1. Introduction
Increased industrial growth has led to the release of effluents with high amounts of heavy metals, such as mercury, which is one of the most dangerous pollutants that can potentially affect human health (Wang et al., 2004). Owing to their non-biodegradable nature and persistent character in the environment, high concentrations of these ions are a great concern in the scientific community. Research has been conducted to overcome this issue by developing materials and methods to decrease the concentrations of this toxic trace in water (Aguayo-Villarreal et al., 2011), including ion exchangers and adsorbents such as zeolites (Morency, 2002; Theron et al., 2008), biopolymers (Kostal et al., 2001; Shafaei et al., 2007; Zhang et al., 2013), magnetite nanoparticles (Hakami et al., 2012), carbon nanotubes (Saleh, 2015), and dendrimers (Niu et al., 2014).
The methods currently in practice for heavy metal removal include a work by Shahrokhi-Shahraki et al. (2021) in which activated carbon derived from pulverized waste tires exhibited great potential to adsorb three synthetic heavy metal ions: Pb2+, Cu2+ and Zn2+ from aqueous solution. Results proved that synthesized activated carbon had a monolayer adsorption capacity with faster sorption rate in comparison to commercial activated carbons. Other research work determined the potential of activated carbon collected from a full-scale water treatment plant on removing heavy metals (Dong et al., 2018). By measuring the initial and final concentrations of metals in a batch system, the study found that the activated carbon had a significant adsorption capacity for Pb2+ with removal rates above 95% and for Cd2+ reporting rates of 86%. This study also provides information on selective adsorption of activated carbon for Pb2+ with no reported changes in the surface morphology of the material. These studies highlight the performance of different activated carbon-based materials and the internal structure effects. However, activated carbons as raw materials are usually based in powdery form and require treating with different binders to achieve accurate shape for its applications.
According to a study made by Chen et al. (2022) activated carbon incorporation in geopolymer synthesis developed a composite based on hierarchical pore structures with ultra-micropores and mesopores. The material exhibited an increasing oxygen-containing groups with high activated carbon content, which facilitates CO2 adsorption capacity and CO2/N2 selectivity. Similarly, (Candamano et al., 2022) reported the formulation of geopolymer/NaX/activated carbon foamed composites for CO2 adsorption/desorption. This study gave an insight about the effect of thermal treatment on the textural properties of the adsorbents and the refining of activated carbon pore size distribution. Obtained materials developed ultra-microporosity and showed high CO2 storage performance compared with commercial geopolymers.
A recent work by Wang et al. (2022) suggested the use of phosphoric acid-activated geopolymer microspheres with activated carbon for a fast adsorptive separation of ReO4-. This result was successful in obtaining environmentally friendly microspheres with high mechanical properties, proper sphericity, and stable structure, achieving a rapid solid to liquid separation and potential application in column-based separation systems. Although one of the most used materials for heavy metal adsorption is activated carbon, the use of geopolymers based on metakaolin and activated carbon for Hg2 removal has been poorly documented in literature.
Geopolymers are three-dimensional systems with an aluminosilicate frame that were first synthesized by Davidovits in the 1970s (Davidovits, 1991; Davidovits, 1994). These systems, also known in many instances as inorganic polymers, have attracted great attention from academia and industry because of their properties, including thermal stability, chemical and fire resistance, encapsulation of radioactive substances, high durability, low permeability, and high compressive strength (Khale & Chaudhary, 2007; Lyon et al., 1997; Van Jaarsveld et al., 1997; Yunsheng et al., 2010).
The mechanism for producing geopolymers is known as geopolymerization and involves several chemical reactions where the aluminosilicate source is dissolved in a highly alkaline solution, and Al-O and Si-O bonds are released to form sialate groups with tetravalent aluminum and pentavalent silicon atoms. Further condensation reactions promote the formation of ortho-sialate molecules, known as the basic units of geopolymer structures, which react via ionic interactions to form a poly sialate chain. After polycondensation, the final structure consisted of an amorphous to semi-crystalline chain of Si-O-Al bonds in tetrahedral coordination (Davidovits, 2008; Xu & Van Deventer, 2000). Geopolymerization can be performed at room temperature; thus, these materials have a low energy use and, therefore, low emissions of nitrogen oxides (NOx), sulfur oxides (SOx), and carbon dioxide (CO2) (Duxson et al., 2007; Roy, 1999; Shi et al., 2011; Van Deventer et al., 2012; Zhang et al., 2021).
Inorganic polymers can be synthesized from natural aluminosilicate sources into industrial byproducts, such as kaolin, metakaolin, fly ash, red mud, sludge, and rice husk ash (He et al., 2013; Lecomte et al., 2003; Panagiotopoulou et al., 2007). However, the most used source is metakaolin, considering its accessibility, workability, and satisfactory properties compared to those of fly ash and slag-based composites. Metakaolin-based geopolymers have been characterized as porous materials; thus, in the present work, we propose a geopolymer composite that supports activated carbon particles to provide a micro-and mesoporous interface, enhancing the available sites for the leaching of Hg2+ ions. These sorbents provide several functional groups that can interact with Hg2+ ions via adsorption and ion exchange mechanisms. Such properties generate a low-cost material that can overcome high temperatures and pressures for industrial applications, solving an environmental problem related to mercury pollution.
2. Materials and methods
2.1. Raw materials
Kaolin clay (KC) used as aluminosilicate source in this study was collected from Michoacan, Mexico. Metakaolin (MK) was prepared by calcining kaolin at 750°C during 3 h using a heating rate of 5°C/min. The obtained metakaolin was passed through No. 200 sieve. Commercial activated carbon (AC) was purchased from CarbUSA LLC, New York. The chemical composition of KC and MK were determined by X-ray fluorescence spectroscopy (Bruker, S2 Ranger spectrometer) measurements used a palladium target. The results are presented in Table 1.
Table 1 Chemical composition of kaolin and metakaolin using the XRF technique.
Composition, wt% | Sample | |
---|---|---|
Kaolin (KC) | Metakaolin (MK) | |
SiO2 | 73.5 | 73.6 |
Al2O3 | 20.5 | 21.21 |
CaO | 0.36 | 0.32 |
P2O5 | 1.46 | 1.42 |
MnO | 0.40 | 0.38 |
Fe2O3 | 0.19 | 0.18 |
SO3 | 1.92 | 1.24 |
K2O | 0.99 | 0.96 |
ZrO2 | 0.14 | 0.14 |
TiO2 | 0.56 | 0.55 |
Alkali-silicate solution was prepared by mixing sodium silicate solution (Na2O = 8.16 wt%, SiO2 = 26.01 wt%, H2O = 65.83 wt%) with laboratory grade sodium hydroxide (NaOH, 96% purity) and deionized water. NaOH pellets were completely dissolved under magnetic stirring. Previously, solutions were stored for 24h until reaching equilibrium of the forming species. All reactants were purchased from Sigma Aldrich and used without further purification.
2.2. Synthesis of geopolymer composite
As shown in the scheme displayed in Figure 1 geopolymers (GP) were prepared by mixing metakaolin and alkaline solution with a certain amount of deionized water, using solid-to-liquid (S/L) ratios of 0.5, 0.6 and 0.7. The paste was under mechanical stirring at 400-450 rpm for 5 min to obtain a uniform mixture. After complete homogenization the geopolymer paste was then cast in cylindrical PVC molds of 25.4 x 50.8 mm. All samples were vibrated for 3 min to remove entrapped air bubbles and then tightly wrapped in plastic film to prevent water loss.
The synthesis of geopolymer composite (GP/AC) followed the above process, using MK and AC in proportion of 50:50 wt%/wt%. Both GP and GP/AC were placed in desiccators at room temperature for 24 h before demolding. Finally, samples were heated at 70 °C for 12 h and stored into a desiccator for 7, 14 and 28 days.
2.3. Adsorption of Hg2+ ions in aqueous solution
The study of Hg2+ removal was carried out using a geopolymer powder dosage of 0.07 g mixed with 50 mL of Hg2+ with initial concentration of 100 ppm. The solution was stirred for 300 min at a temperature of 25 °C. Each experiment was repeated at pH values 3, 5 and 7. The amount of metal uptake capacity by geopolymer was obtained according to the mass balance Equation (1), while the removal efficiency of Hg2+ ions was calculated using the Equation (2):
where q is the amount of metal ions uptake (mg Hg2+/g geopolymer); C 0 and C eq are the initial and final concentrations of Hg2+ in aqueous solution (mg Hg2+/L), respectively; V is the volume of Hg2+ solution (L) and m is the mass of the adsorbent (g).
2.4. Characterization
Chemical structures of geopolymer composites were investigated by Fourier transform infrared spectroscopy (FTIR) using a Perkin Elmer apparatus model Spectrum One in transmission mode. The samples were pressed into KBr pellets (1.5 mg powdered sample mixed with 150 mg KBr). Spectra were recorded at room temperature in the region between 400 cm-1 and 4000 cm-1 using a resolution of 2 cm-1.
Compressive strength was measured using an Instron 1121 universal machine with a load cell for maximum capacity of 300 kN. Following the procedure listed on ASTM C39 (Roy, 1999) three cylindrical specimens with a dimension of 25 mm × 50 mm (diameter × length) were tested after 7, 14 and 28 days of curing at room temperature. Using a crosshead speed of 0.5 mm min-1 the load was applied up until failure. Microstructure and morphology were evaluated by a scanning electron microscope (SEM) using a JEOL apparatus model JSM 6390 LV operated at 25 kV.
Samples were previously coated with a gold layer in a Denton Vacuum Desk IV sputtering system. The concentrations of Hg2+ ions were determined by inductively coupled plasma optical emission spectrometer (ICP-OES) using a Perkin Elmer Optimum 2100 DB. Prior to elemental analysis acid digestion was performed following EPA 3015A method, samples were digested using a PerkinElmer Titan MPS microwave.
pH-meter model pH120 Conductronic was used for pH measurements, previously calibrated with a 2-point calibration method using pH 4.0 and 7.0 buffer solutions.
3. Results
The XRD patterns of samples under S/L of 0.7 are shown in Figure 2. As can be seen AC displays one strong broad peak at approximately 2θ = 23.35° which is attributed to interplanar spacing (200), while a weak peak emerged at 2θ = 44.52° attributed to interplanar spacing (100). These results confirmed the existence of graphite crystalline phase (Liu et al., 2014; Zhao et al., 2009). Previous studies have widely reported the mineralogy of KC (Aziz et al., 2021; Dewi et al., 2018; Kumar & Lingfa, 2020), since raw material for the synthesis of geopolymers is metakaolin, the mineralogy for pure MK is detailed. Figure 2 shows that pure MK displayed a hump between the range of 2θ = 15° - 25°, which is characteristic of abundant aluminosilicates with pozzolanic activity (Wang et al., 2005). The pattern also exhibited peaks at 20.81°, 21.91°, 26.66°, 36.04°, 39.57° and 50.17° attributable to quartz and muscovite impurities (Granizo et al., 2007; Zhang et al., 2013).
After alkaline activation of raw materials under thermal treatment, GP synthesized using an S/L ratio of 0.7 displayed a diffuse halo that shifts to higher diffraction angles than MK sample, in the region of 2θ = 20° - 40°. This can be attributed to the geopolymerization reaction of aluminosilicates and therefore, the amorphous structure formed (Masi et al., 2014; Provis et al., 2009; Temuujin et al., 2011). In comparison, GP/AC exhibited a wider dispersed peak in the range of 2θ = 20° - 45° contributed by the AC surface (002) and leads to the formation of a more disordered phase than GP sample.
On the other hand, peaks of quartz detected in MK did not change appreciably under activation and remained in all geopolymer samples. This may be attributed to a low dissolution of crystalline quartz into alkaline solution (Palmero et al., 2015); however, the presence of these peaks does not inhibit polycondensation. The XRD pattern of GP and GP/AC samples also gave an insight about the presence of sodalite peaks in the range of 2θ = 24°-29°. These results agree with previous studies reported in literature (Provis & Van Deventer, 2014; Rattanasak & Chindaprasirt, 2009; Zibouche et al., 2009).
Figure 3 shows the compressive strength of the geopolymer samples at different ages. It was observed that AC tends to decrease compressive strength of geopolymers because of its high porosity. However, it was found that adjusting the solid-to-liquid ratio from 0.5 to 0.7 leads to denser materials and thus, the structure strength of geopolymers was enhanced. From 7 to 14 days of curing at room temperature, all geopolymer samples displayed slightly increased compressive strength. This can be attributed to an incomplete formation of AlO4 and SiO4 oligomers in tetrahedral coordination, during geopolymerization reaction.
After 28 days of curing, the formation of a more stable network was achieved and the maximum compressive values were observed ranging from 18.73 MPa to 24.31 MPa for GP/AC and GP, respectively. The impact of S/L ratios in the compressive strength of samples is shown in Figure 3, this parameter plays an important role in the mechanical performance of materials under study. Increasing S/L ratio from 0-6 to 0.7 led to higher values of the mechanical property, this can be attributed to the ratio of aluminosilicates in the activator solution.
Previous studies reported that S/L mass ratio affects the compressive strength of geopolymers by interfering with the viscosity of the slurry and geopolymerization set up (Heah et al., 2012; Zuhua et al., 2009), both ratios contribute to the dissolution and polycondensation of species which are important processes during polymerization (Liew et al., 2011; Mert et al., 2019).
For a better understanding of morphology and microstructure, raw materials and geopolymers were investigated by SEM. Their corresponding micrographs are displayed in Figure 4. As can be seen in Figure 4 (a) AC has porous particles with a size distribution of 57-171 μm, and a very stable lamellar structure which can facilitate the adsorption of Hg2+ ions onto the geopolymer surface.
On the other hand, Figure 4 (b) exhibits MK morphology with dominantly plate-like particles of irregular shape, matching with studies previously reported (Fonseca-Páez et al., 2019; Ríos et al., 2009; Youssef et al., 2008).
After alkaline activation the formation of a gel phase is clearly shown in Figure 4 (c), suggesting the successful geopolymerization of aluminosilicates. However, unreacted particles were identified on GP, which reveal the existence of crystallites throughout the network. This consists well with the XRD analysis. In the case of GP/AC morphology, a drastic change occurred as observed in Figure 4 (d). When activated carbon is added, most of the surface displayed less homogeneity than GP sample. Some irregular cracks can be observed, which are attributed to water evaporation during activation. This provided a reduced compact morphology and therefore, less mechanical resistance compared with GP.
The adsorption capacity of Hg2+ ions from aqueous solution using GP and GP/AC was evaluated and the corresponding results are given in Figure 5. In the first 60 minutes high removal occurred as observed in Figure 5a, with a Hg2+ uptake ranging from 21.9 to 23.07 mg/g for GP and GP/AC, respectively. Further sorption time showed a clear difference between the adsorbents, at 300 minutes GP/AC reached 27.05 mg/g while GP removed 21.04 mg/g of Hg2+ ions. Longer sorption times did not display significant removal and therefore 300 minutes was considered as the equilibrium time for the study.

Figure 5 a) Adsorption kinetics of GP and GP/AC b) Effect of pH on the adsorption of Hg2+ ions and c) Removal efficiency of GP and GP/AC.
It should be highlighted that the equilibrium time reported here is smaller compared with some powered and non-powdered materials previously reported (Ma et al., 2009; Sun et al., 2011; Yu et al., 2016). The addition of AC into geopolymer structure has an important effect in adsorption capacity and this could be attributed to the presence of more intense sodalite peaks in GP/AC sample, according to the XRD pattern previously reported in Figure 2. Prior studies reported sodalite as a type of zeolite with a chemical structure of hydrated aluminosilicates forming an arrangement of Al, Si, O and H interconnected in a honeycomb framework (Candamano et al., 2022), the zeolite properties include a high ion-exchange capacity and unique affinity for heavy metal positive ions. Thus, it seems that sodalite crystals in GP/AC sample are increasing the ion-exchange capacity by promoting the presence cavities in negatively charged lattices (Lou et al., 2009). Therefore, the use of geopolymer adsorbents based on metakaolin and activated carbon demonstrate its potential for fast sorption of Hg2+ species and further applications associated to mercury pollution.
Figure 5b exhibits that pH is an important parameter that significantly affects the removal efficiency of Hg2+ ions. It was found that an increase in solution pH from 3 to 5 led to an increase in removal efficiency from 51.2 % to 65 % in GP samples. Similarly, GP/AC samples reported an increase of 67.4 % to 87.6 %. At pH 7 a slight decrease was observed in both sorbents, reporting efficiencies of 64.2 % and 81.6 % for GP and GP/AC, respectively. This result suggests the different species formed according to pH values, in which below pH 5 the main species acquire negative charge causing repulsions and less efficient uptake of Hg2+ ions. The effect of contact time on removal efficiency is depicted in Figure 5c. After 60 min 63.1 % of removal is reached for GP and 71.5 % for GP/AC. After 180 min of contact time, the efficiency increased slowly and achieved the maximum removal with 69.2 % for GP and 87.6 % for GP/AC. This result implies that initially the sorbents possess abundantly active sites, which progressively are occupied by different species, over time the number of active sites decrease as well as free alkalis available for leaching. Consequently, a decrease in adhesion forces is created between the surface and Hg2+ ions, reducing the removal efficiency after longer periods of time. The removal efficiency and adsorption capacities of Hg2+ ions from previous studies were lower than those achieved by the present study (Gupta et al., 2004; Machida et al., 2004; Machida et al., 2005; Rong et al., 2018; Zou et al., 2018).
Figure 6 shows FTIR spectrum before and after the adsorption test. As can be seen all samples exhibited peaks between 3435 cm-1 and 3328 cm-1 which are attributed to -OH stretching and bending vibrations. Figure 6a display recorded for GP sample before adsorption test, it is noted two bands at about 1462 cm-1 and 1390 cm-1 which are attributed to the presence of stretching vibration of the O-C-O bonds. This group can be related to the reaction of sodium hydroxide with atmospheric CO2 during geopolymerization (Chindaprasirt et al., 2009) forming Na2CO3.

Figure 6 FTIR spectra of all samples before and after adsorption of Hg2+ ions. a) GP, b) GP after removal, c) GP/AC and d) GP/AC after removal.
The appearance of peaks at 872 cm-1 and 848 cm-1 verified the bonding of Al-O tetrahedral. After adsorption, a significant difference is observed with O-C-O and Al-O groups as observed in Figure 6b, intensity of the bands almost disappeared. This indicates the interactions between both groups with Hg2+ ions. In the case of GP/AC spectra displayed in Figure 6c before adsorption a characteristic peak at 1460 cm-1 is observed. This band is assigned to the stretching vibrations of C=O related to carboxylic and carboxylate groups. A peak at 851 cm-1 is also observed, revealing the presence of A-O groups in the sample. After adsorption, a decrease in intensity of these bands is clearly shown in Figure 6d. Modification of intensities verified the involvement of these groups during metal removal and revealed the coordination with Hg2+ ions onto the surface of the sorbents proposed.
4. Conclusions
In this work, a metakaolin-based geopolymer containing activated carbon was developed. The composite exhibits the maximum compressive strength at 28 days of curing, and increases from 8.1 to 18.73 MPa, when the S/L ratio is fixed from 0.5 to 0.7. The optimum pH for the maximum adsorption of Hg2+ ions is found to be pH: 5, with initial concentration of 100 ppm, agitation time of 300 min and sorbent dosage of 0.07 g. The addition of activated carbon improved the removal efficiency from 65% to 87.6 %. According to FTIR analysis the adsorption is strongly affected by coordination of carboxylic, carboxylate and Al-O groups with Hg2+ ions. These results show that activated carbon successfully improved the adsorption capacity of metakaolin-based geopolymers, suggesting the possibility of using the porous composite for further applications linked to mercury removal.