*3.2. Properties in Hardened State*

Physical properties of 90-day-old hardened grouts are given in Table 6, in the form of an average value and associated standard deviation. Measured density was the highest for the LS100 grout (average value of 1.45 g/cm3), which contained no glass microspheres, and the lowest for the LS33-GM67 grout (average value of 0.85 g/cm3) with microspheres occupying 2/3 of the filler volume. This shows that reducing the grout's weight by up to (approximately) 40%, in relation to the reference LS100 grout, can be achieved by replacing part of the limestone filler content with the same volume of glass microspheres. When comparing the fresh and hardened state densities of a particular grout, it is obvious that the drying of the grout is responsible for the reduction in density; this reduction is equal to 0.28 g/cm3, 0.31 g/cm3, and 0.29 g/cm3 for the LS100, LS67-GM33, and LS33-GM67 grout, respectively.

**Table 6.** Physical properties of the LS100, LS67-GM33, and LS33-GM67 grouts in the hardened state: density, total, and capillary porosity, water absorption coefficient after 24 hr (W24) and 10 min (W10), and water-vapour resistance factor (μ).


The reductions are in good correlation with the capillary porosities of the grouts (Table 6), which would be expected due to the evaporable water being held in the capillary pores. The average capillary porosities of the three grouts are in a narrow range between 38% and 40%, despite the relatively large differences in their binder-to-water ratios. Said ratios are equal to 1.86, 1.76, and 1.52 for the LS100, LS67-GM33, and LS33-GM67 grout, respectively. The water absorption ability of the two filler materials needs to be addressed to explain these apparent inconsistencies of properties. The water absorption of the limestone filler is equal to 0.5%, and the water content of the product is less than 0.2%. Glass microspheres are nonporous; thus, they do not absorb water. The highest part of the added water was, therefore, absorbed by the filler particles in the LS100 grout and the lowest by the LS33-GM67 grout particles; as a result, the narrow interval of the capillary porosities was obtained. Total porosity is the sum of the capillary pores and air pores. In the grout compositions containing glass microspheres, the spheres with broken glass walls can contribute to the measured air pores. The contribution of the glass spheres to the measured air pores' content only appeared to be significant in the LS33-GM67 grout, which contained a high amount of glass microspheres. It seems that, during the mixing and/or the test execution, some glass microspheres may have become damaged, which is in line with the backscattered electron images in [8].

The amount of capillary water absorbed by the mixtures at the end of the test (after 24 h; W24) is approximately the same for the three grout compositions, resulting in the W24 coefficient average values between 0.42 and 0.46 kg/(m2√min). Obtained values are considerably lower than values given by Veiga [11] for the hydrated lime: sand (1:3) historic mortars, where W24 is in the range between 1.1 and 1.6 kg/(m2√min). However, considering the requirement that the capillary water absorption of the grout must lie between 50% and 100% of the substrate mortar W24 [4], the obtained results are not far from meeting the required values. Another essential property of the grout is the initial water absorption, presented by the coefficient of capillary water absorption after 10 min [1,38]. From the results in Table 6, it is evident that the average initial water absorption of the three grouts (W10) is approximately the same and ranging between 2.11 and 2.20 kg/(m2√min). These coefficients are within the W10 interval for the fine and coarse lime mortars prepared using Slovenian hydrated limes and limestone sands, where values range between 1.10 and 2.60 kg/(m2√min) [1,39].

The average value of the grouts' water-vapour resistance is lower or equal to 16 (Table 6), which is in line with the results obtained for lime-based mortars by Jornet et al. [38]. The grouts LS100 and LS67-GM33, with the highest contents of limestone filler, showed a slightly increased water-vapour resistance (16 and 15) compared to the grout LS33-GM67 (μ = 12). Broken glass microspheres may be responsible for the obtained result.

Compressive and splitting tensile strengths are related to the total porosity; higher total porosity results in lower mechanical strength. That said, the total porosity values of the LS67-GM33 and LS33-GM67 grouts in Table 6 are underestimated due to the test method applied, which was unable to measure actual hollow volume inside of the glass microspheres. A higher actual total porosity than the one measured is evident from the densities of the grouts in the hardened state (Table 6).

The average values for mechanical strengths are presented in Table 7, along with the corresponding standard deviation. As expected, the glass microspheres decreased the compressive and splitting tensile strengths of the grouts considerably, compared to the reference LS100 composition. At the ages of 90 and 365 days, the average compressive strengths of the LS100, LS67-GM33, and LS33-GM66 grouts were 3.5 and 3.8 MPa, 1.8 and 2.3 MPa, and 1.4 and 1.4 MPa, respectively. This means that a reduction in compressive strength between 40% and 50% can be expected when replacing a third of the limestone filler volume with glass microspheres. When the replacement is increased to two-thirds, the same reduction goes up to about 60%. While the reference grout LS100 complies with the proposed range of compressive strengths given by Ferragni et al. [10] for hydraulic lime grouts (Table 1), the two compositions with the glass microspheres fulfil the requirements for repair lime-based mortars given by Veiga [11], where compressive strengths in the range of 0.4–2.5 MPa are proposed.

Moreover, Pasian et al. [8] studied grouts with reduced water content; they were prepared using slaked lime, pumice powder, quartz sand, and soda–lime–borosilicate glass microspheres. At 150 days, these grouts achieved an average compressive strength ranging from 1.15 MPa to 3.08 MPa. These values are in line with the LS67-GM33 and LS33-GM66 compressive strengths in Table 7.


**Table 7.** Compressive strength and splitting tensile strength of the grout mixtures.

The injection grouts for stabilisation of detached plaster layers are expected to fail predominantly due to tensile stresses [17]. Their tensile strength should be lower than the tensile strength of the original plaster in order to prevent the occurrence of damage to the original material [5]. The average splitting tensile strength of tested grouts at the age of 90 days was between 0.08 and 0.16 MPa (Table 7). These values are well below the 0.3–1.2 MPa range proposed by Ferragni et al. [10]. On the other hand, they fulfil the requirement given in [5] and are close to values reported by Pasian et al. [8] for the nonstructural slaked lime grout and Veiga [11] for the rendering and plastering repair mortar for historic buildings.

The main influencing parameter governing the strength properties is the volume of the grout's solid constituents that can transfer stresses inside of the material; this is reflected in the grout's density and porosity. There are, however, additional parameters that contribute to the strength increase. The results show that grouts with higher limestone-filler content possess higher strength, due to their lower total porosity and better interlocking between the lime binder and the filler particles. The limestone filler is a compact carbonate with sharply cornered grains and a rough surface, which can absorb up to 0.5% of the water from the fresh grout. With water, some lime particles can also be absorbed, making the bond strength between the limestone filler and lime binder considerably higher compared to that between nonabsorbent glass microspheres and lime binder. This finding is supported by the study conducted by Lanas and Alvarez [40], where they concluded that the shape of grains, particle size distribution, and chemical and mineralogical composition of the filler influence the strength of grouts.

The mechanical strength and stiffness of the injection grout and historic lime plaster or render should be approximately equal in order to ensure adequate ductility and durability of the system. From the results in Table 7, it is evident that the glass microspheres are an efficient filler that can be used to adapt mechanical properties of the grout to the mechanical properties of historic plaster or render.

In addition to physical and mechanical properties of the hardened grouts, the durability of the grout mixtures needs to be addressed as well. Besides being an efficient weight-reducing filler, glass microspheres can be seen as a means to introduce stable micro air bubbles to the lime grout. These bubbles can increase the grout's resistance to extreme temperature fluctuations, such as freezing and thawing during the winter and heating and cooling during the summer. The comparison of average compressive strengths for mixtures LS100, LS67-GM33, and LS33-GM67, at the age of 90 days and after the accelerated ageing using distilled water or de-icing salt (3% NaCl), is given in Figure 7.

The accelerated ageing of samples in the presence of distilled water shows that the glass microspheres increased the grout's resistance to the freezing–thawing and heating–cooling cycles. While the LS100 grout was damaged during the accelerated ageing and, as a result, the compressive strength was decreased from the reference value of 3.5 MPa to 2.7 MPa, compositions LS67-GM33 and LS33-GM67—which contained the glass microspheres—were not damaged. Following ageing, the average compressive strength of the LS67-GM33 grout increased from the reference value of 1.8 MPa to 2.1 MPa, while that of the LS33-GM67 grout increased from 1.4 MPa to 1.5 MPa. Accelerated

carbonation of the lime binder, due to wetting and drying, is most probably responsible for the observed strength increase. Similar behaviour was observed by Uranjek and Bokan-Bosiljkov [41] for lime mortar exposed to freezing and thawing cycles.

When the de-icing salt solution was used for the accelerated ageing, grout LS100 fell apart due to the combined effect of water freezing and salt crystallisation (Figure 8). Specimens made from the LS67-GM33 and LS33-GM67 grouts, on the other hand, had retained their shape but were damaged. Dusting, swelling, scaling, and formation of cracks appeared in the lower part of the LS67-GM33 and LS33-GM67 specimens after the sixth cycle of freezing–thawing and heating–cooling (Figure 8). The compressive strength of the LS67-GM33 grout was reduced to 1.4 MPa (22% reduction), while that of the LS33-GM67 grout was reduced to 0.9 MPa (36% reduction).

**Figure 7.** Comparison of compressive strengths of injection grouts at the age of 90 days (normal) and after the accelerated ageing using distilled water or the salt solution.

From the obtained results, it is evident that the durability of tested lime grouts is much higher when their ageing takes place in the presence of pure water, compared to using de-icing salt solution; the specimens aged using distilled water did not show visible damages (Figure 8). Moreover, the compressive strengths of the LS67-GM33 and LS33-GM67 grouts improved after ageing. We can conclude that the tested grouts are highly durable solutions that can consolidate detached plasters or renders if salt-induced problems are not present. The combination of ice formation and salt crystallisation is highly detrimental to the three grouts. However, by incorporating air bubbles in the lime grout using glass microspheres, high enough durability can still be obtained for applications where salts are present in the masonry walls. The highest durability in the presence of salts was obtained for the LS67-GM33 grout, which shows that adequate balance of loadbearing capacity and micro air bubbles' volume is needed to provide adequate durability of the lime grout in an environment containing salts. We can conclude that glass microspheres have the same function as air bubbles in aerated cement mortars; they efficiently reduce the stresses arising from water freezing inside the hardened lime grout and, thus, prevent extensive damage to the grout.

The pull-off strengths, with information about the location of failure in the panel sandwich test, are presented in Table 8. The measured pull-off strength of each particular grout is smaller than its splitting tensile strength (Table 7). The pull-off strength of the LS100 grout in the 2 mm air pocket achieved the value of 0.1 MPa, which is lower than the cohesive strength of lime plaster (0.15 MPa). The failure was predominantly within the grout and partly along the interface between the grout and the fine plaster. The LS100 grout fulfilled the pull-off strength requirement given in Table 1.

**Figure 8.** Specimens of mixtures LS100, LS67-GM33, and LS33-GM67, saturated in the distillate water or the de-icing agent (3% NaCl), before and after freezing–thawing and heating–cooling cycles.

**Table 8.** The pull-off strengths of LS100, LS67-GM33, and LS33-GM67 grouts injected into the simulated air pockets with thicknesses of 2 and 5 mm.


The LS67-GM33 and LS33-GM67 grouts injected in the 2 mm air pocket showed a lower average pull-off strength of 0.08 and 0.07 MPa, respectively; the requirement given in Table 1 was subsequently not met. For the two grouts, failure was predominantly along the interface between the grout and the lower rough layer of the lime plaster. This shows that the bond between the lime plaster and the grout is the weakest link in the sandwich specimen consolidated using grouts with glass microspheres.

Pull-off tests carried out on panel sandwiches with thicker air pockets (5 mm) resulted in considerably reduced pull-off strengths; in the case of the LS100 and LS67-GM33 grouts, they were equal to 0.05 MPa. The failure was predominantly along the interface between the base and the rough plaster. The LS33-GM67 grout sandwich specimens already failed during the test disc installation. This suggests that the drilling of the specimens may have damaged the contacts between different layers of the sandwich panels. Subsequently, measured pull-off values can highly underestimate the actual bond strength between the grout and the plaster.

When comparing results in Table 8 with the pull-off strengths reported in comparable studies [6,12], all the values measured in this study were higher. In Pasian et al. [6] the pull-off strength was in the range of 0.032–0.041 MPa at 150 days, and, in Azeiteiro et al. [12], the maximum pull-off strength after 60 and 90 days was 0.015 MPa and 0.04 MPa, respectively.
