1 Introduction

Wood, as a renewable and natural engineering material, is an essential building material for the development of sustainable building systems (Popescu 2017). The sawn timber is limited in both cross-sectional dimensions and length due to the size of the trees (Vallée et al. 2017). Engineered wood products (EWPs) were introduced to the marketplace since the last century, which can not only breakthrough the size limitation of the sawn timber but also possess decreased variability and redistributed natural defects (Crocetti et al. 2011; Milner and Woodard 2016). EWPs are usually produced by adhesively bonding sawn timber, veneers, or wood chips together to frame bigger timber elements (Yadav and Kumar 2021). As a result, the adhesive bonding performance is crucial for the quality of EWPs. Comprehensive studies have been conducted to understand and improve the bonding quality. There are numerous factors that could affect bond line qualities, which have been extensively investigated in the literature (Wang et al. 2016; Liu et al. 2020; Morin-Bernard et al. 2020; Alade et al. 2022; Hänsel et al. 2022; Slabohm et al. 2023). Some key factors mainly involve adhesive types and process-related parameters, e.g., application amount, mixing ratio, the surrounding temperature and relative humidity, assembly time, and pressing methods, etc. These factors should be controlled to some extent, in order to form a stable and robust bond line.

Adhesives are also widely used in timber-to-timber joints or timber-based hybrid structures to create bonded connections. Another way is to mechanically connect the joints by means of dowel-type fasteners. Although mechanical connections are commonly adopted in current timber structures and the design methods for this kind of connections are covered in many timber design codes in detail, the adhesively bonded connections hold the distinct advantages of being more cost-effective, stiffer, and with lower risk of moisture penetration in the timber elements (Ascione et al. 2017; Stimpfle et al. 2021). In mechanical connections, the load transferring between the fasteners and the timber elements would lead to local stress concentration and local crushing of the timber materials around the holes (Zhou et al. 2022). These phenomena would further result in a significant reduction of the global stiffness and may cause serious problems in the serviceability limit state (SLS).

Due to the aforementioned benefits of the adhesively bonded connections, there are some initiatives to design and manufacture adhesively bonded timber structures. One of the most successful applications is the British “de Havilland Mosquito” aircraft developed during World War II. Plenty of wood-plywood structural members were adhesively bonded together (U.S. Govt. 1944). Figure 1(a) shows an early practical attempt in Britain to construct timber trusses by using glued-on Douglas fir plywood plates (Aero Research Ltd. n.d.). Also, a roof system with the brand name “Freespan” was developed (Wiehag GmbH n.d.), where the timber elements were adhesively bonded by laminated veneer lumber (LVL) (see Fig. 1(b). One recent adhesively bonded application in New Zealand used LVL as structural elements. They were slotted and adhesively bonded together in dovetail node joints as shown in Fig 1(c) (RTA Studio Ltd. 2020). Timber components could also be adhesively bonded by steel plates or rods. Figure 1(d) illustrates a wooden wind turbine tower in Sweden, where the prefabricated modular elements were assembled by bonded-in perforated steel plates on-site (Jockwer et al. 2023). One glued-in rod application is displayed in Fig. 1(e) (Roseley 2013). Apart from the applications in timber-to-timber connections, adhesive bonding also shows great potential in timber-based hybrid structures, e.g., timber-concrete composites, timber-glass beams and panels, fiber–reinforced polymer (FRP) rods in timber joints, etc. Studies related to the hybrid bonded connections can be found in the literature (Blyberg et al. 2012; Schober et al. 2015; Tannert et al. 2020).

Fig. 1
figure 1

Applications using adhesively bonded connections: (a) timber truss adhesively bonded by Douglas fir plywood plates in Britain (Aero Research Ltd. n.d.), (b) “Freespan” roof system adhesively bonded by LVL plates (Wiehag GmbH n.d.), (c) LVL plates adhesively bonded in dovetail node joints (RTA Studio Ltd. 2020), (d) a wooden wind turbine tower adhesively bonded by perforated steel plates (Jockwer et al. 2023), and (e) glulam beams connected by glued-in rods (Roseley 2013)

Full size image

Recently, there has been growing interest to replace steel plates in timber-to-timber connections with plywood plates made of birch (Betula pendula). Plywood plates, as an alternative to the slotted-in steel plates, are more environmentally friendly and cost-effective, with less prefabrication demand and higher tolerance during assembly (Wang et al. 2023b). Compared to other wood-based panels (e.g., LVL), plywood plate exhibits less anisotropic in-plane mechanical properties, making it promising for connection applications where stresses in different directions need to be transmitted (Wang et al. 2022). Additionally, birch has superior physical and mechanical properties compared to most softwood species and is widely spread in the Eurasian continent (Hynynen et al. 2010; Yin et al. 2023). A mock-up presenting the concept of using birch plywood gusset plates to either mechanically or adhesively connect spruce (Picea abies) glulam elements is shown in Fig. 2. It can be seen that, the shear force in the bonded area is parallel to the glulam grain direction but not always parallel to the plywood face grain direction.

Fig. 2
figure 2

A mock-up using birch plywood plates for the connection of glulam elements (Säfwe and Wåhlander 2021)

Full size image

The influence of the load-to-plywood face grain angles on the bonding strength has not yet been fully addressed. Glos and Horstmann (1989) adhesively bonded the truss diagonal member directly to the bottom chord member without the connection plates and investigated the effect of load-to-grain angles of the lumber on the load-bearing capacity of glued lap joints. They found that the bonded capacity drops dramatically when the load-to-grain angles increase from 0° (parallel) to 90° (perpendicular). This is expected because, for the same wood species, the rolling shear strength activated at 90° is much weaker than the shear strength at 0°. They also found that the Hankinson’s prediction model fit the test results in a satisfactory manner. Kalina (1965) tested the bonding strength between the plywood gusset plate and the truss element with different bonded areas and load-to-face grain angles. It was concluded that there is a significant size effect on the bonding strength but the face grain angle has less significant influence. However, whether the failure occurred on the truss element surface or the plywood surface was not clearly stated. The bonding strength might be strongly related to the shear strengths of wood adherends. Hence, the conclusions drawn by Kalina (1965) may not be applicable to spruce glulam and birch plywood adhesively bonded connections.

This paper aims to evaluate the angle-dependence of the bonding strength between spruce glulam and birch plywood. The glulam grain direction is parallel to the loading axis while the face grain direction of birch plywood varies from 0º (parallel) to 90° (perpendicular) with an interval of 15°. Moreover, the shear strengths of the wood adherends, i.e., spruce glulam and birch veneer, were also experimentally determined, so as to analyze the relation between the bonding strength and the wood shear strength.

2 Materials and methods

2.1 Materials

2.1.1 Timber

Glulam elements made of spruce (GL28cs) (EN 14080 2013) with cross-sectional dimensions of 42 mm × 180 mm were produced by Moelven (Töreboda, Sweden). They were then cut to the shape of the specimens by using table saw and band saw.

Birch plywood panels are commercial products manufactured by Koskisen (Järvelä, Finland). Plywood panels with two different thicknesses were utilized in this study. Plywood plates with a nominal thickness of 15 mm were used in the tensile shear test of the bond line. For the birch veneer shear test, 30 mm-thick birch plywood plates were employed. The veneers in these two tested birch plywood panels possess an identical thickness. Moreover, it is worth mentioning that, in both types of tests, the shear strength of the birch veneer, instead of the shear strength of the whole birch plywood, was activated. Hence, the different thicknesses of birch plywood plates are not a disturbing factor for the analysis of the relation between the bonding strength and the wood shear strength.

2.1.2 Adhesive

In previous investigations (Wang et al. 2023a), three types of adhesives were tested between spruce glulam and birch plywood when both (face) grain directions were parallel to the loading axis, namely, melamine-urea-formaldehyde (MUF), phenol–resorcinol–formaldehyde (PRF), and two-component polyurethane (2 C PUR). Although all the tested adhesive systems exhibited satisfactory bonding strength, meaning that the bonding strength was close to the shear strength of glulam, the specimens with 2 C PUR showed slightly higher bonding strength than the ones with the other two types of adhesives. This is possibly due to the more evenly distributed shear stress along the bond line for the less stiff 2 C PUR compared to the stiffer MUF and PRF. In this study, the same type of 2 C PUR, i.e., Aro-Bond 925 (Ureka, Bristol, UK), was used for the bond line tests. Table 1 describes the information about the adhesive and the assembly parameters.

Table 1 Information about the adhesive and the relevant assembly parameters
Full size table

It is noted that the information with regard to the mixing ratio, application amount, and the pot life is provided from the technical data sheet. Adhesives were applied on both glulam and plywood surfaces with an application amount of 400 g/m2 on each surface. The assembly time stands for the time from the application of the adhesive to the establishment of the pressure. It is kept within five minutes considering the relatively short pot life. The pressing method is described in detail in Sect. 2.2.1.

2.2 Tensile shear test of the bond line

2.2.1 Specimen preparation

Glulam beams were produced, packaged, and then delivered from the glulam factory to the laboratory. Compared to solid lumber, glulam exhibits better dimensional stability. No warping was observed and thus the glulam beams were not planed before assembling the specimens. However, both glulam and birch plywood were sanded considering that the wetting of the wood surface could be difficult if it is dirty or oxidized.

After applying the adhesives, the specimens were screw-glued together. The screw was partially threaded with a total length of 50 mm, a threaded length of 30 mm, and an outer diameter of 4 mm, fulfilling the requirement in the German standard DIN 1052-10 (2012). The diameter of the screw head was around 7 mm. An extra washer with an outer diameter of 30 mm was positioned beneath the screw head to spread the pressure more uniformly on the bonded region (40 mm × 40 mm). The screwdriver (DeWALT DCD790, USA) with a maximum torque of 60 Nm was switched to the drill mode and setting 1 (low speed/high torque) when tightening the screw. Thus, although the pressure from screw-gluing was not quantified, the applied pressure could be assumed similar on each specimen.

The specimens were adhesively bonded in an indoor environment (temperature (T) = 20 °C and relative humidity (RH) = 50–60%). Moisture content (MC) was measured on spruce glulam and birch plywood leftovers that were stored in the same indoor environment as the adhesively bonded specimens. Oven-dry method according to EN 322 (1993) was employed for the measurement of MC, which was 10.2% for spruce glulam and 9.6% for birch plywood.

Before the tests, the specimens that had been adhesively bonded together were conditioned in a climate chamber with a constant T of 20 °C and RH of 65% for at least two weeks. For spruce glulam and birch plywood, the measured densities were 492.0 kg/m3 and 682.9 kg/m3 and the moisture constants were 11.8% and 10.5%, respectively.

2.2.2 Test method

Single lap shear tests were conducted between spruce glulam and birch plywood. As aforementioned in Sect. 2.1.2, in the previous study (Wang et al. 2023a), the loading axis was parallel to both the grain direction of glulam and the face grain direction of birch plywood. In this study, a similar test set-up was utilized to investigate the bonding strength at different angles between the directions of the applied load and the face grain for better comparison. The test set-up and detailed configuration of the adhesively bonded specimen are shown in Fig. 3.

Fig. 3
figure 3

Bond line tensile shear test set-up and configuration (unit: mm)

Full size image

As displayed in Fig. 3, the bonding area is 40 mm × 40 mm. There are seven test series in the bond line tensile shear tests. The plywood face grain angles change from 0° (parallel to the loading axis) to 90° (perpendicular to the loading axis) with an interval of 15°.

The screw was not withdrawn during the test. A preliminary test campaign was conducted to investigate the effect of the screw on the tensile shear capacity of the adhesively bonded connection. It was found that the bonding strength tested with or without the screw inserted in the specimens has no significant difference. It further implies that the presence of the screw during the test would not influence the interpretation of the test results. In this study, the upper part of the glulam was fixed onto a steel fixture, and the lower part of the birch plywood was gripped. It could be argued that the eccentricity between the bond line and the load application plane may induce tensile peeling stress, which might affect negatively the load-bearing capacity of the adhesively bonded connection. However, neither the test setup adopted in this study nor the ones suggested in the testing standards can exclude the tensile peeling stress (Serrano 2004). Nevertheless, the main purpose herein is to investigate the angle-dependence of the bonding strength and the relation between bonding strength and wood shear strength. Therefore, discussions on different test methods and how the peeling stress would affect the estimation of the bonding strength in pure shear are out of the scope of this study.

Bond line tensile shear tests were displacement-controlled with a constant loading head motion of 2 mm/min. Failure is defined when the maximum force occurs during the loading. The bonding strength is thus the maximum force divided by the bonding area. The mean strength value was calculated for each test series. Moreover, the characteristic value or the 5th percentile value was determined via a statistical analysis software package IBM SPSS Statistics (Version 27) (IBM, USA). Student’s t-distribution was adopted considering the small sample size and with the assumption that the sample comes from the normal population. Wood failure percentage (WFP) was assessed through visual inspection to the nearest 10% of the bonding area for each specimen according to EN 314-1 (2004).

The angle-dependence of the bonding strength was statistically evaluated by means of one-way analysis of variance (ANOVA). The significance level was set at 0.05. If the p-value is lower than the significance level, it is concluded that there is statistical significance. The Tukey post-hoc test can further specify which test series is significantly different from the others.

2.3 Tensile shear test of the wood adherends

To capture the relation between the bonding strength and the shear strength of the wood adherends, tensile shear tests were also conducted to determine the shear strengths of glulam and birch veneer. These samples went through the same conditioning process as the adhesively bonded samples. The shearing area was the same as the bonding area, i.e., 40 mm × 40 mm, to exclude any potential size effect.

2.3.1 Glulam tensile shear test

Glulam tensile shear test set-up and specimen configuration are shown in Fig. 4. Tests were carried out with a constant loading rate of 2 mm/min till failure. Shear strength is defined as the maximum force divided by the shearing area.

Fig. 4
figure 4

Spruce glulam tensile shear test set-up and configuration (unit: mm)

Full size image

2.3.2 Birch veneer tensile shear test

7 test series were carried out to test the shear strength of birch veneer when its grain direction was oriented at different angles to the loading axis. The test set-up and specimen configuration are displayed in Fig. 5. From the side view, it can be seen that there is a saw cut from each side of the specimen, which is extended to the middle veneer. The birch plywood has a nominal thickness of 30 mm and is composed of 21 veneers. The grain direction of each veneer is perpendicular to the adjacent ones. The veneer lay-up can be notated as ‘|-|-|-|-|-|-|-|-|-|-|’, where ‘|’ is one veneer and ‘-’ is another veneer with the grain direction perpendicular to ‘|’. It is thus noticed that the middle veneer has the same grain direction as the face veneer.

Fig. 5
figure 5

Birch veneer tensile shear test set-up and configuration (unit: mm)

Full size image

2.4 Summary of the test series

The test series conducted in the paper are summarized in Table 2.

Table 2 Summary of the test series
Full size table

3 Results and discussion

3.1 Bonding strength and wood failure percentage

Tensile shear tests of the bond line between spruce glulam and birch plywood were conducted with the varying load-to-plywood face grain angles. Test results including the mean bonding strength and the standard deviation are displayed in Fig. 6 and Table 3. It is noticed that, when the load-to-plywood face grain angle increases from 0° to 90°, the bonding strength does not monotonically raise or drop. The bonding strength firstly increases when the face grain angle changes from 0° to 15° and then declines until the face grain angle reaches 45°. After that, the bonding strength fluctuates at around 5 MPa. It is interesting to observe that the bonding strength at 15° is the highest (6.5 MPa), slightly higher than that at 0°. A possible reason is explained in Sect. 3.3. While the lowest one takes place at 45° and 75°, i.e., 4.6 MPa, which is not much lower than the highest one.

Fig. 6
figure 6

Bonding strength in relation to the load-to-plywood face grain angle. The error bars denote the standard deviation

Full size image

One-way ANOVA reveals that the bonding strength is significantly influenced by the load-to-face grain angle. In order to find out which specific groups have significant difference, a follow-up post-hoc test was employed. As shown in Fig. 6, there are some non-capital letters, for example, a, b, c, etc., above the mean value of each test series. If two test series share the same letter, then the difference between these two groups is not statistically significant. It can be observed that the bonding strength tested at 15° to the face grain direction, i.e., the highest one, is significantly different from the bonding strengths at other load-to-face grain angles. The bonding strengths at 45° and 75°, i.e., the lowest ones, are also significantly different from the bonding strengths at 0° and 30°.

Although there are some statistical differences between the groups, the tested bonding strengths range from 4.6 MPa to 6.5 MPa. When the load-to-plywood face grain angle changes from 0° to 90°, the face veneer of birch plywood that is adhesively bonded to spruce glulam is subjected to the shear stress in different directions. At 0°, the face veneer is loaded in shear parallel to the grain, while at 90°, the face veneer is loaded in shear perpendicular to the grain (or rolling shear). It is known that the rolling shear strength is much weaker than the shear strength parallel to the grain (Blaß and Sandhaas 2017). Interestingly, the bonding strength at 90° is not the lowest among all the tested bonding strengths and is only slightly lower than the one at 0°. This finding is favorable for the design and development of adhesively bonded connections in timber structures, especially in truss structures where the plywood plates need to take the stresses in multiple directions. In order to further explain the relatively small difference between the bonding strengths at 0° and 90°, it is necessary to check the failure modes and the shear strengths of the wood adherends.

Some researchers have made profound investigations on the wood-adhesive systems (Serrano 2004; Frihart and Hunt 2010). A wood-adhesive bond line model is often composed of different components in a chain across the adhesively bonded joint including the wood adherends, adhesive, and adhesive-wood interfaces. In this specific study regarding the adhesively bonded connection between spruce glulam and birch plywood, the failure modes can be categorized into failures in cohesion in glulam, adhesion to glulam, cohesion in adhesive, adhesion to plywood, and cohesion in plywood. Considering that in few specimens, failure was observed not only in the face veneer of plywood but also in the second veneer, the cohesive failure in plywood can be subdivided into the cohesive failure in the face veneer and second veneer separately. An illustration of the studied wood-adhesive system and types of failure is displayed in Fig. 7. In general, the cohesive failure in the wood adherends, i.e., in glulam and plywood, can be classified into wood failure. A summary of the mean wood failure percentage (WFP) of each test series is listed in Table 3.

Fig. 7
figure 7

Illustration of the glulam-adhesive-plywood system and types of failure

Full size image

As can be viewed from Table 3, for the test series at 0° and 15°, failure mainly occurred in glulam with the glulam failure percentage of around 80%, implying that glulam material was the weakest component in the bond line system. For the rest of the test series, the specimens tended to fail in a mixture of cohesion in spruce glulam and birch plywood. Among them, the specimens loaded at 75° possessed the most failure in birch plywood and in the face veneer (80%). Only in few cases at 15° and 30° could the failure in the second veneer be observed. The total WFP was above 70% at all the load-to-plywood face grain angles, with the lowest WFP of 70% at 45° and the highest of 95% at 75°. With regard to the parts that did not fail in the wood adherends, the failure can be mostly classified into cohesive failure in adhesive and a very small portion of the adhesive lost the adhesion to birch plywood.

Table 3 Mean and characteristic bonding strengths and wood failure percentage
Full size table

Some typical failure modes of the adhesively bonded specimens at different load-to-face grain angles are shown in Fig. 8. It is evident that, for the specimens loaded at 0° and 15° to the face grain, failure mainly took place in glulam; while for the specimens at 75°, nearly 100% birch plywood face veneer failure was noticed. Taking a look at the specimens at other angles, i.e., 30°, 45°, 60°, and 90°, more complex failure modes were observed, covering a large portion of cohesive failure in glulam and plywood and a small amount of cohesive failure in adhesive.

Fig. 8
figure 8

Typical failure modes of the adhesively bonded specimens at different load-to-face grain angles

Full size image

3.2 Wood shear strength

To gain more understanding and to explain the relatively small difference of the bonding strength at different angles to the plywood face grain, the shear strengths of the wood adherends were tested. Test results are listed in Table 4 and plotted in Fig. 9. Figure 9(a) and Fig. 9(b) present the shear strength of spruce glulam and birch veneer, respectively. Figure 9(c) displays the weaker shear strength between these two wood adherends at each load-to-face grain angle.

Table 4 Mean shear strengths of spruce glulam, birch veneer, and the weaker one between them
Full size table
Fig. 9
figure 9

(a) Shear strength of spruce glulam, (b) shear strength of birch veneer, (c) shear strength of the weaker adherends, and (d) typical shear failure modes of the birch veneer specimens at different load-to-face grain angles. The pink region stands for the mean shear strength of spruce glulam ± one standard deviation. The error bars denote the standard deviation

Full size image

The shear strength of spruce glulam parallel to the grain, or the longitudinal shear strength, was tested. The mean shear strength was determined to be 5.4 MPa. It is mentioned in the handbook (Crocetti et al. 2011) that the characteristic shear strength of spruce glulam is 3.5 MPa regardless of which class it is. The disparity between the experimentally determined shear strength and the characteristic shear strength in the handbook is also noticed and reported by other scholars. A wide range of shear strength data can be found in the literature, mainly between 5 MPa to 10 MPa (all above 3.5 MPa), which might be associated with the different test methods and specimen sizes in different publications (Crocetti et al. 2010; Wang et al. 2015; Karagiannis et al. 2016).

Moreover, the glulam specimens were cut from different glulam beams and the sawing pattern was not selected or checked for the glulam longitudinal shear tests. Therefore, it is hard to determine if the glulam specimens were loaded in longitudinal-tangential shear or longitudinal-radial shear. Nevertheless, these two types of shear are usually not distinguished in structural engineering practice.

In birch veneer shear tests, the load-to-face grain angle was altered from 0° to 90° with an interval of 15°. By considering birch veneer as an orthotropic material, three mutual perpendicular axes in the coordinate system can be defined along the birch veneer’s longitudinal, radial, and tangential directions, respectively. Hence, the change of load-to-face grain angles from 0° to 90° is the change of shear directions from L-R (longitudinal-radial) plane to R-T (radial-tangential) plane. The shear in R-T plane is also referred to as rolling shear. As displayed in Table 4 and Fig. 9(b), birch veneers exhibit the highest shear strength at 0° (in L-R direction). The shear strength drops step by step from 0° to 15° to 30° and then waves between 45° and 90° (in R-T direction) (rolling shear).

The mean L-R shear strength of birch veneer at 0° is 6.4 MPa with a standard deviation value of 0.4 MPa. No literature has been found regarding the L-R shear strength of birch veneer. In the study performed by Pramreiter et al. (2021), in-plane L-T shear strength of birch veneer was tested with an average value of 8.5 MPa. The slightly higher strength value obtained from their study could be attributed to (a) the intrinsic shear strength difference in L-R and L-T directions and (b) the size effect. The shorter shearing length (10 mm) in their study compared to the study herein (40 mm) might result in higher shear strength. Overall, the tested shear strength of birch veneer at 0° was regarded reasonable.

The mean R-T (rolling) shear strength of birch veneer at 90° is 4.6 MPa with a standard deviation value of 0.6 MPa. Ehrhart and Brandner (2018) conducted comprehensive rolling shear tests on several commonly used softwood and hardwood species in Europe including birch (Betula pendula) following the test set-up in EN 408. They found that the mean rolling shear strength of solid birch is around 3.5 MPa, significantly higher than that of spruce (1.9 MPa) but lower than that of beech (5.4 MPa).

By comparing the longitudinal shear strength of spruce glulam (in Fig. 9(a)) and the varying shear strength of birch veneer at different angles (in Fig. 9(b)), it is noticed that the longitudinal shear strength of spruce glulam is lower than the birch veneer shear strength at 0° and 15° but higher than that at other angles. The bonding strength and bond line failure mode are governed by the weaker wood adherends if the adhesive gives satisfactory performance. This can be proven by reviewing the wood failure percentage presented in Table 3 and the typical bond line failure modes in Fig. 8. At 0° and 15°, most failure occurs in bulk glulam due to its weaker shear strength. At other angles, more portions of failure take place in the face veneer of birch plywood. The weaker wood shear strengths at varying angles to the face grain are displayed in Fig. 9(c) and will be compared to the bonding strength in Sect. 3.3.

3.3 Relation between the bonding strength and wood shear strength

The shear strengths of the weaker wood adherends were compared to the bonding strength in Fig. 10. At all the load-to-face grain angles, the bonding strengths are slightly higher than the wood shear strengths.

Fig. 10
figure 10

Comparison between the bonding strength and the weaker wood shear strength

Full size image

Moreover, the bonding strength and the wood shear strength show very similar face grain angle-dependence, implying their strong correlation. By analyzing the experimental data in this study, a linear relationship between the bonding strength and the shear strength of the weaker wood adherend can be found (see Eq. 1):

$${f}_{v,b}=0.88{f}_{v,w}+1.27 \left(MPa\right) ({R}^{2}=0.78)$$
(1)

where \({f}_{v,b}\) is the bonding strength; \({f}_{v,w}\) is the weaker wood shear strength; and \({R}^{2}\) is the coefficient of determination.

It is worth noting that Aicher et al. (2018) carried out a literature review of the bond line block shear strength in relation to the wood shear strength for hardwood glulam materials. The covered species comprised beech, oak, poplar, and some from tropical areas. A substantial number of test data were employed to formulate the linear relationship between bond and wood shear strength: \({f}_{v,b}=0.912{f}_{v,w}+0.628 \left(MPa\right) ({R}^{2}=0.88)\). As a result, the strong positive correlation found in this study is in line with the results in the literature.

Furthermore, owing to the strong correlation between bond and wood shear strength, the relatively small difference of the bonding strength at varying angles can also be explained by the small difference of the wood shear strength. From 0° to 90°, the weaker wood adherend transfers from glulam loaded in longitudinal shear to birch veneer subjected to rolling shear stresses. Due to the hardwood nature and the high density of birch, the rolling shear strength of birch veneer (4.6 MPa) is only slightly lower than the longitudinal shear strength of spruce glulam (5.4 MPa).

As aforementioned in Sect. 3.1, the bonding strength at 15° is slightly higher than that at 0°, even though both specimens at 0° and 15° mainly failed in glulam. This phenomenon could be possibly explained by their different shear stress distributions along the bond line due to the different elastic modulus of birch veneer at 0° and 15° based on the Volkersen theory (Thelandersson and Larsen 2003).

The European standard EN 14080 (2013) provides the minimum requirements on the bonding strength and wood failure percentage for the glulam block shear test. Around one-third of the individual specimens and the average values at 15° to the face grain fulfill the requirements. However, this standard is solely for softwood glulam and the shear test should be conducted along the grain direction, which is different from the case studied herein. In this study, hardwood plywood was adhesively bonded to softwood glulam with varying load-to-face grain angles. Thus, it is not suitable to conform the test results to the requirements in the EU standard. On the other hand, the American national standard ANSI A 190.1-2017 (2017) requires that the bonding strength should equal or exceed 90% of the wood shear strength, irrespective of the softwood or hardwood species. The tested bonding strength is higher than the wood shear strength, thereby fulfilling the requirement of the American standard.

4 Conclusion

In this paper, the bonding strength between spruce glulam and birch plywood was evaluated. The investigated factor is the load-to-plywood face grain angle. Test results indicate that, when varying the load-to-plywood face grain angles from 0° to 90°, the bonding strength changes within a relatively small range, from 4.6 MPa to 6.5 MPa, although there are some statistically significant differences. The weak angle-dependence of the bonding strength is beneficial for the design and the development of the adhesively bonded joints when utilizing birch plywood as gusset plates in trusses and rigid frame corners. In these types of applications, plywood gusset plates ought to take the stresses transferred from other timber components, which are at different angles to the face grain.

To further explain the observation regarding the weak angle-dependence of the bonding strength, the wood failure percentage (WFP) of each specimen was examined, and the shear strengths of the wood adherends were tested as well. The total WFP was above 70% at all the load-to-plywood face grain angles, with the lowest WFP of 70% at 45° and the highest of 95% at 75°. Specifically, at 0° and 15°, failure mainly occurred in bulk glulam; while at other angles, a mixture of cohesive failure in glulam and plywood face veneer was dominant. Only a very small amount of wood failure took place in the second veneer of plywood and its influence can be considered negligible.

The shear strength of spruce glulam was tested parallel to the grain and then compared to the tested shear strength of birch veneer from 0° to 90°. It was found that the longitudinal shear strength of spruce glulam is lower than the birch veneer shear strength at 0° and 15° but higher than that at other angles.

Although birch veneer is subjected to rolling shear when loaded at 90°, the rolling shear strength of birch veneer (4.6 MPa) is only slightly lower than the longitudinal shear strength of spruce glulam (5.4 MPa) due to the hardwood nature and the high density of birch species. A strong positive correlation was observed between bonding strength and the shear strength of the weaker wood adherend.

It is known that bonding strength has significant size effect, possibly attributed to the non-uniform shear stress distribution and the potential defects along the bond line. Therefore, the specific bonding strength values reported herein are only valid for the adopted bonded area (40 mm × 40 mm). In order to implement the bonding strength into design formulas for adhesively bonded connections, size effect on the bonding strength should be systematically studied by scaling up the bonded area step by step in the future.