E ﬀ ect of Alkali Concentration on the Activation of Carbonate-High Illite Clay

: The present study explores the e ﬀ ect of activating solution concentration (4, 6 and 8 M NaOH) on mechanically and thermally pre-treated carbonate-high illite clay (LCR). Pastes were prepared with an alkaline solution / clay (S / B) ratio of 0.55, which were cured at room temperature and relative humidity > 90% in a climatic chamber. At two and 28 days, compressive mechanical strength was determined, and the reaction products were characterised by X-ray Powder Di ﬀ raction analysis (XRPD), Fourier-transform infrared spectroscopy (FTIR) and Scanning Electron Microscopy - Energy Dispersive X-ray spectroscopy (SEM / EDX). Results obtained showed that the presence of reactive calcium in the starting clay induces co-precipitation of a mix of gels: An aluminium-enriched C-S-H gel (C-A-S-H) and a N-A-S-H gel, in which sodium is partially replaced by calcium (N,C)-A-S-H. Pastes prepared with higher (6 or 8 M) activator concentrations exhibit a more compact matrix than the ones prepared with 4 M NaOH. The ﬁndings show that the use of a 6 M NaOH solution yields a binder with two days compressive strength > 20 MPa and 28 days strength of over 30 MPa.

One of the major technological challenges to be faced in alkaline activation is to find precursors that are both widely available and technically suitable. The aluminosilicates most commonly used include burnt kaolinite clay and industrial by-products such as fly ash (FA) and ground granulated blast furnace slag (GGBFS) [11][12][13]. Output of FA and GGBFS is very high at this time and both will continue to be readily available during the twenty-first century. Nonetheless, the environmental challenges facing humanity herald a decline in future production, and the universal geographic availability for industrial scale AAM manufacture is not guaranteed. Pure kaolinite clay (metakaolin) deposits, in turn, are characterised by limited market availability and competition from other industries (paper, fired clay ceramics), which raise the cost of this raw material. Moreover, metakaolin-based AAMs are subject to workability [14] and drying shrinkage [15] problems.
In contrast, clay minerals, outside of very pure kaolinite, constitute an abundant source of aluminosilicates, available worldwide. Studies carried out in recent years on ways to enhance the reactivity of "low purity" clay minerals, with a view to AAM [14,16] and LC3 [7,8] cement manufacture,  Figure 1 shows the particle size distribution for clay MA157 dispersed via ultrasound in a DOLAPIX CE64 1/100 ethanol and dispersion medium, as determined on a SYMPATEC laser sizer with a measuring range of 0.05 to 875 µm.   Figure 1 shows the particle size distribution for clay MA157 dispersed via ultrasound in a DOLAPIX CE64 1/100 ethanol and dispersion medium, as determined on a SYMPATEC laser sizer with a measuring range of 0.05 to 875 µm. MA157 clay pastes were prepared with 4, 6 and 8 M solutions of NaOH at an alkaline solution/clay (S/B) ratio of 0.55, previously found to be the most suitable to ensure good paste workability. The pastes were subsequently poured into prismatic (1 × 1 × 6 cm 3 ) moulds and cured at ambient temperature and relative humidity > 90% in a climate chamber. They were removed from the moulds 20 h later and stored in the climate chamber until the 2 and 28 days test ages, when six samples at each age were tested for compressive strength on an IBERTEST ® test frame (six specimens per test age and material were tested).
The reaction products were analysed by X-ray Powder Diffraction analysis (XRPD) with a Bruker AXS D8 Advance Cu-Kα radiation diffractometer operating at 30 mA and 40 kV across a 2θ range of 5° to 60° and scanning at 0.0197° (2θ) steps. They were likewise characterised Fouriertransform infrared spectroscopy (FTIR), on KBr pellets (1 mg of sample in 200 mg of KBr) with a Nicolet Thermo FTIR spectrometer, recording 64 scans per sample at 4000 to 400 cm −1 . Lastly, the products were analysed along the sputter-coated fracture surfaces of the samples under Scanning MA157 clay pastes were prepared with 4, 6 and 8 M solutions of NaOH at an alkaline solution/clay (S/B) ratio of 0.55, previously found to be the most suitable to ensure good paste workability. The pastes were subsequently poured into prismatic (1 × 1 × 6 cm 3 ) moulds and cured at ambient temperature and relative humidity > 90% in a climate chamber. They were removed from the moulds 20 h later and stored in the climate chamber until the 2 and 28 days test ages, when six samples at each age were tested for compressive strength on an IBERTEST ® test frame (six specimens per test age and material were tested).
The reaction products were analysed by X-ray Powder Diffraction analysis (XRPD) with a Bruker AXS D8 Advance Cu-Kα radiation diffractometer operating at 30 mA and 40 kV across a 2θ range of 5 • to 60 • and scanning at 0.0197 • (2θ) steps. They were likewise characterised Fourier-transform infrared spectroscopy (FTIR), on KBr pellets (1 mg of sample in 200 mg of KBr) with a Nicolet Thermo FTIR spectrometer, recording 64 scans per sample at 4000 to 400 cm −1 . Lastly, the products were analysed along the sputter-coated fracture surfaces of the samples under Scanning Electron Microscopy -Energy Dispersive X-ray spectroscopy (SEM/EDX) a JEOL JSM 5400 SEM equipment fitted with an Oxford Link-ISIS-EDX spectrometer, collecting the images with secondary electrons (SE) at a working distance of 15 mm and a 20 kV accelerating voltage. The EDX (point) analyses were conducted with a 20 µm beam, also at a 20 kV accelerating voltage and a working distance of 15 mm. Acquisition time was 30 s per point analysis. Figure 2 shows the two and 28 days compressive strengths for the MA157 pastes prepared with 4, 6 and 8 M NaOH solutions. In the two days materials, the highest strength, at >25 MPa, was recorded for the pastes prepared with 6 M NaOH. The 28 days pastes exhibited strength of 32 MPa at 4 M NaOH, 33 MPa at 6 M NaOH and 36 MPa at 8 M NaOH. The increase in mechanical strength with time was attributed to the greater degree of reaction, although strength was also observed to depend on reaction product chemical, mineralogical and microstructural composition, as discussed in a later section.

Mechanical Strength
in a later section.
The compressive strength values recorded here were higher than reported by other authors, who worked with thermally treated alkali-activated illite/smectite clay with CaO contents of <5%. Buchwald et al. [25] found compressive strength values of around 13 MPa with 6 M NaOH, (solution/solid ratio = 0.41, curing 20 h at 20 °C in sealed moulds avoiding drying out, after demoulding; seven days at 60 °C in 100% relative humidity (RH), 3 h at 80 °C in 100% RH and drying over 24 h at 40 °C). Essaidi et al. [26], using a potassium silicate solution, observed strength of 25 MPa (initial curing in an opened polystyrene sealed mould, 48 h at 70 °C, subsequently, the mould was closed into oven at 70 °C for 2 h; after demoulding ambient air (25 °C, 40% RH) during 21 days). Seiffarth et al. [19] obtained values of <10 MPa with sodium silicate molar ratio of SiO2/Na2O = 0.31 (initial curing in wet conditions, seven days, 25 °C and 85% RH, followed by curing in dry conditions: Three days, 40 °C, 55% RH).
The higher strength values found in this study were attributed to the high degree of clay reaction, the outcome of the initial mechanical-thermal treatment. As the benefits of that treatment, which induces phyllosilicate amorphism and carbonate decomposition, were described in detail in an earlier paper [21], they are not addressed hereunder. Another factor deemed to have contributed significantly to the good initial mechanical performance of this alkali-activated, ambient temperature-cured clay, was the presence of a fairly high (~30%) proportion of calcium. Like silicon and aluminium, calcium solubility is heavily dependent upon system alkalinity. Given that solution alkalinity played a substantial role in these complex systems (~43% SiO2, 13% Al2O3 and 30% CaO, see Table 1), it was defined as the primary object of the present study.  The compressive strength values recorded here were higher than reported by other authors, who worked with thermally treated alkali-activated illite/smectite clay with CaO contents of <5%. Buchwald et al. [25] found compressive strength values of around 13 MPa with 6 M NaOH, (solution/solid ratio = 0.41, curing 20 h at 20 • C in sealed moulds avoiding drying out, after demoulding; seven days at 60 • C in 100% relative humidity (RH), 3 h at 80 • C in 100% RH and drying over 24 h at 40 • C). Essaidi et al. [26], using a potassium silicate solution, observed strength of 25 MPa (initial curing in an opened polystyrene sealed mould, 48 h at 70 • C, subsequently, the mould was closed into oven at 70 • C for 2 h; after demoulding ambient air (25 • C, 40% RH) during 21 days). Seiffarth et al. [19] obtained values of <10 MPa with sodium silicate molar ratio of SiO 2 /Na 2 O = 0.31 (initial curing in wet conditions, seven days, 25 • C and 85% RH, followed by curing in dry conditions: Three days, 40 • C, 55% RH).
The higher strength values found in this study were attributed to the high degree of clay reaction, the outcome of the initial mechanical-thermal treatment. As the benefits of that treatment, which induces phyllosilicate amorphism and carbonate decomposition, were described in detail in an earlier paper [21], they are not addressed hereunder. Another factor deemed to have contributed significantly to the good initial mechanical performance of this alkali-activated, ambient temperature-cured clay, was the presence of a fairly high (~30%) proportion of calcium. Like silicon and aluminium, calcium solubility is heavily dependent upon system alkalinity. Given that solution alkalinity played a substantial role in these complex systems (~43% SiO 2 , 13% Al 2 O 3 and 30% CaO, see Table 1), it was defined as the primary object of the present study.

Reaction Products
The XRD patterns for clay MA157 and the three (4, 6 and 8 M) pastes are reproduced in Figure 3. The precursor exhibited a significant amorphous component at 2θ values of 20 • to 40 • . The crystalline phases detected in MA157 included quartz, calcite, illite/muscovite, K-feldspar (orthoclase, albite and anorthite) and plagioclase. With the exception of calcite, the intensity of these diffraction lines barely varied after alkaline activation, an indication that the respective phases were inert. The patterns for all the pastes showed a slight shift in the hump located in the precursor at 2θ = 20 • to 40 • to around 2θ = 30 • to 40 • (the hump is a characteristic of amorphous and/or glassy phases) denoting a reaction between the amorphous phases in the clay and the formation of new, low structural order phases, primarily cementitious gels [4]. The lines attributed to the presence of calcite were more intense in the paste activated with 4 M NaOH. Although less calcite was formed at higher alkali concentrations (6 and 8 M), sodium carbonates or calcites such as natron (Na 2 CO 3 ·10H 2 O (PDF # 00-015-0800)) and gaylussite (Na 2 Ca(CO 3 ) 2 ·5H 2 O) (PDF # 00-021-0343)) were detected on their diffractograms.

Reaction Products
The XRD patterns for clay MA157 and the three (4, 6 and 8 M) pastes are reproduced in Figure  3. The precursor exhibited a significant amorphous component at 2θ values of 20° to 40°. The crystalline phases detected in MA157 included quartz, calcite, illite/muscovite, K-feldspar (orthoclase, albite and anorthite) and plagioclase. With the exception of calcite, the intensity of these diffraction lines barely varied after alkaline activation, an indication that the respective phases were inert. The patterns for all the pastes showed a slight shift in the hump located in the precursor at 2θ = 20° to 40° to around 2θ = 30° to 40° (the hump is a characteristic of amorphous and/or glassy phases) denoting a reaction between the amorphous phases in the clay and the formation of new, low structural order phases, primarily cementitious gels [4]. The lines attributed to the presence of calcite were more intense in the paste activated with 4 M NaOH. Although less calcite was formed at higher alkali concentrations (6 and 8 M), sodium carbonates or calcites such as natron (Na2CO3·10H2O (PDF # 00-015-0800)) and gaylussite (Na2Ca(CO3)2·5H2O) (PDF # 00-021-0343)) were detected on their diffractograms. Unlike other alkali-activated binders made with dehydroxylated clay as precursors [25][26][27], the pastes prepared here contained no zeolite crystals. That absence was attributed in part to the fairly low (13.4%) aluminium content in MA157 and/or to ambient temperature curing which would lower Al solubility. A number of authors have reported that zeolite formation as a secondary product in alkaline cements depends on both the reactive aluminium content and the synthesis parameters, Unlike other alkali-activated binders made with dehydroxylated clay as precursors [25][26][27], the pastes prepared here contained no zeolite crystals. That absence was attributed in part to the fairly low (13.4%) aluminium content in MA157 and/or to ambient temperature curing which would lower Al solubility. A number of authors have reported that zeolite formation as a secondary product in alkaline cements depends on both the reactive aluminium content and the synthesis parameters, primarily activator type and concentration [11,12,36], and curing conditions (time, temperature, humidity) [4,37].
The FTIR spectra for clay MA157 and the two and 28 days pastes are reproduced in Figure 4. The spectrum for the starting material exhibited a wide band at approximately 1079 to 1086 cm −1 , associated with the asymmetric stretching vibrations generated by the Si-O-T (T = Si, Al) bonds possibly present in the amorphous component of the material [40]. It also contained vibration bands characteristics of quartz at 1084, 796, 694 and 463 cm −1 [41], a low intensity band at 1438 cm −1 attributed to the presence of calcium carbonate and another at 1635 cm −1 to the bending vibrations generated by the OH groups in water [41].
characteristics of quartz at 1084, 796, 694 and 463 cm −1 [41], a low intensity band at 1438 cm −1 attributed to the presence of calcium carbonate and another at 1635 cm −1 to the bending vibrations generated by the OH groups in water [41].
The Si-O-T (T = Si, Al) stretching band in the pastes was shifted to lower frequencies and its signal grew narrower at longer reaction times. That band appeared at 1010 cm −1 in the two days and at 1014 cm −1 in the 28 days pastes. Further to the literature [4,9,40,41], these variations stem from the initial alkaline-mediated dissolution of the amorphous/reactive component in the clay and the ensuing precipitation of a cementitious gel with a fairly high aluminium content (intermediate reaction product, band at 1010 cm −1 ). The formation of that intermediate product would be explained by the synchronised dissolution of the aluminium and silicon ions in the alkaline medium in the initial stages. As the aluminium content was fairly low, upon its depletion the silicon would continue to react, affording the gel a higher proportion of silicon and shifting the FTIR band to higher values (1014 cm −1 ).
The carbonate band on the spectra for the activated pastes were both shifted to slightly higher or lower wave numbers (1400 to 1500 cm −1 ) and somewhat more intense. A shoulder appearing at 875 cm −1 confirmed the formation of the carbonates detected with XRD in the alkali-activated MA157 clay pastes. More specifically, vibration bands characteristic of calcite were observed at approximately 1432, 876 and 710 cm −1 [42,43] in the 4, 6 and 8 M NaOH-activated pastes, whereas the signal at approximately 1450 cm −1 might be attributed to sodium carbonate [43]. The bands associated with the presence of calcite were most intense on the spectrum for the 28 days sample activated with 4 M NaOH (Figure 4b), a finding consistent with the XPRD results ( Figure 3).  The Si-O-T (T = Si, Al) stretching band in the pastes was shifted to lower frequencies and its signal grew narrower at longer reaction times. That band appeared at 1010 cm −1 in the two days and at 1014 cm −1 in the 28 days pastes. Further to the literature [4,9,40,41], these variations stem from the initial alkaline-mediated dissolution of the amorphous/reactive component in the clay and the ensuing precipitation of a cementitious gel with a fairly high aluminium content (intermediate reaction product, band at 1010 cm −1 ). The formation of that intermediate product would be explained by the synchronised dissolution of the aluminium and silicon ions in the alkaline medium in the initial stages. As the aluminium content was fairly low, upon its depletion the silicon would continue to react, affording the gel a higher proportion of silicon and shifting the FTIR band to higher values (1014 cm −1 ).
The carbonate band on the spectra for the activated pastes were both shifted to slightly higher or lower wave numbers (1400 to 1500 cm −1 ) and somewhat more intense. A shoulder appearing at 875 cm −1 confirmed the formation of the carbonates detected with XRD in the alkali-activated MA157 clay pastes. More specifically, vibration bands characteristic of calcite were observed at approximately 1432, 876 and 710 cm −1 [42,43] in the 4, 6 and 8 M NaOH-activated pastes, whereas the signal at approximately 1450 cm −1 might be attributed to sodium carbonate [43]. The bands associated with the presence of calcite were most intense on the spectrum for the 28 days sample activated with 4 M NaOH (Figure 4b), a finding consistent with the XPRD results ( Figure 3). Figure 5 depicts the two days and Figure 6 the 28 days micrographs for the 4, 6 and 8 M NaOH-activated pastes. A cementitious matrix consisting primarily in Si, Al and Ca was identified in all. The EDX analyses showed that the proportions of these elements, calcium in particular, varied depending on activator concentration and curing time. That finding was associated with the possible formation of a mix of (N,C)-A-S-H/N-(C)-A-S-H gels. The formation of that mix of gels and the high early mechanical strength at ambient temperature were primarily the outcome of the presence of calcium in these dehydroxylated clays [32,44,45]. Activator alkalinity also affected the type and proportion of the C-A-S-H (Al-enriched C-S-H) and/or (N,C)-A-S-H (calcium-bearing N-A-S-H) gels formed, as well as the amount and composition of the secondary reaction products (such as carbonates or AFm phases) (Figure 3).
As a rule the microstructure of the 2 days, 4 M NaOH-activated paste was less uniform than in the other materials (Figure 5a). The phases in the form of Ca-high plates in the Si, Al and Ca-containing cementitious matrix suggested the formation of an Al-tobermorite-like compound. In the 28 days 4 M-activated paste, the matrix was less compact and more granular (Figure 6a). That observation was associated with the significant carbonation in the 4 M NaOH-activated pastes (as revealed by the XRD and FTIR results) that apparently induced decalcification of the C-A-S-H-like cementitious gel. The EDX analyses denoted a lower calcium content in the 28 days than in the two days gels. Similar decalcification was described by Puertas et al. [46] in a study of C-A-S-H gel carbonation in a study of the alkaline activation of blast furnace slag.
At the two higher alkalinities, the microstructure was more uniform and compact. Nonetheless, the excess sodium from the higher concentration of the activating solution (especially in the 8 M NaOH-activated samples) induced calcium and sodium carbonate salt precipitation, as confirmed by the EDX analyses ( Figure 5(c2)). The resulting calcite crystals were primarily located in, and filling, the matrix pores.  As a rule the microstructure of the 2 days, 4 M NaOH-activated paste was less uniform than in the other materials (Figure 5a). The phases in the form of Ca-high plates in the Si, Al and Cacontaining cementitious matrix suggested the formation of an Al-tobermorite-like compound. In the 28 days 4 M-activated paste, the matrix was less compact and more granular (Figure 6a). That observation was associated with the significant carbonation in the 4 M NaOH-activated pastes (as revealed by the XRD and FTIR results) that apparently induced decalcification of the C-A-S-H-like cementitious gel. The EDX analyses denoted a lower calcium content in the 28 days than in the two days gels. Similar decalcification was described by Puertas et al. [46] in a study of C-A-S-H gel carbonation in a study of the alkaline activation of blast furnace slag. The EDX-determined variability in gel composition was graphed on CaO-Al 2 O 3 -SiO 2 and Na 2 O-Al 2 O 3 -SiO 2 ternary diagrams (Figure 7) to study the effect of activator concentration on the type or types of gels formed. The findings showed that in all the cases studied, the alkaline activation of carbonate-high clays with NaOH solutions entailed the co-precipitation of a mix of C-A-S-H-and (N,C)-A-S-H-like gels, which had a higher SiO 2 and a lower Al 2 O 3 content (i.e., a higher SiO 2 /Al 2 O 3 ratio) than other gels described in the literature [32]. Chemically speaking, these gels are more like the ones forming in blast furnace slag [47,48] than in dehydroxylated kaolinite alkaline activation [14,27].
types of gels formed. The findings showed that in all the cases studied, the alkaline activation of carbonate-high clays with NaOH solutions entailed the co-precipitation of a mix of C-A-S-H-and (N,C)-A-S-H-like gels, which had a higher SiO2 and a lower Al2O3 content (i.e., a higher SiO2/Al2O3 ratio) than other gels described in the literature [32]. Chemically speaking, these gels are more like the ones forming in blast furnace slag [47,48] than in dehydroxylated kaolinite alkaline activation [14,27]. When the activator was 4 M NaOH and the time two days, although scattered, calcium clearly prevailed. The points with >60% CaO were associated with the Ca-high, plate-like particles visible in Figure 5a. Given their chemical composition and shape, those phases would appear to be consistent with an aluminium-enriched C-S-H such as Al-tobermorite [32,49,50].
The clusters observed on the micrographs for the two days, 6 and 8 M NaOH-activated pastes ( Figure 7a) were characterised by a non-uniform distribution of CaO, ranging from 25% to 60% and associated with the precipitation of a mix of an Al-enriched C-S-H gel (C-A-S-H) and a calciumbearing N-A-S-H gel (N,C)-A-S-H). With time, higher activator alkalinity (especially visible for 8 M NaOH) induced (N,C)-A-S-H gel evolution toward a C-A-S-H gel. Earlier studies [32] showed that N-A-S-H gel may persist at lower pH (<12), whereas in the presence of calcium, ionic exchange prompts the formation of (N,C)-A-S-H gel, in which calcium may be taken up until it ultimately replaces all the sodium in the cementitious gel's three-dimensional structure. When the activator was 4 M NaOH and the time two days, although scattered, calcium clearly prevailed. The points with >60% CaO were associated with the Ca-high, plate-like particles visible in Figure 5a. Given their chemical composition and shape, those phases would appear to be consistent with an aluminium-enriched C-S-H such as Al-tobermorite [32,49,50].
The clusters observed on the micrographs for the two days, 6 and 8 M NaOH-activated pastes ( Figure 7a) were characterised by a non-uniform distribution of CaO, ranging from 25% to 60% and associated with the precipitation of a mix of an Al-enriched C-S-H gel (C-A-S-H) and a calcium-bearing N-A-S-H gel (N,C)-A-S-H). With time, higher activator alkalinity (especially visible for 8 M NaOH) induced (N,C)-A-S-H gel evolution toward a C-A-S-H gel. Earlier studies [32] showed that N-A-S-H gel may persist at lower pH (<12), whereas in the presence of calcium, ionic exchange prompts the formation of (N,C)-A-S-H gel, in which calcium may be taken up until it ultimately replaces all the sodium in the cementitious gel's three-dimensional structure.
The 28 days, 8 M NaOH-activated MA157 paste exhibited a chemical composition primarily attributable to a C-A-S-H gel, although at 0.24 ± 0.07 the Al 2 O 3 /SiO 2 molar ratio was lower than the 0.3 to 0.5 range normally observed for such gels [32]. That was due the higher SiO 2 than Al 2 O 3 content in the starting clay. In contrast, the matrix for that same paste exhibited a CaO/SiO 2 molar ratio of 0.64 ± 0.17, consistent with the range suggested in the literature (0.6 < CaO/SiO 2 < 1) for C-A-S-H gels [32,48]. The chemical composition of the 28 days, 6 M NaOH-activated pastes also moved toward higher CaO contents in the CaO-Al 2 O 3 -SiO 2 ternary diagrams (Figure 7b) than the two days pastes (Figure 7a).
In contrast, the matrices for the 28 days, 4 M NaOH-activated pastes were characterised by a lower CaO/SiO 2 ratio than the two days materials (Figure 7a,b) and a higher Na 2 O uptake in the gel as the reaction progressed (Figure 7d). Those findings may be attributed to the higher carbonation observed in the latter pastes. As the XRPD and FTIR data showed, the 28 days, 4 M NaOH-activated pastes had a high calcite content, a result of the interaction between atmospheric CO 2 and the calcium ions in the material. Inasmuch as these pastes had no portlandite, the calcium was necessarily sourced from the C-S-H/C-A-S-H gel initially formed [46].
The mechanisms governing alkali-activated materials are presently under study by a number of authors, for many factors are involved, including type of starting materials [51], activator type [47], relative humidity [52] and CO 2 concentration [53]. The reasons underlying the higher carbonation in the 4 M than in the 6 M or 8 M NaOH-activated pastes are fairly complex. Bernal et al. [54] observed the carbonation rate to decline with rising activator concentration in calcium-high alkali-activated systems (such as metakaolin + slag) due to the more intense carbonate salt precipitation generated by the interaction between the highly alkaline solution and the CO 2 in the air. Carbonate salt precipitation would lower the pore volume in the matrix, hindering additional CO 2 ingress in the sample and carbonation progression [54]. As noted earlier, further to the XRD and SEM data, the 6 and 8 M NaOH-activated pastes, rather than developing high calcite contents, precipitated sodium and sodium-calcium carbonates such as natron and para-alumohydrocalcite/hemicarboaluminate.
Briefly, the rise in the degree of reaction with higher activator concentration, proven by the appearance of progressively more compact matrices in the pastes studied ( Figure 5), would explain the higher compressive strength values in the 6 and 8 M than in the 4 M NaOH-activated pastes. The higher two and 28 days strength in the pastes studied here, than by other authors, was associated with a higher degree of reaction over time and the formation of gels with greater silicon and calcium contents.
There are different factors that affect the development of mechanical strength such as: total porosity, size and porous distribution, degree of reaction, type and morphology of the reaction product, structure and composition of cementitious gels, etc. All these aspects are important and depend on the reactivity of the precursor and the alkaline solution used and the curing procedure. In the literature, there are already works that use molecular dynamics simulations to evaluate the fracture toughness of C-S-H [55] techniques. In this work, we have focused on the chemical aspects (instead of the physical characteristics). In addition, the high SiO 2 /Al 2 O 3 ratio and the co-precipitation of C-A-S-H and (N,C)-A-S-H gels observed in all the pastes studied might explain the high strength of our materials (higher than the ones reported for studies using common low-calcium clays) [5,[22][23][24][25].

Conclusions
The most prominent conclusions to be drawn from this study are listed below.
• NaOH (at concentrations of 4, 6 and 8 M) activation of a carbonate-high illite clay contributes to paste hardening at ambient temperature. The presence of reactive calcium in the starting clay induces co-precipitation of a mix of gels: an aluminium-enriched C-S-H gel (C-A-S-H) and a N-A-S-H gel, in which sodium is partially replaced by calcium (N,C)-A-S-H.

•
When the activator is a 4 M NaOH solution, the C-A-S-H gel is more intensely carbonated than with 6 or 8 M activators. • Pastes prepared with higher (6 or 8 M) activator concentrations exhibit a more compact matrix than the ones prepared with 4 M NaOH. Given the comparability of the findings for the 6 and 8 M NaOH-activated pastes and the cost effectiveness of using a lower concentration activator, the 6 M NaOH solution may be deemed to be the most suitable of those studied here. Activating a mechanically and thermally pre-treated carbonate-high illite clay with a 6 M NaOH solution can deliver a binder with two days compressive strength of >20 MPa and 28 days strength of >30 MPa.