4.1 Isothermal calorimetry

Isothermal calorimetry tests were carried out in pastes for both type of cements; the results are shown in Table 4. To evaluate the effect on the heat generated during the hydration reactions, samples were prepared with 0.50% SP and 1.0 and 5.0% NS. The pastes were had a water/cement (w/c) ratio of 0.40 and 0.50 (see Table 4).

Figure 2 shows that the induction period ended after 2.0 hours for the CPO 40, while for CSA, finished before. It is at this time, when the setting period starts and the C-S-H layer breaks, with the continuation of the hydration process. Results shows that lower w/c ratios slightly accelerate the hydration reaction; this effect is more noticeable for CPO 40.

Figure 2 Isothermal calorimetry curves for both cements at different w/c ratios. 

As shown in Figure 2, when adding 5% SP for both type of cements, a shift of the exothermic curve was observed, indicating an extended sleeping period, causing the curve became more pronounced with a lower crest, having a maximum heat release at 15 h for CPO 40 with 0.5% SP, compared to 7 h without SP, i.e. the addition of 0.5% SP increases the starting time for the development of RC two times. A similar behavior is shown for CSA; where the effect of SP appears to be slightly higher, since cement without SP registers its exothermic maximum at 4.7 h, whereas with the addition of 0.5% SP, the exothermic maximum is observed at 13h, i.e., almost three times slower than the plain cement.

On the other hand, at the same w/c ratio, it is observed that the CSA presents a different heat evolution curve compared with CPO 40, where the main peak due to the exothermic reaction of C3S was at 6.5 h for CPO 40, and 4.5 h for CSA. Additionally, the exothermic peak reached a maximum heat flux for CSA of 0.07 W/g, while for the CPO 40 of 0.06 W/g. This indicates that the hydration reaction of cement CSA starts faster, and the initial RC should be greater in CSA; however, the RC values were higher at 24 hours for CPO 40 (18 MPa) compared to CSA (13.1MPa). This could be due to the dilution effect of CSA, which according to the results of the chemical composition, contains a higher content of limestone, compared with CPO 40. The results also show a shorter setting period of CSA with respect to CPO 40. After the pure cements analysis, the effect of NS addition was studied.

These results confirm those reported by Puertas (^{Puertas et. al., 2001} and ^{Puertas et. al., 2005}), indicating that regardless of the nature of the SP added, the hydration reactions of calcium silicates were delayed, extending the setting period. This effect can be explained due to an adsorption of the polymer in the cement grains, forming a layer barrier around grains, and preventing the transport of water molecules to the cement particles, in addition to the formation of complexes between Ca²⁺ ions formed in the initial hydration and the anions of the polymers that affect nucleation and precipitation of Ca(OH)₂. However, the RC of mortars with SP indicated that at 24 h, the RC of pure CPO 40 was lower than those containing 0.5% SP, i.e., 18 MPa and 23.5 MPa respectively.

The isothermal calorimetry results indicated a significant extension of the initial setting time, however, this result does not necessarily imply any relation with the RC expected by these cements, since the setting process and the development of strength process are independent. It is noteworthy that that for CS, the effect was different, and the addition of SP decreased the RC at 24 hours, from 13.1 to 8.2 MPa.

Figure 3 also showed that by adding the nanoparticles to the same w/c ratio the evolution heat curves shifted to shorter times, suggesting that the processes of setting and hydration reactions were accelerated by the addition of the silica nanoparticles. This trend was observed for both types of cements.

Figure 3 Isothermal calorimetry curves for the cements with NS and SP; w/c = 0.40. 

This behavior confirms that reported by (^{Qing et. al., 2007}), they observed that with the increase in the percentage of NS, the workability decreased, and the hydration process was accelerated compared to other pozzolanic materials such as silica fume. Other authors (^{Björnström et. al., 2004}, ^{Morteza et. al., 2014} and ^{Li et. al., 2004}) found that nanosilica also accelerates the process of hydration and C-S-H gel formation due to its particle size; moreover, with the increase of the addition percentage of NS, the heat of hydration of the mixture incremented.

Hence, it is possible also to observe that at higher NS load, the exothermic maximum is reached at lower time. According to curves for CPO 40, form 0 to 1.0% NS, the time at which the maximum was reached moved from 14.5 h to 11 h, and by adding 5% NS time fell from 11 to 7.5 h. For the CSA, the behavior was similar, i.e., with no additions the maximum was observed at 13.3 h, with 1% NS at 8.8 h and for 5% NS at 6 h.

This experimentally observed behavior had a parabolic shape and can be explained through the following equation:

where: f (x) = maximum time of heat release, x = % of nanoparticles; a, b and c are coefficients of each of the systems.

Equation for the system with CPO 40:

Equation for the system with CSA:

From Figure 4 it can be concluded that, while adding nanoparticles to the cementitious matrix accelerates hydration reactions and therefore increase RC, there is a point where the increase in the nanoparticles content will not have an improvement effect in the cement. In Figure 4 it can also be noted that higher that 3.5% contents of NS, the time where the exothermic maximum is reached, increase again. This would confirm that the nanoparticles presence, accelerated the development of the initial RC, but only up to this content.

Figure 4 Maximum heat release time vs % NS. 

4.2. Compressive strength results.

Samples with 0.5%SP and 0.3%, 1% and 5% NS were used to estimate the compressive strength. The results of 24 h, 3, 7, 28, 90 and 150 days are presented in Figure 5. It is worth to notice that, for this first set of results, the water added to the mix was corrected considering the water supplied by NS (which was provided in suspension), in order to keep the ratio w/c of 0.485 constant.

Figure 5 Compressive strengths results of the CPO 40 and CSA, without adjusting the amount of water. 

From the results, it can be observed that for the CPO 40, the best RC was obtained for the mixture with SP. The RC results also showed that all the systems with NS tested at 24h exceed the results obtained by reference (neat CPO 40), with an increase up to 43% (25.8 MPa vs 18 MPa), for sample with 5% NS. After 3 days of hydration, these percentages decreased, observing an increase of 15% of for mixtures with 5% NS (40 MPa vs 34.9 MPa). At 7 days, the improvement was of 11% for CPO + SP while, with the incorporation of 5%NS the RC was 8% higher than that obtained by the reference. At 28 days, the samples with SP developed a RC 17% RC higher that the reference (54.9 vs. 47.1 MPa), while the increase for samples with NS was around 11% compared to the reference (52.1 and 52.3 MPa for samples with 0.3% and 5% NS vs 47.1 MPa for the reference). After 150 days of curing, all samples with NS reaching almost the same value as obtained by reference and only the sample with SP developed a 18% greater 34.9 MPa than the neat CPO.

The effect shown for CSA, was different as that observed for CPO 40, since the addition of SP, did not favor the development of RC, but rather inhibited the hydration processes. Therefore, all the RC results for the mixtures containing 0.5% SP, showed lower RC with respect to the reference, as 63%, 54%, 52%, 62% and 55% at 24 h, 3, 7, 28, 90 and 150 days respectively due to the decrease in the aforementioned reactivity, since SP are designed for ordinary Portland cement and the sulfate phases contained in the CSA may negatively affect the functionality of the SP.

This behavior was observed for CSA cement in all samples, except for the mixture with 5% NS, where it possible to observe an increase in the RC of about 47% at 24h, 19% at 3 days, 16% at 7 days, and 14% after 28 days. According to (^{Ma et. Al., 2014}) when adding SP, the RC for belite-sulfoaluminate cementsat 24 h decreased, however, at 28 days RC were equivalent to the reference and even higher when dosed in percentage of 0.075%. This was to the delay in the ettringite formation after 1 day of hydration, especially at high concentrations, resulting in a poor RC. This could explain that the 0.5% CSA load used in this research could be relatively high for the nature of this cement.

As can be seen in Figure 6, at similar values of fluidity, by adjusting the mixing water for the same fluidity; RC results for CPO 40 with NS change with respect to those previously described, although the best results are observed in presence of SP, the results with the NS are very similar, and even superior at 24 h for samples with 0.3% and 1% NS.

Figure 6 Compressive strength results for the CPO 40 for a constant flow. 

These results are relevant because it raises a question relative with the Mexican standard specifications of consumption of water to use when it comes to assessing ordinary cements. As is known, the standard indicates that for an ordinary Portland cement, the w/c ratio recommended to evaluate RC is 0.485. The results obtained in this research suggest that for systems with SP systems, or nanoparticles, it is best to seek the fluidity and water required to adjust to this fluidity and do not use a fixed w/c ratio.

4.3. Chemical resistance to sulfate attack

The sulfate attack tests were carried out according to the procedure specified in ^ASTM-C-1012. Figures 7 and 8 show the results for each type of cement measured during 550 days, include the standard sets a minimum duration of 12 months. From the results, it is possible to observe that neat CPO 40 and CSA do not have good resistance to chemical attack by sulfates, since the change in length exceeds the allowable expansion value of NMX-C-414-ONNCE-2010 standard of 0.1% at 125 days for CPO 40 and at 50 days for CSA.

Figure 7 Sulfate attack chemical resistance of CPO 40 after 550 days of exposition. 

Figure 8 Sulfate attack chemical resistance of CSA after 550 days of exposition. 

The addition of nanoparticles in percentages of 5%, promotes the chemical resistance to sulfates, for both types of cement. In the case of CPO 40, the benefit of adding nanoparticles is so favorable that the cement changes from a cement without resistance to sulfate attack ("RS" in the nomenclature used in the Mexican standard ^{NMX-C-414-2010-ONNCCE}) to be a resistant to sulfates at least up to 550 days, with contents of 0.5% of SP and 5% NS. The addition of less than 5%NS did not improve the resistance.

For the CSA, the addition of nanoparticles showed a lower resistance to chemical attack by sulfates compared to CPO 40, since CSA without additions exceeded the limit established by ^{NMX-C-414-2010-ONNCCE} standard after only two months of exposure. However, by adding SP and NS, cement remained with characteristics of resistance to chemical attack up to 275 days (9 months). This change was 94.8% for 180 days, which was the maximum age at which the reference specimen cement volume remained stable with respect to neat CSA.

According with the results observed for CPO 40 the addition of both SP and NS, in low dosages for the latter was not beneficial in terms of resistance to chemical attack, however, with all the additions, the CSA resistance was improved compared with pure CSA.

As mentioned above, this behavior can be explained by a densification of the matrix, which increased with the content of NS, and thus a reduction in porosity and sulfate attack tends to be slower than in cement samples without NS. The NS would be acting as nucleation sites for the generation of gel C-S-H and to promote reduction in porosity. This behavior was confirmed by other investigators (^{Quercia et. al., 2012} and ^{Li et. al., 2004}), they reported that the addition of NA increased RC and decreases permeability through the pozzolanic reaction, resulting in an increased denser structure with a further formation of C-S-H. According to (Khater et. Al., 2006), with the use of modified montmorillonites, the permeability was 100 times reduced compared to neat cement.

Thus, durability of the cements could be modified and improved with the presence of NS, due to the filler effect of these nanoparticles, densifying the matrix and decreasing the porosity and permeability, this in turn increases the RC, as for microsilica.