER IN-IN reached similarly high values (43.39 and 42.88%, respectively; Table 3). Generally, bamboo fibers in the outer region are denser than in the inner region, resulting in structural and surface property differences. Compared with the inner side, the outer surface is harder to wet [9]. Since samples had equal surface areas (50 × 50 mm) and water exchange between glued panels is blocked by the glue line, it was expected to be shown to be a similar water absorption behavior for parallel and perpendicular groups.

Figure 7 and Table 4 show the thickness swelling (TS) results. After 24 h, the P IN-IN group (outer layers facing outwards) exhibited the lowest TS (5.50%), whereas the other five groups ranged from 12.61 to 16.90%, with no significant differences among them.

Fig. 7
Fig. 7
Full size image

a Short-term thickness swelling for all six groups b. long term thickness swelling for all six groups

Table 4 The results of thickness swelling for different groups
Full size table

At saturation (576 h), P OUT-OUT and PER IN-IN presented the highest TS values, 35.74 and 34.75%, respectively, following the same trend observed for WA. Interestingly, the perpendicular panels glued with outer layers facing each other (PER OUT-OUT) displayed the lowest TS of all groups, a pattern not observed in P OUT-OUT (inner layers outward). This may be related to the higher dimensional variation of fibers during sorption/desorption cycles, as their thick-walled cells swell more than thin-walled parenchyma cells [29]. In this dimensional stability analysis, however, the entire wall thickness is considered, and the panel arrangement mainly determines the exposed contact area with water. Thus, it was expected that panels with a higher WA presented a higher TS. While WA continued to increase until approximately 432 h (Fig. 6), TS stabilized earlier, around 200 h (Fig. 7).

Mechanical properties

All six groups (P OUT-OUT, P IN-IN, P OUT-IN, PER OUT-OUT, PER IN-IN, and PER OUT-IN) are represented by flexural resistance (a) and shear strength (b) in the Fig. 8, Table 5 and Table 6. A description of the bending properties of the different groups is given in Table 5 along with their modulus of elasticity (MOE), modulus of rupture (MOR), and limit of proportionality (LOP).

Fig. 8
Fig. 8
Full size image

a Flexural curves in all six groups b Shear strength for all six groups

Table 5 Mechanical properties of GFD panels for all six groups
Full size table
Table 6 The shear and tensile stress results of GFD panels for all six groups
Full size table

The matrix and vessels are compressed during bamboo densification, as illustrated by Kadivar et al. [30]. Matrix compressing and crushing have two significant effects on bamboo's mechanical behavior. First, when compression loads are applied, the fiber-to-fiber link is weakened, and each bundle behaves independently, increasing the likelihood of buckling. Second, in the compressed area, longitudinal shear strength between bamboo fiber layers is reduced [6].

Based on the statistical analysis of the MOR tests, Table 5 shows that parallel and perpendicular groups differ significantly, with MOR averages of 181.7, 200.3, and 122.4 MPa for the three parallel groups (P OUT-OUT, P IN-IN, and P OUT-IN), respectively. In contrast, the values obtained for three perpendicular groups (PER OUT-OUT, PER IN-IN, and PER OUT-IN) were 89.2, 119.0, and 128.2 MPa, respectively Table 1. A meaningful difference was also observed in the MOE of the panels glued in parallel versus perpendicular positions (Table 5). During the three-point bending test, it was observed that the perpendicular arrangement of the panels caused a significant reduction in the resistance. For linear elements such as beams, the preferable arrangement for higher resistance is parallel. However, for applications subjected to multi-directional loads, perpendicular panels would provide a more isotropic response.

Sclerenchyma cells with superior mechanical qualities are found in bamboo fibers [31, 32]. The hollow parenchyma cell performs poorly in mechanical properties [33,34,35]. Bamboo fiber serves as the reinforcement and parenchyma as the matrix when considering bamboo as a fiber-reinforced composite material [36]. From the outer to inner culm wall, the fiber content diminishes. Because the exterior side of bamboo contains a larger density of fibers, it frequently exhibits superior mechanical properties, such as greater tensile strength [37].

As expected, in the parallel configuration, the P IN-IN group exhibited the highest strength and elasticity Table 2. This can be attributed to the higher fiber content in the outer layer, which in this arrangement is positioned on both the top and bottom surfaces of the panel. The flexural behavior of the inner, middle, and exterior layers of D. asper bamboo was examined independently by Kadivar et al. [30], who found that the resistance increases from inner to outer layers. Moreover, during bending, fiber orientation at the tension face plays a decisive role: when outer layers are aligned with the tension zone (as in P IN-IN), bending resistance is maximized, whereas in perpendicular groups the discontinuity at glue lines interrupts load transfer, reducing overall strength.

The effect of densification on the mechanical properties for D. asper bamboo, investigated by Azadeh et al. [6], revealed that densification up to 34% results in 81.3% higher MOR and 36.1% higher MOE, respectively. In panels densified without skin, the MOR obtained was 241.3 MPa, whereas in panels densified with skin, resistance reached 311.8 MPa. Similarly, Kadivar et al. [5] reported that one-layered densified D. asper bamboo had a flexural resistance of 226.36 MPa. In contrast, the values obtained in the present study were lower, even for the P IN-IN group, which can be explained by the strong influence of the glue line when the pith ring and outer skin are retained in multi-layered panels. While keeping the outer skin improves flexural resistance in one-layer densified bamboo [5], in multi-layered configurations it reduces the enhancement from densification by concentrating stresses at the adhesive line, where both skin and pith ring hinder effective bonding Table 3. For this reason, in industrial practice they are often mechanically removed during composite manufacturing [38].

The findings of the DIC analysis (Fig. 9) provided new information about the relative stability and strength of fibers versus glue lines under different stress scenarios. Glue demonstrated exceptional strength and resilience in the perpendicular groups, where forces operate perpendicular to the glue lines, preserving its integrity even when fiber bundles separated. This emphasizes the adhesive’s effectiveness under perpendicular loading.

Fig. 9
Fig. 9
Full size image

The glue line of parallel and perpendicular groups after the point-bending test

On the other hand, a distinct behavior emerged in the parallel groups (Fig. 9), where forces were aligned with the adhesive line Table 4. Here, the glue's strength was lower than that of the fibers, resulting in noticeable weakness along the adhesive line under load. This confirms the susceptibility of adhesives to parallel stresses. Furthermore, compared to parallel groups (Fig. 9), which fractured more abruptly, the perpendicular groups—where the glue line was stronger—sustained longer plastic deformation before failure.

Figure 8b and Table 6 depict the shear strength results for both parallel and perpendicular groups. P IN-IN exhibited the highest shear strength (7.16 MPa), while the other two parallel groups (P OUT-OUT and P OUT-IN) reached 6.15 and 6.96 MPa, respectively.

Following shear tests, the fracture patterns of bamboo samples are shown in Fig. 10. In the parallel groups, separation occurred mainly at the glue line. The outer side bonded to the inner side failed most frequently Table 5. Both OUT-OUT and IN-IN fractured surfaces appeared relatively smooth and did not differ substantially.

Fig. 10
Fig. 10
Full size image

Fracture topography of bamboo samples after shear test

The shear strength values of the GFD panels were significantly lower than those of most bonded bamboo products manufactured with other adhesives. For instance, Zheng et al. [9] reported shear strength up to 18.35 MPa for glued strips of undensified bamboo panels, with inner-to-inner bonding being the strongest. Likewise, Guan et al. [39] found that shear strength of panels glued outer-to-outer and inner-to-inner with different adhesives exceeded 10 MPa. Wettability, contact angle, surface roughness, pressing conditions, and adhesive type are among the variables influencing bonding performance [40,41,42]. In this study, the presence of the pith ring and cortex at the glue line likely reduced adhesive effectiveness, as also discussed in the chemical composition section. Chen et al. [10] studied glued curved P. edulis bamboo samples while removing the cortex but retaining the pith ring. Their results showed shear strengths of 9.71 MPa (without pith ring) and 8.12 MPa (with pith ring), which are still higher than those reported here (maximum 7.16 MPa), likely because the cortex was removed in their case.

For perpendicular groups, failure occurred by tension perpendicular to the fibers, rather than at the bonding line (Fig. 11). As a result, for PER OUT-OUT, PER IN-IN, and PER OUT-IN, the tension perpendicular to the fibers was calculated instead shear strength (Fig. 8b and Table 6). This indicates that in perpendicular panels, the bonding line resisted higher stresses than the bamboo fibers themselves.

Fig. 11
Fig. 11
Full size image

Rupture of perpendicular groups A PER IN-IN B PER IN-IN side view

According to the DIC analysis (Fig. 12), the glue line of parallel groups absorbed the greatest amount of energy at the critical stage, leading to its failure, whereas in perpendicular groups the glue line resisted pressure and failure occurred in the bamboo.

Fig. 12
Fig. 12
Full size image

The result of the DIC test for shear and tensile test for two parallel and perpendicular groups. A (at the beginning of the test when panels are glued) B (at the critical state) C (after separating of glued layers for parallel groups and breaking for perpendicular groups)

Scanning electron microscopy (SEM) analysis

Research emphasizes how adhesive penetration impacts bamboo bonding ability, which affects the integrity and functionality of the resultant composite materials in use [8]. Nonetheless, few studies have addressed the mechanisms underlying the best adhesive penetration for bamboo Table 6. Glue is wasted and the bond line becomes starved when adhesive penetrates too deeply into large pores and fissures [43]. Another factor contributing to low bond strength is the inability to penetrate bamboo cells. Polyurethanes (PU), a type of pre-polymerized glue, are composed of large polymer molecules. These adhesives create more flexible connections but are less mobile when adhering to wood or bamboo [44]. Flexible bond lines help reduce the dimensional changes brought on by shrinkage or swelling, which increases the endurance of the connection.

Parallel groups are illustrated in Fig. 13 with glue lines shown. It is evident from (Fig. 13a) that the adhesive is much more concentrated on the inner layer of the bamboo, and that it penetrates a higher amount of the interior of the panels. Based on Marra's model and as evident from (Fig. 13a), the adhesive is much more concentrated within the bamboo interphase, meaning it penetrated the inner layer of bamboo cells [45]. In addition to protecting the culm from pathogens and drying out, the waxy and siliceous materials also prevent or reduce adhesive interactions with bamboo tissue. Additionally, bamboo parenchyma cells contain a lot of starch grains throughout the year [39], which lower porosity and prevent adhesive penetration. As shown in Fig. 13a, PU enters the starch grains in parenchyma cells and prevents the adhesive from penetrating the bulk bamboo cells. Furthermore, several factors influence the bond formation "interphase" zones. These include the adherend's natural cellular structure (porosity, permeability), the presence of cut cells, macro- and microcracks, the surface chemistry affected by organic compounds lining the cell lumen (hydrophilic hemicellulose or hydrophobic extractives), and the element surface preparation (splitting, incising, sanding, crushing, planing, etc.). The level of adhesive penetration [39, 46] and the kind of connections that are generated between the adhesive's functional groups and the surface hydroxyl groups (−OH) during cold/hot pressing have a major impact on the bamboo-adhesive interface. The adhesive phase also determines the penetration depth and the types of bonds created. Key factors include the adhesive type and chemistry, its viscosity, gel/cure and flow dynamics (influenced by fillers or viscosity modifiers), and the curing process. Adverse surface pH or oxidation states, as well as unsuitable curing circumstances including heat, time, and applied pressure, can all influence bond malformation at the adhesive phase.

Fig. 13
Fig. 13
Full size image

The result of SEM for parallel groups. a P IN-IN b P OUT-OUT c P OUT-IN

In the P IN-IN configuration, the adhesive penetrated the bamboo interphase, whereas in P OUT-OUT and P OUT-IN, much of the adhesive remained on the surface, resulting in a weak or incomplete adhesive interphase. This explains the superior mechanical performance of the P IN-IN group and the lowest shear strength of P OUT-OUT. These regions are prone to bond defects, with quality influenced by bamboo’s natural structure (porosity, permeability) and surface preparation (splitting, incising, sanding, crushing, planing) (Fig. 13). Since bamboo lacks transverse tissue, pits are the sole radial pathways [47], limiting PU penetration through the cortex and leaving it concentrated at the bond-line in outer layers.

As can be seen (Fig. 14), the thickness of glue lines is almost similar for all three perpendiculars. Under the conditions that two PER OUT-OUT and PER OUT-IN had at least one outer layer glued to each other, the cortex and a partial part of the adhesive collapsed. According to Marra’s model of the adhesive interphase, the area between the bamboo surface and the point at which the adhesive begins to manifest its bulk properties is the adhesive interphase when there is a high amount of adhesive. The adhesive in perpendicular groups is also more widely distributed in bamboo surface layers as compared to parallel groups.

Fig. 14
Fig. 14
Full size image

The SEM result of perpendicular groups. a PER IN-IN b PER OUT-OUT c PER OUT-IN

Overall, these SEM analyses highlight the critical role of bamboo surface properties and adhesive penetration in determining the bonding efficiency and mechanical properties of the glued flattened-densified bamboo panels. The IN-IN configurations, both parallel and perpendicular, offer superior adhesive infiltration, translating to enhanced bonding strength.

Conclusion

Experimental research was conducted on 6 different arrangements of glued flattened-densified (GFD) bamboo panels made from D. asper, in which the pith ring and outer skin were intentionally preserved. The results demonstrated that the orientation and arrangement of GFD panels strongly influence their physical and mechanical performance. Panels glued with their inner layers facing each other (P IN-IN) exhibited superior dimensional stability, flexural and shear properties, due to better gluability of the pith ring at the glue line and the higher protection provided by the outer skin when in contact with water. In contrast, P OUT-OUT panels showed higher water absorption and swelling. Perpendicular arrangements generally reduce mechanical performance but could provide more isotropic behavior, making them suitable for flooring or wall applications, whereas parallel arrangements are better suited for structural elements such as beams.