1. Introduction
Masonry structures are nowadays designed for a high thermal insulation level to reduce energy losses in buildings. Compared to concrete blocks or clay bricks, autoclaved aerated concrete (AAC) blocks have better insulating properties, but often at the cost of a lower density and hence of a lower compressive strength [
1]. To prevent cold bridges in the wall-floor junction, the first layer of the inner leaf of a cavity wall consists of a thermal break layer, often in the form of such AAC blocks, with concrete blocks or clay bricks on top as can be seen in
Figure 1. The AAC layer thus acts as a thermal break layer ensuring a continuity of the insulation of the wall and floor [
2], which is shown by the red dashed line.
On top of the AAC layer, a damp proof course (DPC) is placed to prevent water and moisture from entering or rising in the inner leaf of a cavity wall. If this damp propagation is not properly addressed, mold formation can occur inside the wall as well as degradation of the masonry units [
3,
4]. However, from a structural point of view, such a DPC layer comes at the cost of a lower adherence between the mortar and masonry units [
5,
6,
7,
8], resulting in planes of weakness [
9]. In the literature, little experimental data are available regarding the shear strength of masonry walls including a combination of DPC and AAC layers. Therefore, this paper presents an experimental study on the shear behavior of such composite wall systems in which multiple parameters are varied: the unit’s material (concrete or clay), the mortar type and thickness in case of clay bricks (glued thin layer mortar (TLM) or a thicker general purpose mortar (GPM) layer), the precompression load, the presence of a DPC layer, and the positioning of such a layer (in the middle of the GPM layer or directly on top of the unit). To the best of the authors’ knowledge, such an extensive multi-parameter experimental campaign has not been performed before.
In a European standardization context, the standard EN 1052-3 [
10] has been adopted as the standard laboratory test to characterize the shear strength for specimens without a DPC layer and has been adopted by multiple researchers [
11,
12,
13,
14,
15,
16,
17,
18,
19,
20]. For specimens with a DPC membrane, a second European standard, EN 1052-4 [
21], is the recommended experiment for determining the the shear strength, using small wallets containing at least six units. However, for the sake of consistent comparison, EN 1052-3 is followed in this study for both specimens with and without a DPC layer. Both standards assume a Mohr–Coulomb friction law (
Figure 2) to calculate the initial shear strength or cohesion, i.e., the shear strength when no normal (precompression) load is present. This law,
assumes a linear relationship between the shear stress
and the normal stress
, where
is the cohesion and
the friction angle from which the friction coefficient
is calculated. When precompression levels reach a certain level, failure is governed by a compression cap where the specimen will not fail in shear but due to pore collapse.
Barrattucci et al. [
12] carried out triplet tests on masonry units with different cement ratios subjected to monotonic and cyclic shear loadings. The failure modes largely depended on the applied precompression levels rather than the mortar composition. Mojsilović performed extensive research on different types of DPC membranes as well as their position [
6]. Except for the bitumen-based DPC membrane, the influence of the DPC position was rather small for the obtained shear strength parameters. The initial shear strength, though, showed hardly any adherence between the mortar and DPC layer. Martens et al. [
5] came to the same conclusions by performing experiments according to EN 1052-4 [
21] using polyester-based DPC layers. They both suggested to use a characteristic initial shear strength of zero as a lower bound value.
To the knowledge of the authors, no research has been performed on the shear behavior of composite masonry consisting of a combination of DPC and AAC layers, while only limited research exists on the compressive behavior of this type of composite masonry. Deyazada [
22] performed compressive tests on duplets and half-scale masonry panels consisting of an AAC layer with clay bricks or concrete blocks on top. A positive influence on the compressive strength was observed in the composite masonry compared to the one of masonry constructed using only the weakest material.
In the current paper, composite triplet specimens are tested in shear using the materials and test method described in the next section. Next, the results are shown and discussed in
Section 3, in which also an evaluation is made on whether the assumed values in Eurocode 6 [
23] for the initial shear strength and angle of internal friction are met.
3. Results and Discussion
Figure 6 shows the obtained force-displacement curves for a selection of the executed experiments. The first column shows all concrete specimens (‘C’), whereas the center column lists all specimens using clay bricks without tongue and groove (‘T’) and general purpose mortar, and the last column displays the specimens using clay bricks with tongue and groove (‘P’) and either TLM (for specimens without AAC and DPC layers) or GPM. The homogeneous concrete specimens (‘CCN’,
Figure 6a) and the clay brick samples with TLM (‘PPG’,
Figure 6c) show a more brittle failure behavior than the specimens with clay bricks using GPM (‘TTN’,
Figure 6b). The brittleness decreases for all samples when an AAC block is placed in the middle (‘CAN’ (
Figure 6d), ‘TAN’ (
Figure 6e), and ‘PAN’ (
Figure 6f)), and appears to be the lowest when a DPC layer is added (‘CAND’ (
Figure 6g), ‘TAND’ (
Figure 6h), and ‘PAND’ (
Figure 6i)). The sudden drop after the peak load of the concrete specimens (
Figure 6a), which indicates a brittle failure, remains visible for all precompression levels. However, at low precompression levels, sliding failure occurs at the mortar–brick interface, whereas at higher precompression levels, sliding failure occurs with cracking within the mortar joints.
Figure 5c shows a specimen with clay bricks at a higher precompression level. These specimens crush internally before/during sliding, thus leading to a mixed failure mode. The triplet tests with the AAC layer (
Figure 6d–f) always slid between the AAC layer and the mortar, whereas with higher precompression levels, the AAC was scratched or crushed.
Handling and setting up specimens with a DPC layer, in particular the ones with a DPC layer put directly on top of the units, was challenging due to a lack of adhesion between the DPC layer and the rest of the specimen (this can be seen in
Figure 5b). This resulted in excluding two experiments from the TTNDU-03 series since their respective shear strengths were six times lower than the other experiments of that series. At low precompression levels, the remaining specimens fail in sliding whereas at higher precompression levels, crushing of the center brick occurs with sliding afterwards (
Figure 5b).
As can be seen from
Figure 6 all specimens behave quasi-linear elastically at the beginning of the curve, followed by various possible nonlinear behaviors:
The peak load is followed by a sudden drop, after which then the center unit starts sliding at a constant load (
Figure 6a,c,d,f).
Softening behavior occurs after reaching the peak load, but still with a considerable drop. This is the case when internal cracking is happening before sliding (
Figure 6b,h,j).
Softening behavior occurs after reaching the peak load, but no considerable drop occurs, only sliding at the maximum load level. A small smooth drop might occur due to differences in thickness of the mortar joint (
Figure 6g,i,k–m).
A first crack occurs already in the linear elastic part on one side of the center unit. The curve continues with a lower stiffness until one of the aforementioned failure modes takes place (
Figure 6e).
Table 3 shows the mean values of the maximum shear stress and precompression stress. These are thus the mean values of the test repetitions recorded for every precompression level of every setup. The position of the DPC layer resulted in different failure modes. When it is placed at the bottom of a mortar joint (indicated with a ‘U’ in the specimen’s name), the layer lays ‘dry’ on the surface of the brick, and the maximum shear strength is significantly lower. This position of the DPC is often implemented in practice because of convenience since no double mortar layer is required to ‘sandwich’ the DPC membrane, yet this is against the recommendations of DPC manufacturers. The maximum recorded values for the specimens without an AAC block were higher than those with an AAC block, and the presence of a DPC layer decreased the maximum shear stress even more. The bed joints from the specimens were not always intact before testing (indicated in the third column of
Table 3), and this influenced the results. The last column of the table shows the different failure modes recorded during the experiments.
All Mohr–Coulomb curves from the experiments can be seen in
Figure 7. The curves with the concrete blocks show the highest initial shear strength, and different scales were needed to indicate the complete scatter of the concrete specimens. The Mohr–Coulomb curve from the ‘PPG’-specimens could not be drawn because the samples did not fail in sliding but a combination of crushing (i.e., pore collapse of the center unit) and sliding, even at very low precompression levels. The scatter from this graph does, however, show a mean initial shear strength of around 0.20 MPa.
All the results from the Mohr–Coulomb curves (
Figure 7) are combined in
Table 4, converted to characteristic values by multiplying them by 0.8 [
10], and compared with the default characteristic values given by
Table 3 and
Table 4 in Eurocode 6 (which are values for homogeneous and not composite walls) [
23]. Except for the ‘TAN’ and ‘PPNDU’ specimens, all friction coefficients comply with Eurocode 6. The initial shear strengths are, however, always lower than the default values from the standard, except for the homogeneous concrete specimens (‘CCN’). The specimens with a DPC layer exhibit hardly any initial shear strength, and therefore, it is recommended to use an initial shear strength of zero. These results are also confirmed by other researchers [
5,
6]. The ‘PPG’ specimens show a characteristic initial shear strength of ±0.16 MPa, again lower than the Eurocode 6 value. These specimens failed in crushing or crushing with sliding and did not follow a Mohr–Coulomb friction law, i.e., no friction coefficient could be determined, as can be seen in
Figure 7c. Therefore, this result might not be an accurate representation of the initial shear strength. A solution to obtain sliding as the failure mode can be achieved by testing the triplets in the out-of-plane transversal direction instead of the in-plane transversal one.