Nowadays, scientific researchers are increasingly looking for optimal solutions that simultaneously mitigate the negative impact of industrial waste and produce ecological materials that contribute to environmental protection. One example of this is the alkaline cements also known as geopolymers, which are recognized as sustainable materials, usually made with secondary by-products as precursors (such as blast furnace slags or coal fly ash), and alkaline sources as an activator. These can play an important role in many fields such as the construction industry, immobilization of hazardous materials, adsorption of contaminants from wastewater, etc. (1-3). Geopolymers have always attracted the interest of researchers for their potential to decrease the carbon dioxide emissions associated with cement production, offering a viable alternative (4). Historically, these materials were known first as alkali-activated materials. These materials have proven not only to be more sustainable from an environmental point of view, but also to have similar or even superior strength and technological behavior to conventional cements, making them the object of various applications (4-6). Traditionally, these cements are generated by reacting an aluminosilicate precursor with an alkaline activating solution (sodium hydroxide, waterglass...), however, recent studies (7) have demonstrated the possibility of using acid solutions (phosphoric acid) generating materials with interesting properties, such as porous matrices capable of acting as immobilizing matrices for dyes or heavy metals (2), or materials for thermal insulation (8, 9). The reaction products of acidic geopolymers differ significantly from those of alkali-activated geopolymers due to the influence of the medium’s pH. Generally, in alkaline geopolymers, the major reaction product is C-A-S-H gel (for high calcium content precursors), N-A-S-H gel (in low calcium content precursors). In hybrid systems, the products is a mixture of different types of gels (3, 10). For acidic geopolymers, the main reaction products depend on the nature and chemical composition of the precursor used. In this context, Li et al., used different mixtures of high-magnesium nickel slag (HMNS) and fly ash (FA) activated with H₃PO₄ (11). Their study focused on various curing methods: curing with fresh-keeping film, curing in water, curing in air, standard curing, and curing at constant temperatures. The curing time were 3, 7, and 28 days. The main reaction products identified were aluminophosphate gel, newberyite gel phase and crystals, and aluminum silicon phosphate gel phase and crystals. Alshaaer et al., conducted a study aimed at producing geopolymers activated with 10M H₃PO₄. The specimens were cured 24 h at 60 °C and then at room temperature for 28 days (12). The main reaction products were -Si-O-P-O-Si- with aluminum phosphate crystals. Lin et al., also studied the activation of metakaolin using H₃PO₄. The major structural units identified were Si-O-Si, Si-O-Al, Si-O-P, and Al-O-P. Moreover, Zhan et al., studied the activation of metakaolinite with H₃PO₄, incorporating Al₂O3 powder into the mixture (13). The main reaction product in this case was Si-O-Al-O-P-O-. In summary, the type of reaction products generated is highly dependent on the nature of the precursor, the acid activator, and the activation conditions.

It is important to mention that most of these studies used precursors with low calcium content. This suggests that the reaction products in the present study will be significantly different because the precursor used is obtained from the wastes generated during phosphate ore extraction, which are rich in CaO. The use of industrial wastes as precursors for the production of geopolymers has attracted considerable research interest. Various wastes, including waste glass, fly ash, and mining wastes such as phosphate waste, have been tested for this purpose (3, 14).

Moroccan researchers have actively focused on valorizing phosphate mining wastes for various applications. These by-products have been successfully used in the formulation of concretes, fired bricks, mortars, and geopolymers with promising results (15-18). The use of such wastes in various applications can significantly reduce their negative environmental impact. Given the large quantities of phosphate mining waste typically deposited near mining sites, this practice often leads to serious environmental problems, including the degradation of the physical landscape and its aesthetic value (3).

The objective of the present work is to evaluate the potential of a secondary mine waste products (red clay) as a precursor for the elaboration of acid geopolymers, paying particular attention to the identification of the reaction products generated and trying to optimize the experimental conditions for the production of geopolymers from phosphate by-products. This approach will also help reduce the negative impacts of phosphate mine waste disposal.

Figure 5 shows the compressive strength values for geopolymers activated with 5M and 8M H₃PO₄ solution after 7, 28, and 90 days of curing. It is distinguished that the strength development is highly dependent on the H₃PO₄ concentration. The systems with 8M H₃PO₄ develop higher strengths attaining 14.5 MPa at 90 days. The effect of phosphoric acid concentration on metakaolin-based geopolymers was previously studied by (30), in the acidic geopolymers prepared with different acid concentrations (4, 6, 8, 10, 12, and 14M) to study its effect on the properties of the final products. The results of this study showed that increasing the H₃PO₄ concentration from 4 to 10M increased the compressive strength values. This was explained by the presence of a small amount of free water remaining in the solution at low concentration values, that does not react during the geopolymerization process. This leads to the formation of pores when the specimens are dried, resulting in low compressive strength values, which was the case in the present study. In addition, the compressive strength values of geopolymers activated with 8M H₃PO₄ increase with the amount of NaAlO₂. After dissolution, NaAlO₂ provide tetrahedral aluminium, that can participate in the formation of reaction products (especially AlPO₄ and the amorphous gel, as it will be shown below), which could lead to increase the compressive strength values.

XRD analysis of both types of geopolymers (activated with 5 and 8 M H₃PO₄ acid) are shown in Figure 6. The results show the presence of certain phases, such as quartz and tridymite, that remained after acid activation (indicating that are not potentially reactive under these conditions). In addition, some new phases were formed due to the reaction between phosphoric acid and lime. Brushite (CaHPO₄·2H₂O) appears as a new crystalline phase in geopolymers activated with 5 M phosphoric acid, while monetite (CaHPO₄) appears in those activated with 8 M. Monetite is the dehydrated phase of brushite (31). The formation of these phases depends highly on the pH values of the pore solution. According to the literature, monetite forms at pH values below 5 (32), while brushite forms at pH values below 6.5 (33). This implies that the pH for monetite formation is lower than that for brushite formation, which is consistent with the present study where the pH of 8M H₃PO₄ is lower than that of 5M H₃PO₄. It has also been reported by (34) that brushite cements have low compressive strength due to their rapid setting, resulting in poor crystal compaction during the setting process, which is in good agreement with the previous findings. In addition, newberyite (MgHPO₄·3H₂O) and aluminum phosphate (AlPO₄) are present in both types of geopolymers. Notably, the crystallinity of newberyite is higher (much more sharp and defined peaks) in geopolymers activated with 5 M phosphoric acid compared to those activated with 8 M phosphoric acid. This suggests that newberyite may be present in both crystalline and amorphous forms in geopolymers activated with 8M phosphoric acid, potentially contributing to the higher compressive strength of these geopolymers. Moreover, the appearance of humps ranging from 18-40° indicates the formation of amorphous gels. Similar to the present study, the influence of phosphoric acid on the properties of geopolymers has been explored by several researchers, with a key difference being the type of precursor used. The study conducted by Pu et al., demonstrated that when fly ash was activated with phosphoric acid, the crystalline reaction products were Berlinite, Monetite, and Brushite, while the amorphous phases included Si-O-Al-O-P, Si-O-P, and Al-O-P units (35). The highest compressive strength, 20 MPa, was achieved with 50% phosphoric acid (L/S = 0.3) after 90 days of curing. They noted that Berlinite enhances geopolymer strength and structure but may not support long-term strength growth in phosphate geopolymers. Similarly, Tchakouté et al., investigated metakaolin as a precursor activated by different concentrations of phosphoric acid and observed the formation of an amorphous phase (Si-O-P-O-Si) along with a crystalline phase (Berlinite) (30). The highest compressive strength was achieved with 10M phosphoric acid, and the presence of AlPO₄ was shown to enhance the overall strength. This emphasizes the key role aluminum content plays in precursor selection. Additionally, Riyap et al., reported that the formation of calcium phosphates in phosphoric acid-based geopolymers can reduce the amount of gel formed (36), resulting in lower compressive strength, which may explain the acceptable strength values obtained in the present study.

FTIR spectra are represented in Figure 7. The bands at 3411 cm⁻¹ and 3485 cm⁻¹ in the spectra of both types of geopolymers (5 and 8 M H₃PO₄) are associated with the stretching vibrations of the H-O-H bonds (2). The bands observed at 1650 cm⁻¹ and 1649 cm⁻¹ correspond to the deformation vibrations of O-H bonds. For the geopolymers activated with 5M H₃PO₄, the presence of a broad band with two shoulder peaks at 1135 cm⁻¹ and 1071 cm⁻¹ indicates the presence of P-O bonds, characteristic of brushite (37). This is consistent with the XRD results, where brushite was noticeable in GP-5M but not in GP-8M. For the geopolymers activated with 8M H₃PO₄, the band shifts to a lower wavenumber (around 1081 cm⁻¹) and is associated with the formation of P-O-Al bonds (38). It is also important to note that these broad bands in both types of geopolymers could result from overlapping P-O bonds not only in crystalline phases, but also in amorphous phase that cannot be distinguished using XRD or FTIR. In addition, the bands around 976 cm⁻¹ and 877 cm⁻¹ are characteristic of P-OH bonds in newberyite, present in both types of geopolymers (39). The bands at 790 cm⁻¹ and 794 cm⁻¹ in the spectra of GP-5M and GP-8M, respectively, are characteristic of quartz, indicating that it was not by the acid activation (40). The band around 532 cm⁻¹ can be attributed to the vibration modes of Al-O-P in (AlPO₄·H₂O) (41). Lastly, the band at 470 cm⁻¹ is attributed to the vibration modes of Si-O-Si (42).

Since there is a tendency for the compressive strength values of GP-8M to increase with increasing sodium aluminate content, they were selected for further study. MIP analyses were performed on GP-8M and the results are shown in Figure 8. As expected for these acid geopolymers, the total porosity in the reference system was quite high (See Figure 8(a)). With increasing amounts of NaAlO₂, the porosity decreases from 39.29% for the GP with 5% NaAlO₂ to 38.12% with 10% NaAlO₂ and to 34.09% with 15% NaAlO₂. In addition, there are changes in the pore size distribution towards smaller pore size diameter (See Figure 8(b). It can be observed that higher amounts of NaAlO₂ result in smaller pore diameter. Specifically, pore volume in the range of 10-100 microns decreases, while those in the range of 0.01-0.1 microns increase. The results obtained are in agreement with the compressive strength values (See Figure 5).

Geopolymers with high compressive strength values were selected as representative materials to be analyzed by BSEM after 28 days. The micrograph of the geopolymer containing 15% sodium aluminate and activated with 8M is shown in Figure 9. The microstructure shows a dense matrix where several phases can be distinguished. The same phases previously identified by XRD were also identified by EDS at various points, including quartz, monetite (CaHPO₄), aluminum phosphate (AlPO₄), and newberyite (Mg(PO₃OH)·3H₂O) were observed. The presence of unreacted remaining clay is also evident in the system, as some particles based on Si and Al are still present. In addition, the identification of a new phase can be distinguished, which could be related to the formation of an amorphous gel containing Si, Al, Ca, and P that can co-precipitate with the crystalline phases, in the acid system.

In order to explore the nanostructure of the reaction products, ²⁷Al, ²⁹Si, and ³¹P MAS NMR analyses were conducted in a selected sample (geopolymers containing 15% NaAlO₂). In the same figure, the ²⁷Al and the ²⁹Si MAS NMR spectra of the thermally treated clay have been including for comparative analysis. ²⁹Si NMR spectrum (see Figures 10(a) and 10(b)), shows the presence of one narrow peaks appearing at -72.8 ppm, that is attributed to the Q¹ sites of remaining gehlenite, (3). The identification of the peaks ranging from -82 ppm to -105 ppm is very challenging, and probably are the result of the overlapping of different Q³(nAl)/Q⁴(nAl) environments, that could correspond with unreacted particles clay’s but also with the amorphous phase previously identified by BSEM. Additionally, the presence of quartz and tridymite (observed in XRD) is characterized by the two resonances at -109.1 ppm and -113.5 ppm, respectively (3). ²⁷Al NMR spectra of the geopolymer (GP-3) show resonances around +47.17, +8.11, and -13.88 ppm (Figure 10-c). The two first peaks are attributed to Al(IV) and Al(VI), respectively in the Si-O-Al-ordered structure (26). According to previous studies, the new peak at -13.88 ppm and the wide hump between +20 and +60 ppm are attributed to Al(VI)-OP and Al(IV)-OP structures coordinating with -O-Si and -OH to form the octahedral and tetrahedral structures (Al(VI)(OP)_x(OSi)_y(OH)_6−x−y) and (Al(IV)(OP)_x(OSi)_y(OH)_4−x−y) (43). Figure 10-d shows the deconvoluted ³¹P NMR spectra of the sample GP-6 where several peaks can be distinguished. According to the literature, peaks at -0.3 and -1.5 ppm are attributed to the presence of monetite (CaHPO4) (44). In the present study, two distinct peaks appeared at about -0.15 and -1.6 ppm, which could confirm the presence of monetite, previously identified by XRD. The signal at -7.4 ppm is attributed to the presence of newberyite (45). At about -13.9 ppm, the broad peak confirms the presence of aluminum phosphate (46). The formation of AlPO₄ can be classified into three types: surface precipitates (as in the case of aluminum phosphate resonating at -13 ppm), inner-sphere surface complexes where the resonance appears between -11 ppm and 0 ppm (as indicated by the bands appearing at -4.76 and -9.07 ppm), and outer-sphere types resonating between 0 ppm and 10 ppm (46). According to a previous study, the phosphorus environment of Q⁰, generated by the orthophosphate group, corresponds to +2 ppm (13). Thus, the presence of peaks at about +1.7 and +2.8 ppm is attributed to the phosphorus environment of Q⁰ (that could correspond with the formation of an amorphous phase containing P).