Introduction

Natural phosphorus compounds are of crucial importance for the production of phosphate fertilizers which are necessary for global agriculture (Becker, 1989; Toama, 2017). Two main geological origins exist for phosphate ores: sediment and igneous (Becker, 1989). H₃PO₄ is an intermediate chemical substance. Several methods were developed to produce H₃PO₄ (Becker, 1989; Kthiri et al., 2022). The most widely used method is the wet process, due to its low price and energy consumption (Delcea & Siserman, 2021; Delcea et al., 2023c). Natural phosphate is dissolved by mineral acids, which could contaminate the product with several heavy metals (Delcea et al., 2023a, 2023b). Nitric acid, sulfuric acid, or hydrochloric acid are used to dissolve phosphate. In terms of volume, wet digestion of phosphate rock using H₂SO₄ is the preferred method (European Commission, 2007).

Phosphoric acid can be used to make phosphate fertilizers and manufacture food additives due to bond purity regulations, detergents for the pharmaceutical industry, paints, and cosmetics (Delcea et al., 2023b). It is also used for the treatment of wastewater, pickling, polishing, and antirusting of metal surfaces (Gonzalez et al., 2002). However, heavy metals and other dissolved impurities in phosphate material could contaminate the final product during the production of phosphoric acid. Therefore, these heavy metals may contaminate phosphorus fertilizers which would then spread to the soil and atmosphere. The growing environmental levels of heavy metals pose a major hazard to human health, living things, and ecological systems (Inglezakis et al., 2018; Saleh & Ali, 2018). The purification of phosphoric acids is so necessary. Thus, the design of an effective technique to eliminate impurities that contaminate H₃PO₄ is necessary for the sustainable production of phosphate fertilizers from wet H₃PO₄. The presence of contaminants including organic waste and heavy metals that were initially present in phosphate rocks is one of the main downsides of wet H₃PO₄ manufacturing. Indeed, cadmium is a minor component in phosphate rock. It is a contaminant. Depending on the deposit, ores containing between 10 and 80 g/ton cadmium may be discovered in both phosphoric acid and phosphate fertilizers. The amount of cadmium brought by phosphate fertilizers depends on the origin of the phosphates used during manufacturing. The World Health Organization (WHO) describes cadmium as a dangerous element that may build up in soils and plants, leak into ground and surface waters, enhance crop absorption, increase levels in animals and food, and could be harmful to human health (Zhao et al., 2017).

Calcination is an efficient way to remove cadmium from raw phosphate, but it needs temperatures between 850°C and 1150°C. It is not thought to be cost-effective (Boughzala & Hidouri, 2020). Several techniques, such as solvent extraction, adsorption, ion exchange, precipitation, flotation, and membrane processes, have been developed to remove cadmium from phosphoric acid solutions. Magnesium ion, found in H₃PO₄, causes a variety of issues such as worsening acid viscosity, raising operational complexity, operating costs, and decreasing fertilizer water solubility (Boughzala & Hidouri, 2020; Bembli et al., 2022). Therefore, Mg²⁺ ion removal from H₃PO₄ will be crucial for practical production.

The present work focuses on the use of abundant natural and H₂SO₄-activated clays to remove cadmium and magnesium ions from a phosphoric acid solution. The solid materials were characterized using Elemental analysis, SEM, FT-IR spectroscopy, XRD and BET. The adsorption process was assessed by varying experimental conditions. The kinetic and isotherm data were modeled using common theoretical kinetic and isotherm equations.

Materials and Methods

Natural Phosphate Preparation

Oum El Khecheb (Gafsa–Metlaoui) deposit is an old mine in the South of Tunisia. The most mineralized and possibly lucrative strata, according to the mineralogical investigation, are layers CII, CV, CVI, and CVIII. Despite having relatively high amounts of MgO and Cd, analyses of the phosphate from CII they were revealed that the deposit has a major economic interest due to its high proportion of P₂O₅. Phosphate was subjected to crushing, washing, and desliming to obtain fine particles. Particles have been thoroughly washed with distilled water before use and then dried in an oven at 105°C.

Preparation of Activated Clays

Different types of clays, collected from Nafta and Tozeur regions (south of Tunisia) were studied in this work. To prepare activated clays, the minerals were subjected to chemical treatments in 5 M sulphuric acid (>95% Prolabo) (10 g/500 mL) at 80°C for 2 h under magnetic stirring. After, the clays were washed with deionized water until the water became neutral. Then, the activated minerals were dried in a drying oven at 105°C. In the following sections, the mineral clays collected from Nafta, Tozeur, and acid-treated clays of Nafta will be labeled as RBC, ABC, and ACHTC (Ta), respectively.

Production of Phosphoric Acid

As previously said, the majority of phosphate rock is converted into phosphoric acid via a wet method. Apatite, which makes up the majority of phosphate rock, reacts with concentrated H₂SO₄ to produce phosphoric acid and an insoluble calcium sulfate salt. This process facilitates the direct filtration separation of phosphoric acid.

The following reaction shows the entire H₂SO₄ attack for producing phosphoric acid (Toama, 2017).

Ca₁₀(PO₄)₆F₂ + 10H₂SO₄ + 20H₂O → 6H₃PO₄ + 10CaSO₄. 2H₂O + 2HF

A variety of phosphate rocks can be processed using the dihydrate method, and the weak acid comprises 26–30% P₂O₅. This weak acid is then evaporated to a concentration of 50–54% to fulfill the needs of fertilizer manufacturers.

Adsorption Experiments

The removal of cadmium and magnesium ions was carried out using natural and acid-activated clays as adsorbents. The adsorbent dosage was varied from 0.02 and 0.1g. The adsorption of different ions was performed in a batch mode, at room temperature, shaking speed of 150 rpm, using 20 mL of phosphoric acid solution. Ions concentrations were prepared by the dilution of stock solution

The adsorbed quantity (q_e, mg/g) was determined using the following equation:

V is the utilized solution volume (L), m is the mass of the clay material (mg), and C₀ and C_e correspond to the initial and equilibrium concentrations of Cd or Mg (mg/L), respectively.

Characterization

An atomic absorption spectroscopy (Perkin-Elmer 3110) was used to determine the chemical elements in natural phosphate as well as contaminants in phosphoric acid. The chemical compositions of the natural and activated clays were determined by X-ray fluorescence spectrometry (X-ray fluorescence spectrometer, Philips brand, and model PW-2400 equipped with a 3Wh power tube). An X'Pert Pro, PANalytical diffractometer running on Cu K radiation was used to perform powder XRD analysis. Using the information provided in the ASTM cards, the mineral phases were identified. A Perkin-Elmer 1283 spectrometer and the KBr pellet method were used to obtain Fourier transform infrared (FTIR) spectra. Using a Seteram Instrumentation SETSYS evolution system with a heating rate of 10°C/min up to 1000°C in an air atmosphere, thermal gravimetric analysis was carried out. Scanner Electronic Microscopy (SEM, FEI Quanta 200) was used to analyze the samples' surface morphology. BET (Brunauer-Emmett-Teller) equation was used to derive the specific surface area values from the nitrogen adsorption isotherms.

Results and Discussion

Characterization of Natural and Acid-Treated Clays

The chemical analysis of the studied clays. The main constituents of these samples are SiO₂ (43.64 - 51.24 %), and Al₂O₃ (19.92 – 31.24%). Other elements are also found such as CaO (0.92 - 3.78 %), Fe₂O₃ (4.92-8.34 %), sodium oxide (0.25 – 1.47 %), and MgO (0.01 – 2.34 %). A high CaO content (3.78 %) is present in clay ACHTC (Ta). The clay samples exhibit a low percentage of MnO (0.04 - 0.13 %).

The specific surface area values of the studied samples range from 74.25.35 to 85.14 m²/g. These values prove that these materials are non-porous. This trend aligns with the results of some other natural clay minerals studied in the literature (Bembli et al., 2022, Sboui et al., 2022).

The size and the shape vary and depend on the nature of the clay. Overall, the morphological behavior of all samples showed a clear resemblance. The agglomerates exhibit irregular shapes and small sizes ranging from 1 to 3 µm. These findings suggest that the morphology of clay powders influences the adsorption mechanism, which is suggested to be more favored when the powders are formed from more developed agglomerates. The diffractograms show the presence of kaolinite, quartz, goethite, and hematite. All samples present many phases, with quartz being the most prevalent one. This mineral is characterized by their basal reflections at 28.04 and 49.14°, respectively.

The results show, in addition to the above-mentioned constituents detected by XRD, the presence of a small amount of organic matter, in the form of compounds of aliphatic chains, characterized by the weak bands located at approximately 2920 and 2843 cm^-1. The deformation vibrations of Si-O-Mg and C-O bonds in carbonates are observed at 539 and 1450 cm^-1, respectively.

The bands at 3620-915 cm^-1 show the predominance of dioctahedral smectite with bands of [Al, Al-OH] (European Commission, 2007). The adsorption bands at 3430-1635 cm^-1, which correspond to the adsorbed water molecule on the clay surface, are assigned to the OH frequencies. The strong absorption bands observed at 1100 and 1000 cm^-1 are attributed to the silicate structure.

Effect of Experimental Parameters on the Adsorption of Mg and Cd Ions from Phosphoric Acid

Table 1 displays the contents of P₂O₅, MgO, and Cd in diluted and concentrated H₃PO₄. Results reveal that the content of these elements is more important in the case of the diluted acid compared to the concentrated one.

Table 1. Contents of P₂O₅, MgO, and Cd in diluted and concentrated H₃PO₄

The effect of the adsorbent dosage on the adsorption of Mg and Cd ions is given in Figure 1. As it is observed, when the mass of the clay increases from 0 to 0.08 g, the adsorption efficiency of Cd ions for RBC clay increases from 0% to 57.6%. Further mass increase, (from 0.08 g to 0.1g), no effect is observed. However, the adsorption efficiencies of Cd ions for ABC and ACHTC (Ta) clays are 46.3% and 31.3%, respectively (Figure 1a). At optimum masses of clays, the adsorption of Mg ions for ABC, RBC, and ACHTC (Ta) are 46.5%, 28.5, and 16.7%, respectively (Figure 1b).

The effect of contact time on the adsorption of Cd and Mg ions using the studied clays is shown in Figure 2. As contact duration increases, the adsorption of Cd and Mg ions increases and eventually reaches a plateau after about 20 min for the Cd and Mg ions. This phenomenon can be interpreted by the fact that there were a lot of sites accessible for Cd and Mg metals at the beginning of the reaction. Near equilibrium, there are fewer sites available, so difficult to access. Additionally, the rate of adsorption is reduced by the repulsive forces between the cations and the solid as well as those in the solution (Wang et al., 2015; Boughzala & Hidouri, 2020; Gmati et al., 2022; Bembli et al., 2024).

Figure 3 shows the change in the adsorbed capacity of Cd and Mg ions, as a function of phosphoric acid concentration, in the presence of the studied clays. It is observed that, at lower phosphoric acid concentrations, the adsorbed quantities of Cd and Mg ions increase significantly as H₃PO₄ concentration increases and further reaches stable values. This is assigned to the saturation of accessible sites on the surface of the studied clays at high H₃PO₄ concentration. For the RBC clay sample, the maximum adsorption capacity of Cd ions is 17.8 mg/g. However, the maximum adsorbed quantities of Cd ions for ABC and ACHTC (Ta) samples are 15.2 mg/g, and 11.8 mg/g, respectively. This difference in adsorption capacity is related to the change in the structural composition of the studied clays.

Similarly, Figure 3b gives the variation of the adsorbed quantity of Mg ions versus H₃PO₄ concentration. The adsorbed amounts of Mg ions for ACHTC (Ta) and RBC are 0.35 mg/g and 0.51 mg/g, respectively. The adsorption capacity is more important for the RBC clay sample, (0.92 mg/g). Indeed, the changes in the materials' surfaces and the availability of binding sites for the cadmium and magnesium metals are responsible for this difference in the adsorption efficiency.

Kinetic Modeling

The pseudo-first-order, pseudo-second-order, Elovich, and intra-particle diffusion models are used to simulate the equilibrium data (Ho & McKay, 1999; Bezzi et al., 2001; Ho et al., 2005; Bembli et al., 2021). Table 2 summarizes the Kinetic parameters of Cd and Mg ions adsorption onto clay samples.

Table 2. Kinetic parameters of Cd and Mg ions adsorption onto clay samples

As is shown, the highest regression coefficient values (0.99) are observed for the pseudo-second-order equation. In addition, the calculated adsorption capacities are close to those determined theoretically. Therefore, it is suggested that the adsorption of Cd and Mg ions on surface clays is chemical (Ho & McKay, 1999; Doğan & Alkan, 2003). Moreover, it was observed that the plots did not pass through the origin when the kinetic data was simulated using the intra-particular diffusion model, showing that intra-particular diffusion is not the only rate-limiting process for the two studied metal ions (R² values are less than 0.68). It was shown that the bigger the intercept, the more the surface adsorption contributed to the rate-limiting step. The C parameter depicts the boundary layer effect or surface adsorption.

Isotherms Study

To better understand the adsorption mechanism, the adsorption equilibrium data of Cd and Mg ions in the presence of the studied clays were simulated using Langmuir, Freundlich, Temkin, and Dubinin-Radushkevich models (Freundlich, 1906; Langmuir, 1918; Al-Ghouti et al., 2005; Al-Ghouti & Da'ana, 2020). Table 3 summarizes the constant parameters for the different studied models.

Table 3. Relative constant parameters of Langmuir, Freundlich, Temkin, and Dubinin–Radushkevich models

The highest regression coefficient values are observed for both the Freundlich and Temkin equations. However, the adsorption of Mg could also follow the Langmuir model (Boughzala et al., 2020). The surface of the adsorbents is expected to undergo multilayer adsorption. For the Freundlich model, the values of n are higher than 1, indicating favorable adsorption. The Temkin constant values, B_T related to the heat of sorption, varied between 137 J/mol and 149 J/mol, in the case of the adsorption of Cd ions which suggests physical adsorption. These constant values (3364 J/mol - 9027 J/mol) are much greater than those calculated for Cd ions. This suggests that the interaction between the clays and Mg ions is chemical (Al-Ghouti & Da'ana, 2020). Also, the high R² values observed for Radushkevich isotherm may suggest that the adsorption mechanism onto heterogeneous surfaces of the adsorbents varies with a Gaussian energy distribution (Saleh & Ali, 2018; Bembli et al., 2022). The results from the middle range of concentrations and high solute activity have frequently been well fit by the model. As globally observed, the adsorption phenomenon remains complex, and physical and chemical interactions are involved.

Comparison of the Removal Performance with Other Reported Techniques

To describe the performance of the current investigated technique, a comparison with other methods described in the literature was assessed. Table 4 gives a comparison of the adsorption capacities of Cd²⁺ and Mg²⁺ ions with other reported methods in the literature. For instance, the removal efficiency of Cd²⁺ ions is considered significant compared to the literature (Yu & Liu, 2010; Khamar et al., 2021; Mahrou et al., 2021; Kouzbour et al., 2022; Samrane & Bouhaouss, 2022; Es-Said et al., 2023; Marszałek et al., 2023) and it is more important compared to the method utilizing precipitation by birnessite-type Na-MnO₂ (Amarray et al., 2023).

Table 4. Comparison of the adsorption capacities of Cd²⁺ and Mg²⁺ ions with other reported methods in the literature