It is well known that some mechanical properties of materials can be linked with simple relations due to their similar dependence on material nano- and microstructure. For instance, the hardness H of ductile metals depends upon the yield stress σ_y, as H ≈ 3σ_y, while for hard and superhard metals and alloys, it depends upon the shear modulus G, as H ≈ 0.15 G [1]. There is reason to believe that similar relations could link mechanical and thermophysical properties (TPP) of materials belonging to the same type. In particular, thermal conductivity λ and diffusivity a ^[2][3][4][2,3,4] of wood and other anisotropic composite materials depend on density, porosity, humidity ^[5][6][7][5,6,7], and specifics of interconnections between microstructural units of the material as hardness, Young’s modulus, and strength do ^[8][9][10][8,9,10]. Once revealed, such relations could allow switching from laborious and material extensively destructive mechanical tests to non-destructive measurements of TPP using express methods, as described in ^[11][12][13][11,12,13]. Typically, they require around 1 min per measurement and do not demand for cutting samples of any specific geometry. Such approach could be used to estimate relative mechanical properties, sorting, and grading of materials and products made of wood, fiber-reinforced composites etc.

There is a lot of information concerning the mechanical and thermal properties of natural ^[6][8][10][6,8,10] and modified [14] wood, wood-based layered materials ^[3][15][16][3,15,16], as well as composites reinforced with artificial ^[5][9][5,9] and natural organic fibers ^[17][18][19][17,18,19]. For example, it is found that the thermal conductivity of bamboo measured by the classical stationary method along the fibers λ_l exceeds the transversal one λ_n by nearly 2 times [20], and both increase linearly with material density. In [21], the mapping of the thermal conductivity of bamboo cellular structure at the nano- and microscale was carried out by means of scanning thermal microscopy. It provided information about the relative thermal conductivity of different layers of the cell wall and revealed that the relative thermal conductivity λ* significantly depended upon the angle between the microfibers forming the specific layer of cell wall and cell long axis. Thermal conductivity and diffusivity measurements of birch wood in longitude and transverse directions were reported and discussed in [4]. Instances of correlations between thermal conductivity and stiffness of specific layers of the cell wall can be implicitly found in [22].

However, the mechanical and thermal properties are measured in separate tests executed on different samples and are not related to each other, except for several papers that use thermal non-destructive testing to estimate structural damage of wood ^[23][24][25][23,24,25], defects ^[26][27][26,27], porosity [28], and anisotropy ^[29][30][29,30] of composites and their possible impact on material mechanical properties. In [31], a micromechanical model for the prediction of effective thermal conductivity in two- and three-phase composites is proposed.

2.Relationship between Thermal Diffusivity and Mechanical Properties of Wood

In the first stage of the experiments, thwe researchers sstudied Brinell hardness HBW, Young’s modulus E (Figure 1a), and the main components a₁, a_2, and a₃ of thermal diffusivity tensor (Figure 1b) dependencies in common pine wood on its humidity w. The absolute and relative values of a₁, a₂, and a₃ are affected by w differently. The a₁ component displayed significant dependence on w along the fibers, while across the fibers, the dependence of the components a₂ and a₃ upon w was weak, and they were statistically indistinguishable. Hence, it was reasonable to introduce the averaged value a_n = (a₂ + a₃)/2 of thermal diffusivity normal to the fibers. The relative values a_l/a_n that characterize thermal diffusivity anisotropy were even more sensitive to humidity change than the absolute ones (Figure 1b). The values of hardness HBW normalized to the hardness HBW₁₀ at w =10% for the lateral and radial faces were statistically indistinguishable too (Figure 2); therefore, they were approximated as a single set by the following linear function (HBW/HBW₁₀) = 1.59(a_l/a_n) − 1.34 with the coefficient of determination R² = 0.75. The hardness measured on the face perpendicular to the fibers was found to be independent of tensor a_ii components, so that it could be used for method and equipment calibration. In the second stage of the experiments, microhardness and Young’s modulus in early and late wood layers, Brinell hardness HBW, and thermal diffusivity tensor main components were measured in dried wood of all three species (Figure 3, Figure 4 and Figure 5). Typical P–h curves recorded in NI for the three wood species are shown in Figure 3a, Figure 4a and Figure 5a. Wood of all three species displayed pronounced periodicity of local mechanical properties (Figure 3c, Figure 4c and Figure 5c). Locations of abrupt changes of H_IT and E_IT coincided with the boundaries of annual growth rings found using an optical method. Changes of H_IT and E_IT in transition regions between early and late wood in each growth ring were found to be gradual in pine and lime but stepped in oak. H_IT and E_IT were linked by a linear relation with the same slope m = 0.017 ± 0.002 for all three species (Figure 3b, Figure 4b and Figure 5b). In other words, the ductile ratio DR = E_IT/H_IT for all studied species appeared to be around 60, which is typical of many wood species. For example, the DR for eucalyptus is between 54 and 68 ^[32][45], with an average of 61. The values of H_IT and E_IT of the EW layer differed only slightly (within 10–15%) both within each ring and for different rings, despite the growth conditions differing significantly. For instance, in 2010, there was a severe drough, so that the width of the growth ring in pine wood was less than half its normal value; however, there was no discernible effect on H_IT and E_IT of EW. The lateral sizes of cells corresponding to different years had no visible difference too, but the wall thickness did differ. Therefore, the variation of the annual growth ring width was attributed to the different number of cells in the layer, while the size and mechanical properties of each cell remained almost the same. Effective hardness and Young’s modulus of the LW layers were several times higher than those of the EW layers due to higher cell wall thickness and lower lumen area. In the third stage of the experiments, tensor a_ij components and Brinell hardness HBW were measured for the three wood species. Effective values of H_IT, E_IT, HBW, and thermal diffusivity anisotropy are presented in Table 1. As can be seen, all mechanical characteristics correlated with the ratio of the main components of the thermal diffusivity tensor, and for E, such correlation was about three times stronger than for H_IT or HBW (Figure 6). It should be mentioned that the mechanical and thermal properties were correlated in all three studied species with significantly different microstructures.