3.2. Mechanical and Physical Properties of the PC
Figure 9a shows the results obtained from the MIP tests in the enrichment phase (positive intrusion). The curves or isotherms of this graph show that their form is similar in all cases, their path being as follows: starting with slight increases from a common origin (the theoretical diameter of maximum pores φ
p max = 500,000 nm, macropores [
62,
63]), they continue to an abrupt inflexion point, located in the range of critical pore diameter 60 ≤ φ
pc ≤ 140 nm. Finally, they continue to an accelerated increase in the intrusion volume, which causes a steep rise until the points of maximum intrusion value are reached, minimum pore diameter φ
p min in the mesopore zone.
The PC sample—NCA 100 C—establishes the most representative pore samples as 7 ≤ φ
p ≤ 110 nm, with its inflexion point located in Vp = 0.01 mL/g. This means that a lower amount of Hg is needed to fill the large diameter (100 ≤ φ
p ≤ 400,000 nm) pores. In contrast, the NCA 100 SF sample shows fewer pores, which are significant in the size range of 6 ≤ φ
p ≤ 90 nm; the reason being the insertion of SF, as this is made up of small particles and may be used as a ‘filler’ for pores in the microstructure of concretes [
64]. The RCA 50 C version establishes the highest number of pores of all the samples analyzed—set at 30 ≤ φ
p ≤ 200 nm—due to the inclusion of the RA; as this material is classified as low density and high porosity [
1] the Hg can pass easily through its porous structure. Finally, it can be seen that adding SF to a sample of PC with RCA (RCA 50 SF) causes a reduction in the number of pores of 6 ≤ φ
p ≤ 80 nm; the cause is similar to that of the NCA 100 SF, although on this occasion with less importance (porous RCA).
To sum up, the order of the increase in porosity of PC with NCA was as follows: first NCA 100 SF and then NCA 100 C; regarding the PC with RA, first RCA 50 SF and finally RCA 50 C. The second group is affected by the RA with high original porosity. In both groups, the inclusion of SF (filling effect) is connected to the reduction of porosity.
The extrusion curves in
Figure 9b help to interpret the morphology of the “trapped porosity”, identified when the intrusion–extrusion gaps were compared (a hysteresis phenomenon or the ability of the material to keep to the initial intrusion curve line as the Hg pressure returns to its initial state in the test). The difference between both curves established the quantity of Hg retained in the sample, a consequence of the number of bottleneck pores in the system [
65,
66]. The extrusion curves shown in the test samples have similar lines and the same relative positioning on the intrusion curves.
Table 6 shows the values of trapped porosity (P
a) obtained after determining the difference in volumes between both curves. It can be seen that the samples with RCA have twice as many bottleneck pores as the NCA samples. This is understandable if we consider that the RCA has open pores which are partially blocked by the new CP cement paste and that the use of C or SF is not significant in either sample group.
Table 6 and
Table 7 show the values of the porous network indicators determined according to the previously established procedures [
51,
67,
68]: maximum pore radius (r
max), minimum radius (r
min), critical radius (r
c), average radius (r
m), total pore area (A
tp), intruded and extruded volume (V
i y V
e) and P
a. Vi = Volume intruded, Ve = Volume extruded.
Regarding the r
max in the intrusion stage, the average radius established in the study samples (244,802 nm) highlights the ability of the PC to cause large pores—up to 40 times bigger than those noted in the usual recycled concretes [
51,
69]. However, in the extrusion stage, the average (6695 nm) is similar to the usual recycled concretes in their intrusion stage. In terms of the varieties studied, the order of radii variation between them is not considered important (less than 1% in both stages of the test). In both stages of the test, and all the varieties studied, r
min is considered a constant (equal to 3.4 nm). This raises two points for consideration: the MIP technique may be able to establish the network of pores in the micropore zone, and this minimum value probably has more in common with the limits or parameters of experimental techniques than with the varieties of this research.
The average values attained for the varieties studied in the case of rc were 110.5 and 235.4 nm (intrusion and extrusion respectively). However, unlike the previous indicators (in the intrusion stage) the NCA 100 SF sample established an rc some 50% less than that of the NCA 100 C, which shows that the use of SF is able to shift the rc toward lower rc values (closing the porous network). With regard to the RCA 50 C and RCA 50 SF versions (intrusion stage), the rc variation between them is not important (a percentage of 50% SF is not important, the change of effect in rc requires a high percentage of SF to be noticeable). Finally, the effect of the type of aggregate establishes that the use of RCA regarding NCA (intrusion stage) causes the former to establish a greater rc than the latter (about 60% more), a result of the high porosity of the RCA itself. Concerning the extrusion stage, the previous trends have the same application, although the values established are greater than in the intrusion stage.
Regarding the indicators of rm, Vi,e, and Atp, the values shown in both stages of the study have, in percentage terms, similar interpretations to those established for rc; it can be seen that these are the most appropriate and sensitive for establishing the correlation between the study varieties and the porous network produced in those PCs.
3.3. Porosity, Anhydrous Material and Hydration Products in Cement Paste by mMeans of IA-SEM
After performing an IA-SEM on the study samples, it was possible to obtain the percentages of porosity, anhydrous material and hydration products (PH) (see
Figure 10). Three specifically different zones of interest were also evaluated for the RCA 50 C and RCA 50 SF samples: (a) ITZ between the new aggregate and the new paste, (b) ITZ between the old aggregate and the old paste and (c) ITZ between the old paste and the new paste.
As for porosity, and by comparing the samples that analyze the ITZ, it can be seen that this is higher in the samples containing 100% CP (possible paralysis of the hydration process); in the case of SF it is prolonged with age reducing the porosity, or is favored by the filling effect of pores which it causes. Among the zones of interest, b) (ITZ between old aggregate and old paste) establishes the greatest porosity (the variety linked to the less dense and more porous RCA concrete). If the use of RCA in the PC is eliminated (NCA 100 C and NCA 100 SF), it can be seen that there are minor variations of porosity between them, with the former being the best (in this case the filling effect of pores by SF is not favored, as the use of NCA does not include them). With regard to the anhydrous material and the PH, the use of SF establishes low percentages of anhydrous material and high ones of PH in the samples that contain RCA (the SF reacts with the CH present in the cement hydration to form more SCH). The sample RCA 50 SF is in the zone of interest b), the best of all (possibly due to the prolonged hydration over time of the old concrete from which the RCA originated). For the samples that use NCA, the use of C in the PC increases the formation of PH (the use of SF does not favor the creation of more PH, due to the possible effects of a blockage in hydration caused by a lack of pores or spaces wherein crystals may form).
3.4. Evaluation of Porosity at Different ITZ Thicknesses by Means of IA of SEM
Figure 11 shows the number of pores (unitary accumulated porosity, µm) determined in the study samples. It also considers the different theoretically possible thicknesses of the aggregate ITZ (ITZ of 10, 20, 30, 40 and 50 μm), which allows the transition from the porous network of the PC to be verified. It can be seen that there is a direct connection in most of the samples between the increase in the percentage of porosity and the increase in the theoretical ITZ thickness. The theoretical thickness is 10 μm, which establishes the best percentage of porosity compared with the other thicknesses evaluated; this confirms that the percentage of porosity of the ITZ decreases as its theoretical thickness increases. Another way to understand the curves is that the ITZ has finished when the curves of the samples tend toward a horizontal asymptote, with a constant—or non-increasing—percentage of porosity equal to that of the paste zone. Significantly, the three evaluated samples which had the lowest porosity in the ITZ are those with 50% of RCA and SF; this confirms that the use of SF leads to the filling effect in the ITZ porosity, making it denser and stronger. The RCA 50 C (a) sample shows the highest percentage of unitary porosity when the theoretical thickness is set at 10 μm (0.0216%/µm); at 50 μm reaches 0.032%/µm (the SEM image analyzed showed that this sample contained considerable porosity). This may have been due to the more porous nature of the RCA. Similarly, the effects of using SF were also observed in the NCA 100 SF, which had lower porosity than the NCA 100 C sample.
Therefore, it can be stated that the use of RCA in making a PC will cause increases in the ITZ porosity. However, the addition of SF may compensate for this, and achieve a denser cement matrix.
Figure 12 shows the percentage of unitary accumulated anhydrous material at different theoretical ITZ thicknesses in the PC samples studied. The trend shows a direct, rising, connection between the increase of this parameter and the increase in the theoretical ITZ thickness. It should be noted that the samples with SF have lower percentages of anhydrous material. The SF can react with the cement hydration products and form new ones, which show low percentages of non-reacting material. The RCA 50 C sample (b), has the most anhydrous material: at 10 μm an important concentration of 0.0062%/µm can be seen, reaching 0.0106%/µm when the theoretical thickness is 50 μm. Additionally, comparing NCA 100 C and NCA 100 SF, the former is shown to contain a far greater percentage of anhydrous material at all theoretical ITZ thicknesses than the latter. Therefore, to sum up, it can be said that the use of SF contributes to the formation of more hydration products in the cement paste; this material can fill pores, thereby leading to lower percentages of anhydrous material.
The PH (the result of subtracting the porosity and the anhydrous material from the paste) that can be detected in the ITZ studied usually has a low value, as porosity dominates in an ITZ.
Figure 13 shows the percentage of unitary pH at different theoretical ITZ thicknesses (%/µm). The RCA 50 C samples (a and b) have the highest percentage of PH, which corroborates previous paragraphs. This sample showed low percentages of porosity and anhydrous material. The NCA 100 C has a high amount of PH and shows a significant increase at theoretical ITZ thicknesses of between 40 and 50 μm.
The samples with SF (NCA 100 SF, RCA 50 SF (a and b)) have the lowest amounts of pH and correlate the results of the mechanical properties (they are the least resistant). Their mechanical properties were low compared with the samples made from Portland.