Four Portland cement types were evaluated in this study: Ordinary (from two origins: O₁ and O₂), Blended (B) (containing approximately 17 % of limestone powder and 12 % of slag) and Pozzolanic (P) (containing approximately 40 % of natural pozzolan). Table 1 presents the physical and mechanical properties and chemical compositions of the studied cements. Table 2 shows the potential Bogue mineralogical compositions of O₁ and O₂.

Table 1

Source: The authors.

Table 2

Source: The authors.

The properties of the chemical admixtures under evaluation, a plasticizer (M), and three superplasticizer (S, P, and V) are presented in Table 3. Two river siliceous sands (fine and coarse) with different fineness modulus were selected (Table 4).

Table 3

Source: The authors.

Table 4

Source: The authors.

For comparative purposes, reference non-admixed mortars were produced for the four cement types so as to evaluate the properties in their hardened state. Then, efficiencies of superplasticizers in varying dosages and with constant proportions of reference mixes were evaluated.

Among these, the mortars that showed the best performance in the fresh state were replicated, including a combination of superplasticizer and the plasticizer (M).

Table 5 shows mix proportions for mortars. Nomenclature corresponds to the cement and type of admixture used (only a letter is indicated for reference mortars, i.e. non-admixed mortars). The mortar mixes were prepared in accordance with ASTM C305. The dose of admixtures is expressed in % w/w of cement.

Table 5

Source: The authors.

Flow and consistency of fresh mixes were determined by the flow test (Fig. 1) and the slump test (Fig. 2), respectively. To evaluate the active period of SPs, these tests were performed immediately after mixing and 30 minutes after the cement got in contact with water.

Figure 1 Source: The authors.

The flow test is commonly applied to select materials, and to proportion pastes and mortars for self-compacting concretes [2, 21-24], but it has not been standardized yet. This method is applied with the aid of a frustum cone. The cone is placed on a plane, smooth, levelled and non-absorbent surface, and filled with the mortar to be tested. The mould is then lifted, and the mortar flows freely. The value of the flow is the average of the two perpendicular measurements of the diameter (Fig. 1).

The slump evaluation was carried out using an “Abrams mini cone” (Fig. 2). This is a scaled version of the standard Abrams cone, corresponding to 1/8 of its volume. The methodology used followed regulation NCh 2257/3 [25].

Figure 2 Source: The authors.

For assessments in the hardened state, prismatic specimens (4 cm x 4 cm x 16 cm) were moulded for each one of the studied mortar mixes. They were then cured under water saturated with lime at 23 ± 2 °C until tested.

Compressive strength at 7, 28 and 90 days as well as density and water absorption at 28 days were evaluated. The determination of density and absorption was carried out according to ASTM C642 and compressive strength was determined following ASTM C349.

Fig. 3 shows flow in relation to the admixture dose. From the comparison of the two mortars with ordinary Portland cement (O₁ and O₂), it can be observed that O₁ shows higher fluidity than O₂, in correspondence with a lower C₃A/calcium sulphate (1.91 versus 2.20 for O₁ and O₂, respectively computed from contents of SO₃ and C₃A, Tables 1 and 2). Therefore, the higher flow value for O₁ may be explained as the result of a higher amount of admixture available to disperse C₂S and C₃S phases.

Figure 3 Source: The authors.

Also in Fig. 3, admixture S shows better performance with cement O₁ and B, while compatibility with the other cement types is limited. This is also in agreement with the fact that the other cement types showed a higher [Al₂O₃]/[SO₃] (however, although for blended cement an estimation of C₃A content from Bogue equation is not possible, a qualitative analysis is still viable).

Therefore, the most probable reason for such compatibility is chemical composition. Moreover, some cement grinding aid admixtures are known to have an important effect on the compatibility with SPs, but this hypothesis could not be corroborated in this study as only commercial cements were analyzed and this information is not provided by manufacturers.

On the other hand, the admixture V with cement P shows increased flow values for 30-minute periods in comparison with those determined immediately after mixing, which shows excellent compatibility with cement P.

When mortars made with plasticizer and SP are compared with those made with SP only, an improvement can be appreciated in the flow parameters for the combinations P-MD (with respect to P-D) (Fig. 3). Regarding the combination P-MV, an incompatibility between admixtures is observed, as the combined action shows less efficiency than that obtained for the P-V series; this is attributed to the different origins of their compositions (Table 3).

Fig. 4 presents the results of the slump as a function of the admixture dose. It can be observed that, for mortars with plastic consistency, such as P-D, P-S, O1-S, O2-MD, and P-MV, flow values rise with increasing admixture dose. This greater fluidity of the mortars with 1 % of admixture, in comparison with those that have a lower admixture dose, was not revealed by the results from the flow test (Fig. 3). This is due to the different range of sensitivity of both methods. For the P-S series, a noticeable decline in the values of the slump is observed at 30min, which also rises with increasing admixture dose.

Figure 4 Source: The authors.

Fig. 5 shows relative slump loss 30 minutes after mixing in relation to the admixture dose. In almost all cases, it may be observed that the smallest slump loss is obtained for a dose of admixture between 1 and 1.2 %, which indicates the saturation point for the cement-admixture systems. Negative values in Fig. 5 show that slump after 30 minutes was higher than the initial value. As observed from the fluidity tests, this is the case of P-MV mortars and, to a lesser extent, P-V mortars.

Figure 5 Source: The authors.

Fig. 6 shows the relationship between the values of slump and flow. It may be observed that for most of the stiff mixes, slump reveals great differences in consistency among mortars, while lower differences for slump values are obtained for increasing values of flow.

Figure 6 Source: The authors.

In the case of concrete, the slump method is considered representative of the level of consistency for values up to 60 % of the cone height (18 cm). The same criterion applied to mortars leads to a maximum slump of 9 cm. As it can be noted even for the outlier points (P-S, O₂-S, and B-S 30m), the results obtained can be consistently divided based on this limit. The use of Abrams mini cone is adequate for mortars with plastic consistency, as it shows higher sensitivity for differences within this range of values, and the flow test method is more appropriate for evaluating the consistency of fluid mortars.

Fig. 7 shows density values for hardened mortars determined at saturated surface dry condition. It is clear that the use of admixtures may cause air occlusion and originate variations in the density of mortars. It can also be noticed that, except for mortars with admixture D, in all cases the reduction in density, compared to the reference mortar (P= 2.16; B= 2.16; O₁= 2.15; O₂= 2.16 g/cm³), is less noticeable with increasing admixture dose. This fact can be explained by the net effect of the competing factors of improved compactability and increased air content, caused by the admixture.

Figure 7 Source: The authors.

Fig. 8 shows values for 24 hours water absorption. In the case of admixed mortars with cement P, water absorption increases for all mixes between 35 % and 70 % more than for the reference mortar (P= 4.94 %). Admixed mortar mixes with cement O₂ increase their absorption up to 83 %, compared with the reference mortar (O₂= 4.30 %), independently of the type of admixture used. As in the case of density, this is attributed to the air occlusion caused by the admixtures, which creates macropores and increases water absorption. On the other hand, for mortars with cement types O₁ and B, the adsorptions values of the mortars containing admixtures were similar to or lower than those from the reference mortars (O₁= 5.27 %; B= 6.12 %).

Figure 8 Source: The authors.

Table 6 shows compressive strength and corresponding standard deviations of mortars determined at 7, 28 and 90 days. Nomenclature includes the admixture dose in ‰ with respect to cement weight. Here it may be observed that series O₂-S-6 shows high standard deviation for the three ages (7, 28 and 90 days), whereas mortars P-S-15, P-D-12, P-D-15, P-V-15 and P-MV-8, P-MV-12 and P-MV-15 show high standard deviation at 90 days, and P-MD-5, at 28 days.

Table 6

Source: The authors.

Series O₂-S-6 shows high standard deviations at the three ages. This suggests inadequate compaction of specimens. A reliable comparative analysis is therefore not possible in this case. What may be observed is that the strength of mortars O₁-S is substantially higher than that of the reference mortar, with decreasing difference for increasing age. This outcome is also presented, to a lesser degree, for P-V mortars in spite of the great compatibility shown in the fresh state. Likewise, the relative strength of P-MD series decreases with test age.

For a more reliable analysis of the influence of admixtures on compressive strength, a statistical analysis through paired values (mortar with admixture - reference mortar) was done for the three ages. The p-values were determined for the null hypothesis of equal mean values for each mortar containing admixture and the corresponding reference mortar. Table 7 shows the results obtained for this analysis. It can thus be observed that mortars O₁-S and P-V show a p-value <0.05 for the three ages assessed, with the same condition for P-MD at 7 and 90 days.

Table 7

Source: The authors.

From the statistical analysis (Table 7), it can be seen that mortars O₁-S and P-V show substantially higher compressive strengths than those of reference mortars at the three ages. This may be attributed to improvements in the compaction degree produced by the used admixture and also due to the possible strength accelerator effect of SPs in the cement-admixture system. Conversely, P-D mortars show significantly lower strengths with respect to the reference mortar at the three ages. This must be attributed to air occlusion in the cement-admixture system, which is consistent with the low density and high water absorption of these mortars in the hardened state. Regarding mortars O₂-S, inconsistent p-values were obtained for the different ages; therefore, analysis of these results is not possible.

Finally, for the rest of the mortars, p-values do not allow one to discard the null hypothesis, i.e., no statistically significant differences in compressive strength due to the incorporation of these admixtures could actually be identified.

On the other hand, it is noticeable that in mortars where absorption increases with respect to the reference mortar, no effects on the compressive strength were found. This fact suggests the need to carry out studies that evaluate durability behavior in connection with SP doses.