Method for determination of beech veneer behavior under compressive load using the short-span compression test

The wood species European beech (Fagus sylvatica L.) was used for the investigations. The manufacture of the peeled veneer was done with a laboratory peeling machine at Wilhelm-Klauditz-Institut WKI (Braunschweig). Veneer with thicknesses of 1 mm and 3 mm was produced. Due to the peeling process, the veneer cross section is pre-damaged and lathe checks occur (Pot et al. 2015; Rohumaa et al. 2018). In Fig. 2, the existing lathe checks of the beech veneer are shown. To make these more visible, they were colored in with black ink. It can be seen, that the thicker the veneer, the larger are the depth and the interval between the checks (Palubicki et al. 2010; Rohumaa et al. 2018). To compare the absolute lathe check values of differently thick veneers, the ratio of lathe check depth to veneer thickness is used. This value was found to be 36% and 52.3% on average for 1 mm and 3 mm veneers, respectively (in total 1517 cracks were measured on a veneer length of 1440 mm).

For comparison also solid wood samples with a thickness of 3 mm were prepared for the investigations. These specimens are cut of the same wood as used for veneer production. The solid wood samples were orientated in the longitudinal-tangential direction (LT) like the main orientation of the peeled veneer.

Prior to testing, the veneer and solid wood samples were conditioned in a climatic chamber at 20 ± 1 °C and 50 ± 3% relative humidity (rh) until equilibrium moisture content (EMC) in desorption had been achieved.

Table 1 shows the mean values of density (measured according to DIN 52182 (1976)) and moisture content (DIN EN 13183-1 (2002)) at 20 °C/50% rh of the sampling material. The density and EMC of the solid wood samples was significantly higher in comparison to the veneer material. The differences in density and moisture content between solid wood and veneer are due to the method of veneer production. During veneer peeling, lathe checks develop in the veneer, which lead to a "loosening" of the material. All geometric ratios being equal, the pure wood material decreases as the number of lathe checks and the size of it increase. Simplified, the more checks in the veneer, the greater the volume of air in the material. Due to this fact, there is less wood material available to store water. Therefore, both the density and the moisture content are highest in solid wood and lowest in 3 mm veneer than in 1 mm veneer.

The method used for the investigations is based on the SCT for paper and is shown in Fig. 3 . The testing device consists of a top part and a bottom part, which are connected to each other via a frame with a two-column support and can be moved linearly in test direction. Each part is equipped with hydraulic, parallel grips for clamping the specimen. The top part of the testing device is connected to the crosshead of universal testing machine (Inspekt 10, Hegewald & Peschke, Nossen, Germany) and transmits the test force. The clamping pressure of the hydraulic grips depends on the sampling material and the expected compression strength and was varied within a range of 60 to 200 bar to prevent slipping during testing and also damage of the specimen. The testing speed was equivalent to a strain rate of 1% min⁻¹. The applied load was measured with a 10 kN load cell. A pre-force of 2 to 10 N was applied to the specimen prior to starting the test.

The clear test length between the grips in loading direction has to be chosen so that no stability failure such as buckling occurs during testing. This length depends on the specimen thickness and can be approximated using Euler`s equation for elastic buckling. Because Euler`s equation is only valid for the elastic range and for long, slender bars, the real clear test length including plastic deformation is significantly shorter. This length depends on material, specimen thickness and loading direction (see Table 2) and was determined in pre-tests. Starting from the calculated elastic buckling length according to Euler as the initial value for the clear test length for each material, specimen thickness and loading direction, this length was reduced stepwise (1 mm per step) in preliminary tests until no stability failure but a strength failure occurred during the force drop (parallel or LT direction) or a defined strain limit (transversal or TL direction) was reached. The aim was to choose a clear test length as large as possible to avoid a potential influence of the clamping on the strain measurement. Buckling of a specimen was defined as a deflection of the specimen in thickness direction of more than 5% of specimen thickness. This criterion was defined empirically during the preliminary tests. Figure 4a shows an example for 1 mm thick veneer in which the selected clear test length of 10 mm is too large, so that the failure of the test specimen during the force drop is caused by buckling. The buckling can be shown in the area of the force drop, where the curve slope for the maximum deflection of the specimen surface in thickness direction shows a sharp increase. In contrast, Fig. 4b shows an example test with the resulted clear test length of 5 mm for 1 mm veneer, where no buckling failure occurs during force drop. Here, the deflection curve does not reach the buckling criteria and also the curve slope does not show a sharp increase as in Fig. 4a.

For the strain measurement, a stereo camera system (Aramis adjustable 12 M, Carl Zeiss GOM Metrology, Braunschweig, Germany) was used which applies the principles of digital image correlation (DIC). A contrast speckle pattern was applied to the area of interest (AOI) of the specimen surface. AOI is defined as the visible specimen surface between the hydraulic grips (see Fig. 3b). The DIC measurement was always done on the closed veneer side (without checks). With this system, both the strain along the loading direction \({\epsilon }_{x}\) and the strain transverse to the loading direction \({\epsilon }_{y}\) of the specimen can be measured. In addition, the out-of-plane deflection of the specimen in the thickness direction can be monitored to check whether buckling occurs.

The specimens were rectangular stripes with a length between 133 and 150 mm in dependence of the material, thickness and specimen orientation (see Fig. 3c). The specimen width was 20 mm for all test series. The thickness was varied between 1 and 3 mm for the veneer and only 3 mm for solid wood. The sample preparation of the veneer was done by a laser cutter to avoid additional damages within the veneer cross section. Two load directions were investigated. The veneer was tested parallel (fiber-load angle 0°) and transversal (fiber-load angle 90°) to the fiber direction. Corresponding to these directions the solid wood specimens were oriented in LT (parallel) and TL (transversal) plane. A total of 120 specimens were tested in 6 test series with 20 replicates per series.

Figure 5a shows an exemplary stress–strain curve. The modulus of elasticity (MOE) is defined as the slope of the stress–strain curve within the elastic range. MOE was determined as a linear regression in a range of 15% to 35% of the maximum stress ((green) shaded area in Fig. 5a). The compression strength \({R}_{m}\) is defined as the maximum stress within the compression test. According to \({R}_{m}\) the equivalent strain \({\epsilon }_{max}\) was also determined. In TL orientation, densification of the material occurs without a defined failure due to a force drop like in LT orientation. For this reason, \({R}_{m}\) was defined at a compression strain of 2%.

Another important characteristic point for describing the material behaviour in compression mode is the transition from the elastic to the plastic range. Until this point is reached, the Hooke`s law can be applied with MOE as proportionality factor. Beyond this point, other relationships are valid. This characteristic point is often called the yield point. In this paper the yield point is defined at a plastic strain of 0.02%. The yield point is described by the yield stress \({R}_{p}\) and the yield strain \({\epsilon }_{p}\).

Figure 5b shows an exemplary transverse strain-strain curve. From this curve, the Poisson’s ratio \(\mu\) can be determined as the slope of the transverse strain-strain curve within the elastic range analogous to MOE. Therefore, the same strain limits were used as for the determination of the MOE ((green) shaded area in Fig. 5b) and the Poisson’s ratio was determined as a linear regression within these limits.

Figure 6 shows the determined material properties of beech veneer and solid wood under compression load in fiber direction as boxplots. The plots for stress (Fig. 6c) and strain (Fig. 6d) contain two boxplots for each veneer thickness and solid wood. The left one corresponds to the yield point where the elastic limit is reached while the rigth boxplot corresponds to the mechanical value when the sample fails.

For the MOE of beech veneer, there are only minor deviations depending on the veneer thickness and also compared to solid wood. Thus, neither the lathe checks nor the material thickness have an influence on the MOE of veneer in fiber direction. The MOEs determined in this paper are slightly higher than the literature values (Table 4). This fact can be explained by the lower EMC of the samples in this study. Young’s modulus decreases with increasing moisture content (Hering 2012a, 2012b, Ozyhar 2013).

The Poisson’s ratios (Fig. 6b) are characterized by a high difference between 1 and 3 mm veneer. The results for 1 mm veneer are significantly lower than for 3 mm veneer as well as for solid wood. This is due to the considerably shorter clear test length of 1 mm veneer (5 mm) compared to 3 mm veneer (15 mm) and solid wood (20 mm). The fixed clamping of the specimen leads to an interference of the transverse strain near the clamping area (see Fig. 7). In the case of short clear test length between the grips, this has an influence on the transverse strain measurement even in the middle of the clear test length. This influence is reduced as the clear test length increases. For 3 mm veneer and solid wood the Poisson’s ratio is equal, so it can be assumed that the clear test length is adequate, so that the transverse strain measurement is not be influenced by the fixed clamping of the specimen. The measured Poisson’s ratios for 3 mm veneer and solid wood with a mean value of 0.36 (Table 3) are in the range of the literature data of 0.24–0.51 (Table 4).

The veneer thickness has no influence on the measured yield stress and strength. In comparison to solid wood the determined yield stresses between veneer and solid wood are similar, while the strength of solid wood is about 10% greater than for veneer. This can be reasoned by the 10–20% higher density for solid wood than that of the tested veneer. The measured strengths in fiber direction in the range of 60–70 MPa (Table 3) are significantly higher than the literature values (45–60 MPa) listed in Table 4. The high density as well as the lower moisture content of the tested specimen can be specified as reasons. The influence of the lathe checks on the mechanical behavior in the elastic range can be neglected, because the yield stress is equal for veneer with varying thickness and for solid wood as well.

Similar relationships to those for yield stress and strength are seen in the evaluation of the measured strains. The yield strain between veneer and solid wood is comparable and the veneer thickness as well as the lathe checks have no influence on this value. The failure strain of solid wood is like the strength about 20% higher (mean value) than that of the tested veneer. This is due to the higher solid wood strength, which leads to a later failure in the stress–strain relationship. The measured failure strains of veneer and solid wood are in the range of the literature values of 0.5%–1.0% (Table 4).

Overall, the applied SCT can be used to determine the mechanical properties, such as MOE, Poisson’s ratio and compressive strength, for veneer in fiber direction with a thickness of at least 3 mm. If the veneer tested is thinner, the Poisson’s ratio cannot be correctly determined because of the short clear test length and the hindered transverse strain near the clamping area. The differences between the determined values of veneer and solid wood in fiber direction are small, so it can be assumed that the lathe checks have no significant influence on the mechanical values of the veneer in fiber direction. This is consistent with the tensile behavior of veneer compared to solid wood. This correlation has already been shown by Buchelt and Pfriem 2011, for example. However, since the determined clear test length (see Table 2) of 3 mm veneer (15 mm) is smaller than that of solid wood (20 mm), the lathe checks nevertheless seem to influence the compressive behavior in the fiber direction.

The lathe checks are of great importance for the mechanical behavior transverse to the fiber direction. They are arranged orthogonally to the load-direction and thus reduce the load-bearing cross section of the veneer. Figure 8 shows the determined material properties of beech veneer and solid wood under compression load transversal to fiber direction as boxplots. Comparable to Fig. 6, the plots for stress (Fig. 8c) and strain (Fig. 8d) contain two boxplots, one for the yield point and one for fracture.

The MOE in Fig. 8a decreases with the veneer thickness, which is due to the lathe checks in the veneer. The thicker the veneer, the deeper are the checks (Fig. 2). The MOE of solid wood is 1.7–2.6 times higher than that of veneer. These differences are also caused by the lathe checks and the resulting reduced load-bearing veneer cross section. The MOE of solid wood is slightly higher than the literature values in Table 4, which is attributed to the lower moisture content of the tested samples.

Transversal to the fiber direction the Poisson’s ratio (Fig. 8b) is in fact very low. Based on the results of the Poisson's ratio in the fiber direction, the measured values for veneer must be fundamentally questioned, as the measurements were done using very short clear test lengths of 3–5 mm (Table 2). Due to the short clear test lengths the measurement of the strain orthogonal to the loading direction is influenced by the fixed clamping of the specimen. For solid wood this length is greater (10 mm), which results in larger Poisson’s ratios. However, compared to the literature (Table 4), the measured mean value for solid wood is 4–5 times lower, so that an influence of the fixed clamping cannot be definitely rejected either here.

The veneer thickness has an influence on the measured stresses. The average yield stress for 1 mm and 3 mm veneer is comparable (Fig. 8c). However, the strength of 1 mm veneer is greater than that of 3 mm veneer, which is due to the lathe check depth increasing with veneer thickness (Fig. 2). The yield stress and strength of solid wood are considerably higher than the veneer values because the veneer cross section is pre-damaged by the lathe checks. A comparison of the measured values with literature values is difficult, because the strength is not affected by the failure of the specimen but by the defined total strain of 2%. Hering et al. (2012b) also specified the ductile failure mechanism of solid wood in tangential direction at a total strain of 2% and 5%. The stress value \({\sigma }_{2\%}\) they determined was 6.0 MPa, which is 40% lower than the average strength of the solid wood tested in this study. The difference can be reasoned by the higher density and the lower moisture content of the tested samples. Likewise, an influence of the sample geometry (bone shape in Hering et al. 2012b) and the measurement method cannot be excluded.

As noted above, the strain value, which corresponds to the stress value for strength, is specified as 2% and thus, this is equal for all test series. The yield strain is similar for 1 mm veneer and solid wood, and above that for 3 mm, but with a high variability. The higher yield strain for 3 mm veneer may be caused by the deeper lathe checks and the wider check opening. Simplified, there is less wood material within the cross section of 3 mm veneer to resist the compressive force.

Thus, it can be assumed, that the SCT method is also applicable to measurement transversal to the fiber direction of veneer and solid wood. As already mentioned for the Poisson’s ratio in fiber direction, the clear test length used is of great importance when measuring in transverse fiber direction. To prevent the measurement from being influenced by the fixed clamping, a clear test length of at least 10 mm is required. This clear test length is not applicable to rotary cut veneer with lathe checks and veneer thicknesses up to 3 mm. Due to the pre-damage of the veneer cross section, the specimen would fail during the test before reaching a total strain of 2% because of stability failure.