The pozzolan (PZ) is quarried in Penja, and the kaolinite clay from Dibamba, all in the Littoral Region of Cameroon. PZ was dried, grounded and sieved below 75 µm. The kaolinite clay was enriched in clay fraction by wet sieving below 75 µm, removing large particles of quartz and other non-clayed minerals. The fine fraction was heated in an electric furnace (Nabertherm Model LH 60/14) with a temperature rate of 5˚/min up to 700˚C for 4 hours, to obtain a powder named MK. This temperature was selected based on previous research indicating its optimal use for the amorphization of kaolinite clays, enhancing their reactivity [14]. The quantification of the amorphous phase was done in previous works [15]. MK was used to substitute PZ respectively at 20, 30 and 40 wt.% to obtain solid precursors that will be denoted as 80/20, 70/30, 60/40. The liquid precursor was an equivolumic mixture of activating alkaline solutions of a 8 M aqueous sodium hydroxide (99%) and a sodium silicate (SiO₂/Na₂O = 44,209; H₂O 55.41%).

To obtain the foam, a commercial foaming agent called “aircrete” with a 0.32 g/L relative density was-mixed with distilled water in the volume ratio of aircrete/water = 4%, and then vigorously stirred with an electrical foam machine to obtain a foam having a density of 45 kg/m³.

The sand used was standardized in accordance to EN 933-1 and EN 933-2 standard [16]. It was a mixture of Sanaga river sand and a fine equatorial sand from Yaoundé. The physical characteristics of sands are within the limits reported for standard sands with a specific particle size distribution. The fineness modulus (FM) is 2.97, the uniformity coefficient, UC = D60/D10 is 6.85, and the coefficient of curvature (CC) is equal to 2.204.

To elaborate geopolymer mortars, each of the three solid precursors (PZ, MK and sand) was mixed to the activating solution, with a liquid (mL)/solid mass (g) ratio of 0.16 (Table 1). The mixture was homogenized for ten minutes using an M&O mixer (model N50-G) to create a paste. Foam was added to the mortar at 0, 1, and 2 wt.% under continuously stirring and the resulting pastes were labeled GM0, GM1, and GM2, respectively. The viscous mortars obtained were molded into cylindrical test discs with a diameter of 20 mm and a height of 40 mm. To remove larger air bubbles, a M&O electric vibrator (type 202 No. 106) was used for vibration. To prevent water evaporation during paste setting and hardening, the specimens were covered with a thin polyethylene film and left in the ambient atmosphere (24˚C ± 3˚C). Demolding was done 24 hours later.

Table 1. Composition of mortars.

The starting materials (PZ and MK) were characterized by ICP for chemical composition, XRD, FTIR and amorphous phase content, spectroscopy method for the structural compositions and specific surface area. Amorphous phase quantification was measured using the method described in previous works (Balde etal., 2021). The M₁ (g) sample was mixed with 30 mL of an 8 M sodium hydroxide solution in a 100 mL beaker in order to eliminate the acidic oxides present in the material. The mixture obtained was washed with distilled water and centrifuged using an Eppendorf type 5840 R centrifuge until a neutral pH is obtained. The residue obtained was neutralized with 30 mL of a 0.5 M hydrochloric acid and washed with distilled water. Filtration was carried out to recover the residue. The latter was dried at room temperature (25˚C ± 3˚C) and in a Heraeus brand dryer (VT 5042EK) at 110˚C for 24 hours. The residue was weighed (M₂) and the amorphous rate in the tested material was obtained by Equation (1):

The synthesized mortars were characterized by setting time according to the EN 196-3 [17] using a Vicat device, linear shrinkage using a digital caliper on cylindrical test discs, apparent density and open porosity with the Archimedes method. Optical microscopy was carried out using an optical microscope (Ceramics Instruments Model 101T-M7). The compressive strength was measured according to standard EN 1961 [18].

Table 2 shows the chemical composition of the pozzolan and the metakaolin used in the formulation

Table 2. Chemical composition of PZ and MK.

It can be noticed from Table 2 that silica (SiO₂) and alumina (Al₂O₃) are the major oxides in both PZ (Pozzolan) and MK (Metakaolin). The Si/Al molar ratio is 2.1 for the pozzolan and 4.4 for the metakaolin. It is in accordance with many works in the literature, proving that metakaolin has a higher capacity to strengthen the geopolymer tridimensional network [19][20].

The Si/Al ratio of the solid precursor is known to be a key factor influencing both the properties of geopolymers and the geopolymerization mechanism [19]. Furthermore, geopolymers with a Si/Al ratio of 4 or higher exhibit significantly greater mechanical properties. This is due to the presence of a large concentration of Si-O-Si bonds acting in the strengthening in the structural arrangements [21].

Figure 1. X-rays patterns of (a) MK and (b) PZ (A: Anatase/Augite, B: Biotite, An: Anorthite Q: Quartz).

Figure 1 shows the X-ray patterns of metakaolin and pozzolan. The identified phases in metakaolin are quartz (SiO₂, PDF 89-8934) and anatase (TiO₂, PDF 21-1272). In the pozzolan powder, augite (Ca, Na) (Mg, Fe, Al) (Si, Al)₂O₆, PDF 24-201), anorthite (CaAl₂Si₂O₈, PDF 41-1486) and biotite (K(Mg,Fe)₃AlSi₃O₁₀(F,OH)₂, PDF 2-45) are detected. In the metakaolin pattern a dome in the range 18 - 28 2θ is evidenced, and for pozzolans, it is seen in the range 5 - 24 2θ. It is a clear indication of the presence of an amorphous phase (Figure 1).

Figure 2(a) presents the infrared spectrum of pozzolan. It shows an absorption bands at 1641 and 3349 cm⁻¹, which respectively express the deformation vibrations of the H-O-H valence bonds and of the O-H bonds of water molecules [22]. The absorption bands at 944 cm⁻¹ corresponds to the symmetric and asymmetric elongation vibrations of Si-O-Si and Si-O-Al bonds, while the band at 730 cm⁻¹ reflects the presence of the elongation vibrations of Al-O.

Figure 2(b) shows the IR spectrum of metakaolin (MK). The spectrum has a broad band at 1044 cm⁻¹ reflecting the presence of the asymmetric and symmetric elongation vibrations of the Si-O-Al and Si-O-Si bonds. It demonstrates the presence of amorphous silica in metakaolin. The band at 546 cm⁻¹ is attributed to the vibration of Al-O bond. The band between 789 and 776 cm⁻¹ reveals the presence of the stretching vibration of Al-O. The band at 429 cm⁻¹ is attributable to the deformation vibrations of the Si-O-Si and Si-O-Al bonds [23].

Figure 2. FTIR Spectra of (a) PZ and (b) MK.

The values of the specific surfaces areas (SSA) are 8.6 and 20.5 m²/g for PZ and MK respectively. Since the specific surface area of metakaolin is greater than that of pozzolan, proving the higher reactivity of metakaolin compared to pozzolan [24].

The mass percentage of the amorphous phase is 45.1% and 65.2% for PZ and MK respectively and this is in accordance with the observation of the X-ray patterns exhibiting a dome around 2θ = 22˚ for MK. The relative quantity of the amorphous phase is equivalent to the reacting phase content with the activating alkaline solutions. The amorphous phase improve the geopolymerization process, resulting in a greater mechanical resistance after setting [25].

It was observed that the shaped discs of mortars containing 3 wt.% of foam are very porous and weak and will not be considered for further investigation. The results presented below are for mortars with 0; 1 and 2 wt.% of foam.

It was observed that the shaped discs of mortars containing 3 wt.% of foam are very porous and weak and will not be considered for further investigation. The results presented below are for mortars with 0; 1 and 2 wt.% of foam.

Figure 3 shows the variation of the initial and final setting times of mortar pastes as a function of both substitution of pozzolan by metakaolin and percentage of foam.

It can be seen that the initial and final setting times of geopolymer mortars increase with the amount of the foam. For instance, the initial setting time increases from 180 min to 255 min when the foam quantity increases from 0 to 2 wt.% for PZ/MK = 60/40. Similarly, the final setting time increases from 420 min to 540 min with the addition of 0 to 2 wt.% of foam for PZ/MK = 80/20.

Figure 3. Initial and final setting times for the compounds 80/20 (a), 70/30 (b) and 60/40 (c) versus the foam content.

But the setting time decreases with the percentage of metakaolin for a fixed quantity of foam in the mixture. For the addition of 1 wt.% of foam, and when pozzolan is replaced by increasing quantities of metakaolin, the initial setting time decreases from 370 to 240 min, and the final setting time decreases from 480 to 300 min. For the addition of 2 wt.% of foam, a similar replacement of pozzolans by metakaolin lead to the reduction of the initial setting time from 480 to 255 min, and of the final setting time from 540 to 316 min. The results align with the reactivity of metakaolin thank to its higher alumina and silica contents. An increase of Si and Al contents results in a higher quantity of the geopolymeric gel during the initial stage of the setting process [21]. Since the gel plays a critical role in the geopolymerization reactions, it strongly influences both the reaction rate and the initial and final setting times. Additionally, the addition of foam may reduce the viscosity of the mortar paste, leading to a decrease in paste consistency. This is also expected to increase the mobility of species and the dissolution rate of aluminosilicates. However, foam induces the creation of voids from bubbles distributed in the volume [26][27]. These physical discontinuities limit the diffusion process of species, and lower the setting rate. These statements are in on the same line with previous research that examined the effect of foam on the consolidation process by analyzing the kinetics of the reaction mechanisms [28][29].

Figure 4 presents the IR spectra of the different geopolymer mortars obtained by mixing PZ and MK with the PZ/MK ratio of 60/40 with different foam dosage of 0, 1 and 2 wt%.

Figure 4. FTIR spectra of GM0, GM1 and GM2 (with PZ/MK = 60/40).

The FTIR spectra of the three geopolymers are similar proving that the effect of foam dosage is not noticeable as per structural composition. Anyway, the absorption bands at 3684 - 3200 cm⁻¹ and at 1780 - 1650 cm⁻¹ respectively reflect the presence of elongation and deformation vibrations of the O-H and H-O-H bonds of water molecules [30]. The absorption bands at 825 - 1263 cm⁻¹ are attributed to the asymmetric stretching vibrations of the Si-O-Si and Si-O-Al bonds as well as those observed at 856 - 779 cm⁻¹, 759-702 and 570 - 546 cm⁻¹ [31]. In fact, the ability of Al-Si oxides to inter-react with the other species characterizes the geopolymerization system and is also linked to the degree of polymerization alongside with the hardening properties.

Figure 5 displays the correlation between the linear shrinkage of the mortar specimens and the foam content. The shrinkage is also changed by the substitution of pozzolans by metakaolin, and a more pronounced effect occurs at a higher foam content (2 wt.%). Undoubtedly, during setting, water from the foam is firstly removed by a more or less effective combination with hydrolyzed aluminosilicates. A large amount of water is mostly removed by diffusion towards surfaces during drying. It contributes to an internal readjustment of the system, resulting in an increased shrinkage when more foam is added. This physicochemical phenomenon contributes to obtain a more consolidated and hardened material [32]. It is also evidenced that the shrinkage rates are similar to those of common structural materials [33]. The specific role of metakaolin is related to its effective dissolution process at contact with the activating solution. This process leads to the formation of a structural tridimensional network [6]. It may limit the influence of metakaolin on the shrinkage during the hardening process, even though the solid/liquid ratio of precursors is predetermined.

Figure 5. Linear Shrinkage of the mortar specimens at day 28 for 80/20, 70/30 and 60/40 PZ/MK ratios.

The decrease of the apparent density with increase of the amount of foam is evidenced in Figure 6(a). Test discs with a ratio of 60/40 for PZ/MK, have a density of 1.7 g/cm⁻³ for 1wt.% of foam and a density of 1.51 g/cm⁻³ for 2 wt.% of foam. The values are similar to those of lightweight materials of other findings [13][34]. Foam is used to aerate the microstructure by creating voids through the formation of bubbles during the setting process. These voids form a 3D network that increases porosity, resulting in an increased open porosity at a higher foam content. Correspondingly, Novais etal. (2020) found that voids can contribute to both closed and open pores [29]. Figure 6(b) also shows the variation of the apparent density versus the foam content. The occurrence of either open or closed porosity depends on various factors, such as the foam formation technique and concentration in the starting formulation. In this work, a direct foaming technique was used, meaning that foam was added into the geopolymer paste after the mixing of components. It results in a more pronounced open macro porosity with an accentuated interconnectivity [35].

Figure 7 displays micrographs of test discs made with different mortars. In all cases, the morphology shows a homogeneous microstructure with sand grains distributed throughout the matrix, resulting in a compact microstructure. As foam is added gradually, the microstructure takes a honeycomb appearance with rounded cavities of varying sizes, becoming more pronounced with increasing foam content. This type of microstructure defines a material having a lower density and a high mechanical strength since the compressive stress is distributed along the rounded cavities [36].

Figure 6. Apparent density (a) and Open porosity (b) of the mortar specimens after 28 days for 80/20, 70/30 and 60/40 PZ/MK ratios.

Figure 8 shows the variations of the compressive strength of the geopolymer mortars as a function of the foam content and the rate of substitution of pozzolans by metakaolin. It is evidenced that the compressive strength of the reference mortar without foam is greater than that of porous mortars. Increasing the foam quantity results in decreasing the density and the compressive strength of mortar [3]. Since the density of foamed concrete is generally correlated with strength, a balance between these two parameters could be found depending on the intended application. Generally, a denser foamed concrete has a higher compressive strength and a lower volume of voids, and a low density has a decreased strength. A good compromise is achieved by optimizing the composition and by selecting the appropriate quality and quantity of foaming agents having a significant foaming capacity. Generally, strength decreases with an increase of the number of bubbles, but it depends on the distribution of bubble size and in the foam stability with time. Besides, the increase of strength with the increase of the metakaolin relative content is explained by its specific contribution since it is the most active phase in comparison with that of pozzolan. More precisely, metakaolin promotes a local microstructural consolidation [37], particularly near pores where a decohesion process with the geopolymer matrix can occur.

Considering the chemical compositions, the Si/Al ratio of metakaolin is 4.4 whereas that of pozzolans is 2.1 which results in a greater quantity of MK leading to a greater mechanical strength. In addition, metakaolin has a high specific surface in comparison to that of pozzolans, and metakaolin also has a high amorphous phase content. These parameters favor a larger quantity of geopolymeric gel during geopolymerization, and promote obtaining a more consolidated material having better mechanical resistance. This is in agreement with the work of Bondar and al. 2011 and others [38] who showed that mineral additives to pozzolan promote an increase in mechanical resistance.

Figure 7. Optical micrographs of mortar after 28 days with 0 wt.% (a)-(c), 1 wt.% (d)-(f) and 2 wt.% (g)-(i) foam content.

Many researchers also reported that the structural stability of geopolymers increases with the content of silicate species. This is due to the formation of silicate oligomers having long structural arrangements, with the occurrence of Al-O-Si complexes. This could also explain the delay in the intake observed when the metakaolin content decreases [39].

Figure 9 reports the relationship between compressive strength and density for various foamed concrete compositions [40], including data from our present work. A very large distribution of data is obtained from the numerous parameters linked to the fabrication of foamed concrete. Despite the wide range of data, there is a clear trend of increasing strength with density. Our data fall just below the average strength value of the distribution that validates the process used in this work [3][41].

Figure 8. Compressive strength of the mortar specimens at day 28 for 80/20, 60/40 and different PZ/MK ratios.

Figure 9. Compressive strength values from this work (black line) in comparison to data (gray line) from the literature [40].

Table 3 reports the comparison of some properties of the geopolymers from this study with that from different published works. The data are focused on values of compressive strength, apparent density, linear shrinkage and porosity.

Similar strength values are observed when the density is close to 2 g/cm³. However, the strength is increased or decreased significantly when the compositional domain is altered by adding volcanic ash or rice husk ash. This demonstrates the significant role of each component in the behavior of mortar. Additionally, it proves that metakaolin is an efficient component due to its high reactivity and availability at a moderate environmental cost.

After comparing the characteristics of various geopolymer mortars found in the literature, it can be concluded that the mortars used in this study are of the lightweight range.

Table 3. Comparison of mortar properties after 28 days of setting.