Introduction

Acid activation of calcined clays for cementitious purposes is a relatively recent approach compared to traditional alkaline activation (Gao et al., 2020; Tchakouté; Rüscher, 2017). However, it’s noteworthy that the insolubility of the products formed by this technique was already recognized in agronomy. Agricultural studies revealed that mixing phosphate fertilizers with acidic soils, especially clayey ones, led to the formation of hardened and insoluble soil, which were unsuitable for agriculture (Medina; Guida, 1995). Moreover, laterite soils, prevalent in tropical regions and rich in iron and aluminum oxides, have found application in civil engineering. These soils can be stabilized with phosphoric acid, making them suitable for various purposes, like soil stabilization for roads and paving (Eisazadeh; Kassim; Nur, 2012; Michaels; Tausch, 1961). For application purposes in civil engineering, the activation of metakaolin (MK) using H₃PO₄ still registered at limited literature (Alvi et al., 2024; Derouiche; Baklouti, 2021)

According to Pereira (2003), the ionic exchange between the structural interlayer cations of clays and the acid’s anions during acid activation results in significant changes in the crystalline structure and original properties of clays. These structural and property changes can lead to different characteristics of the final products, depending on the specific chemical composition of the clay and the type of acid used. For instance, products may exhibit properties like adsorption or filtration capacity (Le-Ping et al., 2010). The ability to adjust the characteristics of final products by varying the chemical composition of clay and acid opens up possibilities for applications in different fields, such as the development water purification adsorbents or industrial effluent treatment filters.

Geopolymerization produces an inorganic polymer developed by similar reaction kinetics in acid or basic environment. The acid-based polymerization reaction involves the bonding of Al–O (from metakaolin) and P–O tetrahedral units (from H₃PO₄ solution) (Alvi et al., 2024). [AlO₄]⁵-, [SiO₄]⁴-, and [PO₄]³- are the fundamental groups evolving into acid activation by phosphoric acid, where [PO₄]₃- ions replace alkali metal ions in alkaline systems (Li et al., 2022). Louati, Baklouti and Samet (2016) indicate that the positive charges of ionized H₃PO₄ are balanced with [AlO₄]⁵-, as the positive charges of dissociated NaOH are balanced by [AlO₄]⁵- (Pacheco-Torgal et al., 2015). Thus, the bonds formed are of the Si type /Al - O - P. Furthermore, the degree of polymerization is influenced by factors like the Si/P molar ratio and curing temperature, similar to alkaline systems (Zribi; Samet; Baklouti, 2020).

The acid activation products of aluminosilicates are also termed phosphate-based geopolymers. These composites, with high phosphate content, can be incorporated into traditional aluminosilicates with low acid concentrations to enhance reactivity. They exhibit characteristics like low efflorescence/carbonation, dimensional stability against acid attack, and proportional dielectric properties to mechanical performance (Derouiche; Baklouti, 2021). Ongoing research on these composites enables the utilization of other by-products, more specific applications, and the development of micro-nanoparticles with higher affinity in acidic environments.

Materials and methods

Metakaolin (MK) and Fly Ash (FA) were obtained from a national industry and a thermoelectric plant, respectively. X-ray fluorescence (EDX7000, Shimadzu) was performed to determine chemical compositions and the results are shown in Table 1.

Laser diffraction (Microtrac S3500) was used to determine the particle distribution, as shown in Figure 1a. The mean diameters (µm) of MK and FA were 16.26 and 51.06, respectively. The mineral phases were identified, see Figure 1b, using XRD measurement (Miniflex, Rigaku).

For acid activation, phosphoric acid at a concentration of 85%, diluted into deionized water, resulting in a molar concentration of 6.12 mol/L of H₃PO₄ was used. This concentration was chosen considering the highest value from He et al. (2023) and Tchakouté and Rüscher (2017). For alkaline activation, the solution employed was composed by sodium silicate and sodium hydroxide, getting a concentration of 6.26 mol/L of NaOH, this concentration was used by the research group (Luz et al., 2019). Both solutions were prepared on a magnetic stirrer and allowed to stand for 24 hours before use. The solution was added to the power material during the mixing process according to the mass proportions specified in Table 2. The mixture was prepared using a mechanical mixer at 600 rpm for 4 minutes, tested for mini-slump test, isothermal calorimeter at 65 °C (I-Cal 2000HPC, Calmetrix), and then molded into cylinders with a diameter of 19 mm and a height of 24 mm. These molds were subsequently placed into an oven for curing at 65 °C for three days.

The rheologic measures were obtained from Haake Mars rheometer (Vane geometry – pre-shear at 50 s^-1, rise ramp from 0.1 to 100 s^-1 and from 90 to 0.1 s^-1 for four minutes), using a similar routine used by Taborda-Barraza et al. (2022) and in its hardened state by the compressive strength test (Instron model 5569) press and a load rate of 4000 N/min. Furthermore, a microstructural analysis was carried out using X-ray diffraction (XRD) (Miniflex, Rigaku) and Fourier Transform Infrared Spectroscopy (FTIR) after 28 days of reaction. KBr pellets were used to prepare the samples (1 mg of sample/ 300 mg of KBr) (4200 FTIR, Jasco) based on the following parameters: analysis range of 400-4000 cm^-1; resolution of 160 cm^-1; and 52 accumulations. X-ray diffraction analysis was also applied for the samples. The precursors also were analyzing through zeta potential (Zetasizer NanoZS, Malvern).

Results and discussion

Figura 2 shows the zeta potential at different pH for both precursors. The results indicate that raw material samples exhibited greater stability at alkaline pH levels (>7) while instability was observed at more acid pH levels. ±30mV indicates the most unstable zone for the particles, a state with tendency to aggregation or agglomeration (Kriegseis et al., 2020). Specifically, FA samples showed less stability than MK due to their lower value of PZ modulus, i.e, a system predominantly negative charge (Gunasekara; Setunge; Sanjayan, 2015). However, their stability starts to be manifested even at lower pH (>4), and pH alkaline arrived better stabilization. Conversely, MK sample demonstrated few stabilities for acid pH values. This means that both precursors probably register precipitation and few dissolutions in acid conditions but great behavior on alkaline conditions.

Evaluation on fresh state

The appearance of the fresh state of various samples, via mini-slump, has been documented in Figure 2. Specifically, the acid activation system was found to be easy to mix, but the consistency of the resulting mixture was firm, showing minimal spreading compared to alkaline activation. Additionally, it was observed that, as the fly ash content increased in the samples, the scattering increased in both systems. This was attributed to spherical shape of ash particles that favor flow. Both systems shared a common trait: their final setting time tended to be longer, typically exceeding 3 days, when compared to cement systems based on Portland cement (Xie et al., 2023). Consequently, thermal curing is often applied (Kaze et al., 2021). This delayed setting time is a consequence of the slow dealumination of particles in acidic environments at room temperature. Similarly, in alkaline environments, a low rate of particle dissolution is observed at room temperature, particularly in mixtures with low calcium and sodium hydroxide contents.

In addition, the rheology of the pastes was evaluated, and the results were summarized in Figure 3. Both systems exhibited a reduction in viscosity as the shear rate increased, indicating pseudoplastic behavior. The representative equations for the rheological models are summarized in Table 3. Calculating the yield stress of the systems, it was noted that not all samples could be fitted into the same model. This was because the yield stress value appeared to be negative in some cases, indicating a fluid that was sufficiently easy to spread. This behavior was observed particularly when fly ash was incorporated into a hybrid system, a phenomenon also documented by (Aboulayt et al., 2018).

In the analysis, representative models for each system were highlighted in yellow. Consistent with the mini-slump test, the incorporation of ash resulted in a reduction in yield stress. Both individual systems required energy with values close to each other to initiate flow within the system. When evaluating the systems using the Herschel-Bulkley model, it was found that acid activation exhibited greater consistency, indicating visual firmness. On the other hand, when assessing viscosity using the Bingham model, it was observed that alkaline activations were more viscous compared to acidic ones.

The same groups were also defined using a calorimetric test, i.e., 100MK and 40MK-60FA. The results are summarized in Figure 5. Figure 5 shows the heat flow at 65 °C (Figure 5a) and the total heat accumulated (Figure 4b). The first observation is related to the higher intensity of dissolution (Peak I) in acid systems compared with alkaline systems. This behavior was discussed by Bu et al. (2024), who explains that in acid environments the solubility on MK surfaces is higher compared to alkaline environments, due to Al-O bonds being weaker in this condition. Even when the presence of FA led to a reduction in this intensity for 100MK group, the dissolution value was higher than in alkaline activations. Consequently, the total heat accumulated for 40MK-60FA was reduced in 78,68% at 48 hours, compared with 100MK, after contact between precursor and activator. This phenomenon also indicated that FA did not contribute significantly to the kinetic reaction under these conditions.

The second observation is related to peak II, that appears to alkaline activated systems nevertheless for acid systems, based on a tendency of transformation of a concave curve to a convex curve between 1-2 hours. Xie et al. (2023) registered a decrease in the intensity of dissolution peak when the concentration H₃PO₄/H₂O was reduced, after eight hours no peak was registered. Few studies discuss the calorimetry in acid systems. Finally, alkaline systems were negatively affected by the presence of fly ash, reducing the intensity of the dissolution peak and the final energy of the system.

Evaluation on hardening state

In Figure 6 the results of compressive strength are presented. It was observed that acid activations performed better in systems based on 100MK and 80MK, showing an increase of 45% and approximately 90%, respectively, compared to the same systems in alkaline activation. However, for combinations with a lower proportion of MK, such as 40FA and 60FA, alkaline activations exhibited superior performance over acid activations. The 60FA system showed zero strength, as consequence of demold failure due to soft consistency. These findings align with similar results obtained by Louati, Baklouti and Samet (2016), and Xie et al. (2023) for acid activation of MK. The data represented in Table 2 highlights the influence of material composition on compressive strength, with acid and alkaline activations demonstrating variable performance depending on the specific proportions of materials used.

As the curing period progressed, alkaline activations exhibited a greater formation of rigid structures, resulting in a substantial increase in strength. Specifically, there was an approximately 155% and 80% increase in systems incorporating 40% and 60% FA, respectively. However, 100MK systems showed a drop at 28 days, this behavior is common on MK-paste due to shrinkage related to reaction products with low CaO (Huang; Wang, 2024). On the other hand, acid activations showed a more consistent increase in strength over time. For systems based on 100MK and 80MK, there was a constant increase of approximately 15% at 28 days. However, for the group with 60MK, although there was a significant increase in strength, it remained lower at 3.50 MPa. These observations underscore the distinct temporal evolution of compressive strength in acid and alkaline activation systems, with alkaline activations demonstrating a more pronounced improvement over time due to the formation of rigid structures.

Finally, XRD and FTIR spectrums were analyzed from samples at 28 days. The results are represented in Figure 7a and 7b for acid groups. About XRD characterization, Figure 7a, shows the amorphous halo linked to the formation of structures, that turn wider after activation, compared with the precursors in all groups, i.e, it is the indicator of phase phosphate aluminosilicate (Zulfikly et al., 2021). Similarly to Guo et al. (2021), mixtures containing fly ash record this amorphous halo slightly wider than the 100MK group. However, it was also observed that certain crystalline phases as quartz and illite remain after activation, showing that they unaffected by the treatment acid (Karuppaiyan; Jeyalakshmi, 2024). The Berlinite phase, or AlPO₄, appears together with quartz remaining (André et al., 2023; Morsy et al., 2019).

In addition, Figure 7b records the results obtained from FTIR analysis. Therein, several band regions named to #1 until # 5 were identified. Band #1, located approximately between 3226-3558 cm^-1, corresponds to the O-H group band. Band # 2 located at 1618 cm^-1 is associated with angular deformation vibrations of the H-O-H band. As occurs in alkaline activations, this band indicates the presence of water molecules that are superficially absorbed or trapped in the large cavities of the polymeric network of the geopolymeric gel (Nobouassia et al., 2022). Band # 3, located at approximately 1080 cm^-1, corresponds to the P-O asymmetric stretching and PO₂ asymmetric stretching bond band of the P=O doubly bonded oxygen vibration of phosphoric acid (Zribi; Baklouti, 2021; Louati; Baklouti; Samet, 2016). According to Zribi and Baklouti (2021), the region between 1300 cm^-1 and 900 cm^-1 is very informative, as it can include different vibration modes of Si-O- P, Al-O-P. Band # 4, located between 980 cm^-1 and 780 cm^-1, represents the vibrations of the P-O-P bond, characteristic of the acidic geopolymeric system (Li et al., 2022). Finally, band # 5 corresponds to vibrations between 600 cm^-1 to 550 cm^-1 and corresponds to stretching vibrations of P-O-P and Al-O-P. The appearance of these new peaks indicates that the reaction of MK with H₃PO₄ transforms the Si-O-Al structure in MK into Si-O-P, Si-O-Si and Al-O-P structures in the resulting acidic activation matrix (Lin et al., 2021).

Regarding the XRD spectra of alkaline activations, see Figure 8a, the amorphous halo resulting from the formation of structures was recorded (Fernandez-Jimenes et al., 2017). However, a significant contrast was also recorded. Initially, due to the intensity of the crystalline phases, which was greater than in acid activation, in the case of quartz, mullite and hematite. On the other hand, the illite peak disappeared under these conditions. This behavior indicates that even though the precursors are the same, the nature of the activation will lead to the reduction of certain phases, which depends mainly on the dissolution of the particles in the medium, which is a consequence of the concentration of the activator, the temperature, specific area and the energy of activation, as the barrier to be overcome to produce ionic exchange (Huang; Wang, 2024; Lin et al., 2023).

Figure 8b shows the vibration peaks of the alkaline activation samples. As the same way, band # 1 and #2 correspond to O-H bond (Li et al., 2023). As occurred in acid activation, the absorption band at 1611-1625 cm^-1 is associated with bending vibrations of carbon molecules. water H-O-H (Tchakouté; Ruscher, 2017). The band located in the area #4, at 1450 cm^-1, corresponds to stretching of O-C-O, which implies formation of carbonate into geopolymers due to the reaction of free Na⁺ ions with atmospheric CO₂ (Gao et al., 2014). Band #5 of vibration at 1025 cm^-1 is associated with the Si-O-T bond (where T can refer to Al, Fe or Si). According to Tantono et al. (2023), this band corresponds to the presence of the alkaline activation gel. The Si-O-Al and Si-O-Si bands and can be associated with the Silica tetrahedron (SiO₄) (Tambara Junior; Cheriaf; Rocha, 2018).

Conclusion

The acid activation of unitary and binary systems based on metakaolin and fly ash exhibits hardening and rheological behavior similar to alkaline activation systems. However, systems based on acid activation do not develop a second peak associated to polymerization at isothermal calorimetry. The presence of fly ash decreases substantially the kinetics reaction. Incorporating up to 20% fly ash is considered suitable for both systems. However, replacements greater than 40% fly ash are not recommended in an acid activation system because, under the evaluated conditions, they resulted in low or no compressive strength. In contrast, for alkaline activations, significant increases in compressive strength are observed over time, starting from the 20% incorporated content of fly ash in the system. This suggests that alkaline activations may be more resilient to higher levels of fly ash incorporation compared to acid activations.

References

Publication Dates

History