Effect of basalt fiber length on the behavior of natural hydraulic lime-based mortars

2.1 Mortars: Materials and mixture

The characteristics of the materials used are detailed below.

Natural hydraulic lime NHL 3.5 from the Saint-Astier company, which meets all the conformity criteria established by the standard EN 459-1 [16].

Silica sand was in accordance with EN 13139 [55]. Figure 1 shows the particle size distribution of aggregate obtained by sieving according to EN 1015-1 [56]. It is verified that it has a maximum fines content of 3%.

Figure 1

Particle size distribution of aggregate.

Water for mixing and curing was taken from the drinking water network of the Madrid City Council that complies with the sanitary quality criteria suitable for human consumption according to RD 3/2023, of January 10, and with the technical requirements for mixing and curing concrete and mortar [57,58].

Superplasticizer, water reducer, and accelerator of the mechanical resistance of mortar, were of commercial brand Dynamon NRG 1015 from Mapei that complies with the requirements established by EN 934-1 for concrete and mortar [59]. The superplasticizer is used to facilitate the workability of the mixture during the tests and to accelerate its hardening so that conclusive results can be obtained for the analysis under study at early ages (28 days). The use of this type of additives for the physical-mechanical improvement of the final product is theoretically and experimentally supported [24].

Basalt fibers were obtained from a unidirectional fabric called MAPEWRAP B UNI-AX from Mapei. In order to test the influence of fiber length on the properties of the resulting mortar, the fabric was cut into five different lengths: 6, 12, 18, 24, and 30 mm (Figure 2). Table 1 shows the characteristics of the fiber used.

Figure 2

Basalt fibers (a) cut for use and (b) fabric prior to cutting.

All specimens have been prepared with a lime:aggregate ratio by weight (L/A _w) equal to 1:3, commonly used in conservation and rehabilitation works of historical constructions [60,61]. Based on this ratio, the dosage of NHL specimens shown in Table 2 has been determined from the bibliographic update [14,15,17,42] of the experimental research carried out by Bustos [54]. Volume proportions of basalt fiber (1%) were converted in weight to avoid measurement imprecision on mixing process.

2.2 Testing program

The tests are carried out according to the current European standard [56,62,63,64,65,66,67,68,69,70,71] in order to obtain results that can be easily compared with other research already carried out or under development, although Veiga et al. [72] pointed out the limitations of some of these tests to adequately characterize the lime mortars used in the conservation of historic buildings. The specified quantities of materials (Table 2) were weighed and mixed in a mechanical mixer. Standard metal molds were filled with the mixture and cured according to the laboratory conditions of each test.

Characterization tests of fresh mortars were carried out in laboratorial controlled standard conditions (temperature (T) = 20 ± 3°C and relative humidity (RH) = 65 ± 5%). Characterization tests of hardened mortars were carried out at 28 days; the freshly molded mortar samples were placed in a humid chamber for 2 days for initial curing with T = 20 ± 2°C and RH = 95 ± 5%; after the first two days, the samples were demolded and continued to be inside humid chamber up to 28 days.

Tests on fresh mortar were performed on a sample of each mix. Tests for hardened mortars were performed on a minimum of three samples of each mix except for dry bulk density and thermal conductivity test for which only one sample of each mortar was used, and water absorption coefficient with two samples per mix; a total of six samples per mix were used in the compression test. Table 3 summarizes the test program for each mix of hardened mortar.

2.2.2 Physical tests on the hardened mortar

The following physical tests were carried out on the hardened mortar: bulk density, capillary water absorption, and thermal conductivity. These properties are directly related to the durability of the mortar.

The dry bulk density on hardened mortar was determined according to EN 1015-10 [66], by weighing the specimens in an oven until a constant weight was reached and immersing the saturated specimens in water to determine the apparent volume by hydrostatic weighing.

The water absorption coefficient was estimated in accordance with the standard EN 1015-18 [67] although Veiga et al. [72] recommend the use of the standard EN 15801 [73] with some improvements, especially in the case of air lime mortars.

The thermal conductivity was determined according to EN 12664 [68] by the heat flow meter method. For this purpose, specimens were manufactured and cured for 28 days before being placed in an oven at 100°C to obtain a constant weight over time. Then, they were then placed on one side of a closed enclosure at a constant temperature, generated by an internal heat source and a fan, and insulated from the outside with an expanded polystyrene plate. Finally, to determine the conductivity, the temperatures of the interior and exterior air, as well as of the interior and exterior surfaces of the mortar (contact between the plate and the insulation), and of the expanded polystyrene were measured. With these values, the thermal conductivity of the mortar was determined using Eq. (1) (Figure 3).

(1) Ø = Q / t = λ × d T d x = λ × T 2 – T 3 e ,

where Ø = Q / t is the heat flow (W·m⁻²), T ₂ is the temperature of the interior surface of the mortar (K), T ₃ is the temperature of the exterior surface of the mortar in contact with the insulating material (K), e is the plate thickness (m), and λ is the thermal conductivity of the material (W·(m⁻¹·K⁻¹)).

Figure 3

Test enclosure for obtaining thermal conductivity.

2.2.3 Mechanical tests on the hardened mortar

The mechanical tests carried out on hardened concrete are as follows: surface hardness, adhesion, flexural strength, and compressive strength of hardened concrete.

The surface hardness was obtained according to UNE 102042 [69], changing the scale to Shore D hardness. A total of ten measurements are taken from each test tube (five per side).

The adhesion test was performed according to EN 1015-12 [70]. A 10 mm layer of mortar was applied and leveled on a simple hollow brick substrate with dimensions of 400 mm × 200 mm, previously moistened to prevent it from absorbing the water from the mixing. Before hardening, specimens were marked with a truncated cone-shaped cylinder of 50 mm diameter. After 28 days of curing, the mortar specimens were bonded to the tensile pads with an epoxy adhesive. Finally, once the adhesive has hardened, tensile force is applied on the tablets until the cylindrical specimen is detached from the brick substrate (Figure 4). The bond strength is calculated using Eq. (2).

(2) f u = F u A ,

where f _u is the adhesive strength (N·mm⁻²), F _u is the breaking load (N), and A is the area of tested surface (mm²).

Figure 4

(a) Adhesive strength test and (b) surface hardness test.

The fabrication, curing, and testing of the specimens to determine the flexural and compressive strength of the hardened mortar were carried out in accordance with EN 1015-11 [71]. The standard metal molds were filled with the fresh mix in two batches using an IBERTEST CIB-801 compactor. Testing of the specimens at 28 days was performed with the Ibertest universal press model MIB-60/AM instrumented with displacement transducers. The tests were performed by controlling the displacement at a rate of 0.05 mm·min⁻¹ for the bending test and 0.5 mm·min⁻¹ for the compression test.

The bending tests were performed on prismatic specimens by applying, with the aid of the corresponding adapter, a point load at the center of the span of the upper part of the specimen supported on two points on the lower face 100 mm apart (Figure 9a). The test was performed with a preload of 5 N to avoid premature breakage of the specimen. The value of the flexural strength is obtained with Eq. (3).

(3) σ f = 1.5 F · l b · d 2 ,

where σ _f is the flexural strength (N·mm⁻²), F is the maximum bending load recorded (N), l is the distance between supports (mm), b is the specimen section width (mm), and d is the specimen section depth (mm).

Compression tests were performed on each of the resulting halves of each specimen tested in bending (Figure 9b) with the same universal press, but changing the adapter used. The compressive strength and strain were obtained from Eqs. (4) and (5).

(4) σ c = Q S ,

(5) ɛ = L – L 0 L 0 ,

where σ _c is the compressive strength (N·mm⁻²), Q is the maximum compressive load recorded in the test (N), S is the cross section of the specimen (mm²), ɛ is the compressive strain (dimensionless), L ₀ is the initial length of the specimen (mm), and L is the final length of the specimen (mm).

3.1 Tests on fresh mortar

Table 4 shows the results obtained in the tests carried out on fresh mortar. The consistency of the fresh mortar is totally related to the workability of the mortar. The standard EN 1015-3 [64] establishes three types of consistency: dry consistency for values below 140 mm, plastic consistency for values between 140 and 200 mm, and soft or fluid consistency for values above 200 mm. Experience shows that there is a different demand of the consistency based on the use of the mortar in the structure and on the binder [6]. The laboratory workability conditions under which the water quantities of most NHL mortars have been fixed in the ranges of 140 mm [74], 150 mm [17], 134–155 mm [75], 150–155 mm [76], and 140–200 mm [77]; the values obtained for the reference mortar (142 mm) are within these ranges.

There is a clear difference between the consistency of the reference mortar and the consistencies of the mortars with fibers, so that the addition of fibers means a significant reduction in the value of the mortar slump, which is around 110 mm for mortars with fibers compared to 142 mm in the reference mortars. The reason for this decrease is due to the presence of fibers resulting in the formation of a network-like structure that restricts the segregation of the mixture together with the need for a greater amount of binder to envelop all the fibers due to their large specific surface area [78]. As the fiber length increases, the value of the slump decreases, although this decrease is very small.

The bulk density in the fresh state is directly related to the density of the raw material components of the mortar and to the occluded air content. Although the density of basalt fibers (2,750 kg·m⁻³) is higher than that of the reference mortar (2,107 kg·m⁻³), the density of reinforced mortars (variable from 2,041 to 2,099 kg·m⁻³) is lower than that of the reference mortar. This is because the fibers tend to trap air inside the mixture [4,45,79,80].

The results obtained for the reference mortar (2,107 kg·m⁻³) are similar to those obtained by other authors. Apostolopoulou et al. [17] found an average value of 2,067 kg·m⁻³ for NHL 3.5 and B/A _w = 1/3. As the length of the added fiber increases, there is a slight progressive decrease in the value of the mortar density despite the fact that the amount of added fibers is the same in all dosages. This can be explained by a lower compaction of the mortar as the length of fiber added increases.

The air content in fresh mortar is an indicator of the porosity and, therefore, of the long-term behavior of the hardened mortar. The presence of air inside the mortar may be due to air entrapment between the fibers, poor mixing, or inadequate compaction of the mortar. Experimental results show that the occluded air content in reinforced mortars is higher than in reference mortars and that it increases with the increase in the fiber length. The air content amplification factor (relationship between the characteristic values of reinforced and unreinforced mortars) varies from 1.038 for samples with 6 mm fiber length and 1.417 for 30 mm. Microscopic studies show that the inclusion of fibers disrupts the continuity of the mixture [15,39,40].

The joint analysis of the three properties studied for fresh mortar is congruent. Mortars with fibers present lower slump, lower densities, and higher amounts of air content than reference mortars and in addition, the slump and bulk density decrease with fiber length while the occluded air content increases.

Figure 5 shows the relationship between the factors of reduction of consistency and bulk density in relation to the amplification coefficient of the air content with respect to the reference mortar. It is found that there is a good correlation between the increase in air content and the decrease in density (R ² = 0.979), while the relationship with the decrease in slump is poor (R ² = 0.381).

Figure 5

Relationship between air content amplification factor, consistency reduction factor, and bulk density reduction factor.

3.2 Physical tests on the hardened mortar

The results of the dry bulk density, water absorption, and thermal conductivity tests performed on the hardened lime mortars are shown in Table 5.

The dry bulk density of the mortars reinforced with basalt fibers is lower than that of the reference mortar and is related to the air content of the mixture. Other authors also found lower bulk density values in basalt fiber reinforced NHL 3.5 mortars (B/A _w = 1/3; river sand; 28 days-humid chamber) [15,42,81]. A relationship between fiber length and dry bulk density value is not appreciated. In any case, for the reinforced mortars, the results obtained are very similar, with an average value of 1,929 kg·m⁻³, which is a decrease of 4.4% with respect to the value of the reference mortar.

Water absorption coefficient is very important for NHL [82] because it provides valuable information about the size and volume of the open pore network inside the mortar [83]. When mortars are exposed to the action of atmospheric agents, the capillary water absorption coefficient is a determining property for assessing the durability of the mortar, since the inclusion of water in the pore network accelerates the degradation of the material. At the same time, higher water absorption coefficients in NHL mortars are indicative of larger pore sizes and lower risk of degradation by salt crystallization and freeze–thaw [72]. The analysis of Table 5 shows that the water absorption coefficient is higher in the reference mortars than in the reinforced mortars. These results agree with those obtained by other authors with basalt fiber reinforced NHL 3.5 mortars (B/A _w = 1/3; river sand; 28-days humid chamber) [4], although it is not possible to establish a trend of relationship with the length of the fibers.

The thermal conductivity is a measure of the mortar’s capacity as a thermal insulator. Reinforced mortars have a higher thermal conductivity than reference mortars (Table 5). In addition, an increase in the coefficient can be seen as the length of the added fiber increases. Figure 6 shows the relationship between the amplification or reduction factors with respect to the reference mortar of thermal conductivity vs dry bulk density and water absorption coefficient. The decrease in air content has also been added as a comparison variable. It is observed that there is no relationship between thermal conductivity and water absorption coefficient (R ² = 0.0132), but there is an inverse relationship between thermal conductivity and dry bulk density (R ² = 0.1767), and air content (R ² = 0.9823). These data are contrary to what was expected and could be explained by a variation in the compaction times of the tested specimens, a factor that has a great impact on the values of the thermal conductivity coefficient [84].

Figure 6

(a) Relationship between thermal conductivity amplification factor, water absorption reduction factor, and dry bulk density reduction factor and (b) relationship between thermal conductivity amplification factor and air content amplification factor.

3.3 Mechanical tests on the hardened mortar

Table 6 summarizes the mean values and standard deviations of the mechanical tests carried out on the mortars cured at the age of 28 days.

Despite the large dispersion of the results in the hardness tests, it is observed that the lime mortars reinforced with basalt fiber present slightly higher values than those of the reference mortar, although no relationship can be established between the length of the fiber and the hardness of the reinforced mortar. The relationship between the characteristic values of reinforced and unreinforced mortars (amplification factors) ranges from 1.01 for NHL-30 mm ln to 1.05 for NHL-12 mm ln. Barbero-Barrera and Flores [85] also found that polypropylene fiber reinforcement of NHL lime pastes (NHL 5; water/blinder = 0.60) slightly increased surface hardness at 90 days when the percentages were small (up to 2%). In contrast, Morón et al. [86] obtained slightly lower hardnesses when reinforcing HL mortars with glass, carbon, and basalt fibers (1% by volume) (CL-70; cement/aggregate = 1/3 and ¼, and 28 days-humid chamber).

The adhesive strength of NHLs reinforced with basalt fibers is slightly lower than that of the reference mortar with reduction factors ranging from 0.56 for NHL-18 mm ln to 0.80 for NHL-6 mm ln and NHL-8 mm ln. No clear relationship, however, can be established between increasing fiber length and decreasing adhesive strength in the range of long fibers (>20 mm). Morón et al. [86] also found slightly lower results for reinforced HL mortars with cement/aggregate ratios of 1/4 and similar at 1/5 ratios. Zanotti et al. [87] found the reverse effect in cement mortars. Spadea et al. [43] also confirmed that reinforcement increases the adhesive strength of cement mortars (Figure 7).

Figure 7

Failure of specimens in adhesive strength test.

Table 7 shows the results of the flexural test, indicating the maximum force resisted by the mortar, the corresponding maximum stress and the displacement reached by the specimen for the maximum stress level, as well as the energy absorbed up to the point corresponding to the maximum stress (A _max) and the total energy absorbed in the test, corresponding to the integral of the load–displacement curve up to the breakage of the specimen (A _u). The quotient between both values of energy A _u/A _max is an index of the ductility in terms of energy of each mix.

The average flexural force–displacement curves for each dosage are shown in Figure 8a. The average flexural strength of the control samples was 0.909 N·mm⁻² at 28 days. The maximum strain was reached at an average displacement of 0.063 mm. The specimens show brittle breaks and sudden drops in strength after reaching the maximum load (0.388 N). The behavior is linear up to that point, but the appearance of a first crack quickly triggers failure (Figure 9a).

Figure 8

Mechanical behavior of NHL mortars (a) flexural behavior and (b) axial compression behavior.

Figure 9

Failure of specimens in mechanical tests (a) flexural strength and (b) compressive strength.

Reinforcement with basalt fibers increases the flexural strength and ameliorates the sharp decrease in strength after the peak by showing more ductile modes of failure. Specimens with 6 mm fiber lengths have an average flexural strength of 1.11 N·mm⁻², an increase in strength of 22% over the reference mortar. Specimens with longer fiber lengths show very similar peak values, with strengths on the order of 1.30–1.33 N·mm⁻², approximately 45% higher than the reference mortar. The curves show a linear section of similar slope up to the peak load and failure modes that become more ductile as the reinforcement fiber length increases. The drop in strength after the peak is sharper in the specimens with 6 mm fiber length than in the rest, with very similar failure patterns observed between the 12–18 and 24–30 mm specimens. The energy absorbed until reaching the peak load of the reinforced specimens increases between 90% for specimens with 6 mm length and 160% for the rest of the cases. The total energy increases by 340% on average and the ductility by 80%. In spite of the homogeneity of the failure patterns, there is considerable dispersion in the failure points of the specimens.

Other studies have also shown that reinforcement with basalt fibers increases flexural strength, although with much lower strength amplification factors. Lucolano et al. [15] found almost no variation for specimens with 1% basalt fiber of 4–5 mm length. Similar values were obtained by Asprone et al. [42]. Santarelli et al. [4], on the contrary, provided a strength decrease coefficient of 0.87 for mortars with 3% fibers of 6.35 mm length. Similar losses were obtained by Demircan et al. [81] for specimens reinforced with 1.20% of 6 mm long basalt fiber (flexural strength reduction factor of 0.85). There is unanimity, however, that reinforcement with basalt fiber provides more ductile failure patterns [4,15,42,81]. The disparity in these results, all of them carried out with the same type of lime (NHL 3.5) and aggregate (river sand), at a dosage of B/A _w = 1/3, and curing conditions (28 days-humid chamber), can be explained by the indeterminacy of the technological properties of both materials and by the sensitivity of the interfacial transition zone between the lime matrix and the basalt fiber to factors such as the type and roughness of the fibers, the characteristics and chemical composition, the mineralogical characterizations and particle size of aggregate, and the mixing method. For other types of lime and fiber reinforcement mortars, the results also show a large dispersion: in the study by Rosato et al. [88], a small decrease (NHL mortar reinforced with nano-structured cellulose fibers), in the study by Pachta and Goulas [89], a decrease at 28 days and an increase at 90 days (LP and NHL grouts reinforced with glass and polypropylene fibers), and in the study by Kesikidou and Stefanidou [90], an increase (CL mortar reinforced with jute, coconut, and kelp).

Figure 10 presents the flexural strength amplification factor as a function of basalt fiber length (ratio between the characteristic flexural strength of reinforced and unreinforced mortar). Eq. (6) proposes a prediction model based on the obtained experimental results that shows a good fit. For small fiber lengths, less than 12 mm, the flexural strength increases almost linearly as the fiber length increases; for larger fiber lengths no major increases are observed and the optimum theoretical length would be 23 mm. This effect can be explained because long fiber sizes would produce non-uniform distributions that would affect the continuity of the solid particles, increasing the demand for lime and causing an adverse effect on the workability, porosity, and strength of the mortars.

Figure 10

Strength amplification of bending and compression in NHL.

Table 8 shows the results obtained in the compression tests: maximum stress and strain, ultimate stress and strain, and maximum and ultimate energy densities. The energy density is related to the ductility of the material in compression, and for its calculation, the procedure detailed in the UNE 83508:2006 standard [91] has been carried out, obtaining the area enclosed by the diagram in the test up to the corresponding stress point. The ratios between the ultimate strain and the strain reached at the point of maximum stress, and the ratio between the ultimate energy density and the maximum energy density have also been obtained. These last two values serve as dimensionless indices to compare the ductility of the six dosages in terms of deformation and in terms of energy.

The average axial compressive stress–strain curves for each dosage are shown in Figure 8b. The compressive strength of the control samples was 3.56 N·mm⁻² at 28 days. The maximum strain was reached at an average displacement of 0.013 mm. The specimens show ductile breaks with smooth drops in the strength after reaching the maximum load (2.55 N·mm⁻²).

Reinforcement with basalt fibers increases the compressive strength and does not improve the strength drop after the peak showing slightly less ductile modes of rupture. Specimens with 6 mm fiber lengths have an average compressive strength of 4.13 N·mm⁻², an increase in strength of 21% over the reference mortar. The specimens with longer fiber lengths show very similar peak values, with strengths in the order of 5.12–5.59 N·mm⁻², approximately 51% higher than the reference mortar. The curves show a similar elastic section and a plastic section that increases with the increase in the fiber length up to the maximum stress, and ductile failure modes that hardly vary with the increase in the fiber length. The equal slope in the elastic section is representative of the fact that the addition of fibers does not substantially alter the modulus of elasticity of NHL mortars, a very important aspect in terms of compatibility between repair mortars and historical mortars. The energy density absorbed until reaching the maximum load of the reinforced specimens increases between 90% for specimens with 6 mm length and 177% for the rest of the cases. The total energy density increases by 28.5% for the 6 mm specimens and between 64 and 87% for the rest of the cases. Consequently, the ductility of the reinforced specimens decreases by 34.7% on average. Homogeneity is observed in the failure patterns and in the points at which the specimens fail.

Other experimental works have determined compressive strength losses in mortars reinforced with basalt fibers. Demircan et al. [81] obtained strength losses of 17% in specimens reinforced with 1.20% of 6 mm long basalt fiber. Deteriorations of 36% were found by Lucolano et al. [15] and Asprone et al. [42] for specimens with 1% basalt fiber of 4–5 mm length. Santarelli et al. [4], on the other hand, also found 40% improvements in compressive strength for mortars with 3% fibers of 6.35 mm length. As mentioned before, the lack of consensus on the influence of synthetic reinforcements on the mechanical properties of NHL mortars is explained by the indeterminacy of the parameters that determine the characteristics of the interfacial transition zone between the lime matrix and the basalt fiber. For other types of lime and fiber reinforcement mortars, the results also show a large dispersion: similar at 28 days and increase at 90 days (LP and NHL grouts reinforced with glass and polypropylene fibers) in the study by Pachta and Goulas [89], and an increase (CL mortar reinforced with jute, coconut, and kelp) in the study by Kesikidou and Stefanidou [90].

Figure 10 presents the compressive strength amplification factor (SAF_c) of reinforced lime mortars as a function of basalt fiber length. Eq. (7) proposes a prediction model based on experimental results, and also shows a good fit. The compressive behavior does not differ substantially from the flexural behavior, and for fiber lengths less than 12 mm, the compressive strength increases almost linearly as the fiber length increases; for longer fiber lengths, no major increases are observed and the optimum theoretical length would be 20 mm.