The process of calcination of raw clay in an air atmosphere rotary furnace at 800 °C for 30 min was used obtain metakaolin (MK) and calcined illite clay (IC) in the laboratory setting. The specific surface area of ground MK was 15.86 m²/g and of IC – 21.42 m²/g. The average diameter of MK particles was 743.1 nm and of IC – 709.5 nm. According to XRD analysis, MK used is largely amorphous, with minor crystalline phases such as quartz (SiO₂), mica KAl₃Si₃O₁₁ and muscovite (K_0.82Na_0.18)(Fe_0.03Al_1.97) (Figure 1a). Calcined IC had more crystalline structure and, in addition to quartz (SiO₂) and illite K(AlFe)₂AlSiO₃O₁₀(OH)₂·H₂O, it also contained microcline (KAlSiO₈) (Figure 1b). Chemical composition of both calcined clays is given in Table 1.

The results of FTIR spectroscopy of calcined clays show the characteristic bands of MK and IC (Figure 2a, b). The mode at 913 cm⁻¹ corresponds to Al–OH band and Si–O–Al^VI band is characterized with frequency at 540 cm⁻¹ (Figure 2a) (14,15). Band at 800 cm⁻¹ could be related to the change from octahedral coordination of Al³⁺ in kaolinite to tetrahedral coordination in metakaolin (14,15). Wavenumbers at 1031, 1008 and 469 cm⁻¹ (Figure 2a, b) correspond to Si–O band, and also quartz interference could occur at 1085 cm⁻¹ (Figure 2a). The shoulder at 1100 and 1200 cm⁻¹ (Figure 2a, b) is attributed to amorphous SiO₂ (14). Band at 685 cm⁻¹ could be attributed to the spinel phase (MgAl₂O₄) which occurs during the calcination of the selected clays (16). Calcined IC has infrared vibration modes at 1620 cm⁻¹ (Figure 2b) associated with bending vibrations of H–O bond. The doublet at 777–795 cm⁻¹ is characteristic of the quartz (17,18), which was not detected for MK.

The ASRW is final waste product at the aluminium scrap recycling factories. Chemical composition of ASRW is given in Table 1. The elemental analysis by ICP-OES, AAS and PTA of the ASRW showed the following content: aluminium – 34.4%, silicon – 4.4%, magnesium – 2.44%, calcium – 1.3%, sodium – 1.7%, potassium – 2.3%, sulphur – 0.1%, chlorine – 4.2%, iron – 3.6%, copper – 1.0%, lead – 0.1%, and zinc – 0.6%. According to the XRD data, the ASRW contained metallic aluminium (Al), iron sulphite (FeSO₃), aluminium nitride (AlN), corundum (Al₂O₃), aluminium iron oxide (FeAlO₃), magnesium dialuminium (MgAl₂O₄), quartz (SiO₂), aluminium chloride (AlCl₃) and aluminium hydroxide (Al(OH)₃) (Figure 3).

The LSG was obtained from fluorescent lamp recycling process, which includes lamp classification, glass separation, cleaning from harmful components and glass grinding. The chemical composition of LSG is given in Table 1. The results of grading analysis showed that the additionally ground material contains a wide range of particle sizes: from 8 μm to 30 μm with the average grain size of 26.3 μm. The specific surface area of LSG was 1.126 m²/g.

The AAM were prepared by using alkali activating solution with silica modulus M_s 1.67 (SiO₂ to Na₂O ratio) defined with previous research performed on alkali activation solution properties on AAM structure (19). Alkali activators were prepared by using commercially available sodium silicate solution from “Vincents Polyline” Ltd. characterized by the silica modulus M_s = 3.22. To obtain alkali activation solution with the required chemical composition, the modification with an addition of sodium hydroxide was done. Commercially available sodium hydroxide flakes from “Tianye Chemicals” (China) with 99% purity were used.

The structures of MK-A1-G1 and MK-A1-G0 were „x-ray amorphous”. Both of XRD patterns had a halo in the 2 θ = 15 – 40° region, which is characteristic for amorphous and/or vitreous compounds (Figure 4). The broad peak observed from 20 to 30° 2 θ in MK (Figure 1a) has become wider to 40° 2 θ after the reaction with activator M_s 1.67 and shifted to higher angle (Figure 4a), as is known to occur during geopolymer formation from MK. This shift demonstrates the formation of new amorphous phases, usually described as geopolymeric gels (21,22). The presence of the halo can be explained by formation of alkali aluminosilicate gel as the primary reaction product in activation of MK as well as presence of LSG in the composition of MK-A1-G1, where the halo is more pronounced. Generally speaking, the only crystalline phase, which could be identified clearly by XRD analysis, was quartz. Significant difference in the mineralogical composition between AAM made with or without LSG was not determined. The structure of AAM made with IC was not completely X-ray amorphous comparing with AAM made of MK (Figure 4). The halo for IC-A1-G0 and IC-A1-G1 was not strongly marked in the diffractograms in comparison with MK which can be explained by the crystalline mineralogical composition of IC calcined at temperature 800°C.

The infrared spectra (FTIR) for all the samples studied are given in Figure 5. All AAM specimens with and without glass waste have mode at 446 cm⁻¹ which corresponds to Si–O–Si or Al–O–Si bands (22,23). Mode at 540 cm⁻¹ is typical only for MK-A1-G0 and IC-A1-G0 and it is associated with Si–O–Al bands, where Al is present in octahedral coordination. In the FTIR spectrums doublet at 777 – 795 cm⁻¹ (Figure 5b) is attributed to Si–O–Si bands recognized as quartz (17). The small modes at 688 cm⁻¹ could be attributed to the symmetrical stretching vibration of Si–O–Si or Al–O–Si bonds (Figure 5a) (23). The bending vibration of the same bonds was identified at 457 cm⁻¹ (Figure 5b) (18).

The vibrational band for calcined MK and IC with a maximum at 1085 cm⁻¹ (Figure 2) shifted toward lower frequencies for AAM made with MK and IC, whereas that at 800 cm⁻¹ it disappeared after the alkaline activation. Both modes indicate that the original aluminosilicate structure of the MK and IC was significantly depolymerized and dissolved by the alkaline solutions. For MK-A1-G1 the maximum of the vibrational band appeared at 1020 cm⁻¹ (Figure 5a), whereas for IC-A1-G1 the maximum located at 1016 cm⁻¹ (Figure 5b). These bands have broad shape due to the overlapping of various types of bonds, which include Si–O, Si–O–Si, Si–O–Al, and probably Si–O–Al–(Na), resulting from the fine intermixing of phases (24). The shoulder at 1100-1200 cm⁻¹ is assigned to amorphous SiO₂ which indicates that polymerization was not fully completed or amorphous SiO₂ remained unreacted for AAM made with glass (MK-A1-G1 and IC-A1-G1). It seems that presence of the LSG prevents complete reaction of raw materials in alkali media and formation of Al–O and Si–O bands. A lower degree of polymerization of silicates within the remaining gel was observed for samples with LSG. It can be explained by vitreous structure of specimens containing LSG which limits reaction between activator and calcined clay minerals. Therefore the NaOH is available for crystallization on the pore surface and it turns to carbonates in the open air atmosphere.

The most important finding in this research was related to modes at interval from 1436 cm⁻¹ which correspond to the stretching vibration of O–C–O bonds in the carbonate group (CO₃^2-) (17, 25). Mode at frequency 1436 cm⁻¹ is typical for MK-A1-G1 and for IC-A1-G1 (Figure 5). The mode at 881 cm⁻¹ is associated with sodium bicarbonate deformation band (18). Obviously, the presence of the LSG in the mixture contributes to the formation of carbonate species in AAM. The identification of carbonate phases in a number of studied samples, especially in all samples containing LSG, is particularly relevant to the interpretation of the results. Such strongly exhibited bands at modes 1436 cm⁻¹ and 881 cm⁻¹ were recognized only in the specimens containing higher amount of LSG (Figure 5). The sodium carbonate or sodium bicarbonate was not found in XRD patterns which could be explained by non-crystallized or semi-crystallized form of carbonates concentrated within pores in the form of pore solution or by dissolving and leaching of carbonate species.

FTIR analysis was performed for all specimens after 20 days of immersion in deionized water (MK-A1-G1-W and IC-A1-G1-W) with the purpose to determinate changes regarding modes at 881 cm⁻¹, 1436 cm⁻¹ which correspond to the stretching vibration of O–C–O bonds in the carbonate group (CO₃^2-). The intensity of modes corresponding to O–C–O bonds decreased or disappeared due to leaching of alkali ions and dissolving of carbonates or bicarbonates in the water for the all types of AAM containing LSG as raw material (Figure 5, curves IC-A1-G1-W and MK-A1-G1-W). It was associated with solubility of carbonates and bicarbonates in water media. The interpretation of infrared spectra (FTIR) for all the samples studied is summarized in Table 3.

There are four main parameters characterizing the leaching rate of AAM: porosity of AAM (i), amount of ASRW (ii) and LSG (iii) in the composition as well as to the type of clay used (iv). The typical microstructure of AAM is given in Figure 6. The highest amount of hydroxide ions leached after 20 days (0.0371 mol/l/g_AAM) was for IC-A0.1-G1 with open porosity 16% and total porosity 78% (Table 4 and Table 5). The increase of ASRW (IC-A1-G1) reduced leaching rate to 0.0268 mol/l/g_AAM while porosity of AAM increased (24% open porosity and 87% total porosity) (Figure 7a). Similar tendency was observed for AAM based on IC without LSG - the leaching rate for IC-A0.1-G0 (open porosity 23%, total porosity 69%) was 0.0172 mol/l/g_AAM and it decreased to 0.0123 mol/l/g_AAM for IC-A1-G0 (open porosity 26% and total porosity 84%). MK-A1-G0 provided similar results with regard to porosity and leaching. The leaching of OH- ions was slightly higher for samples with higher open porosity (29%) – 0.0077 mol/l/g_AAM for MK-A1-G0 and 0.0052 mol/l/g_AAM for MK-A0.1-G0 (26%). The material porosity did not influence the leaching rate for AAM based on MK significantly. Different results were obtained for AAM samples based on MK with LSG (Figure 7b). The total porosity of AAM decreased from 84 to 79% by reduction of ASRW in mixture composition. The open porosity of MK-A1-G1 increased from 16 to 21% and the leaching of OH^- ions was 0.0110 mol/l/g_AAM. Leaching rate for MK-A0.1-G1 was slightly lower - 0.0090 mol/l/g_AAM.

The equipollent relationship between leaching rate and porosity was not found, thus it can be concluded that the leaching rate depends on the composition. It seems that ASRW influences mixture composition and the leaching rate of hydroxide ions; lower amount of ASRW increased leaching rate of alkali ions for AAM made of IC. Most likely it can be explained by geopolymerization process of raw materials where ASRW provides increased amount of Al₂O₃. The fact that leaching rate for AAM based on MK was not strongly influenced by the amount of ASRW in the composition could be explained by several factors: the amount of ASRW added to mixture composition did not have significant impact on material porosity; MK has higher amount of Al₂O₃ compared to IC; and aluminium content of the AAM does not play such an important role in removing alkali ions from the pore solution; total amount of Ca in MK based AAM system is higher compared to AAM made of IC.

The general tendency for all samples analyzed in this research is that leaching rate decreased for AAM made of both clays without LSG compared with samples with LSG.

The leaching rate could be attributed to the SiO₂/Al₂O₃ and Na₂O/Al₂O₃ mass ratio of AAM (Table 5 and Figure 8). MK has higher Al₂O₃ content compared with IC (Table 1) therefore the AAM made with MK has lower SiO₂/Al₂O₃ ratio compared with AAM based on IC and lower amount of total leached OH^- ions. By reducing ASRW in the composition, both SiO₂/Al₂O₃ and Na₂O/Al₂O₃ ratio and leaching rate increased significantly for AAM samples made with IC, while for AAM based on MK the effect of ASRW on leaching rate was reversed and leaching rate decreased. By using LSG in the composition, the SiO₂/Al₂O₃ ratio increased; therefore the leached amount of OH^- ions was higher for all AAM made with LSG. The relevance of main oxide ratios to the leaching of OH^- ions was calculated and leaching model was obtained (Figure 8). The leaching rate was analyzed on 1^st and 20^th day after leaching test. More precise correlation was observed at early age (on 1^st day) and scattering increased on 20^th day. The correlation between main oxide ratios and leaching could be regarded as acceptable while the results at later age can be affected by the physical properties of AAM.