X-Ray diffraction

The compounds present in the Mexican clay were determined through X-Ray Diffraction and through diffractograms on a Bruker brand D8 Advance diffractometer using Cu-Ka radiation (λ= 0.15406 nm) with a range of 10-80° in 2θ. The particle sizing was performed using the Scherrer equation. For the analysis of the sample there was no need for special preparation, since it is not required for the X-ray diffractometer.

X-Ray fluorescence

For X-ray fluorescence there were two types of sample preparation: First. Loose powder. The clay was ground into samples of sizes less than 60 micrometers. The powders were placed in a glass for liquid samples, using a 6-micron thick Mylar film as a carrier with an area of analysis of approximately 30 mm in diameter. Second. Cast bars. The samples were analyzed for their loss on ignition (LOI) by melting 1 g of sample in 6 grams of lithium tetraborate flux (Li₂B₄O₇), to ensure that flux to sample ratios reflects only non-volatile elements that will be present after fusion (^{Zipkin, Ambrose, Lundstrom, Bartov, Dwyer & Taylor, 2020}). To analyze this process, an Axios brand advanced model with analytical PAN wavelength sequential spectrometer equipped with an X-ray tube containing Rhodium anode was used. For analysis, a sequential spectrometer wavelength PAN analytical brand Axios Advanced model with X-ray tube with Rhodium anode was used.

Nitrogen physisorption

By Nitrogen physisorption the textural properties of the Mexican Clay were determined from the Nitrogen adsorption isotherms recorded at 77 K with a Quantachrome iQ2 apparatus. The samples were previously degassed at 423 K for 24 h under a vacuum (10-4 mbar) to ensure a clean, dry surface, free of any loosely bound adsorbed species. The specific area of the sample was calculated according to standard BET procedure using nitrogen adsorption data collected in the relative equilibrium pressure interval of 0.03 < P/P0< 0.3. Pore size distributions were calculated from the desorption branches of the corresponding nitrogen isotherm using the BJH method (^{Betancourth, Gómez, Mosquera & Mejía, 2010}). The total pore volume (V _total) was estimated from the amount of nitrogen adsorbed at a relative pressure of 0.99.

Membrane preparation

Mexico is facing water scarcity problems that will increase in the coming years. The selected site in Querétaro (La Solana), where sampling and characterization were done, is a suitable place to collect clay for membrane preparation due to its low price and availability (industrial zone of San Nicolás) as well as its composition (silico-aluminates), which allows chemical affinity. For membrane preparation using the pressing technique, a series of steps were required to obtain a proper membrane with suitable physical and chemical properties. First, milling and sieving of the clay was performed. Followed by the selection of the binder, selection of the lubricant, pressing, heat treatment, and finally the membrane was obtained. The mold used was designed with flat type steel using the following dimensions: length = 2 cm, outer diameter = 5 cm, and inner diameter = 3 cm. The membrane preparation technique that was developed in this study was through compaction and subsequent sintering of the clay. In the pressing of the clay, additives such as lubricants and binders were necessary for obtaining a proper membrane. To evaluate the efficiency of the membranes in reducing water hardness, acrylic reactor was constructed. In this reactor, the membrane was mounted in the central point using an O-ring to avoid leakage.

Scanning electron microscopy (SEM)

For scanning electron microscopy, a JEOL-JSM-7800F was used. It was necessary to split the membrane and mount it in the sample holder, which was covered with a layer of graphite (because the sample is not electrically conductive). The accelerating voltage used in the microscope was 20 kV and the images were formed from secondary electrons. The samples were coated by an amorphous carbon thin film to avoid electrostatic charge accumulation using an EMS 150T sputter. Finally, the sample holder was placed in the scanning electron microscope to be observed.

Membrane evaluation

Figure 1 shows the diagram used in the study to filtrate the water. To evaluate the membrane, three separate readings for each sample and blank were taken. The method consisted of an assay using a visual indicator endpoint, the eriochrome black T, which is red in the presence of Calcium and Magnesium and turns blue when they are bound or absent. The analysis of water hardness was conducted in accordance with the provisions of the ^{Official Mexican Regulations and Standards NMX-AA-072-SCFI-2001}. This draft Mexican standard establishes the analysis method for determining total hardness in natural and wastewater. Finally, using the membrane obtained the analysis of water samples and quantitative calculations were made as established by the Official Mexican Regulations and Standards. The pump used was a ProMinent, Model GALA1801PPB200UD012000 (Heidelberg - Germany).

Figure 1 Diagram of the equipment used in the study. 

Total hardness expressed as where:

A are the mL of EDTA titration spent on the sample;

B mL of EDTA spent on the titration in the blank;

C mg CaCO₃ equivalent to 1 mL of EDTA;

D mL of the sample.

Mexican clay characterization

X-Ray diffraction

The raw material was characterized by X-ray diffraction at four different temperatures (25, 500, 800, and 1,000 ° C). The Mexican clay used in this study was basically composed of feldspars [K (AlSi₃O₈)] corresponding to the series of plagioclase mainly andesine. It also contains Fe₂O₃ and SiO₂ (tridymite and quartz) (Figure 2). The heat treatment causes a phase transition from andesine to anorthite. At high temperatures the formation of anorthite is favored, but it depends on the amount of Calcium or Sodium present in the sample. Moreover, there is a phase transition due to heat treatment with SiO₂ minerals. Quartz has a higher thermodynamic temperature stability at low temperatures, (present in the sample at 25 and 500 °C), and it transformed to tridymite at high temperatures (Figure 3).

Figure 2 XRD spectrum of Mexican clay (25 °C). 

Figure 3 XRD spectrum of Mexican clay (500 °C). 

The heat treatment caused the removal of some amorphous compounds, possibly of an organic nature (^{Lahsini, Bentama, Addaou & Rafiq, 1998}). This is observed in the change of the shape of the base line of the diffractograms at 25 °C and 500 °C. In these diffractograms, few curves were observed between 10-15° and 20-30° in 2θ, and similarly were not present in the diffractograms of 800 °C and 1,000 °C. The heating phase transitions originated in the andesite state, which then transformed into anorthite, however, Quartz, became tridymite. In Potassium feldspar and Iron oxide compounds, no changes occurred due to heat treatment (Figure 4 and Figure 5).

Figure 4 XRD spectrum of Mexican clay (800 °C). 

Figure 5 XRD spectrum of Mexican clay (1,000°C). 

Nitrogen physisorption

The N₂ adsorption-desorption isotherms of the Mexican clay corresponded to the type II isotherms according to IUPAC classification (^{Zelenka, Horikawa & Do, 2023}). The isotherm was reversible at low relative pressures and hysteresis loop at higher relative pressures, which corresponded to the H3 type hysteresis loop. This type of hysteresis loop is characteristic for aggregated and agglomerated plate particles forming slit shape pores (^{Vuković, Milutinović- Nikolić, Krstić, Abu-Rabi, Novaković & Jovanović, 2005}). The enlargement of the hysteresis loop implies no uniformity in shape and size of the pores (^{Gelves, Monroy, Sánchez & Ramírez, 2013}). The clay exhibited a unimodal distribution; the average pore diameter was 3.42 nm, and the specific surface area was 48.801 m²/g. In Figure 6 adsorption isotherm types are showed (^{Zelenka et al., 2023}).

Figure 6 Adsorption isotherm types. 

Membrane preparation

The technique developed during this study for the preparation of the membrane was through the pressing the raw material. Posteriorly, the membrane was treated at different temperatures allowing it to harden to obtain the final membrane (^{Pérez-Moreno, Arellano & Ramirez, 2004}). First, the Mexican clay was crushed and screened. During the pressing of the raw material (clay), it was necessary to take additives into consideration, such as choosing adequate lubricants and binders (stearic acid and alumina). When additives were not used, the membranes broke easily. However, when stearic acid and alumina were used, good textural property membranes were obtained. The global density was determined against the uniaxial pressure for the combination of stearic acid and alumina with the clay. The best properties for obtaining the membranes (membranes without rupture problems) were found using stearic acid (5%) and alumina (15%), taking into consideration the influence of the clay granule (300 μm of size particles). ^{Anbri et al. (2008)} reported 56% of silico-aluminate in the clay of Meknes (Morocco), the same percentage as the one reported in the present study. Also, ^{Bentama et al. (2002)} reported the same average of silico-aluminate in Fés (Morocco).

Membrane characterization

X-Ray diffraction

Figure 7 shows the diffractogram of the membrane made of Mexican clay and shows that the reflections present are consistent with the applicable standards of the following phases: Sodium anorthite, potash feldspar [K (AlSi ₃O₈)] and two polymorphs of SiO₂ (quartz and tridymite). This indicates that the nature of the sample is fiery in nature. They are composed of mostly tectosilicates group, plagioclase and orthoclase. The presence of the mixture of quartz and tridymite is because the sample was thermally treated and moderately cooled.

Figure 7 XRD spectrum of membrane made of Mexican clay. 

Scanning electron microscopy (SEM)

In Figure 8, the membrane made from Mexican clay (80%), alumina (15%), stearic acid (5%), compacted to 300 Kg/cm³ and sintered at 800 °C is shown. In the figure crystals presented are pooled. The pores are estimated to be 1 µm. This is in accordance with the results of the X-ray diffraction performed at different temperatures with clay calcination. The material undergoes changes in its structure and therefore its morphology. Pérez-Moreno et al. (2004) show that as the sintering temperature increases the pore size decreases, thus the temperature of 800 °C used in this study was suitable for the preparation of membranes.

Figure 8 Scanning electron microscopy of the membrane made of Mexican clay. 

Nitrogen physisorption

The adsorption-desorption isotherms of Nitrogen of the Mexican clay present in the main formulation of the membrane are shown in Figure 9. The curve corresponded to the type II isotherms according to the IUPAC classification. The isotherm had a reversible relative low-pressure part and the hysteresis loop at high relative pressures corresponding to an H3 type hysteresis loop. This type of hysteresis loop is characteristic for aggregated and agglomerated plate particles forming slit shape pores (^{Rouquerol et al., 1994}). The elongation of the hysteresis loop implied the presence of non-uniform pore shape and size. The clay exhibited a unimodal distribution; the average pore diameter is 3.408 nm, total pore volume of de 0.096 cm³/g and the specific surface area was 55.922 m²/g. The specific areas of the samples were calculated by applying the BET method to the adsorption branch of the nitrogen isotherm within the 0.005 <P/P0< 0.25 range. The pore size distribution and cumulative pore volume were obtained by the Barret-Joyner-Halenda method (BJH) and from the isotherms at P/P0 = 0.99, respectively. This slight increase in the surface area in the membrane compared to clay is due to the generation of textural porosity (interparticles), during the process of sintering in the membrane preparation (^{Bouzerara, Harabi, Achour & Larbot, 2006}; ^{Saffaj, Persin, Younsi, Albizane, Cretin & Larbot, 2006}; ^{Yoshino, Suzuki, Nair, Taguchi & Itoh, 2005}).

Figure 9 Nitrogen physisorption of membrane made of Mexican clay. 

Membrane evaluation

The membrane was washed for 24 h in a solution of 10% hydrochloric acid then rinsed with distilled water, then further washed for 30 min in a solution of 10% sodium hydroxide and finally rinsed for 30 min with distilled water.

After the membrane was washed, it was placed in the center of two acrylic molds and then applying a pressure gradient, water was passed through it. The flux was 28 mL/min.; the membrane area was 7.07 cm² and the transmembrane pressure of 1.1 bar. The water was assessed before and after passing through the reactor containing the membrane to determine the total hardness by means of the technique specified by Official Mexican Regulations and Standards (^{NMX-AA-072-SCFI-2001}). According to this procedure the initial hardness of the water was 280 mg/L of CaCO₃. After passing the water through the reactor containing the membrane, the total hardness decreased by over 40%. This is shown in Figure 10 where the graphical evaluation is presented of how hardness in water decreased using the membrane elaborated in this study.

Figure 10 Graphical evaluation of hardness in water of membrane made of Mexican clay.