128 samples of residual soils formed from metamorphic rock were taken according to macroscopic criteria and were characterized and evaluated as Pozzolans. The parental rock contained around 43.3% quartz of 0.1-0.6 mm in size, garnet (2.3%), tourmaline (1%), epidote (1%), 20.2% subhedral biotite and 10.4% muscovite crystals of 0.1-0.5 mm in size, microcline crystals (3.4%) and albite (12.1%), anhedrals and subhedrals of around 0.1-0.6 mm in size, with sericitation, argilization and other accessory minerals such as apatite (1%) and zircon (1%) (Figs. 1-3).

Figure 1 Source: The Authors.

Figure 2 Source: The Authors.

Figure 3 Source: The Authors.

The physical, chemical and mineralogical characterization of 128 clay samples were carried out using density, specific surface area, X-ray fluorescence (XRF), X-ray diffraction (XRD), and thermogravimetry analysis TGA.

For the X-ray fluorescence studies, fused Pearl in Malvern Panalytical’s Axios FAST was used. Diffraction analysis was carried out in an aCubiX3 range XRD from Malvern Panalytical, with an X’celerator detector and an X-ray tube anode made of copper at 6-70° (° 2theta), a step size of 0.015 (° 2Theta) and a step time of 0.5 sec. For the mineralogical analysis and quantification, a calibration curve using pure minerals of quartz, mica and kaolinite according to the reflections in plane (00l ) was made in a mixture, and is shown in Table 1.

Table 1

Source: The Authors.

The thermogravimetric study was carried out in a TA instruments SDT series Q600, into which 30 mg of the sample was introduced, where it was heated to 900 °C at a rate of 10 °C/min in a nitrogen atmosphere.

Density was measured in Micromeritics’ AccuPyc II 1340 Series helium pycnometer, with three measurements per sample. A Micromeritics Gemini V pore size analyzer was used to measure specific surface areas using the Brunauer-Emmett-Teller (BET) method with nitrogen.

To calcine the clays in this study, the common calcination temperature for kaolinite was used, and so clays were calcined at 750 °C for 30 minutes. Coal at 1.5 % was used, as previous essays have shown that this can reduce red coloration. Afterwards, the clays were milled until 10% was retained in a 325-sieve to simulate the size distribution of cement. The strength activity index (SAI) was used as a measurement of reactivity, and the ASTM C 311 method was employed. Finally, Frattini tests were carried out on the same mixes that were used for SAI, after 7 and then 28 days, following the UNE EN 196-5 standard [7], for the evaluation of the pozzolanic reactivity of the calcined clays.

The chemical composition of clays when crude and calcined is presented in Fig. 4.

Figure 4 Source: The Authors.

The principal oxides were measured, and it was found that the silicon oxide in clays was around 60%, aluminum oxide 23% and iron oxide presented values of close to 10%. In the raw state, the loss of ignition was around 10%, and so these chemical compositions comply with the requirement for pozzolanic materials as stated in ASTC C618. When these values are compared with values found in the clay in its calcined state, there were no large differences in the variables’ median values, except for loss of ignition, which was close to zero after calcination. The values of silicon oxide and aluminum oxide were similar to contents reported in other studies that tested materials as possible clinker replacements such as that of Rashwan et al., who found that metagrabbo was made up of 48% silicon oxide and 15% aluminum [8]. In another study, Avet et al. evaluated the amount of reacted metakaolin in different calcined clay blends [9].

The mineralogical composition of samples is shown in Fig. 5. Like the chemical analysis, mineralogical analysis, where possible, observes the effect of the elimination of volatile compounds on the concentration of minerals. The analysis principally looks at quartz content, but also reveals that minerals such as kaolinite and illite tend to be reduced to zero after calcination. Kaolinite content in the samples was below 40% which is unusual in these kinds of studies because high kaolinite content has been associated with good pozzolanic activity[10-12]. For example, in studies such as that carried out by Castillo et al., clays with low kaolinite content were used for thermal activation and the kaolinite was decanted in order to increase its concentration [13]. However, this process is not relevant to the focus of this study. Other factors can also affect pozzolanic activity, such as the physical effects of specific surface areas and densities [14]. Thus, these variables were measured in the calcined clays and are shown in Fig. 6.

Figure 5 Source: The Authors.

Figure 6 Source: The Authors.

Density was measured to be around 2.7 g/cm³ and specific surface area was around 26 m²/kg, both values are similar to those found by Fernandez et al. in different kinds of clays such as montmorillonite, illite, and kaolinite [12].

SAI values are show in Fig. 7.

Figure 7 Source: The Authors.

The strength activity index (SAI) used in the ASTM C 311 method indicates that values of SAI larger than 75% at 7 or 28 days indicate a pozzolanic material. In Fig. 7 it is possible to see that most of the SAI values obtained after 7 and 28 days of curing are greater than 75%, so it can be confirmed that almost all the samples evaluated behave like pozzolans. At 7 days, 75% of the samples exceed the limit, while at 28 days, 85% surpass it. The pozzolanic activity of these clays increases the longer the curing time, which can be clearly seen by comparing the number of samples that have a SAI greater than 100% at 7 and 28 days. Thus, it can be stated that it is possible to obtain materials with very good pozzolanicity (SAI greater than 100%) from clays from residual soils formed from paragneiss with kaolinite contents of below 40% or with a high silicon oxide content associated with a high quartz content.

The Frattini test is a chemical method, the results of which are shown in Fig. 8, and indicate that most of the data obtained at both 7 and 28 days for the clays calcined at 750 °C is located below the saturation curve (pozzolanic region). Due to this, the tests show pozzolanic reactivity to be positive.

Figure 8 Source: The Authors.

Additionally, the uncalcined clays (room temperature), both at 7 and 28 days, are above the saturation curve, and so the result is negative. This allows us to conclude that these clays develop pozzolanic properties only when calcined.

The test obtained two vectors which are shown in Fig. 8, and each implicitly retains a relationship with the calcination temperature. The vector obtained under natural conditions, that is, at room temperature, is above the Frattini saturation curve, from which it can be concluded that raw gneissic clay cannot be categorized as pozzolan under these conditions. However, when the other vector is studied, it can be seen that the consumption of [CaO] and [OH^-] tends to increase considerably, causing the vector to appear below the saturation curve, thus the vectors resulting from 750 °C conditions show that thermally activated gneissic clay can be determined to be pozzolan.

Hollanders et al. evaluated the pozzolanic activity of pure clays of kaolinitic, illite, and smectite and found that kaolinite clays presented the highest values for pozzolanic activity at a lower temperature than the two other types of clay evaluated and that are globally well known [15]. It is, however, interesting to note that the clays with poor kaolinite content or that were found to be mixed with other materials presented good SAI values. This was also evidenced by Yanguatin et al., who evaluated the pozzolanic activity of excavated waste clays, and found satisfactory results that indicated that the best pozzolanic activity was found in materials with high kaolinite content [16]. Uchima et al. obtained good results that showed pozzolanic activity in biomass and kaolinitic clay [17].

The present study could suggest that illite and kaolinite could have a synergic effect because a study carried out by Msinjili et al. that compared illite and kaolinite clays as supplementary cementitious materials concluded that all of these kinds of materials have pozzolanic activity but differ in the temperature at which they activate [18]. It is important to note that in this study, it was possible that the illite had a disordered structure-a normal characteristic of this mineral [19]-and so activated with kaolinite in the same temperature range.

On the other hand, when the correlations between SAI and the characterization variables were evaluated, it was found that the only characterization that had a relationship higher than 75% was kaolinite content. It is important to note, however, that although many researchers have studied the relationship between SAI and kaolinite[16,3,6,17], it was also relevant to the present study, although these kinds of materials showed low values of kaolinite and had a high quartz content, which could suggest that these kinds of materials’ pozzolanic activity are related to the clay content more than the kaolinite content, which has been suggested by Msinjili et al., referring to the pozzolanic activity of illite clays [18].