Introduction

In recent years, a worldwide increasing awareness of human-related environmental impact has led to new research in sustainable and environmentally friendly materials. Bamboo is an eco-friendly material with many advantages, such as a higher strength-to-mass ratio than concrete (Youssefian and Rahbar 2015); a high volume fraction of cellulose fibres by mass of 48% (Moradbak et al. 2015). Moso bamboo is among the fastest growing bamboo species in the world; in its accelerated growth stage, its culms can achieve heights of over 20 m in just 45–60 days (Chen et al. 2022). These advantages combined with the low cost and high tensile strength of bamboo make it suitable as a replacement material for other wood and steel which can be applied widely in engineering and civil construction (Raghavan and Khaligh 2012; Silva et al. 2020; Yadav and Mathur 2021). Furthermore, reinforced composites based on hierarchical bamboo structures have attracted attention in new material fabrication processes mimicking bamboo structures for weight reduction (Kanyilmaz and Berto 2019), and smart structures for electrochemical energy storage and conversion (Ding et al. 2020; Xiang et al. 2020).

Due to its advantageous properties, previous studies have investigated the structure and mechanical properties of bamboo. Fibre bundles, which are the main part of bamboo and the biggest contributor to the mechanical properties, also have a vital function for the health of the plant by transporting water and nutrients (Li and Shen 2011). The extraordinary mechanical characteristics of fibre bundles mainly originate from a composite structure of every fibre, which is a multilayer structure of differently oriented crystalline cellulose fibrils connected by a matrix of lignin and hemicellulose polymer. The elastic modulus of the cellulose fibril (167.5 GPa) is forty times larger than that of hemicellulose (4 GPa) and lignin (2 GPa) (Li et al. 2020). Thus, each component plays a specific role in the formation of bamboo. Interactions between cellulose, lignin, and hemicellulose in bamboo (Besombes and Mazeau 2005; Lahiji et al. 2010; Dri et al. 2013; Sinko et al. 2014; Wegst et al. 2015) consisting of friction, electrostatic, and van der Waals forces as well as chemical bonds such as covalent and hydrogen bonding have been extensively investigated (Ren et al. 2017; Marcuello et al. 2020; Khodayari et al. 2021). Especially lignin plays an important role as it connects the moisture-sensitive cellulose and hemicelluloses (Wegst et al. 2015). On a larger scale, it is normally found that the outer part of the bamboo is harder than its inner parts, which can be related to two reasons. One is the increased density of fibre bundles, and the other is the increased amount of cellulose in every fibre in the outer parts of the bamboo (Li et al. 2020).

Mechanical properties of bamboo fibre are mainly investigated using independent systems such as indenters (Yu et al. 2011; Tan et al. 2011; Youssefian and Rahbar 2015) and microtensile testers (Li and Shen 2011; Javadian et al. 2019). The elastic modulus derived from these two methods shows a large range of values. Yu et al. (2011) for example reported the mechanical characterization of single fibres with nanoindentation of 18–21 GPa, which is only 55% of the value of 32–34.6 GPa measured by microtensile testing. Major challenges of those methods are fibre gripping, alignment of fibres to the tensile direction, and determination of the cell wall of single broken fibres (Yu et al. 2011). Moreover, they tend to average over several microstructures and growth directions of bamboo cells and destroy the test sample. On the other hand, atomic force microscopy (AFM) is an effective tool to reveal the mechanical properties of biomaterials at the nanometer scale without damaging the sample. In a recent study (Nguyen et al. 2021), we compared several AFM-based mechanical testing methods for their reliability and demonstrated their capability for investigating biomaterials.

For most applications using bamboo, the material is used in a dried state to lower the moisture content and thus control the risk of fungi and insects as well as material cupping, warping, or delamination. The majority of existing research focuses on freshly-cut samples with higher moisture levels, or the dried bamboo fibers were treated with alkalis or acids (Yu et al. 2011; Tan et al. 2011; Li and Shen 2011; Wang et al. 2012; Ren et al. 2017; Mannan et al. 2017; Osorio et al. 2018; Javadian et al. 2019; Malekzadeh et al. 2021; Shih and Young 2022), which can affect the material’s mechanical properties. Thus, there is a lack of mechanical studies on un-treated dried bamboo in atmospheric environments to determine its sturdiness for future applications.

Here we use three different AFM-based methods to study micro-and nanoscale elastic and plastic mechanical properties of commercially dried bamboo. We show that the elastic modulus of the fibres reduces from the outside to the inside of a bamboo culm and fibre stiffness varies between its interior and the fibre wall in nanoscale regions. In particular, the relationship between the adhesion force from tip/sample interaction, the elastic modulus, and the lignin level in the fibre walls is demonstrated and explained. The results present new ways to analyse the nanoscale mechanical properties of the biomaterials using AFM-based techniques. This is a preliminary step towards promoting the design of new bio-sourced materials and composites for nanoscale biomimetic materials.

Experimental methods

A tonkin cane bamboo sample (Pseudosasa amabilis, China) with a length × diameter of approximately 6.80 × 7.84 mm2 was cut from dried bamboo sticks from Beihai Hiyi Trading Company Limited. These commercially bought bamboo sticks were air-dried for 12 weeks and had their coat removed. No chemicals were used in the fabrication process. After cutting the samples, they were embedded into epoxy resin for polishing and grinding with a Struers LaboPol-5 polishing machine. Samples were ground at 250–300 RPM with polishing papers with grit sizes of 320, 800, and 1200. Further polishing was done with 3 μm and 1 μm diamond particle suspensions at 50 RPM. In between polishing steps as well as after the whole process the samples were cleaned in an ultrasonic bath consecutively with ethanol and distilled water. The completed sample was stored in a desiccator with a relative humidity of 25–30% and a temperature of less than 25 °C.

Optical microscopy (OM) images of bamboo culm were taken with a Nikon Eclipse ME600 Inspection Microscope. In particular, the scanning electron microscopy (SEM) images were taken with a Hitachi TM3000 instrument with an electron accelerating voltage of 15 kV to show the microstructure across the vascular bundles of the outer and inner areas.

All AFM scanning measurements were performed using an MFP-3D Infinity system (Asylum Research, USA), while AFM-based nanoindentation was carried out using a Smart SPM 1000 system (AIST-NT, USA). Both systems were operated using various cantilevers at room temperature under ambient conditions. The spring constant for each cantilever was determined before each measurement using the thermal noise method (Cleveland et al. 1993; Sader et al. 1999). To ensure the accuracy of our result, we repeated each measurement at least 10 times.

Contact Resonance Force Microscopy (CRFM): CRFM experiments were conducted using uncoated AC160TS-R3 silicon probes with a spring constant of k = 33 N/m (Asylum Research, USA). Additionally, the samples were mechanically excited by an actuator (V104-RM UT 2.25 MHz from Olympus, USA). To track the contact resonance frequency of the tip-sample system we used the Dual AC™ Resonance Tracking (DART) provided by the Asylum Research AFM system.

F-D maps: Force-distance maps were generated by performing force-distance curves at each point of a 100 × 100 points array. The elastic modulus is extracted from fitting every curve by using the Derjaguin–Muller–Toporov (DMT) model. This is done directly in the Asylum software (version 16.05.193). The DMT model is a modified Hertzian model, which is applicable for systems with low adhesion (Johnson and Greenwood 1997; Ciavarella et al. 2019). The experiments were performed using an Ir/Ti coated ASYELEC.02 silicon probe with a force constant of k = 45 N/m (Asylum Research, USA). Before the experiments on the bamboo substrates, the cantilever was calibrated on a polystyrene standard sample (Bruker, USA).

AFM-based nanoindentation: Nanoindentation tests were carried out using a single cycle indentation (load-unload) test method with a matrix of 5 × 5 points. In every cycle, a maximum loading force of 30 µN was applied by using NM-RC-C diamond cantilevers (k = 470 N/m, Bruker, USA), which were pre-calibrated using standard fused silica. To verify the right location and produced impact of the indentation, contact mode AFM scans were performed to image the topography.

Results and discussion

Tonkin bamboo micro- and nanoscale structure

An optical microscopy image of the bamboo sample surface after completing the polishing process is shown in Fig. 1a. The bamboo culm consists of three regions: culm wall, vascular bundles, and sponge structures. The cross-sectional view of the bamboo culm reveals the expected graded distribution and hierarchical structure of every vascular bundle (Tan et al. 2011). Figure 1b shows an SEM image of one vascular bundle having a typical polyhedral geometry. Every vascular bundle consists of (1) vessels, (2) phloem, (3) sclerenchyma sheath, and (4) fibre bundles, which are surrounded by parenchyma cells (5) (Osorio et al. 2018). The vascular bundles shown in Fig. 1a are large, numerous, and concentrated in the outer area, while in the inner area, they are smaller and spaced wider, which also changes the volume-fraction of the fibre bundles that play a large part in the mechanical properties of bamboo. Figure 1e, f illustrate the cross-sectional profile corresponding to the white dashed line in Fig. 1c, d, respectively. The vascular bundle size of the outer area is approximately 200–300 μm and decreases to 100–200 μm in the inner area. The main part of a vascular bundle is the fibre bundles made out of singular fibres that are our main focus in this study. It can be seen that the overall fibre density in a vascular bundle is higher in the outer area. In particular, singular fibres have an approximately hexagonal shape with a small hole in the centre named lumen (see 3D AFM images in Fig. 1c, d. The average size (statistics value from 25 fibres) of a single fibre in the outer area is 16.62 ± 3.57 μm and larger than that of the inner area of 9.12 ± 2.04 μm. Therefore, the cellulose density is also proportionally higher in the outer areas.

Fig. 1
figure 1

a Optical microscopy image of the sectioned bamboo culm surface. b SEM image of vascular bundles with the composition description of (1) vessel, (2) phloem, (3) sclerenchyma sheaths, (4) parenchyma cells, and (5) fibre bundle. c, d Topography in 3D view images using AFM measurement of elementary fibre of outer and inner areas. e, f The cross-sectional profiles along the white dashed line in (c, d)

Full size image

Mechanical characterization

The internal mechanical properties of each fibre in both outer and inner areas can be determined using different AFM-based methods. The first method we used is CRFM which tracks the contact resonance frequency of the AFM cantilever. A shift in contact resonance frequency to lower or higher frequencies represents a change to softer or harder mechanical properties of the material (Hurley et al. 2005; Hurley and Turner 2007; Hurley 2009). To induce an oscillating cantilever the sample is mechanically excited by an actuator (Nguyen et al. 2021). Figure 2a, b illustrate the surface topography of the outer and inner fibres, respectively, whereas Fig. 2c, d show the CRFM maps measured at the same contact areas in Fig. 2a, b (we used the same AFM probe and the same conditions for comparability). In both CRFM images, the centre of the fibre region (lumen), as well as the fibre walls, shows a lower contact resonance frequency compared to the rest of the fibre indicating that they have a lower elastic modulus. The corresponding histograms of each CRFM image are shown in Fig. 2g, h.

Fig. 2
figure 2

a, b Topography, c, d CRFM maps, e, f cross-sectional profiles along the white dashed lines in (a–d). g, h Gaussian peak fitting to the histogram of the CRFM maps in (c) and (d) of both outer and inner fibre regions while the blue and red lines correspond to the fibre and fibre wall, respectively. Note the lower frequency, and hence lower elastic modulus at lumen (fibre centre) and fibre walls

Full size image

The contact resonance frequency of the outer fibre is 844.53 ± 0.27 kHz and higher than that of the inner fibre with a value of 841.62 ± 0.32 kHz. The elastic modulus can be extracted from the CRFM measurements by using the D.C. Hurley method (Jackson et al. 1988; Hurley et al. 2005; Huey 2007; Hurley and Turner 2007; Hurley 2009). The elastic modulus values of the outer and inner fibre are 7.46 ± 0.41 and 6.08 ± 0.32 GPa, respectively. This result indicates the inner fibre bundles have a lower elastic modulus than the outer fibre. Moreover, the fibre boundary, which is mainly formed by the hemicellulose/lignin matrix to connect the fibres, has the lowest contact frequency with the value of 843.76 ± 0.65 and 841.03 ± 0.42 kHz for the outer and inner areas. Figure 2e, f show the line cuts across the topography height and CRFM maps along the white dashed line in Fig. 2a–d. The inner and outer fibres present the same behaviour as the lowest contact resonance frequency value is at the lowest parts of fibre walls. Interestingly, we do not see any upper bound due to strong interference in the CRFM signals at the lowest point of topography. As our previous report (Nguyen et al. 2021), due to the low tip/sample contact area in our CRFM measurement compared to the fibre wall thickness which can exceed up to 1–1.5 μm, the surface topography is less impact on our CRFM results.

Apart from the dynamical measurements involving contact resonances, a method to measure the nanoscale quasi-static mechanical properties of bamboo is to measure force–distance (F–D) maps with an AFM system. This method allows extracting the elastic modulus directly using the DMT fitting model (see supplementary information). As we mentioned before, each bamboo fibre is a combination of cellulose fibrils and polymer lignin/hemicellulose matrix (Youssefian and Rahbar 2015). The volumetric fraction of that polymer matrix is about 23% (Youssefian and Rahbar 2015) and can reach up to 50% (Li et al. 2020) depending on the humidity of the bamboo. Therefore, to characterize the adhesion properties, and the behavior of the polymer matrix on the mechanical properties of bamboo, the more appropriate model is the DMT instead of the Hertzian model. The DMT model is based on the basic Hertzian model but includes adhesion (Fig. S2 in the supplementary information). It allows to consider the role of noncovalent interactions, including van der Waals forces and the hydrogen-bonding interactions, on a different surface of cellulose crystals and lignin/hemicellulose matrix. Every F–D map consists of 100 × 100 points measured with the same maximum force of 1000 nN. As seen in the maps, the elastic modulus has similar behavior in both outer and inner areas. The interior of the fibre has a relatively homogenous higher elastic modulus than the fibre boundaries and the lumen. The adhesion force maps of both outer and inner bamboo fibre areas (Fig. 3c, d) at the same location as the elastic modulus maps show an opposite contrast compared to the elastic maps in (a) and (b). The cross-section profiles displayed in Fig. 3e, f represent the higher elastic modulus and lower adhesion force in the fibre compared to the boundary.

The histograms of the elastic modulus maps and adhesion force maps in Fig. 3 of both outer and inner regions are presented in Fig. S1 of the supplementary. The elastic modulus values extracted from the histograms are 8.89 ± 2.72 GPa and 4.13 ± 0.75 GPa for the outer and inner fibre, respectively. Whereas the adhesion force histograms show a reversed behavior with lower values of 25.84 ± 2.12 nN in the outer fibres compared to 29.86 ± 4.68 nN for the inner fibres of the culm. Additionally, the fibre wall also shows the same behaviour, as the area has the lowest elastic modulus and highest adhesion force. Interestingly, the areas with higher elastic modulus show a lower adhesion force. This reveals an inverse relationship between elastic modulus and adhesion force, which reflects the tip/sample surface interaction and energy. Moreover, the area below the Gaussian fitted peaks of the CRFM and FM histograms can tell us the contribution amount of fibre and fibre wall regions. The surface topography of the outer and inner areas consists of 74.53% and 76.69% of the fibre, and 25.47% and 23.31% of the walls, respectively. In the histogram of CRFM (Fig. 2g, h), the fibres are 82.66% and 85.98%, and they are 17.33% and 14.02% for the fibre walls of the outer and inner areas, respectively. Whereas in the FM measurements (Fig. 3a, b, we get 98.27% (fibre) and 1.73% (fibre wall) of the outer area, and 98.4% (fibre) and 1.6% (fibre wall) of the inner area. This result demonstrates that the fibre is more prevalent than the wall area and the CRFM measurement is more responsive to the surface topography than the FM measurement.

Fig. 3
figure 3

a, b Elastic modulus maps, and c, d adhesion force maps of single fibres in the outer and inner regions. e, f Cross-sectional profiles along the black lines in (a) and (c), and blue lines in (b) and (d)

Full size image

In order to test plastic deformation behaviour at the nanoscale a possible method is nanoindentation (Jäger et al. 2011; Tan et al. 2011; Osorio et al. 2018; Gangwar and Schillinger 2019). However, conventional indentation methods require large diamond probes (up to a few hundred nm depending on the probes (Yu et al. 2011; Wang et al. 2012) and high loading forces, and in addition it is not possible to easily choose precise locations on samples (for example individual fibres). In one of our previous work (Nguyen et al. 2021), we have shown that AFM-based nanoindentation with a diamond cantilever (radius ~ 15 nm) can provide accurate indentation results without the above drawbacks of conventional nanoindentation.

Figure 4a, b show the topography of the outer and inner bamboo fibres after applying 30 µN of loading force during an AFM nanoindentation loop. The difference in depth and width of the holes are shown in Fig. 4c, d by the cross-sectional profiles along the black, red, and blue lines in Fig. 4a, b. The indentation in the inner area is wider and deeper than in the outer regions. The statistical values for the depth and width of 25 indents of inner areas are 58.17 ± 7.91 nm and 0.42 ± 0.04 μm compared to those of the outer areas of 55.11 ± 6.67 nm and 0.39 ± 0.03 μm. This is the first indication from the indentation measurements that the inner fibre is softer.

The elastic modulus and hardness values can be determined from the load–unloading curve measured during the indentation using the Oliver–Pharr method (Pharr et al. 1992; Oliver and Pharr 1992, 2004), The results of the inner and outer areas are presented in Fig. 4e. They show that both elastic modulus and hardness values are significantly different in both locations. The values of the outer fibre are 8.05 ± 0.37 GPa (elastic modulus) and 0.82 ± 0.004 GPa (hardness) and are higher than the values for the inner fibre with an elastic modulus of 5.70 ± 0.11 GPa and hardness of 0.39 ± 0.008 GPa.

In addition, we investigate the plastic, elastic, and total energies that are involved in the nanoindentation process (see Fig. 4f, which quantifies the elastic/plastic deformation of the material (the detail of those energies is explained in our previous report (Nguyen et al. 2021). The total energy UT of the inner fibre bundles is 94.21 ± 2.27 × 10−2 pJ and larger than that of the outer ones with 86.74 ± 6.21 × 10−2 pJ. While the elastic energy UE is almost similar in the outer and inner areas with respective values of 48.88 ± 2.42 × 10−2 pJ and 48.27 ± 1.41 × 10−2 pJ. The plastic energy UP, however, shows a slightly higher value in the inner fibre of 45.94 ± 2.22 × 10−2 pJ compared to the outer fibres of 35.54 ± 7.40 × 10−2 pJ. The results indicate that both areas show the same elastic deformation, but with more plastic deformation in the inner area which leads to a larger and deeper indent. Thus, as in the other two mechanical measurements, the elastic modulus of the outer bamboo fibre is slightly higher than the inner one.

Fig. 4
figure 4

a, b Topography images of outer and inner fibre after AFM nanoindentation at 30 µN. c, d Cross-sectional profiles along the black, red, and blue lines in (a) and (b). e Elastic modulus and hardness, and f energies fitting from the load-unloading curves in both outer and inner areas

Full size image

It is useful to compare our results of the outer and inner fibres with higher elastic modulus values for the outer ones to values from bamboo in the literature using indenters and microtension testers to see if there is a difference in measuring modes (Yu et al. 2011; Tan et al. 2011; Li and Shen 2011; Youssefian and Rahbar 2015; Javadian et al. 2019). Figure S3 summarizes the elastic modulus values from our AFM-based methods and the reported literature. The overall elastic modulus values measured in our report vary between 7.46 and 8.89 GPa and 4.13–6.08 GPa for the outer and inner bamboo fibre. Although we use three different techniques, the obtained elastic modulus values are very similar and are close to the range of elastic modulus values found in the literature (E = 9–13.5 GPa for the outer fibre bundles and 2.5–8 GPa for the inner one). Our values of the inner fibre bundles correspond very well with the literature and follow the common observation that outer fibres have higher mechanical properties (Tan et al. 2011; Li and Shen 2011; Youssefian and Rahbar 2015). Nonetheless, our measurements of the outer ones are slightly lower leading to a smaller difference in mechanical properties between inner and outer fibre bundles. This could come from the small diameter of our used bamboo culm (less than 8 mm) compared to larger used bamboo in the literature (Moso bamboo with a culm diameter of 8–12 cm, depending on the bamboo age (Zhang et al. 2014; Gu et al. 2019)). As a result, the fibre density in our bamboo sample is lower than in the literature and it could lead to the lower stiffness of outer fibre.

Additionally, the lignin level in the bamboo culm varies with higher levels in the fibre boundaries. Higher lignin levels also lead to higher combined van der Waals forces in the material, and thus more adhesion. Therefore, areas with higher lignin levels should show stronger adhesion with the measuring AFM tip. Thus, the adhesion maps are a representation of the lignin levels. Youssefian and Rahbar (2015) reported the adhesive interaction energy was accordingly reduced with the reduction of lignin level. Moreover, in the DMT model, the adhesion force is identified as the lowest point in the retrace curve. This result illustrates that the area with higher adhesion force may exhibit higher tip/sample attraction forces, including higher van der Waals forces and more hydrogen bonding. It leads to a higher concentration of lignin/hemicellulose polymer matrix (i.e. lignin level), and tends to soften the fibre, increasing the plasticity and resulting in a lower elastic modulus (Jajcinovic et al. 2018; Shih and Young 2022).

The difference in adhesion values between the inner and outer fibres could also be a result of the cellulose content, which is higher in the outer fibres. In a recent study, it was found that the adhesion is higher between two lignin/hemicellulose matrices compared between cellulose and a lignin/hemicellulose matrix. Thus a higher cellulose content means more cellulose–lignin/hemicellulose interfaces and thus a lower adhesion in the material (Wang et al. 2012). Thus, a lower adhesion value in the outer fibres found in our experiments could indicate a higher cellulose content in the fibres as predicted in the literature. Thus, the adhesion force maps in Fig. 3c, d are in good agreement with the elastic modulus maps as the adhesion force increase with the decrease of the elastic modulus. As higher cellulose content is linked to higher elastic modulus, this explains the higher elastic modulus in the outer fibres, which is also found in our measurements.

Further, with the higher plastic energy in the inner fibre indentation in Fig. 4e, the plastic deformation processes increase from the outer layer to the inner fibre. As we mentioned, the degree of lignification also shows a similar trend. Therefore, these results present the relationship between plasticity, lignin level, and adhesion force. The higher lignin level tends to create a plasticizer to connect the cellulose and hemicellulose/lignin matrix, and leads to lower elastic energy recovery and a reduced hardness and elastic modulus in the inner fibre.

Moreover, one of the original contributions of our study here is to compare and discuss three different mechanical AFM characterization methods for their applicability and suitability to study the nanoscale mechanical properties of composite materials. However, there is a major limitation in this study that could be addressed in future research. The chemical composition analysis on the lignin level and cellulose content of the inner and outer fibres would need a further focus using bio AFM measurement techniques to support the aforementioned discussion.

Conclusion

In this research, we use different mechanical AFM characterization methods to analyse the mechanical properties of dried tonkin cane bamboo samples at the micro- and nanoscale. The results present the relationship between nanoscale elastic properties, plasticity, lignin level, and adhesion force. The elastic modulus of fibre in the outer area range between 7.5 and 8.9 GPa compared to 4.1–6.1 GPa of the inner fibre. The outer fibre exhibits higher fibre density and lower lignin level, resulting in weaker attractive interaction and adhesion forces. This leads to lower plastic deformation and higher elastic modulus and hardness in the area. In contrast, the inner area reveals a reverse behavior which shows lower stiffness. Our result aligns with prior macroscopic studies, confirming that the outer bamboo region possesses a superior elastic modulus. The difference in adhesion values between the inner and outer fibres could also be a result of the cellulose content, which is higher in the outer fibres. The implications of the results are discussed for the application of bamboo as a high-performance, renewable, and sustainable material, providing insight for the design of new bio-sources and composites for biomimetic materials.