Treatment of seawater and wastewater using a novel low-cost ceramic membrane fabricated with red clay and tea waste

2.3.1 Characterization of powder

An X-ray fluorescence (XRF) spectrometer, specifically the Epsilon 4 model from Malvern Panalytical, was employed to ascertain the chemical compositions of the analyzed samples. For XRD examination, a Bruker-AXS D8 powder diffractometer, utilizing Cu Kα radiation (λ = 0.154186 nm), was operated within a 5° to 70° (2θ) range. Additionally, FTIR spectroscopy was conducted in transmission mode for all raw materials using a Nicolet 5700 spectrometer, with a spectral range of 400 to 4000 cm⁻¹.

Characterization of thermal properties was performed using setaram SETSYS 24 equipment for differential thermal analysis (DTA) and thermogravimetric analysis (TGA), with α-Al₂O₃ serving as a reference material. Under the experimental conditions, the samples were heated to 1100 °C at a rate of 10 °C/min.

2.3.2 Ceramic membrane characterization

To examine the changes in the sintered membrane's apparent porosity and apparent density, water was used as the test fluid, following the Archimedes principle. A detailed examination of the membrane surface's morphology and texture was carried out using scanning electron microscopy (SEM). After acquiring SEM micrographs, we used by ImageJ (Version 1.44e) to analyze the images and determine the pore-size distribution of the membrane.

The evaluation of mechanical strength was conducted through the employment of Brazilian tests and three-point bending tests, utilizing (Lloyd Instruments, UTM Ametek). Stress-displacement curves were acquired at a rate of 0.1 mm/min, providing valuable insights into the material's mechanical properties.

Bending strength (σ_F) and tensile strength (σ_T) were determined using Eqs. (1) and (2):

In this context, F_F and F_T denote the applied force at the fracture point (N), with L = 31 mm the spacing between support, w signifies the membrane width (mm) and t is the thickness (mm) of the rectangular samples. h and d represent the thichness and diameter of the cylindrical samples, respectively in (mm). To ensure the reliability of the test results, three specimens were produced for each addition of TW (0–30 wt%).

The chemical stability of membranes was tested using NaOH (pH = 13) and HNO₃ (pH = 1) solutions. To determine chemical stability, the weights of sintered membranes were measured before and after ten days immersed in 100 ml of acidic and basic solutions.

3.1.1 Raw material physicochemical and mineralogical properties

In order to assess the suitability of clay for use in the ceramic industry, we conducted an investigation of raw clay powder from various perspectives, including chemical, mineralogical, and technological analyses. Our results, presented in Table 1, provide a breakdown of the oxide content and ignition loss of the raw clay powder. As expected, this clay has a typical composition that is rich in silica (56.203%) and alumina (15.109%). The presence of clay minerals and quartz in the clay has a significant impact on the refractoriness and mechanical strength of the final product, as noted in previous research (Sadik et al., 2013). However, our analysis also revealed a substantial amount of iron oxide (9.695%) present in the clay, which contributes to a darker hue in the fired pieces (Dondi et al., 2014; El Yakoubi et al., 2006). In addition, the clay contains high levels of fluxes, including CaO and K₂O, as well as smaller amounts of MgO (1.215 wt%), P₂O₅ (1,455 wt%), and TiO₂ (1.416 wt%). Negligible amounts of Na₂O and SO₃ were also detected. Overall, our results indicate that the clay sample has the necessary components for use in the ceramic industry.

The results presented in Fig. 2 highlight the abundant presence of key minerals suitable for ceramic production, namely quartz (Q), kaolinite (Kaol), calcite (C), muscovite (M), and lepidocrocite (L).

Close

Fig. 2

X-ray diffraction of red clay.

Export to PPT

FTIR spectroscopy was employed to characterize and identify the constituents present in tea waste and red clay samples. This analytical technique has been widely used by researchers to study diverse materials. The FTIR spectrum of tea waste (TW) is depicted in Fig. 3, where a broad absorption band is observed at 3405 cm⁻¹. This feature has been linked to the O–H and N–H stretching vibrations associated with polyphenols (Loo et al., 2012; S R Senthilkumar and Sivakumar, 2014). Saif et al.(Saif et al., 2016) reported that the extensive absorption band observed at 3310 cm⁻¹ can be attributed to the presence of hydroxyl functional groups (OH) typically found in alcohols and phenolic compounds. Additionally, a pronounced band at 1636 cm⁻¹ has been associated with the stretching vibrations of C = C bonds in aromatic rings and C = O bonds in polyphenolic compounds (Dubey et al., 2017; S R Senthilkumar and Sivakumar, 2014). The C–H stretch and O–H stretch in alkane and carboxylic acid, were reported to surface at 2924 and 2854 cm⁻¹, respectively (S R Senthilkumar and Sivakumar, 2014). A band has also appeared at 1030 cm⁻¹ due to C-O stretching in amino acids (Dubey et al., 2017; S R Senthilkumar and Sivakumar, 2014). J.Cai et al. (Cai et al., 2015) discovered identical FTIR bands in numerous types of tea, including oolong, green tea, and black. In previous research, it was determined that the IR bands of polyphenols in tea appeared at 3387 cm⁻¹, 1635 cm⁻¹, and 1037 cm⁻¹; these correspond to N–H/O–H, C = C, and C–O–C stretches (Dubey et al., 2017; S R Senthilkumar and Sivakumar, 2014; Weng et al., 2013). The primary functional groups in tea residue are carboxylic acid, polyphenols, and amino acid, as indicated by the FTIR spectrum.

Close

Fig. 3

FTIR spectrum of tea waste (TW) and red clay (RC).

Export to PPT

The FTIR spectrum analysis of the red clay sample in Fig. 3 reveals characteristic absorption bands. A broad band ranging from 3200 to 3700 cm⁻¹ is associated with the stretching vibrations of hydroxyl groups (–OH) linked to the water of constitution. The presence of bands at 2834 and 2932 cm⁻¹ can be attributed to the low organic matter content in the sample. An intense absorption band centered at 1023 cm⁻¹ with additional peaks at 900 and 1225 cm⁻¹ signifies the valence vibrations of Si-O bonds (Saikia et al., 2016). Furthermore, the absorption bands between 785 and 688 cm⁻¹ are indicative of Si-O-Al bonds (Miyah et al., 2017). The presence of quartz in the red clay sample, as confirmed by X-ray fluorescence, is supported by the absorption band at 1030 cm⁻¹. Additionally, the absorption bands observed at 466 and 542 cm⁻¹ are consistent with the characteristic features of quartz (U.O.Aroke and A.Abdulkarim R.O.Ogubunka, 2013).

3.1.2 Thermal analysis

Fig. 4 presents the DTA/TGA results for a 20 wt% tea waste with clay mixture and pure tea waste. As the temperature increases, a continuous and irregular weight loss is observed, culminating in a total weight loss of 30.30 wt% (Fig. 4 a). An initial weight loss of 4.27 wt% at temperatures below 100 °C is attributed to the endothermic evaporation of free water. Subsequently, a weight loss of 9.41 wt% between 100 and 315 °C is likely due to the removal of adsorbed water and the outgassing of volatile organic compounds. A further weight loss of 12.18 wt% is observed within the 315–560 °C temperature range, which can be attributed to the decomposition of residual organic compounds with tightly bonded carbonyl and hydroxyl functional groups, as well as the dehydroxylation of clay minerals (lepidocrocite, kaolinite) and the transformation of α-quartz into β-quartz (Mouiya et al., 2022; Vizcayno et al., 2010). An endothermic peak accompanied by a 4.44 wt% weight loss is detected at approximately 761 °C, which can be ascribed to the thermal decomposition of mineral carbonates. No noticeable reduction in weight occurs beyond 900 °C. The thermal evaluation suggests that ceramic membrane sintering should take place at temperatures of 900 °C or higher.

Close

Fig. 4

The DTA/TG curves of (a) clay-tea waste mixture containing 20 wt% TW and (b) pure TW.

Export to PPT

As illustrated in Fig. 4b, the ATD/TGA analysis of TW at a heating rate of 10 °C/min may be separated into 3 different phases: (1) drying at 185 °C, (2) devolatilization between 185 and 537 °C, and (3) mineral degradation above 537 °C. The weight loss of TW during the first stage is primarily attributed to water evaporation at lower temperatures. In the next phase, which encompasses the main devolatilization process, the weight loss corresponds to the temperature range of 185 to 537 °C. The total amount of weight lost of 76.62% in this primary stage is due to the enhanced release of volatile compounds resulting from the breakdown of organic macromolecules such as celluloses, partial lignin, hemicelluloses, and additional macromolecular substances, along with a decrease in the fixed carbon content of the sample. In the main devolatilization stage, the process of thermal breakdown was split into three smaller parts using the OriginPro 9.5.1 peak fitting tool, similar to the approach employed for the thermal decomposition of petrochemical wastewater sludge, oily sludge, and hydrotalcite (J. Chen et al., 2015; Liu et al., 2018). In the initial sub-phase, weight loss is predominantly attributed to hemicellulose decomposition, with the primary degradation temperature ranging from 185 to 311 °C (Fig. 4.b). Hemicellulose is a complex composition of polymerized monosaccharides with a low degree of polymerization, including glucose, xylose, arabinose, galactose, and mannose (Ma et al., 2015). During the second sub-stage (311–390 °C), the primary decomposition of cellulose and a fraction of lignin occurred, resulting in a decrease in weight. Cellulose is a high-molecular-weight molecule characterized by a long linear chain formed by diglucosyl groups (Lin et al., 2017), Its crystalline structure, composed of organized microfibrils, renders its thermal degradation more difficult compared to hemicellulose (Gu et al., 2013). In the third sub-stage (390–537 °C), lignin and macromolecular compounds, due to their intricate composition, exhibited slower degradation rates than hemicellulose and cellulose. As described by Liang (Liang et al., 2018), hemicellulose, cellulose, and lignin primarily undergo pyrolysis within the temperature ranges of 200–300 °C, 300–400 °C, and 200 °C to the end of the thermal process, respectively. The final stage (>537 °C) seemed to be a consequence of the gradual thermal decomposition of residual materials, such as chars, minerals, and ash, which constituted the ultimate solid residues (Chandrasekaran et al., 2017).

3.2.1 X-ray diffraction analysis

The XRD patterns shown in Fig. 5 for samples sintered at 1100 °C reveal that the primary phase present is quartz (SiO₂). Additionally, a considerable amount of hematite (γ-Fe2O3) is observed, which is formed through the dehydroxylation of lepidocrocite (γ-FeO(OH)) originally present in the raw clay. Furthermore, minor amounts of mullite (3SiO₂·2Al₂O₃) and cristobalite (SiO₂) are detected, both of which are products of metakaolinite decomposition at temperatures exceeding 900 °C. The appearance of small amounts of albite implies a release of amorphous silica and alkaline elements within the clay matrix. Furthermore, after firing, the peak for calcite (CaCO₃) was no longer there because CaCO₃ broke down into CaO and CO₂. In the same way, the kaolinite peak disappeared from all of the graphs. This is because kaolinite changed into metakaolinite. But the peak of quartz is the most intense and stays the same in every diffractogram. This demonstrates the heat stability of the quartz powder.

Close

Fig. 5

XRD spectra of samples sintered at 1100 °C (Q = quartz; Mu = mullite; C = cristobalite; H = hematite; Al = albite.

Export to PPT

3.2.2 Porosity and density

Fig. 6 shows the correlation between the apparent density, open porosity, and the proportion of waste in terms of weight. The results indicate a positive relation between TW loading and porosity. A porosity increases from 22.26 to 49.13% was seen in a TW content increase from 0 to 30 wt%, which was attributed to enhanced gas generation induced by the TW breakdown at 185–537 °C (Fig. 3(b)). In contrast, the density decreased to a minimum of 1.35 g cm⁻³ when the TW loading climbed to 30% by weight. This drop in apparent density is a result of the solidification of the supports, the breakdown of carbonates, and the creation of fluxing agents (Fe₂O₃, MgO, K₂O, and Na₂O) during the firing process. The decrease in apparent density with tea waste content could be the result of decarbonization and the combustible of organic materials (Anjum et al., 2019). Moreover, the porosity observed in samples containing 30 wt% tea waste in this work surpassed the reported values for porous zirconia ceramics fabricated with 40% commercial starch (Albano et al., 2009).

Close

Fig. 6

Relationship between oppen porosity, apparent density, and percentage by tea waste of weight in clay at 1100 °C firing temperature.

Export to PPT

3.2.3 Microstructural analyses and pore size distribution

The surfaces of the membranes sintered at 1100 °C were evaluated by scanning electron microscopy (SEM), (Fig. 7) to provide additional evidence for the Archimedes' test results in ceramic membranes, which demonstrated an increase in pore structure openness with increased of tea waste loading. It appears in Fig. 6 that the three membranes A, B and C with 10, 20 and 30% by weight of tea waste respectively had quite homogenous surface. The increase in TW loading, resulting in porous structures after the burning of TW in the sintered samples. In addition, tea waste loading was found to increase the amount of interconnectivity between pores, which was expected to lead to open porosity Fig. 7(C). This observation underscores the successful creation of a more porous structure within the ceramic matrix, a consequence of adding tea waste. This enhancement can be attributed to the combustion of the organic matter within the tea waste during the firing process. As this combustion occurs, it releases CO₂ gas, which escapes from the material, promoting the development of a more extensive and intricate porous network within the ceramic matrix. The high TW-loading sample (30 wt% TW; Fig. 7(C)) exhibited open porous microstructures, consistent with the outcomes of Archimedes test (Fig. 6). This underscores the significant role of tea waste as a pore-forming agent and highlights its ability to increase porosity and modify the pore structure of the ceramic supports.

Close

Fig. 7

SEM analysis of sintered samples with TW contents of (A) 10, (B) 20, and (C) 30 wt%, and distributions of pore size (D).

Export to PPT

The surface pore-size distribution of ceramic membranes sintered at 1100 °C exhibits a unimodal distribution pattern in pore diameters, as depicted in Fig. 7 (D). The distribution varies between membranes due to differences in the amount of tea waste (TW) incorporated. Three distinct pore size distributions were identified, corresponding to the varying concentrations of TW. Ceramic membranes containing 10 wt% and 20 wt% TW exhibit average pore sizes of 1.20 μm and 2.8 μm, respectively. For membranes with a 30 wt% TW content, the average pore diameter is approximately 5.8 μm, with a maximum pore diameter of around 7.9 μm. Notably, about 95% of the pores possess diameters smaller than 5 μm. It is postulated that the incorporation of significant quantities of tea waste in the samples reduces the amount of contact sites between clay granules. This reduction in contact points leads to an increased formation of pore interconnections during the sintering process, ultimately promoting the development of larger pores.

3.2.4 Chemical stability

A crucial element of the membrane filtration procedure is the cleaning and disinfection stage. This process generally entails submerging the membranes in a solution containing appropriate chemicals, such as alkalis or acids (Gan et al., 1999), to effectively eliminate contaminants. Consequently, assessing the flat ceramic membrane's resistance to chemical degradation during the cleaning process is of paramount importance to ensure its durability and effectiveness. In this study, we investigated the chemical corrosion behavior of ceramic membranes by immersing them in solutions of HNO₃ (pH = 1) and NaOH (pH = 13) for varying periods. Fig. 8 shows the mass loss of selected ceramic membranes during this immersion process. The results reveal that the ceramic membrane exhibits a high resistance to basic corrosion, with a mass loss of less than 0.15 percent during 250 h of exposure to a pH = 13 solution of NaOH. In contrast, the mass loss can reach up to 4% in a concentrated HNO₃ solution, indicating a lower resistance to acid corrosion.

Close

Fig. 8

Weight loss of the optimal sintered ceramic membrane in HNO₃ solution at pH = 1 and NaOH solution at pH = 13.

Export to PPT

The fast breakdown of Al and Mg components made from clay is to blame for the manufactured membranes' poor acid-corrosion resistance, followed by the formation of free amorphous silica in very acidic circumstances (Dong et al., 2007). Consequently, chemical cleaning using an alkali solution at room temperature is preferable for these membranes, despite the extended cleaning duration compared to acid cleaning. Overall, the results of this study provide important insights into the chemical resistance of ceramic membranes during cleaning and disinfecting processes, which can be useful for optimizing membrane filtration processes in various applications.

3.2.5 Mechanical properties

The mechanical resistances of samples sintered at 1100 °C as a function of TW content are depicted in Fig. 9. The bending strength load displacement curves in Figs. 9-A indicated an elastic phase at first, followed by a densification step in which the bending strength increased abruptly just prior to fracture. This phenomenon was attributed to the increase in density caused by the progressive disintegration of struts during bending strength loading. In accordance with the behavior of brittle materials, the tensile strength curves Figs. 9-B for samples with a low TW content exhibited severe linear increases until rupture. However, samples with greater porosity exhibited a linear evolution with greater elasticity prior to rupture, which resulted from their enhanced elasticity. Fig. 9 demonstrates conclusively that TW content significantly affects the mechanical strength of sintered samples. Increasing TW loading from 10 to 30 wt% resulted in decreased flexural and tensile strengths from 23.25 to 12.71 MPa and from 13.68 to 6.98 MPa, respectively. This decrease was related to an increase in porosity. Overall, the results highlight the importance of carefully controlling the TW content in the sintered samples to achieve the desired mechanical properties.

Close

Fig. 9

Flexural strength (A) and indirect tensile strength (B) of the ceramic membranes fired at 1100 °C with different contents of tea waste.

Export to PPT

Close

Fig. 10

Water flux through the ceramic membrane as functions time (A) and transmembrane pressure (B).

Export to PPT

Close

Fig. 11

Flux of permeation of tannery effluent and seawater as a function of time and applied pressure.

Export to PPT

Close

Fig. 12

Rejections of permeate turbidity, COD, color, pH, and conductivity as a function of filtration time.

Export to PPT

Close

Fig. 13

Photograph of tannery effluent before (A) and after (B) filtration through the ceramic membrane.

Export to PPT

3.3.1 Characterization of feeds solutions

Tannery industries produce approximately 30–35 L of effluent for every kilogram of processed skin, resulting in wastewater with significant levels of suspended particles and salinity. Throughout the leather processing steps, from preservation to finishing, various chemicals, including chromium, sulfide, and chloride, are utilized, leading to the presence of harmful substances in the effluents. This study assessed membrane performances in the treatment of tannery wastewater and raw seawater. Table 2 summarizes the parameters of the solutions. The turbidity of tannery effluent (180 NTU) was found to be higher than that of raw seawater (60 NTU), attributed to the large quantity of suspended material (6.22 g/l) generated during the beamhouse process, which may also be to blame for the color of the wastewater. The higher COD in tannery effluent (1400 mg/l) compared to raw seawater (12.20 mg/l) was due to the increased presence of inorganic and organic substances produced during the delimitation processes. The conductivity of raw seawater (30 ms/cm) was significantly higher than that of tannery effluent (1.60 ms/cm), as a result of the chemical salts present in seawater. Lastly, both solutions exhibited alkaline pH values, ranging from 11.60 for tannery effluent and 8.10 for raw seawater.

3.3.2 Filtration of the feed solutions

During the filtration experiments, the feed solutions under consideration were passed through a filtration pilot at a pressure of 0.14 bar for a duration of 90 min. To maintain a constant feed and avoid concentration effects, the feed tank was refilled every 30 min. Fig. 11 displays the changes in permeation flux of the tannery effluent and seawater with respect to time and transmembrane pressure. As observed from the plots in Fig. 11A & C, the permeation flux of both tannery effluent and seawater gradually decreased over time, stabilizing at t = 45 min for the former and at t = 60 min for the latter. This decrease occurred in two stages (Achiou et al., 2017; Manni et al., 2020):

1-Internal fouling resulting from the capture of particulates that are marginally smaller than the membrane's pores.

2-The buildup of suspended matter on the membrane's surface, which ultimately results in external fouling.

Significantly, both the feed characteristics, such as the presence of microorganisms and colloidal particulates, and the properties of the ceramic membrane's internal and external surfaces substantially influence the variations in filtration flow. The progressive accumulation of particulates on the membrane surface leads to the formation of a fouling layer that obstructs the pores. Prior studies have sought to elucidate the fouling mechanism through chemical characterization (Almandoz et al., 2015; Zhou et al., 2019; Zou et al., 2021), and modeling in mathematics (Manni et al., 2020). Chen et al.(Q. Chen et al., 2015) established that particle build-up on the membrane surface causes a cake layer to build, which creates more frictional resistance in the water, obstructing the fluid's permeation. Again observing Fig.11A & C, a deviation exists between the curves for each pressure in Fig. 11A&C, signifying a strong correlation between flow and pressure. To elucidate the relationship between permeation flux variations and transmembrane pressure, we conducted a systematic analysis by monitoring the variations in relation to the pressure, as illustrated in Fig. 11 B&D. The obtained graphs exhibited an almost linear relationship, whereby the slopes of the curves decreased with increasing filtration time. These results suggest that an increase in filtration time results in a reduction in the volume of filtrate that passes through the membrane wall. Hence, it can be concluded that an optimal synergistic value between the applied pressure and filtration time is crucial for continuous and efficient filtration.

In contrast to the ceramic membranes we developed in our prior research (Mossaab Mouiya et al., 2019), the ceramic membrane produced in this study demonstrates a noteworthy reduction in pore size while preserving an impressive flux rate. Furthermore, we conducted a comparative analysis of the pure water flux and turbidity rejection of our newly obtained ceramic membrane against other ceramic membranes and data from the literature (as summarized in Table 3 and Table 4). In summary, our as-prepared ceramic membranes consistently outperformed membranes of similar pore sizes, displaying both superior pure water flux and remarkable turbidity rejection capabilities.

The efficacy of membrane filtration was assessed through the analysis of permeate characteristics, including turbidity, chemical oxygen demand (COD), total suspended solids (TSS), electrical conductivity, and pH. Fig. 12 illustrates the consistency of permeate pH throughout the filtration process, signifying the stable quality of treated tannery wastewater and seawater. Notably, the pH of tannery wastewater decreased from an initial value of 11.60 to 9.88 after filtration. This reduction in pH can be primarily attributed to the exclusion of biologically-derived substances, which typically possess basic properties. Additionally, Fig. 12 presents the rejection percentages for more parameters, such as turbidity, conductivity, COD and color, over a 90-minute filtration period.

Tannery wastewater is characterized by high turbidity, indicative of a substantial concentration of suspended matter such as suspended and colloidal particles (Mossaab Mouiya et al., 2019). As demonstrated in Fig. 12A & C, the ceramic membrane achieved turbidity rejection rates of 99.16% and 99.76% for tannery effluent and raw seawater, respectively. This suggests that the primary mechanism for turbidity removal by the ceramic membrane is sieving, because the rejected particles are bigger than the holes in the membrane (Achiou et al., 2017; Majouli et al., 2012). Consequently, the ceramic microfiltration membrane (MF) demonstrates high efficiency in removing turbidity from tannery wastewater and raw seawater. Regarding the chemical oxygen demand (COD), the ceramic membrane achieved rejection values of 76.43% and 87.70% for tannery wastewater and raw seawater, respectively (Fig. 12A & C). This confirms the ceramic membrane's ability to significantly reduce the COD of both tannery wastewater and seawater. It is well-established that the color of raw tannery wastewater is attributable not only to solids, but also to dyes that can be dissolved in water. In the present study, just color arising from soluble dyes was determined, with the rejection of color results presented in Fig. 12-A. The membrane was found to remove approximately 83.33% of color, indicating its effectiveness in addressing this aspect of wastewater treatment. The incomplete removal of soluble dyes can be attributed to their small molecular size, which hinders their complete rejection by the membrane filtration process. It is important to note that the color of the wastewater under investigation primarily arises from colored suspended particles rather than soluble compounds. Due to the small size of ions, soluble salts cannot be efficiently removed by microfiltration (MF) processes, resulting in low conductivity rejection (Saja et al., 2018). Consequently, it is imperative to develop asymmetric ceramic membranes that facilitate UF or NF processes, which are more suitable for the removal of soluble dyes and salts (Saja et al., 2018). It should be emphasized that the removal efficiency of turbidity, COD, and color only marginally increases over the course of the filtration process. This marginal improvement is mainly attributable to the buildup of rejected particles on the membrane's surface, which forms a cake layer that acts as an additional selective barrier. Such a phenomenon is commonly observed in MF processes (Majouli et al., 2012).