Eco-Friendly Ceramic Membranes from Natural Clay and Almond Shell Waste for the Removal of Dyes and Drugs from Wastewater

Abstract

This study investigates the influence of sintering temperature (850–950 °C) and almond shell content (2–10 wt.%) on the structural, mechanical, and functional properties of natural-clay-based ceramic membranes. Several membranes were prepared by incorporating different proportions of almond shell powder and 2 wt.% lime as additives and sintered under controlled thermal conditions to optimize their performance. The results demonstrate that both sintering temperature and almond shell content significantly affect membrane porosity, mechanical strength, and water permeability. Among all of the tested samples, the membrane designated MP2-900, composed of natural clay, 2 wt.% almond shell powder, and 2 wt.% lime, sintered at 900 °C, exhibited the most balanced performance. It showed high mechanical strength (≈28 MPa), low shrinkage (<5%), and good water permeability (35 L·h⁻¹·m⁻²·bar⁻¹). When tested for the removal of crystal violet (CV) dye and paracetamol (PCT) from synthetic wastewater, the MP2-900 membrane achieved a removal efficiency of 87% for both pollutants. Overall, the MP2-900 membrane represents the optimal configuration, providing an excellent balance between mechanical robustness, porosity, and separation performance. These findings highlight the potential of sustainable clay-based ceramic membranes derived from agricultural by-products for the efficient removal of recalcitrant pollutants from wastewater.

1. Introduction

The continuous deterioration of surface water quality, mainly due to the discharge of untreated industrial effluents and the accumulation of various pollutants, has led to increasing concentrations of toxic substances in aquatic ecosystems. Among these, emerging contaminants such as dyes and pharmaceuticals pose serious environmental and public health concerns. These pollutants, introduced through both anthropogenic activities and natural processes, are often found at levels exceeding regulatory limits. Due to their high chemical stability, many of them are resistant to biodegradation, persist in aquatic environments, and bioaccumulate through the food chain [1]. These challenges underscore the pressing need for efficient, robust, and sustainable water treatment technologies that can effectively remove chemically stable and persistent organic contaminants.

In recent years, membrane separation has emerged as a promising approach for sustainable water purification, offering high removal efficiency, operational simplicity, and low energy consumption [2,3]. Despite their compactness, selectivity, and environmental advantages, the performance and long-term viability of these systems remain strongly influenced by the membrane material properties and associated production costs [4,5,6]. However, the effectiveness and sustainability of membrane-based processes strongly depend on the nature of the membrane material and the feasibility of its large-scale production.

Inorganic ceramic membranes present significant advantages over conventional polymeric membranes, including excellent chemical, thermal, and mechanical stability, high permeability, and strong fouling resistance, making them particularly suitable for advanced water treatment applications [7,8,9]. However, their large-scale use remains limited by high production costs, mainly due to the price of raw materials and the elevated sintering temperatures required. Conventional metal oxide membranes such as alumina (Al₂O₃), titania (TiO₂), zirconia (ZrO₂), and silica (SiO₂) exhibit excellent performance but must be sintered at 1400–1600 °C, making fabrication energy-intensive and costly [10,11,12]. These limitations have stimulated increasing interest in alternative ceramic membrane materials that can maintain high performance while significantly reducing fabrication costs and energy consumption.

To address these economic and environmental challenges, recent research has focused on the use of abundant and low-cost natural clays—such as kaolin, smectite, bentonite, zeolite, and sepiolite—as alternative raw materials for ceramic membrane fabrication [9,13]. These clay-based membranes have demonstrated comparable separation performance while markedly reducing production costs and energy consumption [6,9,13,14,15,16]. Nevertheless, most reported studies still rely on significant amounts of organic additives and relatively high sintering temperatures, which partially offset the economic and environmental benefits of clay-based ceramic membranes.

Building on these recent developments, the present study focuses on the fabrication of eco-friendly flat ceramic membranes using natural kaolin combined with almond shell powder and lime as affordable, locally available raw materials. To the best of our knowledge, the combined use of kaolin, almond shell powder, and lime for the fabrication of flat ceramic membranes has been rarely explored, particularly in the context of sustainable water treatment. The proposed formulation aims to minimize the use of organic additives (e.g., plasticizers and binders), which can release volatile organic compounds during sintering. Additionally, the incorporation of almond shell powder and lime is expected to lower the sintering temperature, thereby reducing both energy demand and overall production costs. This approach directly addresses the need for low-cost, low-energy, and environmentally benign ceramic membrane fabrication routes.

In this context, the present study aims to develop eco-friendly flat ceramic membranes based on natural kaolin, almond shell powder, and lime, and to evaluate their structural properties and separation performance. The simultaneous removal of a cationic dye (crystal violet) and a pharmaceutical compound (paracetamol) is investigated to demonstrate the versatility and applicability of the developed membranes for treating complex organic pollutants.

2. Materials and Methods

2.1. Materials

Flat composite ceramic membranes were fabricated using natural clay sourced from northwestern Algeria. The clay consists of 42.96 wt.% silica (SiO₂) and 37.7 wt.% alumina (Al₂O₃), with smaller amounts of TiO₂, MgO, Fe₂O₃, CaO, and K₂O. The physicochemical properties of this clay have been detailed in a previous study [17]. Lime (CaO) and almond shells were incorporated as additives in the membrane formulation. Lime was produced locally by calcining limestone in a traditional kiln. Almond shells, collected from agricultural waste, served as a natural pore-forming agent. Before use, the almond shells were washed with distilled water, dried, and ground into a fine powder [18]. All raw powders were sieved through a 100 μm mesh to ensure uniform particle size prior to mixing.

Scheme 1 presents the chemical structures of Crystal Violet (CV) and Paracetamol (PCT), highlighting their distinct molecular features:

Crystal Violet (CV) is a cationic triphenylmethane dye widely used in textile, biological, and analytical applications. In aqueous solution, it predominantly exists in its monocationic form (CV⁺), where the positive charge is delocalized over the aromatic structure through resonance involving the dimethylamino substituents. This cationic nature strongly influences its interaction with charged membrane surfaces, particularly under conditions where the membrane exhibits a negative surface charge, thus enhancing electrostatic attraction and retention.

Paracetamol (PCT), also known as acetaminophen, is a widely used pharmaceutical analgesic and antipyretic. Structurally, it is a neutral organic molecule consisting of a para-aminophenol core with an acetamide functional group. Unlike cationic dyes such as CV, paracetamol does not carry a net charge at neutral pH, and its removal during filtration processes is mainly governed by adsorption mechanisms, hydrogen bonding, and size-exclusion phenomena rather than electrostatic interactions.

2.2. Preparation of MK Membranes

To investigate the influence of mixed clay, lime, and almond shell content on the properties of the ceramic membranes, four different composite membranes were prepared under identical processing conditions (

Table 1. Composition of the prepared membranes.

Membrane	Clay (wt.%)	Lime (wt.%)	Almond Shell (wt.%)
MP-2	96	2	2
MP-4	94	2	4
MP-6	92	2	6
MP-10	88	2	10

). The choice of four compositions, with a maximum almond shell content ≤ 10 wt.%, was based on preliminary tests indicating that higher contents caused excessive shrinkage and swelling, compromising membrane morphology.

Scheme 2 schematically illustrates the main steps involved in the MK membranes preparation process. Here, clay, lime, and almond shell powder were homogeneously mixed with a minimal amount of water using the dry mixing method. The mixture was compacted at 940 MPa for 5 min in a cylindrical mold to form flat, disk-shaped membranes. The resulting green membranes (50 mm in diameter and approximately 3 mm in thickness) were dried in an oven at 100 °C to remove residual moisture. The dried membranes were sintered at 850, 900, and 950 °C to evaluate the influence of thermal treatment on structural and mechanical properties. The optimal sintering temperature was determined based on visual appearance, shrinkage behavior, and mechanical strength.

Sintering was performed in a programmable furnace following a four-step heating profile. An intermediate dwell at 300 °C for 2 h was maintained to ensure complete removal of organic matter and volatile impurities. The final sintering temperature ranged from 850 to 950 °C, depending on experimental conditions. To minimize thermal stress and prevent cracking, a slow heating rate of 2 °C·min⁻¹ was applied throughout the process [19].

Although the combined use of kaolin, almond shell powder, and lime for the fabrication of ceramic membranes has been scarcely reported, several closely related studies can be found in the literature. Ahmed and Mir demonstrated that the incorporation of almond shell powder as a natural pore-forming agent in kaolin-based ceramic membranes leads to enhanced porosity and water permeability while maintaining high dye rejection efficiencies [20]. Furthermore, previous investigations on kaolin-based ceramic membranes incorporating agro-industrial wastes such as rice husk or sawdust have confirmed the suitability of bio-derived additives for developing cost-effective and environmentally sustainable membranes for water treatment applications [21]. Nevertheless, the specific formulation combining kaolin, almond shell powder, and lime, particularly in the context of sustainable water treatment, remains largely unexplored, thereby underlining the novelty and scientific relevance of the present work.

2.3. Membrane Characterization

The shrinkage rate (S, %) was determined by measuring the membrane diameter before (D₀, mm) and after (D₁, mm) thermal treatment using Equation (1):

$$S\left(\%\right)={\displaystyle \frac{D_0-D_1}{D_0}}\times 100$$

(1)

Porosity was determined using Archimedes’ principle [22]. The procedure involved the following steps: The membrane was first weighed to obtain its dry mass (M_d). Next, the submerged mass (M_sub) of the membrane immersed in distilled water was measured. Finally, the membrane was removed from the water, its surface gently blotted with tissue paper, and the saturated mass (M_sat) recorded. The porosity (P, %) was then calculated according to Equation (2) [3,9]:

$$P\left(\%\right)={\displaystyle \frac{M_{sat}-M_d}{M_{sat}-M_{sub}}}\times 100$$

(2)

Mechanical strength was evaluated using a three-point bending test with a Lloyd mechanical testing machine (model LRX, Lloyd Instruments, Fareham, UK). Chemical resistance was assessed by measuring weight loss after three days of immersion in strongly acidic (pH 2, HNO₃) and basic (pH 10, NaOH) solutions [13].

Crystalline phases were identified using X-ray diffraction (XRD) on a Shimadzu XRD-6000 diffractometer, manufactured by Shimadzu Corporation, located in Kyoto, Japan. The ICMPE facility is located in Thiais, France equipped with Cu Kα₁ radiation (λ = 1.5406 Å), scanning from 4° to 60° with a counting time of 2 s per step. Fourier transform infrared (FTIR) spectroscopy were performed using a PerkinElmer Spectrum 100 spectrometer, manufactured by PerkinElmer, based in Waltham, MA, USA. was used to investigate chemical bonding and functional groups within the membrane matrix.

Thermal stability was evaluated by thermogravimetric analysis (TGA) using a simultaneous thermal analyzer (SDT Q600 V20.9 Build 20, TA Instruments, New Castle, DE, USA). Measurements were conducted in air with a heating rate of 10 °C·min⁻¹ over a temperature range of 22–1000 °C.

Surface morphology and microstructural features of the membranes were examined using scanning electron microscopy (SEM) with a Merlin microscope (Carl Zeiss, Baden-Württemberg, Germany) and a Hitachi S4800 microscope (JEOL JS 6010LV, Tokyo, Japan). Elemental composition and distribution were further analyzed by energy-dispersive X-ray spectroscopy (EDX).

2.4. Filtration Performance of MK Membranes

Filtration tests were performed using a laboratory-scale dead-end filtration setup (Figure 1). The flat membrane, with an effective filtration area of 19.62 × 10⁻⁴ m², was immersed in distilled water for 24 h before installation in the module. Filtration experiments were carried out in a pressurized system using compressed air at room temperature and operating pressures ranging from 1 to 4 bar.

The permeate flux (J) was determined by measuring the time required to collect a fixed volume of permeate, as expressed in Equation (3):

$$J={\displaystyle \frac{V}{t\times A}}$$

(3)

where V is the permeate volume (L), t is the filtration time (h), and A is the effective membrane surface area (m²).

Water permeability tests were conducted with deionized water at room temperature by varying the transmembrane pressure (ΔP). The membrane permeability (L_p, L·m⁻²·h⁻¹·bar⁻¹) was calculated according to Darcy’s law [20], as shown in Equation (4):

$$J_v=L_p·\Delta P with \Delta P=({\displaystyle \frac{P_{inlet}+P_{outlet}}{2}})-P_f$$

(4)

where J_v is the permeate water flux (L·m⁻²·h⁻¹), and ΔP = P_inlet − P_outlet is the transmembrane pressure (bar). In this system, P_outlet ≈ P_inlet, and the filtrate pressure (Pf) is considered negligible (P_f = 0).

The removal efficiency of crystal violet (CV) and paracetamol (PCT) was evaluated by UV–Vis spectrophotometry. The absorbance of the feed and permeate solutions was measured at their respective maximum wavelengths (λ_max = 590 nm for CV and 243 nm for PCT). The removal efficiency (R, %) was calculated using the following equation:

$$R\left(\%\right)={\displaystyle \frac{(C_0-C_p)}{C_0}}\times 100$$

where $C_0$ (mg·L⁻¹) is the initial concentration of the solute (crystal violet or paracetamol) in the feed solution, and $C_p$ (mg·L⁻¹) is the solute concentration in the permeate after filtration.

2.5. Evaluation of Fouling Resistance

The fouling resistance of the kaolin membrane system was evaluated at a transmembrane pressure of 1 bar using both synthetic contaminated solutions. Two indicators were determined to assess membrane fouling behavior: the Flux Recovery Ratio (FRR) and the Flux Decline Ratio (FDR). They were calculated using Equations (5) and (6):

$$FRR(\%)={\displaystyle \frac{J_W-J_C}{J_W}}\times 100$$

(5)

$$FDR(\%)={\displaystyle \frac{J_{WA}}{J_W}}\times 100$$

(6)

where J_W is the initial pure water flux, J_C is the stabilized permeate flux during filtration of the contaminated solution, and J_WA is the water flux measured after cleaning with distilled water. A higher FDR value indicates a greater flux decline due to fouling, while a higher FRR reflects better flux recovery and cleaning efficiency, thus demonstrating the membrane’s resistance to fouling [23].

3. Results

3.1. Characterization of Flat Ceramic Membranes

The prepared flat circular membranes exhibited a diameter of 50 mm and an average thickness of approximately 3.5 mm. Sintering was carried out for 2 h at temperatures of 850, 900, and 950 °C. This thermal protocol was selected to optimize key parameters, including porosity, surface morphology, and mechanical strength of the membrane supports.

3.1.1. Shrinkage

Diametral shrinkage was measured for the membranes MP 2%, MP 4%, MP 6%, and MP 10% at various sintering temperatures (see Figure 2). At 850 °C, all membranes exhibited low shrinkage (≈3–4%) with minimal variation among compositions. Shrinkage increased moderately at 900 °C, particularly for samples with higher porogen (almond shell) content. At 950 °C, shrinkage rose sharply, reaching approximately 11% for MP 10%.

Generally, increasing both the sintering temperature [24] and porogen content enhanced membrane shrinkage, with composition exerting a particularly strong influence at higher temperatures. Sintering at 950 °C led to visible structural deformation, reducing dimensional stability and mechanical integrity. Consequently, these membranes were excluded from further characterization due to their inadequate performance and unsuitability for organic effluent treatment.

3.1.2. Porosity and Mechanical Strength

Figure 3 presents the porosity and mechanical strength of ceramic supports containing varying amounts of almond shell and sintered at different temperatures. Open porosity was measured using Archimedes’ principle with water as the wetting liquid, allowing water to penetrate the membrane pores under atmospheric pressure.

All membranes (MP2, MP4, MP6, and MP10) exhibited a decrease in porosity as the sintering temperature increased, with values dropping from 23.61%, 23.96%, 26.8%, and 28% at 850 °C to 21.25%, 22.39%, 23.96%, and 26.39% at 900 °C, respectively. This reduction is attributed to the densification process promoted by fluxing oxides such as CaO and MgO, which facilitate crystallite rearrangement, grain growth, and the formation of a glassy phase [20]. The impact of almond shell content is particularly notable: higher porogen loadings (MP10) generate more extensive and interconnected pore networks upon burnout during sintering, leading to higher porosity but reduced mechanical integrity. Conversely, membranes with lower almond shell content (MP2) retain smaller, more uniformly distributed pores, which favor structural consolidation and higher mechanical strength.

In contrast, mechanical strength improved with increasing sintering temperature and decreasing almond shell content (from 10 wt.% to 2 wt.%). This enhancement results from a higher degree of sintering and densification of the ceramic matrix, leading to a more consolidated structure, as well as from the thermal transformation of kaolinite during heat treatment [13]. MP2, containing the minimal almond shell content, served as the reference formulation, providing a baseline for assessing the influence of porogen incorporation on both structural and mechanical properties. Increasing almond shell content from 2 wt.% to 10 wt.% progressively reduces strength due to the formation of larger voids and thinner pore walls, which act as stress concentration sites under mechanical loading.

Overall, these results indicate that a sintering temperature of 900 °C is optimal, achieving a good balance between energy efficiency, minimal shrinkage (<5%), and high mechanical strength (>25 MPa), These observations align with previously reported trends for clay-based ceramic membranes incorporating organic porogens, reinforcing the validity of our findings and providing a robust framework for tailoring membrane microstructure through controlled porogen content [25].

3.1.3. Chemical Resistance

Membrane separation processes, particularly ultrafiltration (UF), require effective disinfection and cleaning protocols, making chemical resistance a critical property for long-term durability [3,20]. The newly developed membranes (MP2-850, MP4-850, MP6-850, MP10-850, MP2-900, MP4-900, MP6-900, and MP10-900) were evaluated for chemical stability at ambient temperature (25 °C) over 72 h in acidic (HNO₃, pH 2) and basic (NaOH, pH 11) environments [3].

As shown in Figure 4, the weight loss data indicates minimal degradation under alkaline conditions, with a maximum loss of 1.1 wt.% after 72 h in NaOH. In contrast, exposure to acidic conditions caused more pronounced weight loss, which is consistent with the composition of the ceramic membranes. Being mainly inorganic, these materials are more susceptible to acid attack. Acidic solutions can dissolve certain mineral constituents, such as calcium and magnesium, leading to higher mass loss, whereas alkaline environments may favor the precipitation of these components, thereby reducing dissolution [26,27].

Nevertheless, the observed weight losses remain minor and are considered acceptable, particularly given the prolonged exposure time and the low production cost of the membranes [20].

After chemical resistance tests, the SEM images for the lowest and highest almond shell composition prepared at both sintering temperature of 850 and 900 °C, show no significant cracks or degradation of the membrane structure indicating good stability in aggressive chemical environment. These results are consistent with that reported by Xue et al. 2025 [28]. Therefore, this ceramic membranes show strong potential for applications requiring high chemical durability.

3.1.4. XRD Analysis

Figure 5 presents the X-ray diffraction (XRD) patterns of eight membranes: MP2-850, MP4-850, MP6-850, MP10-850, MP2-900, MP4-900, MP6-900, and MP10-900. The diffractograms for the membranes sintered at 850 °C are shown in Figure 5a, while those for the 900 °C series are displayed in Figure 5b.

Quartz is identified as the dominant crystalline phase in all membranes after sintering [29], with characteristic diffraction peaks at 21.4°, 24.8°, 34.5°, 55°, and 78°. Lime (CaO) is also present, exhibiting distinctive peaks at 18.8°, 39.2°, 44°, and 62°, consistently across all sintering temperatures and almond shell contents. Minor variations in the phase composition are attributed to differences in thermal treatment and the degree of crystallization [20,30].

Lime plays a key role as a fluxing agent in ceramic membrane formation, promoting densification and the development of a glassy phase that binds the particles together. It enhances mechanical strength, helps control shrinkage during sintering, and contributes to dimensional stability. During thermal processing, reactions of lime with moisture or CO₂ are reversed, allowing CaO to fully participate in the sintering reactions and to strengthen the structural integrity and overall performance of the membranes, consistent with findings from previous studies [20].

3.1.5. FTIR Analysis

The FTIR spectra of the MP2, MP4, MP6, and MP10 membranes sintered at 850 °C and 900 °C reveal the structural evolution of the materials during heat treatment (see Figure 6). A distinct band between 3679 and 3682 cm⁻¹ corresponds to the O–H stretching vibrations of hydration water within the clay lattice, while the bands at 1545 and 1475 cm⁻¹ are attributed to H₂O bending vibrations in the kaolinite interlayers [31]. The peaks around 1010 cm⁻¹ are assigned to SiO₄ and AlO₄ tetrahedral stretching [32], whereas weaker bands near 678 cm⁻¹ correspond to external symmetric stretching modes. Vibrations in the range of 457–526 cm⁻¹ are linked to Si–O–Mg, Si–O–Al, and Si–O–Fe bonds [33].

The deformation band at 915 cm⁻¹ indicates Al–Al–OH vibrations, confirming the non-dioctahedral character of the structure [34]. Lime-related absorption bands appear at 1485, 1795, 872, and 711 cm⁻¹ [29], whereas residual organic components originating from almond shell powder generate characteristic C–H and C–N stretching bands at 2917 cm⁻¹ and 1009 cm⁻¹, respectively [35].

When comparing the membranes sintered at 850 °C and 900 °C, higher sintering temperatures led to the disappearance of the organic-related bands, indicating almost complete combustion of organics and the formation of a stable silicate-based ceramic network. The thermal decomposition of almond shell additives promoted pore generation and enhanced structural consolidation, thereby improving the ceramic integrity and permeability of the membranes.

The influence of almond shell content and sintering temperature on membrane structure was further confirmed by FTIR analysis. At 850 °C, increasing almond shell content resulted in stronger absorption bands of organic functional groups, suggesting the presence of more residual organic material before complete sintering. O–H bands associated with adsorbed water and kaolinite were also observed. Increasing the sintering temperature to 900 °C significantly reduced the intensity of the organic bands, reflecting the decomposition of lignocellulosic components and the partial removal of structural water. The persistence of Al–O–Si and Ca–O bands confirmed the stability of kaolinite and lime at elevated temperatures.

At constant almond shell content, higher sintering temperatures promoted densification and controlled pore formation through the thermal degradation of organic matter. In summary, almond shell powder acted as an effective porogen, generating porosity upon combustion, while lime remained stable and contributed to particle cohesion and mechanical reinforcement of the membrane structure [20].

3.1.6. Thermogravimetric Analysis

Thermogravimetric analysis (TGA) was used to identify the temperature ranges associated with weight loss and structural transformations during membrane fabrication. Heat treatment may induce the release of H₂O, CO₂, and residual organics, as well as modifications in the mineral structure of the raw materials.

The TGA curves of MP2-850, MP4-850, MP6-850, and MP10-850 membranes are shown in Figure 7a. All samples exhibit an initial weight loss below 200 °C, attributed to the removal of adsorbed moisture and volatile species [20,36]. MP2-850 shows the largest mass loss (~2.5%), suggesting higher porosity or hydrophilicity, while the other samples lose less than 2%. Between 200 and 600 °C, the membranes remain thermally stable, with no major structural degradation. Above 600 °C, a slight weight decrease (1–3%) occurs, mainly due to the α- to β-quartz transformation [29] and the decomposition of oxygen-containing groups. The total weight loss decreases with increasing almond shell content, following the trend MP10-850 < MP6-850 < MP4-850 < MP2-850, indicating improved thermal stability at higher additive ratios.

Similarly, Figure 7b presents the thermal behavior of MP2-900, MP4-900, MP6-900, and MP10-900 membranes. The initial weight loss below 250 °C corresponds to the desorption of moisture and volatile surface species [20,36]. This initial weight loss corresponds primarily to physically adsorbed water and surface-bound volatile compounds; structural or chemically bound water in the clay matrix is not removed at this stage and occurs at higher temperatures. MP2-900 exhibits the highest mass loss (~3%), whereas the others display smaller variations (~1.5%). Between 250 and 640 °C, all membranes remain largely stable, with only minor weight decreases above 600 °C, again associated with the α- → β-quartz transition and the degradation of residual oxygenated groups. The total mass loss follows the same decreasing order (MP10-900 < MP6-900 < MP4-900 < MP2-900), confirming enhanced thermal stability at higher almond shell content.

Overall, membranes sintered at 900 °C demonstrate greater thermal robustness than those sintered at 850 °C, reflecting better structural consolidation and improved resistance to thermal variations. This makes the 900 °C membranes particularly suitable for applications requiring long-term mechanical and thermal durability.

3.1.7. Morphology Analysis

SEM micrographs of the MP-850 samples with increasing almond shell content (Figure 8a–d) show a clear evolution of surface morphology and porosity. At low almond shell content (MP2-850, Figure 8a), the surface is compact with few pores, indicating limited gas release during pyrolysis. As the almond shell content increases (MP4-850 to MP10-850, Figure 8b–d), the structure becomes progressively rougher and more porous, with a higher number of interconnected and uniformly distributed pores [6]. This trend suggests that higher almond shell content at 850 °C promotes greater volatilization, advanced carbonization, and the development of a more open and heterogeneous microstructure, likely to enhance the specific surface area and adsorption capacity of the carbon material.

SEM micrographs of the MP-900 samples (Figure 8e–h) reveal changes in morphology and porosity compared with the MP-850 series. At low almond shell content (MP2-900, Figure 8e), the surface appears slightly more porous than MP2-850, reflecting increased volatilization at the higher sintering temperature. However, as the almond shell content increases (MP4-900 to MP10-900, Figure 8f–h), the pores become fewer and finer than in the corresponding MP-850 membranes, despite a rougher surface. This reflects the effect of the higher sintering temperature, which promotes the tightening of the porous structures and partial densification of the material. Thus, the MP-900 membranes exhibit a higher degree of carbonization and a more homogeneous microstructure, but overall smaller pores compared with the MP-850 membranes.

SEM micrographs of the membranes sintered at 850 °C and 900 °C (Figure 8) highlight the combined influence of almond shell content and sintering temperature on surface morphology and porosity, reducing porosity and potentially enhancing mechanical strength [20,37]. At low almond shell content, MP2-850 shows a compact surface, while MP2-900 exhibits slightly increased porosity. As the almond shell content rises, both series (at 850 and 900 °C) develop progressively rougher and more porous structures, but the MP-900 membranes have finer and less abundant pores than the MP-850 ones due to densification caused by the higher temperature. Overall, increasing almond shell content promotes an open and heterogeneous microstructure, while increasing the sintering temperature to 900 °C leads to partial densification, producing membranes with a higher degree of carbonization but smaller pores compared with those sintered at 850 °C.

Table 2. EDX analysis of MP-900 at different percentages of almond shell. MP2-900, MP4-900, MP6-900, and MP10-900.

	MP2-900	MP4-900	MP6-900	MP10-900
O	61.4	66.5	61.2	62.6
Al	14.1	17.0	14.7	14.4
Mn	13.7	-	12.6	-
Si	7.6	15.4	9.1	12.8
Ni	2.0	-	1.4	-
Ca	0.8	0.9	0.6	1.9
S	0.4	0.2	0.2	0.4
C	-	-	-	7.9

presents the EDX analysis of flat membranes sintered at 900 °C, showing pronounced signals corresponding to Si, Al, C, O, Ca, S, K, and Cl which are typically associated with clay-based compositions. The presence of Ca reflects the lime incorporated into the membranes formulation, while the carbon (C) signal mainly originates from the almond shell used as an organic additive. A weak nitrogen (N) peak was also detected; however, this is not related to the membrane composition but rather results from the metal coating (gold/nitrogen sputtering) applied to the samples prior to EDX observation. In addition, trace elements such as Mg, Fe, and Na were identified, reflecting the complex mineralogical composition of the raw materials.

3.1.8. Permeability Test

As shown in Figure 9, the sintering temperature of the kaolinite membranes and the proportion of almond shell used as a porosity agent significantly influence the water permeability of the composite membranes. It was observed that permeability decreases from 46 to 35 L·h⁻¹·m⁻²·bar⁻¹ as the almond shell content decreases from 6 wt.% to 2 wt.%. This decrease in permeability can be attributed to a decrease in the number and size of pores, resulting from the densification of the membrane surface during sintering. This interpretation is supported by the less porous morphology observed in the SEM micrographs for MP2-900.

3.1.9. Determination of the Isoelectric Point (pH_iso) of the Prepared Membranes

The isoelectric point (pH_iso) represents the pH at which a surface carries no net electrical charge. In this study, the pH_iso of the membranes sintered at 850 °C and 900 °C (MP2-850, MP4-850, MP6-850, MP2-900, MP4-900, and MP6-900) was determined by varying the pH from 1 to 12 (Figure 10). This analysis provides insights into membrane–wastewater interactions. For the kaolinite-based membranes, the pH_iso ranged from 6.9 to 7.2, with the optimal membrane (MP2-900) exhibiting an isoelectric point of approximately 6.4. Below this value, the membrane surface is mainly positively charged, whereas above it, the surface acquires a negative charge.

3.2. Application for Crystal Violet and Paracetamol Removal from Aqueous Solution

3.2.1. Novel Membrane Performance

The performance of the clay membranes was evaluated by considering the removal of crystal violet (CV) dye and paracetamol (PCT) from aqueous solution at 1 bar and ambient temperature. The permeate flux over 60 min of filtration exhibited a gradual decline during the first 30 min for both solutions, and then stabilized at the following values:

• For CV: 34, 31, 41, 34, 43, and 39 L·h⁻¹·m⁻² for membranes MP2-850, MP2-900, MP4-850, MP4-900, MP6-850, and MP6-900, respectively;
• For PCT: the stabilized flux was of 35, 32, 42, 36, 44, and 43 L·h⁻¹·m⁻² (Figure 11a,b). This trend can be primarily attributed to the deposition and accumulation of the recalcitrant molecules on the membrane surfaces, which depend on the characteristics of the membranes.

This fouling effect is further exacerbated under dead-end filtration conditions. As shown in Figure 11a,b, the ultrafiltration process achieved very high removal efficiencies for both PCT and CV dyes.

The retention of PCT is mainly due to the adsorption onto the functional groups present on the surfaces of MP2 membranes [38]. In contrast, CV retention is predominantly governed by electrostatic interactions. Crystal Violet exists in solution as a cationic dye (CV⁺: see Scheme 1), owing to the positively charged triphenylmethane structure stabilized by its dimethylamino groups. Above the membrane isoelectric point (pHiso ≈ 6.4), the surface becomes negatively charged, which promotes strong electrostatic attraction toward CV⁺, enhancing retention. This mechanism, combined with steric hindrance and partial adsorption, explains the high removal rate observed for CV [39].

Based on the experimental results, the performance of the membranes in removing crystal violet (CV) and paracetamol (PCT) was evaluated. As shown in Figure 11c, the MP2-900 membrane exhibited the highest retention for both CV and PCT of 87%, indicating its higher separation efficiency. The MP2-850 membrane also demonstrated high retention of 83% for CV and 87% for PCT, although slightly lower than MP2-900. Membranes MP4-850 and MP4-900 showed moderate performance, with CV retention below 72%, while MP6 membranes exhibited even lower values (<70%). These results suggest that the MP2-900 membrane can be retained for low ultrafiltration of both CV and PCT solutions, combining high removal efficiency with reliable performance.

In terms of water permeability, the performance of the various membranes reveals an important trade-off between flux and solute retention. While the MP2-900 membrane exhibits both high retention (87% for CV and PCT) and relatively high flux (31 L·h⁻¹·m⁻² for CV and 32 L·h⁻¹·m⁻² for PCT), the MP4-850 membrane demonstrates even slightly higher permeate flux (41 L·h⁻¹·m⁻² for CV and 42 L·h⁻¹·m⁻² for PCT) but lower retention values (69% for CV and 83% for PCT). Therefore, the MP2-900 membrane can be selected considering both rejection and permeate flux, providing a balanced performance suitable for ultrafiltration applications.

3.2.2. Determination of the Fouling Coefficients and Membrane Regeneration

Determination of Fouling Coefficients

Membrane fouling by organic contaminants remains a major challenge in membrane technology, as it reduces both permeate flux and separation efficiency. To restore membrane performance, cleaning procedures are required, with frequency and intensity depending on the type of fouling. In this study, MP membranes were rinsing with deionized water after filtration to assess the permeation resistance associated with fouling, the results for the various membranes are summarized in

Table 3. Fouling study of MP2-850, MP4-850, MP6-850, MP2-900, MP4-900 and MP6-900 membranes using Crystal Violet (CV) according to Equations (5) and (6).

Membrane	Jw1 (L·m−2·h−1)	Jwf (L·m−2·h−1)	Jw2 (L·m−2·h−1)	FRR (%)	Rt (%)	Rr (%)	Rir (%)
MP2-850	38.0	35.0	34.0	89.5	7.9	2.6	10.5
MP4-850	42.0	41.5	41.0	97.6	2.4	1.2	1.2
MP6-850	46.0	44.0	43.0	93.6	6.5	2.2	4.3
MP2-900	35.0	32.0	31.0	88.3	11.4	2.9	5.6
MP4-900	39.0	36.0	34.0	87.2	12.8	5.1	7.7
MP6-900	42.0	40.0	39.0	92.9	4.8	2.4	7.1

and

Table 4. Fouling study of MP2-850, MP4-850, MP6-850, MP2-900, MP4-900 and MP6-900 membranes using for PCT according to Equations (5) and (6).

Membrane	Jw1 (L·m−2·h−1)	Jwf (L·m−2·h−1)	Jw2 (L·m−2·h−1)	FRR (%)	Rt (%)	Rr (%)	Rir (%)
MP2-850	38.0	34.0	33.0	86.8	13.2	2.6	10.5
MP4-850	42.0	41.0	40.0	95.2	4.8	2.4	2.4
MP6-850	46.0	43.0	42.0	91.3	8.7	2.2	5.4
MP2-900	35.0	31.0	30.0	85.7	14.3	2.9	6.5
MP4-900	39.0	34.0	33.0	84.6	15.4	2.6	12.8
MP6-900	42.0	39.0	37.0	88.	11.9	4.8	7.1

.

The flux recovery rate (FRR), an indicator of the membrane’s antifouling capability, ranged from 86 to 98% for membranes sintered at 850 °C, demonstrating that initial permeation performance can be effectively restored with simple physical cleaning. Increasing the sintering temperature to 900 °C, a slight reduction in FRR values was observed. However, MP2-900, MP4-900, and MP6-900 still exhibited excellent recovery rates of 88.3%, 87.2%, and 92.9%, respectively. The high value of FRR is attributed to the looser structure of the cake layer formed on the membrane surface, which is more easily removed during cleaning. This behavior indicates a balance between fouling susceptibility and clean ability.

Consequently, membranes derived from natural clay as matrix and almond shell as porosity agent, sintered at relatively low temperature of 850 and 900 °C exhibit both high FRR and efficient removal performance, demonstrating their suitability for multiple ultrafiltration cycles.

Determination of Membrane Regeneration

Membranes were subjected to repeated backwashing with deionized water following filtration of textile and paracetamol-contaminated water. Each cycle, performed three times, aimed to maximize foulant removal. Pure water permeability was measured after each cycle to assess membrane regeneration, with flux changes compared to the initial virgin membrane values and corresponding recovery rates calculated (Figure 12a,b). The study shows that water backwashing is sufficient to recover over 90% of the initial permeability, highlighting its effectiveness in mitigating fouling primarily. This finding is consistent with earlier research on ultrafiltration membranes fabricated from kaolinite clay [13,20].

3.2.3. Membrane Cost Estimation

Membrane fabrication costs were estimated based on the prices of raw materials and the energy required during ultrafiltration (UF) processing. Conventional ceramic membranes made from metal oxides such as alumina or zirconia are significantly more expensive than those produced from natural materials like sand, clay, or zeolite. This difference arises mainly from the high cost of synthetic precursors and the substantial energy demand associated with high-temperature sintering.

The cost of commercial α-Al₂O₃ membranes typically ranges from about $500 m⁻² (alumina) to $3000 m⁻² (stainless steel supports) [40]. In contrast, the clay-based ultrafiltration membranes developed in this study were sintered at 900 °C, whereas conventional alumina membranes are usually fired above 1400 °C [13].

Using raw material unit prices and electricity consumption data reported by Bahrouni et al. [20], the total estimated fabrication cost of the present clay membranes is approximately $3.36 m⁻². This value closely matches the $3.5 m⁻² estimated by Salek et al. [15] for flat phengite-based membranes sintered at 1050 °C. These findings confirm that natural clay-based membranes represent a highly cost-effective and energy-efficient alternative to conventional ceramic membranes for large-scale applications [41].

3.3. Comparative Study

Table 5. Details of cost estimation for the fabrication of the flat clay membrane.

Price of Raw Materials
Material	Unit per Kg ($)	Amount of Raw Material (g)	Price ($)
kaolin powder	0.5	19.2	0.010
Lime	-	0.4	0.008
Almond shells	-	0.4	-
Total raw materials cost for the fabrication of 1 membrane			0.018
Energy cost (Based on power consumption)
Dry oven			0.027
Presse			3.245
Furnace			0.086
Total production cost for the fabrication of 1 membrane ($)			3.358
(Surface of membrane = 1.7 × 10−3 m2)
Total production cost of the (MP2) Kaolin membrane ($ m−2)			3.376

presents a comparative analysis of the developed MP ceramic membranes with other low-cost ceramic membranes reported in the literature. This comparison highlights high performance of the novel developed membranes in terms of porosity, water permeability, mechanical resistance and overall recalcitrant pollutant removal efficiency.

Table 6. Various low-cost ceramic membranes for pollutant removal.

Reference	Raw Materials of Ceramic Membrane	Fabrication Technique	Sintering Temperature (°C)	Porosity (%)	Permeability (L·h−1·m−2·bar−1)	Flexural Strength (MPa)	Applications and Removal
[16]	Moroccan red clay with tea waste	Uniaxial pressing	1100	39.1	1249	14.8	Seawater: 99.76%
[14]	ball clay, quartzite waste, and corn starch		1000	35.0	203	8.6	Treatment of domestic laundry wastewater: >91%
[15]	Natural Moroccanphengite clay		1050	34.5	43.5	26.7	pretreated real wastewater (RWW3) from local clothes washing: 100%
[42]	Ceramic Membrane From Clay and banana peel		1100	40.3	550	19.2	Not specified
This work	Algerian clay, Almond shell, and Lime		900	12.5	35.0	25.0	Treatment of PCT: 87% Treatment of CV: 87%

compares the performance of the ceramic membranes developed in this study—based on Algerian clay and almond shells—with other low-cost ceramic membranes reported for tannery wastewater treatment.

The starting materials used in previous studies vary considerably: Rakcho et al. [16] employed Moroccan red clay with tea waste, Franca et al. [14] synthesized membranes from ball clay, quartzite waste, and corn starch, Lagdali et al. [15] fabricated membranes using natural Moroccan phengite clay, and Mouiya et al. [42] incorporated clay with banana peels as an organic additive.

All these studies used uniaxial pressing as a shaping technique—common for producing flat ceramic membranes at the laboratory scale—but their sintering temperatures ranged between 1000 and 1150 °C. In contrast, the membranes developed in the present study were sintered at significantly lower temperatures (850–900 °C), which considerably reduces energy consumption and overall production costs.

Regarding the physical characteristics, the porosity of the reference membranes reported in the literature ranged from 34.5% to 40.3%, whereas the porosity of the MP900 membranes in this study did not exceed 12.5%, indicating a denser structure that favors higher mechanical strength. Concerning hydraulic performance, the literature values of pure water permeability ranged from 43.5 to 1249 L·h⁻¹·m⁻²·bar⁻¹, while the MP2-900 membrane in this work achieved 35.0 L·h⁻¹·m⁻²·bar⁻¹. Although this value is relatively modest, it remains well suited for ultrafiltration applications where tighter pore structures enhance selectivity and mechanical stability.

From a mechanical standpoint, the flexural strength of the MP2-900 membrane reached 42.0 MPa, which is markedly higher than the 8.6–26.7 MPa range reported in previous studies. This confirms that the membranes developed in this work possess superior mechanical robustness, a key parameter for ensuring durability and long-term operational stability in wastewater treatment.

Looking forward, several directions can be envisioned to further enhance the performance of these ceramic membranes. The optimization of pore size distribution through controlled porogen content or alternative biodegradable additives could improve water permeability while maintaining mechanical strength. Surface functionalization or coating with selective layers may enhance rejection of specific pollutants, including emerging contaminants. Additionally, investigating hybrid membrane designs that combine clay-based ceramics with polymeric or metal oxide components could offer further improvements in flux, selectivity, and fouling resistance. Long-term stability tests under realistic wastewater conditions and scale-up studies will be essential to validate these modifications and guide practical applications.

4. Conclusions

In this study, low-cost flat ceramic membranes were fabricated from natural clay, almond shell waste, and lime, without added organic binders. Among all compositions, the MP2-900 membrane (2 wt.% almond shell, 2 wt.% lime, sintered at 900 °C) showed the best overall performance, with high mechanical strength, low shrinkage (<5%), and suitable water permeability (35 L·h⁻¹·m⁻²·bar⁻¹). SEM analysis confirmed a homogeneous microstructure with well-distributed pores. The membrane achieved high removal efficiencies for crystal violet and paracetamol (≈87%) and exhibited good antifouling properties (FRR > 88%) after simple rinsing.

The use of abundant raw materials and a reduced sintering temperature lowers production costs and energy demand, confirming the environmental and economic relevance of this approach. Compared with other low-cost ceramic membranes, MP2-900 demonstrates superior mechanical and functional performance, making it a strong candidate for treating industrial wastewater containing recalcitrant pollutants. Overall, this work provides a simple and sustainable route for producing efficient clay-based ceramic membranes using agricultural by-products.