The Effect of Rays on the Mechanical Behaviour of Beech and Birch at Different Moisture and Temperature Conditions Perpendicular to the Grain

Abstract

The plastic deformation of wood perpendicular to the grain is gaining increasing importance due to advancements in forming technologies and the densification of wood. This study investigates how two hardwood species, i.e., beech (Fagus sylvatica) and birch (Betula pendula), respond to compression in the radial direction and examines the structural changes they undergo during both elastic and plastic deformation. Stress–strain curves at different moisture contents (dry to wet) and temperature conditions (20 to 140 °C) were recorded. In-situ observations at high moisture content and temperatures by means of different microscopic techniques are practically unfeasible. Therefore, the specimens were analysed ex-situ microscopically after the test. In addition to the compression of transversely oriented fibres and vessels, special attention was paid to the deformation behaviour of the wood rays. The results suggest that the wood ray cells carry a relatively higher proportion of the load in the radial loading direction than the surrounding vessels and fibres. This observation is supported by the higher percentage of deformed vessels, seen in the microscopy, in areas where the rays developed kinks, usually in the early wood of beech and anywhere in the cross-section of birch. The weaving of rays around big vessels introduced shear strains under compressive stresses at the kinked rays’ area. Thus, shear deformation is more evident in early wood than in late wood regions of wood. However, when the wood was tested at elevated moistures and temperatures, the material demonstrated a ductile response, namely the absence of localised shear deformations. Notably, wet beech and birch specimens heated to 100 °C and above exhibited pronounced thickness recovery and there was slightly irreversible buckling of rays and vessel deformations. Therefore, under such conditions, wood behaves like a “sponge” and is expected to be successfully processed without introducing clear damage to the material. This characteristic holds promise for replication in the development of bio-based energy-absorbing materials.

1. Introduction

Wood is the most abundant natural resource, holding significant importance for the construction and furniture sectors. Wood processing operations, e.g., 3D shaping, mechanical pulping, board manufacture, and densification, usually involve subjecting wood to compressive stresses perpendicular to the grain at elevated temperatures. Hence, understanding the mechanical behaviour of wood under such conditions is crucial for optimising its use. Furthermore, the behaviour is scientifically interesting with regard to the relationship between structure and mechanical properties.

During the last decades, enormous efforts have been devoted to characterising the stress–strain response of wood under compression perpendicular to grain [1,2,3]. The results indicate that three distinct regions can be identified by examining the stress–strain curves: (i) Wood, at the beginning of loading, exhibits elastic behaviour which terminates by cell wall collapse [1,2,3,4,5]. (ii) During this stage, strain increases rapidly with little change in stress, which is often referred to as a plateau. (iii) Under further compressive loading, cell cavities are removed, and the wood is densified, making it resist higher loads which causes the stress to sharply increase [1,3,4,5,6,7]. This represents a simplification of the actual behaviour and lacks the intricacies required to provide a comprehensive understanding of wood’s response in its non-elastic stages, which is influenced by the microstructure of wood.

While there is a lack of agreement in the literature on the effect of rays on the mechanical response of wood in the radial direction [2,8,9], their role is less understood when wood is tested under different moisture and temperature levels.

Reiterer et al. [10] described fracture characteristics in Mode I in the radial direction and were able to show that wood species with multi-row wood rays exhibited significantly higher fracture toughness. The results were also confirmed for freshly cut samples [11]. The anatomical influence of the wood ray cells was also demonstrated for dynamically loaded samples [12].

Using real-time monitoring of wood microstructure during radial compression tests, it was reported that the yield point on the stress–strain curve of aspen (Populus tremula) [8] and of balsam poplar (Populus balsamifera) [2] corresponded to the deformation of the largest vessels and the uniseriate rays began to deform and then fail simultaneously with the vessels. Huang et al. [2] claimed that very fine rays do not act as spaced columns in supporting compressive loading. However, Sanabria et al. [13] used an innovative individual cell tracking (ICT) approach, that allows direct quantification of deformation mechanisms of various cell types (tracheids, wood rays), to examine the role of rays in Norway spruce (Picea abies) during radial tension. They discovered three signs proving how wood rays reinforce the wood in radial direction: (i) slightly smaller local strains in wood rays, in radial direction, compared to tracheids, (ii) a decrease in wood ray density at the fracture line, and (iii) a notable increase in local strain in regions with fewer wood rays. Consequently, it is hypothesised that thicker wood rays in hardwood species like beech play a more prominent role, and their influence should be observed in compression at different moisture contents and temperatures.

This study examined the compressive response and failure behaviour of beech (Fagus sylvatica) and birch (Betula pendula) wood along radial and tangential directions at different moisture and temperature levels. Both tree species are diffuse-porous hardwood species with comparable density and similar wood anatomical structures, including the size of vessels and wood cells. However, beech has very wide rays, whereas birch is characterised by narrow rays. Therefore, the test material appears to be particularly suitable for investigating the influence of wood rays in the radial direction at different wood moisture contents and temperatures.

The study involved analysing stress–strain patterns and examining failure modes through visual observations and microscopic analysis. The goal is to provide insights into how wood fractures under conditions that resemble those occurring in wood processing, with a special focus on the behaviour of rays, all with the aim of optimising their performance.

2. Materials and Methods

2.1. Quasi-Static Compressive Tests

Two diffuse porous hardwood species, beech (Fagus sylvatica) and birch (Betula pendula Roth.) were selected for this study. Defect-free cubic specimens, each with an edge length of 20 mm, were prepared from wooden boards and stored for six months in an environmental chamber (20 ± 2 °C and relative humidity of 65 ± 5%) according to the ISO 554 standard [14], prior to moisture and temperature conditioning. Beech samples with strictly parallel growth rings to the edge of the sample were selected. Birch is known to have a smaller trunk diameter than beech and hence, obtaining numerous samples with horizontal rings was difficult. Therefore, this limit was relaxed for birch.

The experimental series included different nominal moisture contents: oven-dry (zero MC), 10 ± 2, and water-saturated conditions and three different temperature levels: 20, 100, and 140 °C for specimens oriented in radial and tangential directions. Given the anticipated impact of rays on the behaviour of the wood along the radial direction, additional moisture contents of 13 ± 2% and 25 ± 5% were investigated for radially oriented samples.

The cubic samples were then conditioned according to their individual temperature level (except testing sets of 20 °C) with a laboratory oven for 7–15 min, determined from preliminary studies performed on equivalent samples to the testing series. Further details concerning sample preparation, moisture, and temperature conditioning are given elsewhere [15].

After preheating, the compression tests on beech and birch specimens were conducted in a universal testing machine (Zwick/Roell Z100, Ulm, Germany) equipped with a thermal chamber. The force was recorded by a machine load cell with a capacity of 100 kN while the displacement was determined from the crosshead movement. To measure the displacement as accurately as possible, the deformation of the machine was also considered. For this purpose, the testing machine was loaded up to a force of 40 kN after closing the compression plates. Then, the machine deformation was determined and subtracted from the crosshead movement. The specimens were loaded with a rate of 10 mm/min until reaching a displacement of 5 mm. In this study, 256 samples were tested along the radial direction, while 176 samples were tested along the tangential direction. Therefore, a total of 432 samples were tested in this study. Five to ten replicates were used for each variant.

Beech and birch specimens were weighed three times: before and after preheating and after testing. The last weight was directly considered for dry density calculations of dry samples, whereas the dry density of other specimens was calculated after drying them in an oven. The average dry density, obtained according to the ISO 13061-2 standard [16], ranged from 594 to 757 and from 561 to 645 kg/m³ for beech and birch, respectively. The moisture content (MC) was determined before and after preheating as well as after testing according to the ISO 554 standard [17]. The moisture content presented in the following paragraphs is based on the average moisture content before and after the compressive tests, as given in

Table 1. The details of moisture contents (MCs, %) of beech and birch tested at 20, 100, and 140 °C.

Orientation	[°C]	Beech		Birch
		* MCbp	** MCavg	* MCbp	** MCavg
Radial	20	0.00	0.00	0.00	0.00
		8.27	8.27	9.69	9.69
		11.32	11.32	12.28	12.28
		25.47	25.47	28.98	28.98
		>FSP	>FSP	>FSP	>FSP
	100	0.00	0.00	0.00	0.00
		8.71	5.49	9.56	7.70
		12.87	12.03	14.69	13.36
		21.43	18.38	22.40	18.39
		>FSP	>FSP	>FSP	>FSP
	140	0.00	0.00	0.00	0.00
		9.44	1.83	9.59	3.84
		12.11	6.91	12.78	6.84
		20.52	9.46	21.45	10.29
		>FSP	>FSP	>FSP	>FSP
Tangential	20	0.00	0.00	0.00	0.00
		8.29	8.29	9.6	9.6
		>FSP	>FSP	>FSP	>FSP
	100	0.00	0.00	0.00	0.00
		7.65	4.49	9.71	7.83
		>FSP	>FSP	>FSP	>FSP
	140	0.00	0.00	0.00	0.00
		9.22	2.36	4.55	3.78
		>FSP	>FSP	>FSP	>FSP

* Moisture content before preheating (index bp). ** Mean moisture content (index avg) for after preheating (before test) and after testing.

.

2.2. Microscopic Tests

In-situ measurements are strongly recommended for observing the influence of wood microstructure on the stress–strain relationships [2,8]. However, such in-situ measurements are only possible to a limited extent and with disproportionately high experimental effort. Microscopic in-situ investigations of wood require very small samples and labour-intensive preparation methods [18]. However, compression tests have shown that very small samples dry out or cool down very quickly [15]. Compression tests with small specimens at high MCs and elevated temperatures would, therefore, have to be carried out in a climate chamber. This in turn would make observation with optical methods impossible due to corresponding image artefacts. Measurements with X-ray methods such as computed tomography are limited by spatial resolution or sample volumes, and testing with high-resolution magnetic tomography is not possible as the magnetic field strength does not permit metallic testing equipment such as a universal testing machine.

The only reasonable option is to visually inspect samples after compression. Artefacts due to sample preparation and spring back of the material must, therefore, be considered when assessing and interpreting the images.

A representative specimen from various testing conditions was selected for performing microscopic examinations. They include specimens with 10% MC tested at 20 °C; 17% MC tested at 100 °C; and wet samples tested at 20, 100, and 140 °C as well as untested beech and birch samples for comparison purposes. To aid in the surface preparation process, a few water drops were applied to the specimen surface. After preparing the surface of samples with a conventional slide microtome, micrographs were recorded with incident light microscopy (Olympus digital microscope DSX 1000, Shinjuku, Tokyo, Japan). To show the overall deformation, a few wooden samples were only sanded with sanding paper (grit 300). The surface of the specimens was also characterised using a scanning electron microscope (SEM; Hitachi, TM3030, Tokyo, Japan) at 15 kV.

3. Results and Discussion

3.1. The Compressive Behaviour along the Radial Direction

The average stress–strain curves of beech and birch are presented in Figure 1 for different moisture contents and temperature levels. The curves of each wood species were given in two graphs based on their strength level to enhance clarity and ease of interpretation. Wood samples with higher strength levels are depicted in Figure 1a, while those with lower strengths are shown in Figure 1b. A detailed discussion about the effect of moisture content and testing temperature on the perpendicular-to-grain compressive strength (f_c) of beech and birch is given elsewhere [15] and will only be briefly discussed here.

As shown in Figure 1, beech featured higher compressive strength than birch at the tested moisture and temperature conditions. For instance, at 20 °C and around 12% MC, beech reached approximately 1.6 higher compressive strength than birch. The average density of the beech wood samples was 688 kg/m³, while the average density of the birch wood samples was only 587 kg/m³. This means that beech is approximately 1.17 denser than birch. Like almost all other mechanical properties of wood, compressive strength perpendicular to grain shows a linear relationship to density [19]. Therefore, the difference between beech and birch in their strength cannot solely be attributed to density. Corresponding to Reiterer et al. [10], it is assumed that higher mechanical properties of beech wood, with multi-row wood rays, are likely associated with the reinforcing effect of rays, as will be explained in more detail later.

As mentioned, it is well known that, at standard testing conditions (20 °C/65% rel. humidity), the typical stress–strain curve of wood in the transverse direction is characterised by three consecutive stages: compression starts with (i) a linear elastic stage up to the collapse of early wood cell walls (in the case of softwood) or the collapse of big vessels (as in the case of hardwood), then, the load continues to rise but deviates from linearity. This stage is followed by (ii) a plateau stage where the cells continuously buckle and finally (iii) a strain hardening stage resulting from the densification of the cells. Thereby, it was observed, that the densification stage starts at relatively smaller deformation levels for wood species with relatively higher density values (0.2–0.3 strain) compared to those with lower density values (around 0.6 strain for aspen wood (Populus tremuloides Michx.) compressed in radial direction [4]).

Overall, except for the stress–strain curves of very dry wooden samples, the behaviour of beech and birch at different moisture and temperature levels followed the typical behaviour of wood described in the literature. However, inspection of the micrographs of beech and birch tested at standard laboratory conditions (Figure 2) reveals responses specific to each wood type.

When loading the specimens in the radial direction, the load is initially assumed to be carried by the rays because of their higher stiffness than the surrounding tissue and the material shows a linear elastic response. To better understand the material behaviour under compression, the wood was simplified as a two-component system (Figure 3) where the rays are connected to the fibres and vessels in parallel when the loading direction is radial (along the main axes of rays). This means that there is a homogeneous strain situation for radial direction and a homogeneous stress situation for tangential loading.

Based on that, the following “rules of mixture” model can be applied:

$$Ew=E_f{V}_f+E_m{V}_m$$

(1)

where E_w, E_f, and E_m are the modulus of elasticity in GPa of gross wood, rays, and a combination of fibres and vessels, respectively; V_f is the rays’ volume and V_m is the volume of the rest of the wood tissue.

It was reported in the literature [20,21] that ray cells, extracted from oak wood, have a modulus of elasticity of 4 GPa and the tensile strength of rays, extracted from beech, is approximately 75 MPa [22]. The wood ray content is on average 12.7% for birch and around 20% for beech [23]. Although no exact values are available for beech, it can be assumed here that the modulus of elasticity and the strength of the wood ray cells of those hardwood species are of the same order of magnitude. If those values are inserted in Equation (1) with a 2 GPa for the mean transverse Young’s modulus of fibres [24], then the resultant E_w is 2.4 GPa. This value is in good agreement with the stiffness of bulk beech wood in the radial direction of 2.34 GPa reported in previous studies [25], which strongly supports the above assumption.

Sanabria et al. [13] examined the role of rays in Norway spruce (Picea abies) during radial tension using the ICT approach. Strains were measured independently for wood rays and demonstrated consistent distributions with tracheids, which supports the results of the two-component model presented above, suggesting unit deformation of a composite loaded along the radial direction.

With further loading, the wood rays, adjacent to big vessels in the early wood regions of beech, developed kinks and the load deviated from linearity, but continued to rise as the integrity of the wood was not significantly affected. Simultaneously with early kinks development, large vessels in early wood showed deformation. This can be noticed in Figure 4. which shows kinked ray cells and deformed vessels of beech wood. This scanning electronic microscope (SEM) micrograph was captured at 170 µm of small beech wood blocks after compression in the radial direction (from Müller et al. [18], unpublished results). Müller et al. [18] demonstrated a sound agreement of the plastic deformations between in-situ observations using incident light microscopy and ex-situ inspections by means of SEM. Thus, in the opinion of the authors, it is acceptable for this publication to utilise ex-situ micrographs to assess the plastic deformation behaviour.

Under standard laboratory conditions (i.e., 65% relative humidity and 20 °C), the wood is expected to be glassy and stiff. Therefore, rays deformed in a brittle manner and collapsed and even crushed into the adjacent vessels (Figure 5), which from this point onwards, started, with the fibres, to share more loading as the stiffness of rays decreased and, consequently, they deformed. It can be seen in the video of in-situ microscope inspections (in the Supplementary Materials (Video S1), from Müller et al. [18], unpublished results) that at kinking locations, usually in early wood regions of beech, the collapsing and subsequent densification of vessels is much more pronounced than in the areas where the rays are still intact (see Figure 6a), which provides additional evidence that the rays carry more loads in radial direction prior to their kinking.

The fibres and vessels provide lateral support for the rays at the beginning of compression. Therefore, kinking initiates in the early wood region in beech as there is less lateral support provided to the rays from the surrounding structure in this area, whereas, for birch, kinking and early vessel deformations can happen anywhere in the ring (Figure A1 in the Appendix A), usually wherever some large vessels are aligned near each other. In that case, a crack linking them usually appears (Figure 6b). This behaviour agrees well with the observation reported by Tabarsa and Chui [14] for radially compressed aspen specimens. Like aspen, birch is more homogeneous than beech and demonstrates smaller variation in density across the growth ring.

At any rate, for both wood species, the deformation and collapsing of the vessels corresponded to the plateau stage at the stress–strain curves of the material. After that, the densification of vessels led to strain hardening (see Figure 1).

It is worth noting that the fibres around the vessel appear to deform plastically but show no clear collapsing at the load levels applied in this study. The vessel diameters are about 5–8 times larger than those of the fibres, and their walls are thinner and pitted, making vessels weaker than fibres. Therefore, vessels dominate the cellular structure of beech and birch.

The embedment of fibres and vessels and the rays in solid wood microstructure introduces a geometric imbalance, and thus to lateral deformations at higher loads, which leads to the development of shear strains [13]. For instance, under compression along the longitudinal direction, the misalignment of fibres around the rays initiates fibre micro-buckling, with the consequent shear deformations in the buckling area [7,26,27,28]. Likewise, along the radial direction, the weaving of rays around big vessels introduced shear strains at the kinked rays’ area. Hence, shear deformation is more evident in early wood than in late wood regions (see Figure 2a where the localised shear deformations are more pronounced for beech). The behaviour of rays (along the radial direction) can be viewed to be analogous to the behaviour of wood fibres loaded along the grain. It is worth noting that for birch, the whole cross-section appeared to undergo global shear deformations (see Figure 2b) as the ring layers were not perfectly parallel to the edge of the sample.

When the testing temperature was increased (e.g., beech with 12% MC and birch with 13% MC tested at 100 °C), the material exhibited less brittle behaviour compared to the material tested under standard conditions. For instance, there was slightly less kinking of rays in early wood layers of beech (Figure 7 and Figure 8) and, therefore, less pronounced localised shear deformation, yet more noticed global shear deformations (Figure A2 in the Appendix A). No strong kinks were observed for birch. For both wood species, there were no wide cracks across the tested specimens. It is expected at such testing conditions that the cell walls were only partially plasticised [29].

The material at elevated temperature, therefore, behaves much more ductile than the material at room temperature. Figure 6 (room temperature), therefore, shows more angular kinks of the wood rays, which indicate a brittle failure. In Figure 7 (100 °C), however, the wood ray cells show a wavy deformation behaviour, which indicates much more ductile behaviour.

The permanent plastic deformation appears to be less pronounced in birch than in beech (Figure 7 and Figure 8), which can only be explained by a stronger spring back in the case of birch. It is assumed that the spring back of the vessels and fibres in birch also leads to a partial reversal of the plastic deformation i.e., stretching of the wood ray cells. As the following figures show, increasing plasticisation of the material by increasing the MC and elevated temperature leads to even stronger ductile behaviour and reversible deformations.

When wood was tested at elevated temperatures and relatively high MC, for instance, beech and birch with approximately 18% MC heated to 100 °C, the flexibility of the polymer network of the wood increased. Hence, the thinner rays buckled rather than kinking into the surrounding vessels (Figure 9), resulting in slightly less noticeable deformation (Figure 10), and the material demonstrated a ductile response, namely the absence of localised shear deformations (See Figure A3 in the Appendix A).

Figure 11 and Figure A4 (in the Appendix A) show the post-test behaviour of wet beech and birch heated to 100 °C. There was a noticeable partial thickness recovery, also referred to as the spring-back effect [30], due to the relaxation of internal stresses upon load removal while the wood was still wet and hot. Beech specimens exhibited no clear localised shear failure, whereas birch samples showed only globally small shear deformations (Figure A5 in the Appendix A). Likewise, wet beech and birch specimens heated to 140 °C demonstrated pronounced thickness recovery and there was slightly more irreversible buckling of rays and vessel deformations (Figure A6 and Figure A7 in the Appendix A). As such, processing of beech and birch at these conditions would result in no damage to their microstructure.

To compare the stress–strain curves of wet wood at different temperatures, their stress values were normalised against their compressive or yield strength (see Figure 12). In this regard, Uhmeier et al. [31] reported that the shape of the normalised stress–strain curves of wet spruce when compressed in the radial direction at three different temperatures, 2.5, 50, and 100 °C, was similar. Following the same procedure, the normalised stress–strain curves of wet beech and birch are presented in Figure 12. As shown, while the shape of the curves was rather similar, each wood exhibited a different slope after reaching its compressive strength. For instance, the normalised stress values of birch were almost the same at 20 and 140 °C; whereas, for beech, they continued to increase with temperature. As mentioned before, the deformation and collapsing of the vessels correspond to the plateau stage at the stress–strain curves and the densification of vessels corresponds to the strain-hardening stage. This would mean that, for beech, the densification was enhanced with temperature as it started at lower strain levels while the optimum densification for birch seemed to be around 100 °C.

At any rate, when wet beech and birch were compressed along the grain [26], the stability of the material appeared to improve with temperature for beech, but not for birch. Birch specimens suffered more damage and density-dependent behaviour at 140 °C in comparison to their behaviour at 100 °C.

3.2. The Compressive Behaviour of Wood along the Tangential Direction

The average stress–strain curves of beech and birch are shown in Figure 13 for different moisture contents and temperature levels. As with radial f_c, a detailed discussion about the effect of moisture content and testing temperature on the tangential compressive strength (f_c) of beech and birch is given elsewhere [15]. Overall, the material is weaker and softer in the tangential direction compared to the radial direction. This agrees well with observations on different wood species, tested in standard laboratory conditions, reported in the literature [4,32,33]. In addition, this also corresponds to the lower expected strengths and stiffnesses of a sandwich structure subjected to a homogeneous stress situation, as shown in Figure 3.

As shown, overall, beech and birch exhibited similar stress–strain characteristics, under the different testing conditions, along radial and tangential directions. Yet, there were no visible indications of failure in the case of tangential compression (Figure 14), whereas in the case of radial compression, signs of failure were observed (Figure 4 and Figure 5), even at an early stage, when testing wood at standard laboratory conditions. Note here that rays play only a minor role in this case, as they are oriented perpendicular to the loading direction. When the wet material was tested at 100 °C, less deformation of vessels can be seen here compared with samples tested in standard laboratory conditions, especially for beech (Figure 14c).

Now, regarding technological applications, successful 3D shaping of veneers dictates avoiding a combination of lower moisture contents and high temperatures to prevent microstructural damage resulting from rays kinking and shear deformations as explained above for compression along the radial direction. When beech and birch were tested under high moisture and temperature levels, they exhibited a behaviour akin to that of a “sponge”, with the rays becoming notably soft. This allowed applying significant compressive strains without causing any permanent damage to the material. Therefore, wood should be moulded or processed under such conditions. It also means that wood would absorb or dissipate a lot of energy when compressed without being damaged as it possesses substantial toughness and ductility. This can be exploited in developing biobased components that can well replace foamy or honeycomb materials [34] in energy-demanding applications. While having high moisture and temperature levels is not practical, impregnating wood with substances such as polyethylene glycol may induce the same swelling effects [35], yielding excellent energy-absorbing materials. This aspect will be the focus of subsequent studies.

Corresponding to densification, it may be better to densify diffusive porous hardwood species like beech and birch along their tangential direction when using wood with relatively low moisture contents to minimise microstructural effects resulting from rays kinking explained above for radial compression.

Bekhta et al. [36] reported the development of cracks in cell walls and ray kinking when densifying birch and alder veneers at 150 and 200 °C and low moisture contents. It should be emphasised here that compression along tangential direction would probably result in buckling of late wood cells into early wood cells for softwood species and, therefore, deemed inappropriate.

4. Conclusions

A combination of compressive tests and microstructural investigations was conducted to understand the effect of moisture and temperature on the compressive response perpendicular to the grain and capture the failure behaviour of beech and birch.

The experimental tests were performed at three temperatures, 20, 100, and 140 °C, on specimens with various moisture contents. The results suggested that, when beech and birch were compressed in standard laboratory conditions, the rays carried a higher load proportion, in the radial direction, than vessels and fibres and the stress–strain curve was linear. Then, with further loading, the rays kinked into the adjacent vessels which simultaneously deformed with rays and the stress–strain curves deviated from linearity, but the stress continued to rise. This observation is supported by the higher percentage of deformed vessels, seen in the SEM, in areas where the rays developed kinks, usually in the early wood of beech and anywhere in the cross-section of birch.

The fibres and vessels provide lateral support for the rays during compression. Therefore, kinking is initiated in the early wood region in beech as there is less lateral support provided to the rays from the surrounding structure in this area. As birch has a more homogeneous microstructure than beech and demonstrates smaller variation in density across the growth ring, kinking and early vessel deformations in birch can happen anywhere within the annual ring, usually wherever some large vessels are aligned near each other. At any rate, for both wood species, the deformation and collapsing of the vessels corresponded to the plateau stage at the stress–strain curves of the material.

The wood experienced localised shear deformations when subjected to compression loading, primarily because of geometric imbalances caused by the weaving of rays around large vessels. Under standard testing conditions, the rays crushed and exhibited angular kinks. However, when subjected to elevated temperatures, the wood ray cells displayed wavy deformations.

When the material was tested at elevated temperatures and relatively high MCs, it demonstrated a ductile response, namely the absence of localised shear deformations. Notably, wet beech and birch specimens heated to 100 °C and above exhibited pronounced thickness recovery and there was only slightly irreversible buckling of rays and vessel deformations. Therefore, at such conditions, wood shows a “spongy” like behaviour, and is expected to be successfully processed without introducing clear damage to the material. This enhanced ductility could potentially be reproduced by employing swelling agents to develop bio-based energy-absorbing materials.

Unlike compression along radial direction, the wood had no visible indications of failure in the case of tangential compression when tested under standard laboratory conditions and the wood showed even less vessel deformation when it was tested in wet conditions at 100 °C and above. It should be emphasised here that rays play only a minor role when the wood is tested along the tangential direction, as they are oriented perpendicular to the loading direction.