Chemical and microstructural analyses for heavy metals removal from water media by ceramic membrane filtration

There are around thirty-five metals posing a threat to human health due to residential and/or occupational exposure, 23 of which are heavy metals (Mosby et al. 1996). Along with the industrial revolution, a great amount of industrial wastes discharging into sewage networks led to increasing heavy metals content in wastewater. Moreover, heavy metals in drinking water form a threat to human health. Populations are exposed to heavy metals through water consumption while some of these metals can bio-accumulate in the human body (Chowdhury et al. 2016). Among the heavy metals, As, Cd, Pb, Cr, Cu, Hg, and Ni are of major concern, mainly due to their presence with relatively high concentrations in drinking water and their effects on human health (ATSDR 2015). To minimize the risks of heavy metals, regulatory agencies have proposed the maximum allowable limits in drinking water (WHO 2011; USEPA 2014). Toxic chemicals such as heavy metals, if not treated properly, cause harmful effects to the organisms and the whole environment (Nziku & Namkinga 2013). Toxicity studies confirmed that they can directly impair the mental and neurological functions, influencing neurotransmitter production and utilization, and altering numerous metabolic body processes (Ogunfowokan et al. 2008). As, Cd, and Pb are the most produced metals from various sources and more problematic metal pollutants, they were extensively studied for public health impacts (Chowdhury et al. 2016).

The treatment of water polluted with heavy metals has become a challenging task for environmental engineers. There are several methods adopted to remove heavy metals from aqueous systems. The most widely used methods for removing heavy metals from wastewater are chemical or electrochemical precipitation, both of which pose a significant problem in terms of disposal of the precipitated wastes (Ozdemir et al. 2005). Ion-exchange treatments, as well, do not appear to be applicable due to their high costs (Pehlivan & Altun 2006). In addition, it has been reported that some aquatic plants (Keskinkan et al. 2004) and microorganisms (Tewari et al. 2005) have the capacity to adsorb and accumulate heavy metals, but the use of aquatic plants and microorganisms increases the chemical oxygen demand of wastewater (Argun et al. 2007). Brown algae are among the biosorbents more appointed for industrial applications due to their large uptake capacity (Ahluwalia & Goyal 2007). Dry alginate beads have been found to have a good potential for heavy metal biosorption from a real effluent and the metal ions could be easily desorbed (Silva et al. 2008).

There are some parameters affecting the metals removal such as aqueous solution pH, initial metal concentrations, the overall treatment performance compared to other technologies, and environmental impacts as well as economical parameters such as the capital investment and operational costs. The technical applicability, plant simplicity, and cost effectiveness are the key factors that play a major role in the selection of the suitable method for heavy metals removal. The factors mentioned above should be taken into consideration in selecting the most effective method for the metals removal process (Bobade & Eshtiagi 2015). Using membrane separation techniques in the treatment of contaminated water give high removal efficiency but they require a great membrane area and energy, so it has become an urgent need for developing new membranes based on low cost materials such as clays and apatite. Using such materials in wastewater treatment is an efficient solution, especially when they are burned to form ceramic membranes (Masmoudi et al. 2005; Khemakhem et al. 2007).

Ceramic membranes are categorized as one of the industrial porous materials. According to the International Union of Pure and Applied Chemistry, the porous materials are classified into three grades depending on the diameter of their pores (d). These grades include macro-porous (d > 50 nm), meso-porous (50 nm > d > 2 nm), and micro-porous (d < 2 nm). Moreover, depending on the pore diameter, the filtration process can be classified as; filtration (d > 10 μm), micro-filtration (10 μm > d > 100 nm), ultrafiltration (100 nm > d > 1 nm), nano-filtration (d = 1–2 nm), and reverse osmosis (d < 1 nm). In filtration and micro-filtration, the separation is accomplished mainly by sieving effect where particles whose size is larger than the pore size are trapped. While, in ultra-filtration, nano-filtration, and reverse osmosis where pore size is small, fluid permeability depends on the affinity of solute and solvent with the porous materials as well (Ohji & Fukushima 2012). For structural applications of ceramic materials, pores act as fracture defects that degrade the structural reliability, and therefore, ceramics should be sintered to their full density to attain high mechanical strength. On the other hand, such internal pores give advantages for filtration, absorption, catalysts, structural light weight, and thermal isolator (Messing & Stevenson 2008).

Thermal, chemical, and mechanical properties of ceramic membranes give them more advantages than polymeric ones (Chan & Brownstein 1991). Ceramic filters are in different shapes and sizes, which include disk filters and hollow candles (Matteiletea 2006). The main raw materials used commonly for manufacturing ceramic filters are clay and burnout materials. There is a number of innovative processes for making macro-porous ceramics, namely partial sintering, sacrificial fugitives, replica templates, and direct foaming (Ohji & Fukushima 2012). In partial sintering process, particles of a compacted powder are bonded due to the surface diffusion or evaporation–condensation processes and enhanced by heat treatments to form a homogeneous porous structure. In order to provide the desired pore size, the raw powder size should be geometrically in the range of two to five times larger than that of the coveted pore. In the process of sacrificial fugitives, porous ceramics can be obtained by mixing appropriate amounts of sacrificial fugitives as pore forming agents with the ceramic raw powder and evaporating or burning them out before or during sintering to create pores. Classification of agents forming pores divides into inorganic matters as fly ash (Ding et al. 2007), synthetic organic matters as organic fibers (Colombo & Bernardo 2003), natural organic matters as potato starch (Luyten et al. 2003), and liquid as water (Jia et al. 2007). In replica templates process, macro-porous ceramics having interconnected large pores or channels and open cell walls have been frequently fabricated. The templates need to have adequate flexibility, homogeneous open cell structure, and shape recovery ability. In the direct foaming technique, the ceramic membrane is foamed by incorporating air or gas and dried, and subsequently is sintered to obtain a consolidated structure.

Porous ceramics with high specific surface area are employed for absorptive and catalytic applications, where large area is needed for contacting with reactants. The applications of these porous ceramics have been applied for variety of filtration, adsorptions, and catalysts (Tolba et al. 2015). There are some investigations on combination of processes for the removal of heavy metal ions. Generally, it was found that combination of processes is more effective than an individual process. For example, the combination of chemical precipitation and ion exchange treatments for the heavy metals removal from polluted solutions was adopted (Papadopoulos et al. 2004). The results showed that the combination of precipitation and ion exchange processes caused a higher nickel removal with an efficiency ranging from 94.2% to 98.3% while using the individual ion exchange treatment led to a lower removal of 74.8% (Feng et al. 2000). Another study relied on a single adsorption process for heavy metals removal by acid activated kaolin and the results revealed moderate removal efficiency (El-Molakap et al. 2013). Ceramic filters are effectively used for eliminating suspensions and Bacillus from wastewater because of their higher flux capability, sharper pore size distribution, better durability, and higher damage tolerance than those of organic hollow fibers (Jedidi et al. 2009).

The development of a high performance ceramic membrane technology has been stimulated by the need for submicron filtration of contaminated fluids in cost sensitive environmental applications. The purpose of this study is to evaluate the performance of ceramic membranes made from local low cost materials of clay and sawdust at three different mixing ratios in removing heavy metal ions from water media by filtration. The influence of pH, initial driving pressure head, and concentration of heavy metals on their removal efficiency by ceramic filters were investigated. Chemical and microstructural analyses on polluted water samples and ceramic membranes were carried out before and after the filtration process.

The slip casting method for making porous ceramic membranes was implemented in this work according to the specified procedure in previous studies (Kabagambe 2010; Tolba et al. 2015). Sawdust was used as a pore forming agent based on the sacrificial fugitives process (Ohji & Fukushima 2012). The clay material was collected from an agriculture field near the Nile River in Kom-Ambo area (40 km north of Aswan city, Upper Egypt). A homogeneous clay mass was collected from a depth of 5 m below the ground surface. The clay mass was crushed, dried in an oven at 100 °C for 24 h, and ground to a fine powder (particles size >425 μm). For mineralogical characterization, a powdered raw clay sample was scanned in the range 10–80 ° on X-ray diffractometer. In addition, the chemical composition of another powdered sample was studied by X-ray fluorescence (XRF) analysis. The sawdust was brought from a local wood workshop, dried in fresh air, and sieved through 425-μm openings sieve. A sensitive balance model ADAM EQUIPMENT ACH-30 was used in the preparation process. Different three mixes of clay and sawdust with sawdust percentages of 0.5%, 2.0%, and 5.0% by weight were prepared.

The dried powders of clay and sawdust were mixed well for 15 min. Then, an appropriate amount of water (20% by weight) was added into the mixture. Thereafter, the wet mixture was folded upon itself repeatedly until a smooth knead was formed. The knead mixture was casted in a 15 × 15 cm mold, compacted to perform the green body of the ceramic sheet with a thickness of 2 cm, and its upper face was formed to produce an ambient edge with a thickness of 1 cm. The green sheets were dried in fresh air away from direct sunlight for 15 days so that the internal moisture was expelled. The sheets were then sintered at a fixed temperature of 1,000 °C for 6 h to produce ceramic membranes. The furnace was allowed to cool slowly through 12 h. The sintering process intensifies the compact and affects the microstructure of the ceramic. The fabricated ceramic membranes were found to have a sintered bonding that ensures a long life span; this bonding will not be separated even under challenging feed water conditions (Matteiletea 2006).

Batch filtration experiments were carried out using volumes of 500 and 750 mL from heavy metal solutions to produce initial pressure heads of 5.0 and 7.50 cm, respectively. The solutions were poured into the glass prisms on the top side of the ceramic membranes. The tests were performed under falling head conditions that started at 5.0 or 7.5 cm height over the ceramic filter and decayed to zero at the end of the test. Plastic pots were placed under the ceramic membranes to receive the effluent. The initial flow rate through the ceramic filters was estimated from the volume of the collected filtrate within the first hour of the run. The effect of pH on heavy metals removal by ceramic membrane filtration was conducted by adopting three different values of pH, namely 5, 7, and 9. The desired pH was adjusted by the addition of 0.1 N nitric acid or 0.1 N sodium hydroxide ⁠. The effect of pH and heavy metals concentration on their removal by ceramic filters were studied under an initial driving head of 5.0 cm. On the other hand, the influence of the initial pressure head on both heavy metals removal and filtration rate through the ceramic filters was investigated at initial heads of 5.0 and 7.50 cm with heavy metals concentration of 5 mg/L. At the end of each run, the ceramic filter was subjected to a backwashing process with distilled water to remove the deposited ions. Accordingly, the same ceramic filter could be reused several times.

Surface morphology of the ceramic membranes was investigated by scanning electronic microscope (SEM) images (JEOL JSM-5400 LV) before and after the filtration tests. To comply with the constraints imposed on the specimen size for SEM analysis, a small sample was cut from the membrane center and coated with gold under a vacuum pressure. The gold coating increases the conductivity of the materials. In the SEM chamber, each specimen was allocated in an electron beam to take its image at the desired magnification scale.

Figure 2(a)–2(c) illustrate that ⁠, ⁠, and ions were effectively removed by ceramic filters with different sawdust ratios, especially at the higher values of pH (7 and 9). In general, the removal efficiency of these metals was more than 97% and the best efficiency with the different sawdust ratios occurred at pH 9. A similar study on the removal of heavy metals from aqueous medium using vermiculites achieved around 100% metals removal in alkaline conditions (Sis & Uysal 2014). The reason was partly attributed to the metal hydroxide formation and, consequently, the surface precipitation under alkaline conditions (Katsou et al. 2010). In addition, the affinity between the positive heavy metal ions and negative surface charges resulted from alkaline conditions greatly enhances the removal efficiency (Abdus-Salam & Adekola 2005). For ceramic membrane filtration, the observed high metals removal at pH 9 came from the double effect of two treatment mechanisms as the precipitation of insoluble metal hydroxides along with the membrane filtration. The results proved that the equilibrium capacity of the metals removal increased significantly as the pH of the solution increased. It is concomitant with a previous study on the removal of heavy metal ions from wastewater, which reported that the equilibrium capacity of copper and zinc removal by different adsorbents increased significantly, as the pH of the solution increased (Salam et al. 2011). Previous studies showed a formation of occurs between pH 6 and 10 (Bradl 2004), precipitation of starts at pH 6.3 (Pare et al. 2012), and formation of happens between pH 8 and 12 (Ayres et al. 1994). Accordingly, to exclude the interference between surface ions precipitation and purifying mechanisms provided by ceramic filters, tests of heavy metals in the next parts were performed at pH 5 for and and at pH 7 for ⁠.

The water pressure head acting on the surface of the ceramic membrane directly affects the filtration rate. The solution flow rate presents an important performance parameter for ceramic filters because their applicability is greatly relying on it. To check the influence of the driving pressure head on the filtration rate and removal of heavy metals by ceramic filters, two experiments were carried out with different initial pressure heads of 5.0 and 7.5 cm. The volumes of the aqueous solutions to be filtered in the experiments were 500 and 750 mL. The tests were carried out under falling pressure head conditions, starting at the initial head and terminating at zero head. To exclude the effect of the surface precipitation, values of pH were adjusted to 5 for both and and 7 for ⁠. The experiments were conducted in a constant temperature room (29 ± 2 °C) and with influent metal concentration of 5 ppm. Tests were carried out on ceramic filters with sawdust contents of 0.5%, 2.0%, and 5.0%.

The initial flux through the ceramic membranes was estimated from the filtrate volume within the first hour of the filtration test. Under initial gravitational pressure head of 5.0 cm, membranes with 0.5%, 2.0%, and 5.0% contents of sawdust showed initial filtration rates of 41.36, 45.28, and 55.27 mL/h, respectively. At this initial head, the ceramic filters produced 500 mL of filtrate through 24.2, 22.0, and 18.1 h for 0.5%, 2.0%, and 5.0% sawdust contents, respectively. In case of 7.50 cm initial head acting on filters that have 0.5%, 2.0%, and 5.0% sawdust contents, the initial water fluxes were found to be 50.66, 55.46, and 67.69 mL/h, respectively and the runs continued for 29.6, 27.0, and 22.2 h to produce 750 mL of filtrate. In a previous study, gravitational ceramic filters that had cylindrical shapes were used to treat grey water and showed filtration runs with lengths varied from 0.74 to 2.0 days. The filters were backwashed and used several times for a period more than six months (Hasan et al. 2015).

Figure 5(a) shows the surface structure of a membrane with 0.5% sawdust at magnification of 200× with the biggest number of pores on the surface. The size of the pores was relatively small with diameters less than 8.5 μm. Figure 5(c) illustrates the surface morphology of another membrane with 2.0% sawdust at magnification scale of 200×. The number of pores diminishes, whereas their size seems to be bigger with an average diameter of 15 μm. The morphology of 5.0% sawdust ceramic membrane at magnification of 200× has a different shape with the minimum number of pores and the largest size with diameters up to 50 μm as depicted in Figure 5(e). By increasing the magnification scale to 500 × , a small and narrow micro crack can be observed in Figure 5(d) on the surface of the membrane with 2.0% sawdust. For the membrane with 5.0% sawdust, a long micro crack line with a bigger width can be seen in Figure 5(f). Practically, it was observed that the prevalence of wide cracks through membranes with 5% sawdust led to a fabricated fragile membrane with less lifespan and more possibility of damage. In contrast, micro cracks disappear from the image of the membrane with 0.5% sawdust as in Figure 5(b). In addition, images with the magnification scale of 500× show an agglomeration of the membrane fabricating materials due to the manual compaction. These observations are assured by the reported data about depressions and micro cracks that were found on most of the surface of ceramic filters manufactured from Seeta and Ntaawo clay (Kabagambe 2010).

Ceramic membranes that were manufactured locally from natural clay and sawdust with different microstructure properties were found to be effective in removing heavy metals from aqueous solutions. The most important findings are as follows:

• (1) The ability of ceramic membranes with different sawdust percentages for removing different concentrations of ⁠, ⁠, and by filtration was tested at different values of pH and showed high removal efficiency (above 94.21%).

• (2) The removal of heavy metals by ceramic filters declines slightly with increasing the metal concentration but enhances at the higher initial driving pressure head.

• (3) Ceramic filters showed high removal efficiency relying on two mechanisms, namely precipitation and filtration processes without the need for chemical additions.

• (4) SEM images indicated that the morphology and both size and shape of internal pores of ceramic membranes could be controlled by changing the ratio of the burn out materials in the mixture.

• (5) Increasing the sawdust contribution in ceramic membranes substrates produces fragile ceramics with significant interior cracks, while not frequently affect the metals removal efficiency.