Figure 2 shows the topography and indentation modulus mapping obtained through AFM PF-QNM measurement of flax, nettle, jute, kenaf, hemp, and date palm fibers. These images not only provide values of the indentation modulus but also information about the morphology of the middle lamellae. In References [40,41], the authors made a distinction between the middle lamella (ML) and the compound middle lamella (CML), especially referring to the wood samples, but this distinction is also applicable to other plant fibers [13]. In fact, the middle lamella between two primary walls of two different cells is often not distinguishable [13], and the small tripartite layer appears as a single layer, such as in Figure 2b or Figure 2d. For this reason, in the present paper, the indentation modulus is measured in the tricellular junctions, usually called the cell corner middle lamella (CCML), where the middle lamella is well discernible. However, it should be taken into account that the biochemical composition of the CCML may differ from the ML [13].

In Figure 2a, flax shows an interesting fracture (decohesion highlighted with white arrows), which probably occurred between the secondary S₂ and S₁ cell wall layers during the handling of the green flax stem. In several other images of the same section of the flax stem, other cells showed this kind of fracture between S₁ and S₂ secondary layers, which was previously noted by Arnould et al. [39] and Goudenhooft et al. [38]. In addition, from Day et al.’s (2005) [42] pictures, it is possible to identify fractures, which occurred at the CML-S₁ or S₁-S₂ level and their observations are in line with our results and conclusions. Le Duigou et al. [43] also observed peeling failures of the external fiber layer of flax fibers during a microdroplet debonding test, proving the low cohesion degree between structural layers.

In addition, the images collected confirm the observations done by Zamil and Geitmann [13] and Raghavan et al [15] in the ML and S₁–S₂ region. The authors reported that the middle lamella is stronger than the primary or secondary wall layers and not damaged from a mechanical stretch; on the contrary, fractures take place in the adjacent primary or secondary cell-wall layers. This fact reinforces the generally accepted assumption in the plant fiber community that the mechanical behavior of the ML is far from the brittle behavior associated with the cell-wall layers. Middle lamellae in flax are well distinguishable and with an indentation modulus clearly lower than the other cell-wall layers, as can be expected given that this layer has a random organization and constitutes non-cellulosic polymers. For the same variety of flax investigated by our team [36,38], a similar contrast was highlighted between values of indentation modulus obtained in S₂ and ML, even if measurements were done at a larger scale and not especially dedicated to the investigation of the ML. Possible differences in thickness or arrangement of the ML area can be due to the position of the fibers inside the same plant. In fact, one can find changes in the lumen, the cell, and the size of the middle lamella according to the considered stem cross-section along the height of the stem [44].

Nettle is reported in Figure 2b, and its middle lamella clearly shows a high indentation modulus, almost comparable to the surrounding cell walls. In general, the indentation modulus is highly influenced by the anisotropic character of the considered material; thus, in this case, of quasi isotropic non-cellulosic polymer (NCP) middle lamellae, the indentation modulus can be in the same range as the S₂, which is underestimated due to the highly anisotropic nature of the secondary cell wall [33]. However, the indentation modulus in the middle lamella is remarkably high here and could be linked to a specific biochemical composition that must be investigated.

The indentation modulus in the middle lamella is globally homogeneous, but one can note small areas where the indentation modulus is lower and other areas where it is higher, creating an irregular matrix like a grid or globular aspect. Considering the size of these nodules, of a few tens of nanometers (Figure A1, Appendix A, where zoomed-in view of the ML region are shown, not corrected for the effect of tip dilation), they might be attributed to the lignin [31,45,46], but the variations of the modulus measured in that area are more likely due to a topographic effect. As the roughness in that area is similar to the tip radius, it induces a strong correlation between the topography and the indentation modulus (Figure A1, Appendix A), which induces an apparently higher indentation modulus between two nodules and a lower one on top of them [47]. Bourmaud et al. [36] hypothesized that the degree of lignification is an important factor in the middle lamella morphology; however, although it is known in the literature that this layer is enriched in lignin, a direct quantification remains a challenge because of its sub-micrometric scale. In addition, in nettle, as previously mentioned, the primary cell wall is not well distinguishable from the middle lamellae, and they appear as a single unitary layer between two fibers.

On the other hand, the middle lamellae of jute fibers are clearly distinguishable from the surrounding cell walls (Figure 2c) thanks to their lower indentation modulus. In addition, the indentation modulus in CCML appears homogeneous with a fine texture that does not represent disconnections or asperities when compared with those of nettle. However, one can note that, in jute, there is a small difference between the indentation modulus of the central area of the CCML and its edge (Figure 3) that is not due to an effect of topography. The indentation modulus gradually increases from the edge to the center of the tricellular junction; this singular phenomenon was noted on each jute CCML investigated but not for other middle lamellae investigated in the present paper, such as nettle and palm sheath fibers. The AFM PF-QNM method, thanks to its high resolution, is able to clearly highlight this kind of gradient in a restricted area (Figure 3), confirming its potential for ultra-local mechanical investigations.

In Figure 2d, the primary walls of kenaf fibers show strong inhomogeneities with a highly globular aspect. In the present case, even more than in the case of nettle, the topography of the nodules induces a strong effect on the apparent indentation modulus. Moreover, the distinction of the middle lamella and the primary walls is difficult compared to other plant fibers. The origin of these nodules is not clear but might be linked to the cutting behavior and, therefore, to the mechanical properties of the different components of the cell wall layers during the sample surface preparation. Softer components make them harder to cut and leads to more irreversible deformation when cutting. These differences in behavior induce variations in topography [29,48] such as steps between layers of the cell wall (as between S₁ and S₂ or between S₂ and the embedding resin) and roughness or nodules, as in the case of lignin, within a layer. However, biochemical investigations, like Raman spectroscopy, must be done to better understand this phenomenon.

Figure 2e illustrates a tricellular junction between three hemp fibers. The middle lamella of hemp, together with the nettle, shows the highest values of indentation modulus that are also well correlated with values found by Bourmaud et al. [36] on several plant fibers studied by AFM PF-QNM mode. Nevertheless, in both nettle and hemp, not only the middle lamella but also the indentation modulus of hemp and nettle fiber cell walls are high. The matrix inside the middle lamella of hemp is similar to the grid texture that was observed in nettle; however, in addition to this morphology and contrarily to the nettle, the primary cell wall is clearly distinguishable from the ML.

In Figure 2f, date palm CCML is demonstrated to have the lowest values and an extreme homogeneity of the indentation modulus with a clear separation of the layers from each other. This is especially visible in Figure 4 where the middle lamellae of kenaf and date palm fibers are compared and underlined. Both graphical representations and distributions of the indentation modulus (Figure 4c) underline significant differences in terms of indentation modulus between the two plants considered.

In kenaf, the non-uniformity of color mapping clearly shows discontinuities with higher and lower values of the indentation modulus. These heterogeneities are partly due to a topography effect of the mechanical measurements caused by the huge globular aspect of the primary wall and possibly due to the cutting effects during the sample preparation between cell-wall layers or within a layer. On the other hand, the indentation modulus in the date palm middle lamella is more uniform, although the surface area analyzed is larger than that for kenaf.

For some fibers such as flax and palm, it is possible to trace and identify all the layers present in the fiber bundles, which are underlined thanks to their different modulus in Figure 5, highlighting the typical hierarchical structure.

Average numerical values of the indentation modulus calculated in a selected area of each middle lamella are summarized in Figure 6 and Table 1 with their respective standard deviation.

As previously mentioned, kenaf has the highest standard deviation since these measurements reflect the important inhomogeneity (see Figure 4a) of the CCML region. Nevertheless, average values from flax, jute, and kenaf are comparable. Hemp and nettle have the highest indentation modulus for the middle lamellae. The morphology of these two ML is similar, probably because of the presence of a large amount of lignin. On the other hand, palm fiber CCML regions show the lowest indentation modulus.

The different origin of the fiber elements is also visible in their overall chemical composition (Table 2), which is related to their botanical functions or roles for the plant. Among the six plant species considered, some fibers ensure the role of conduction of raw or elaborated sap, and others have the function of supporting the stem and ensure its stability. For example, kenaf fibers are located in the bast (cortical layer) and core (woody) region. The bast fibers constitute around 40% of the total amount of the fibers. Primary phloic fibers (PPF) from procambium in the protophloem region and secondary phloic fibers (SPF) from cambium are both developed in both jute [49] and kenaf. Thus, in these two plants, the chemical composition of the fibers is highly dependent on the tissue from which they come. This kind of cellular heterogeneity, coming from the tissue origin, does not exist for flax and nettle, which have gelatinous and poorly lignified fibers that can be extracted only from the primary phloem region. In the case of hemp, a possible mix between primary and secondary fibers may occur; these two kinds of fibers have a role in the mechanical support of the plant. Consequently, their compositions and mechanical properties are fairly similar and the main difference is the length, which is significantly shorter for the secondary fibers because of their delayed growth that occurs after the structuration of the tissues.

For flax, hemp, and jute, middle lamellae are often reported as lignin-enriched domains, which are often highlighted to explain the mechanical stiffness and cohesion of the fiber bundles employed as composite reinforcement. Nevertheless, cell-wall lignification generally appears after seed maturity and, according to the cultivation management, can be strongly limited. For flax pulled out after flowering or seed maturity, Day et al. [42] reported a low lignin labeling in the CML area, absent in the cell corner and tricellular junctions, regardless of the epitope considered. Flax samples investigated in our paper come from the plant stem and they were collected before the lignification of the middle lamella; in addition, no retting or scutching processes were performed contrary to the hemp fibers. In this case, the lignification and the retting effect could explain the higher indentation modulus of hemp fibers than of flax.

As previously reported, the amount of lignin and the quality of links between the several layers influence the behavior of the fiber bundles and, in the specific case of flax where decohesion between CML-S₁ or S₁–S₂ level is noted, our results are in line with the figure reported in Reference [42]. These damages at the level of the middle lamella also have an impact on the biocomposite material when flax fibers are used as reinforcement. A second polymer well known to be an influent actor of the cell cohesion is the pectin family, precisely located in the ML. It was demonstrated that the mechanical properties of the ML may vary with the degree of the esterified form of homogalacturonan [50]. This pectin is then more or less able to generate interchain interactions, for instance, by calcium cross-linking, if homogalacturonans are de-esterified. From our panel of phenotype samples, the backbone of the pectin methylation degree is arguably different, and some contrasted mechanical properties can be, therefore, expected due to susceptibility to cross-linking.

Dhakal et al. [37] investigated the damage mechanisms after impact testing for palm/PCL composites and they reported that, even after testing, the morphology of the date palm bundles remained unaltered. This is due to the high content in lignin and the homogeneity of the middle lamella (Figure 2f) that strongly links the single palm fibers together, more than in other plant fibers, despite a soft middle lamella (Figure 6 & Table 1). The silica-reinforced external paravascular or parafibrovascular parenchyma of bundles [51] is arguably too strong to be damaged, even after a high shear-rate process. In Reference [52], opposite results were highlighted on flax bundles and significant cracks were visible after transverse tensile tests on the CML area. Indeed, the individualization of the fibers can be more or less pronounced, according to the process used to produce the composite material [18] and the scutching or hackling degree of the fibers [53] during the fiber extraction process. Consequently, all these factors can introduce significant changes in the mechanical performance of the final composite. In addition, when the composite contains a large number of flax fiber bundles, one can notice that, during loading, the damages occur preferably in the bundles, especially in the CML area [19,20]. In this case, the fiber–matrix interface is not the most critical, even when the polyolefin matrix, with a weak fiber–matrix interface, is used. Thus, even if the indentation modulus of flax ML is higher than the palm one (Table 1), the individualization of flax fibers is facilitated thanks to the specific organization of the bundles and preliminary action of retting and scutching. However, as reported in Reference [13], ML in flax are not directly involved in the failures, but the weak area seems to be between the different layers (S₁–S₂) and not inside the middle lamella, even if ML exhibits a lower stiffness. A hypothesis for this result is that the softness and the cohesion of the ML are probably an advantage for the energy dissipation, and they limit the cell-wall damage during mechanical loading or shear rate stress. To support that assumption, a parallel can be done with the nacre, where the microstructure displays some homologies with the lignocellulosic bundles. Indeed, stiff and large structures of ceramic, as the fiber cell wall in our study, are embedded in a thin and weak mortar made of unfolded proteins and chitin biopolymers. The latter is the middle lamella of the bundles. During loading, in the nacre, the thin interface is able to deform, with most of the fracture occurring at the interfaces, providing more inelastic regions and favoring more energy dissipation and higher work of fracture, thereby giving nacre an elevated damage tolerance [54].

In our plant samples, palm sheath fibers, which have CCML with a lower indentation modulus, are probably less damaged in contrast to the flax fibers.

On the other hand, if we consider the jute fibers, it is clear that they have an indentation modulus similar to flax as reported in this paper (see Table 1), but their mechanical behavior inside a composite material is very different. This is probably due to their high lignin content, as well as to their short length, and, as a consequence, high volume content of middle lamellae. The cohesion in bundles is strong, inducing a specific behavior and a flow orientation in extrusion or injection molding [55]. Moreover, a high lignification is in favor of a good fiber–matrix interface, and Graupner et al. [56] noted that lignin may be an interesting adhesion promoter. Thus, flax, jute, and kenaf have similar CML indentation modulus (Table 1), but their mechanical and morphological behavior differs widely.

Finally, other fibers, such as hemp, have a strong tendency to be divided or fibrillated during the composite production process. This phenomenon was highlighted in extruded or injected compounds [57,58,59]. Interestingly, the difference between CML and S₂ modulus is lower for hemp compared to other plants, and this might introduce a strong link between these two layers, whereby a weaker interface could possibly be identified at the intra S₂ layer level.

Nevertheless, it is important to keep in mind that the morphology of the fibers may be involved in the breaking mechanism, and that the length of fibers and the ratio between the cell wall and lumen area may affect the critical shear rate, leading to irreversible damage. In addition, the indentation modulus is not necessarily correlated to the structural cohesion between layers, and the hardness investigated by the nano-indentation technique can be a better indicator even though it has a lower resolution than AFM PF-QNM. It is interesting to note that Wimmer et al. [28] studied the relationship between the longitudinal indentation modulus and hardness of CCML and S₂ in spruce wood, and they found a strong relationship between the indentation modulus of the CCML with its hardness, i.e., stiffer denoting harder, unlike the S₂. In this article, the indentation modulus of the CCML ranges from 4 to 12 GPa, whereas the hardness ranges from 0.12 to 0.47 GPa with a mean value close to that of the S₂. As the hardness depends on the inelastic behavior of a material, it is linked to its plastic behavior and property at break. If we assume that these results can be transposed here, it would mean that nettle and hemp fiber middle lamellae have a higher hardness in relation to a higher secondary wall longitudinal property, contrary to palm leaf sheath fibers. This seems to make sense mechanically, but needs to be confirmed by additional hardness measurements in our samples and mechanical strength tests of ML [23] or bioinspired films [31].

For this reason, our measurements revealed huge differences in the indentation modulus of the middle lamella according to the plants considered, but it is difficult to correlate them directly to the differences in their mechanical behavior or damage mechanisms when fibers are employed in biocomposites. In fact, these complex phenomena are impacted not only by the nature and cohesion of this middle lamella, but also by the morphology and the biochemical composition or structure of the fibers and bundles.