Adsorptive removal of Cr (VI) from wastewater using magnetite–diatomite nanocomposite

Uncontrolled waste discharge of heavy metals from different industrial sectors causes bioaccumulation in the food chain, leading to serious human and animal health issues (Gaur et al. 2021). Chromium and its compounds are used in electroplating, metal processing, leather tanning, pigments, and metallurgy (Matei et al. 2011; Wang et al. 2016), and their waste discharge is a serious concern due to potential health effects and mobility in aquatic environments (Gorzin & Bahri Rasht Abadi 2018). Chromium exists in two stable oxidation states, Cr(III) and Cr(VI) (Zhang et al. 2019) and occurs most frequently in the form of the oxyanions ⁠, ⁠, and (Huang et al. 2015). In addition, under certain conditions, oxidation by bacteria or manganese oxides can convert Cr(III) to Cr(VI), which is approximately 100 times more toxic than Cr(III) forms (Shahadat et al. 2015). Cr(VI) is exceptionally mobile in aquatic systems at a wider pH range and has been identified as a causative agent of lung cancer, kidney, liver, and gastric damage (Duranoĝlu et al. 2012; Balali-Mood et al. 2021).

Several Cr(VI) removal techniques – physical, chemical, and biological – have been reported. These include chemical precipitation (Ramakrishnaiah & Prathima 2012; Minas et al. 2017), photo-reduction (Du et al. 2018), membrane filtration (Abbasi-Garravand & Mulligan 2013), electrocoagulation (Azis et al. 2021; Villabona-Ortíz et al. 2021), phytoremediation (Kiran Marella et al. 2020), and adsorption (Baehaki et al. 2020). However, these removal techniques lead to incomplete removal, thus involving excess monitoring; they are energy-intensive and require high initial capital costs. Among the aforementioned methods, adsorption is a promising technology and has received increasing attention owing to its high efficiency at trace amounts, easy operation, low cost, and excellent adsorbent recycling performance (Yan et al. 2018; Song et al. 2019). Besides high efficiency and simple operating design, the wide accessibility of natural materials is crucial for adsorption technology (Ahmed et al. 2022).

Nowadays, low-cost adsorbents, including clay, agricultural wastes, and their composite, are among the widely reported adsorbents used for sustainable wastewater management. In this regard, Ethiopia is rich in natural inorganic minerals such as the diatomite (DM) clay of lacustrine origin present in deposits of more than 40 million tons, mainly along the Rift Valleys of Adami Tulu, Gade-Mota, Chefe Jilla, and Abiyata (Weldemariam et al. 2019), the Afar depression, and in some basin catchments (Kurkura et al. 2012). Diatomites are eco-environmental functional materials with very promising and effective adsorption capacities due to their porous structure (Wu et al. 2016; Touina et al. 2021), high specific surface area (Zhao et al. 2019), increased stability (Touina et al. 2021), low water solubility (Le et al. 2020), chemical inertness (Khaldi et al. 2018), potential for reuse (Gao et al. 2021; Zhang et al. 2021), and surface silanol groups(Si-OH) (Chu et al. 2012) that account for their good adsorption characteristics.

The DM, or amorphous hydrated silica (SiO₂·nH₂O), also contains some impurities including silica sand (Flores-Cano et al. 2013), clay mineral (Galal Mors 2010; Touina et al. 2021), carbonates, metal oxides ((Tsai et al. 2006; Cabrera et al. 2017), and organic matter (Ahmed et al. 2022). However, impurities that block some of its pores and other surface properties limit the commercial application of natural DM as an adsorbent.

Even though different techniques for removal of impurities in DM, such as calcination, acid/alkali leaching, and thermochemical treatment, have been reported (Weldemariam et al. 2019; Reka et al. 2020), its adsorption capacity is still less competitive as compared to seolite, kaolinite, and montmorillonite (Yuan et al. 2013; Zhao et al. 2019; Le et al. 2020). Besides, different clay-based composite adsorbents have been reported to be effective for adsorptive removal of Cr(VI) ions from wastewater; for example, montmorillonite nanocomposite (Wang et al. 2016), glutaraldehyde cross-linked chitosan–bentonite composites (Altun 2020) and clay–biochar composites (Wang et al. 2019) have reasonable adsorption capacity from wastewater. However, the production of these composite adsorbents involves high costs of surfactants for clay–composite synthesis. The problems associated with used adsorbents’ separation/regeneration from aqueous solution hinder their large-scale application.

In the current study, magnetite (Fe₃O₄) nanoparticle modification of purified DM to enhance its adsorption capacity and ease of separation/regeneration of used adsorbent was investigated. Fe₃O₄ nanoparticles have excellent properties (higher magnetic properties than other iron oxides) and can be easily recycled through magnetic separation (Chen et al. 2016). Even though Fe₃O₄ has a strong tendency to agglomerate due to Van der Waals forces or magnetic attraction forces, its composite form shows high stability with polymers, chitosan, carbon nanotubes, and clay minerals (Pandit et al. 2022).

The main objective of this research was to study the synergistic effects of Fe₃O₄ loaded into thermochemically modified DM (TMD) through co-precipitation methods for Cr(VI) adsorption removal from synthetic and tannery wastewater. The novel aspect of this study is to synthesize nanocomposite materials with magnetic susceptibility for easy separation/regeneration of used adsorbent by means of an external magnetic field and superior Cr(VI) adsorption properties of composite materials. The response surface methodology (RSM) of multiple factors at the optimal time using a three-level Box–Behnken design (BBD) was used for operating variables like adsorbent dose, initial Cr(VI) load, pH, and contact times for Cr(VI) removal in a batch adsorption process to reduce the number of experimental runs at wider ranges of stated parameters.

Chemicals used were ferric chloride hexahydrate (FeCl₃^. 6H₂O, 97%, AR/ACS, Loba Chemie, PVT.LTD) and ferrous chloride tetra-hydrate (FeCl₂^. 4H₂O, ATICO, India), potassium dichromate (K₂Cr₂O₇, 99 + %, Sigma-Aldrich), ammonium hydroxide (NH₄OH, ATICO), and hydrochloric acid, 37%, Sigma-Aldrich). All analytical grade chemicals were used without any purification, and all solutions were prepared using ultrapure water. The DM used for this experiment was obtained from the Great Rift Valley of Oromia Regional State, Ethiopia.

The DM was ground to powder and sieved with 825 mesh size to obtain uniform particle size. Acid purification of DM was performed using the methods of Turkten (2022) with some modifications. Fifty grams of sieved DM was added to the flask and mixed with 500 mL of 10% HCl. The sample DM was shaken at 200 rpm in a water batch for 2 h at 60 °C. The solution was filtered using filter paper, and repeatedly washed with ultrapure water and oven-dried at 60 °C overnight. The oven-dried DM was calcined at 900 °C for 1 h. The magnetite–DM nanocomposites were synthesized according to Li et al. (2020) with some modifications. Six grams of thermochemically modified DM was dissolved in 50 mL of deionized water and slowly poured into a 250 mL flask containing 10 g of FeCl₃^. 6H₂O and 3.5 g of FeCl₂^. 4H₂O. The mixture was adjusted to a pH of 8–9 using 0.5 M HCl and 30% NH₄OH. The solution was ultrasonicated at 85 °C for 2 h using a shaker at 200 rpm. It was centrifuged at 2,500 rpm for 5 min to separate the sediment, and the residue was washed with ultrapure water until neutral pH and dried in an oven at 60 °C overnight. The dried material was calcined at 500 °C for 1 h.

The X-ray diffraction analysis (XRD) patterns of natural and modified DM were investigated using a Rigaku Ultima IV diffractometer, with Cu K radiation operated at 40 kV and 40 mA in the scan range 3–80° 2 and a scan speed of 10° min. The Fourier-transform infrared spectroscopy (FTIR) spectroscopy analysis was used to detect the functional groups present in the adsorbent materials using a Spectrum 65 Perkin Elmer spectrometer in the range 4,000–400 cm⁻¹ and spectral resolution of 4.0 cm⁻¹. The chemical composition of adsorbents was determined using complete silicate chemical analysis, LiBO₂ fusion, hydrogen fluoride (HF) attack, gravimetric methods, and colorimetric and atomic adsorption spectroscopic techniques. The surface area, pore volume, and average pore diameter were determined using Quanta Chrome version 11.0 of N₂ – Brunauer–Emmett–Teller (BET) analysis at 77.3°K using a SA-9600 Series Surface Area Analyzer (Horiba Instruments, Inc.). The samples were degassed at 100 °C for 2 h and the specific surface area was determined by the multipoint BET method in an N₂ gas relative pressure (P/P⁰) range of 0.05–0.30. To understand the magnetic properties of the magnetite–diatomite nanocomposite (MDM), the magnetic susceptibility study of the material was determined using Sherwood's scientific magnetic susceptibility balance.

The pzc of the MDM nanocomposite was determined by using salt addition methods (Mahmood et al. 2011). A sample of adsorbent (0.2 g) was added to 40 mL of 0.1 M NaNO₃ solutions and the pH of sample solutions was adjusted with 0.1 M HCl and 0.1 M NaOH to obtain the appropriate pH ranges of 2–9 (± 0.02). The pH of the adjusted solution was denoted as pHi and the experiment was repeated with 0.05 M NaNO₃ solutions and each set of experiments was performed in triplicate. The sample was shaken for 24 h at 170 rpm in a shaker. After settling, the pH values of each sample solution were measured as pH_final. Thus, pHpzc of MDM was obtained from the plot of ΔpH (ΔpH = pH_initial − pH_final) against pH_initial.

The preliminary Cr(VI) adsorption removal and adsorption capacity of MDM were measured in a batch adsorption process. The preliminary investigation results confirmed that MDM showed a remarkable adsorption capacity for Cr(VI) ions without competing ions and was chosen for further adsorption studies. The batch adsorption was conducted at room temperature and using a water bath shaker (WS17-2 SHEL LAB) at a speed of 200 rpm. A stock solution of 100 mg L⁻¹ of Cr(VI) was prepared using 0.28286 g of potassium dichromate (K₂Cr₂O₇) in 1 L of ultrapure water, and different standard solutions of Cr(VI) were prepared by serial dilution from the stock solution.

RSM coupled with BBD was applied to design Cr(VI) adsorption on MDM experiments. The design expert software (version 13.0) was used to perform statistical analysis and the three-level BBD was applied to investigate the independent variables pH, adsorbent dose, initial Cr(VI) concentration, and contact time with their coded levels −1 (minimum), 0 (medium point), and +1 (maximum) as listed in Table 1.

The purely whitish colour of the DM sample (Supplementary material, Figure S1) was collected from Adami Tulu, Oromia Regional State, in the Great Rift Valley, Ethiopia. Using LiBO₂ fusion, HF attack, and gravimetric, colorimetric and atomic adsorption spectroscopic methods, the chemical composition of the material was analyzed. As shown in Table 2, raw DM is primarily composed of SiO₂, Al₂O₃, Fe₂O₃, and CaO, with other oxides, such as MgO, Na₂O, K₂O,P₂O₅, TiO₂, found at low concentrations (Weldemariam et al. 2019). Upon acid treatment and calcination of DM, slight decomposition and leaching of soluble oxides from the surface of DM were observed. Consequently, the fractional composition of silica content increased from 76% for DM to 83.68% for TMD. Whereas, for magnetite (Fe₃O₄) intercalated into TMD to form a composite material, both composite fractions SiO_2, and Fe₂O₃ increase from 76.00 to 79.22% and 0.46 to 4.68%, respectively. The slight leaching of inorganic impurities from the DM surface may account for the increase in SiO₂ in the TMD. The enhancement of iron oxide composition in TMD to form nanocomposites may develop composite materials with magnetic properties due to ferromagnetic intercalation into TMD layers. Furthermore, in the thermochemical purification of DM, significant loss on ignition was observed at 900 °C as previously reported (Bello et al. 2014), and this may account for increasing the purity of DM, for which the percentage composition of SiO₂ is one of the main factors in determining the degree of purity, as previously reported (Mohamed et al. 2018).

The XRD result of the calculated crystal size for MD, TMD, and MDM is shown in Supplementary material, Table S1. Thermochemical purification shows a decrease in crystal sizes from 124 nm DM to 71.205 nm: this accounts for the increase in its specific surface area for Cr(VI) adsorption uptake. This result is in agreement with previous reports (Yusan et al. 2014). Whereas, Fe₃O₄ nanoparticle incorporation into the DM surface shows more decrease in composite crystal size from 71.205 nm for TMD to 53.24 nm. This may be due to inter-grain porosities of magnetite over TMD surface, and the positive charge of magnetite was adsorbed by the negatively charged DM surface. Similar results were reported for DM modified with Fe(SO₄)₂·7H₂O (Knoerr et al. 2013) and 3-aminopropyl tri-ethoxy silane and EDTA with magnetic DM (Wu et al. 2021). Chemical composition analysis shows that Ethiopian-origin DM contains more quartz, kaolinite (Al₂Si₂O₅(OH)₄, and alunite KAl₃(SO₄)₂(OH)₆ in polymorphs of quartz, as shown in Table S2 (Supplementary material).

The moderately pure DM from Ethiopia shows remarkable surface areas: this may be due to the low fractional composition of CaO and high concentration of Al₂O₃ as shown in the chemical silicate analysis (Table 2). Thermochemical modification (acid washing and calcination) of DM enhances the BET-specific surface area from 357.835 m²·g⁻¹ for DM to 364.627 m²·g⁻¹ for TMD: this is due to removal of impurities (loss on ignition) that mask natural pore volume and removes alkali oxides to make available new micro and nanopores on the surface of modified DM, as reported by other authors (Goren et al. 2002; Yuan et al. 2013; Zelentsov 2017; Dim et al. 2021). As shown in Table 3, magnetite surface modification of DM (forming MDM) synergistically increases the surface area of TMD from 364.627 to 387.707 m²·g⁻¹ and the pore volume from 0.0571 to 0.0673 cm³/g, respectively. This is due to the blocking of large mesoporous DM through the formation of a uniform thin layer of Fe₃O₄ through in situ reaction, which was also reported for DM modified with aluminum silicate (Zelentsov 2017) and DM modified with MnO₂ (Li et al. 2014). However, high concentration ratios of the magnetite precursor to DM show a significant decrease in the surface area and this may be due to the physical alteration of pore structure and the development of a thicker layer on the surface of DM that may block the fine pores formed by thermochemical modification of DM and blocked (re-occupied) with Fe₃O₄ nanoparticles, as reported by Li et al. (2020).

Magnetic susceptibility indicates the degree of magnetization of materials when an external magnetic field is applied (Yamato & Kimura 2020). According to their magnetic properties, all substances are divided into three groups that differ in magnetic susceptibility range: diamagnetic, paramagnetic, and ferromagnetic (Bangari et al. 2020). The paramagnetic substance has gram susceptibility in the range 10⁻⁵–10⁻³ and is attracted by magnetic fields (Gaeta et al. 2021). In this study, the magnetic susceptibility of MDM was studied by using a Sherwood scientific magnetic susceptibility balance, and the composite materials result shows from the minimum of 2.617 × 10⁻⁵ g susceptibility to above the range of the instrument. The gram susceptibility analysis confirms the magnetization of composite materials with a high potential to be attracted by external magnetic fields. This is due to the ferromagnetic coupling of Fe₃O₄ with TMD that may create spin-polarized carriers with a high molecular field for magnetization of MDM composite materials, as previously reported for magnetic montmorillonite (Cabrera et al. 2014) and MDM composites (Li et al. 2020). In addition, complete chemical silicate analysis results show an enhancement of Fe oxides composition in the MDM materials.

For the statistical evaluation of input effective variables (A, B, C, D) and in order to validate the empirical models, several parameters were considered. Analysis of variance (ANOVA) is an analytical technique that is used to identify the validity and adequacy of the models using Fisher's, F-test, and Student's t-test. In this case, F-values and P-values suggest the design factors, pH, adsorbent dose, initial Cr(VI) concentration, and contact time are highly significant for the quadratic model, as shown in Table 4. The Cr(VI) adsorption removal responses of the experimental design show optimal conditions at pH = 3, MDM = 0.6 g, initial Cr(VI) concentration = 16 mg·L⁻¹ and contact times of 60 min with 98.89% removal efficiency and 26.36 mg·g⁻¹ adsorption capacity (see Supplementary material, Table S3).

In Table 4, the P-value of the quadratic model is <0.005, which shows that the quadratic model fits the Cr(VI) adsorption data on MDM adsorbent. The F-value of lack of fit of 1.63 indicates that the lack of fit is relatively low in comparison to the absolute errors. The model test shows that the design model accurately illustrates the data behavior for the experiments and the factors investigated have very significant effects on removal efficiency (Qiu et al. 2014). In the plots of actual vs predicted values for Cr(VI) adsorption data on MDM (Supplementary material, Figure S2), the predicted values are represented by straight lines inclined from the origin, while the actual points are also aligned with predicted values.

The BBD for Cr(VI) removal optimized model suggested a quadratic model, with adjusted and predicted regression (R²) values of 0.9902 and 0.9719, respectively, as shown in Table 5. The model is more acceptable when R² approaches unity (Mondal et al. 2017). These factors have very large effects on removal efficiency (Qiu et al. 2014). In general, to validate this empirical model, several parameters were considered, such as probability < 0.05 would show the significance of the model and lack of fit should be not significant (>0.05), and adequate precision (signal-to-noise ratio) is ≥4, as shown in Table 4.

The pH of a solution changes the chemical states of active groups and the charge distribution on the adsorbent surface and the dissolution of pollutants in the solution, whereby all of them affect the adsorbent's adsorption affinity (Bilgiç & Karapınar 2022). As depicted in Figure 4(a), at low pH, MDM adsorbs protons (H⁺) and its surface is positively charged due to the presence of Si-OH functional groups (Yang et al. 2017). Cr(VI) ions exist mostly in three forms, i.e. oxyanions in aqueous solution such as ⁠, ⁠, and ⁠, depending on the solution pH. In an extremely acidic environment (pH < 1), Cr(VI) ions exist as H₂Cr₂O₇ and where the pH is in the range 2–6, Cr(VI) exists predominantly as oxyanions (Belibağli et al. 2020). Among Cr(VI) oxyanions, is the most prominent species to be adsorbed on the positive surface (Sun et al. 2017a). This may account for maximum Cr(VI) adsorption values on DM and TMD that were obtained at pH 3. This result is in agreement with previous reports (Memedi et al. 2021). Also, Fe₃O₄ functionalization of TMD does not significantly change the pH of MDM for maximum Cr(VI) adsorption, allowing maximum adsorption performance at a pH of 3. Similar results were reported for Cr(VI) removal using activated carbon–attapulgite clay composite (Hlungwane et al. 2018), clay–biochar composite (Wang et al. 2019) and DM supported by zero-valent iron nanoparticle (Zhang et al. 2019). While the pzc of MDM was obtained at pH 6, as the solution pH increases from 3 to 5, the adsorption capacity decreases. This is explained by the fact that, as the solution pH approaches the pzc, there is a decrease in the positive charge density on the MDM surface and therefore less electrostatic interaction with ions. In contrast, increasing solution pH above pzc, the MDM surface is negatively charged and adsorption of Cr(VI) ions on the negative adsorbent surface leads to electrostatic repulsion interaction between the negatively charged surface of MDM and Cr(VI) ions. In addition, the Cr(VI) ion is competing with OH⁻ ions above pzc (pH > 6) (Bilgin & Tulun 2015).

Figure 4(b) illustrates that removal efficiency increases as the initial Cr(VI) ion concentration increases and reaches its maximum at 16 mg·L⁻¹ with 98.89% removal. However, further increase in Cr(VI) initial concentration up to 30 mg·L⁻¹ shows a decline in removal efficiency and this may be due to the ratio of active adsorption surface sites on MDM surfaces that are occupied with Cr(VI) ions in the solution. An increase in the initial concentration above the maximum level cannot favor the adsorption process (Jayasree et al. 2021).

As shown in Figure 4(c), increasing the adsorbent doses from 0.30 to 0.60 g L⁻¹ enhances the Cr(VI) adsorption removal and this may be due to an increase in the number of binding sites on the adsorbent surfaces. MDM shows rapid uptake of Cr(VI) with increasing adsorbent doses and attained equilibrium at 0.60 g·L⁻¹ with more than 98.89% of adsorption removal. Also, increased concentration of MDM to 1 g·L⁻¹ decreases the Cr(VI) adsorption and this may be due to more adsorbent–adsorbent (hole–hole) interaction than with Cr(VI) ions (Acharya et al. 2017).

As shown in Figure 4(d), rapid initial adsorption of MDM is observed in the time-dependent kinetic adsorption data. This is likely to be due to initial excess active adsorption sites, and Cr(VI) ions easily interacting with active sites (Yang et al. 2017). As contact time increases from 45 to 60 min, MDM adsorption rates reach the maximum of 89.98% at 60 min. However, with a further increase in contact time of Cr(VI) ions with MDM to above 60 min, the results show a slowdown of adsorption rates and reach a stagnant state. This occurs because all available active adsorption sites are occupied with Cr(VI) and a further increase in contact time does not facilitate the adsorption process.

In contrast, the Freundlich isotherm model assumption was based on heterogeneity of adsorption surface and the active sites difference gives different adsorbate and adsorbent interfaces based on multilayer adsorption properties (Foo & Hameed 2010). The Freundlich isotherm model defines the behavior of adsorption as non-ideal, reversible, and non-restricted to monolayer formation, and assumes non-uniform distribution of adsorption heat and affinities over the adsorbent surface (Alam et al. 2021).

Clay-based composites have received much attention for the adsorption removal of different pollutants due to their high stability, low cost, and high adsorption capacities. The maximum adsorption capacity of MDM adsorbent obtained was 88.49 mg/g⁻¹ of Cr(VI). This makes MDM a very promising adsorbent with 76.04% of Cr(VI) removal from a real industrial tannery effluent containing Cr(VI). The high adsorption capacity of MDM can be ascribed to the electrostatic attraction between negatively charged Cr(VI) oxyanions and positive charges of MDM surface at optimized values. Table 9 summarizes the Cr(VI) adsorption capacity reported for various clay-based composite adsorbents. In this study, MDM was found to be effective for the removal of heavy metal contaminants and compared to other adsorbents, it has high adsorption efficiency at low adsorbent dose and room temperature. MDM can be synthesized from cheap and easily available clay minerals, so its use may contribute to sustainable waste treatments. Moreover, MDM can be easily separated from the solution using an external magnetic field with reuse capacity.

In this case, Cr(VI) adsorption occurs: first, the MDM adsorbs protons in acid conditions (pH = 3) and is positively charged. Then, the electrostatic interaction between the positively charged surface and the predominant Cr(VI) species in acidic environments (pH = 2–5) is ⁠, which accounts for high Cr(VI) adsorption capacity (Zhang et al. 2019). Fe₃O₄ immobilization on DM increases the functional group density on the MDM surface and accounts for the effective chemical entrapment of Cr(VI) ions (Li et al. 2014; Sun et al. 2017b). At high pH of alkaline solution, the Cr(VI) ions can be desorbed from the MDM surface and the adsorbent materials regenerated for reuse as shown in the proposed mechanism in Figure 10.

MDM nanocomposite was fabricated as a low-cost adsorbent for effective Cr(VI) removal from wastewater. Batch experiments were performed to investigate kinetics, equilibrium isotherm, and thermodynamics of the Cr(VI) sorption process on MDM adsorbent. The synergetic effects between Fe₃O₄ and TMD composite show 98.89% removal efficiency and 26.36 mg·g⁻¹ adsorption capacity of Cr(VI) at BBD-RSM designed optimal parameters. The optimal Cr(VI) adsorption on MDM adsorbent was obtained at adsorbent dose = 0.6 g·L⁻¹, pH = 3, initial Cr(VI) concentration = 16 mg·L⁻¹, and contact time of 60 min. The adsorption process is best described by pseudo-second-order kinetic models, while both the Langmuir and Freundlich adsorption isotherm best describe the adsorption process of MDM with a maximum adsorption capacity of 88.49 mg·g⁻¹. Thermodynamic experimental data revealed the adsorption process is exothermic and spontaneous for Cr(VI) uptake by composite adsorbents. MDM is highly stable for Cr(VI) removal over five reuse cycles; only a 12% decrease in its removal efficiency was observed after five adsorption–desorption cycles. When the ferromagnetic Fe₃O₄ is intercalated into the TMD layer, it enhances adsorption capacity and improves adsorbent recovery by an external magnetic field. In addition, the MDM is a suitable adsorbent for removing Cr(VI) from real tannery wastewater. On this basis, MDM composite could be an efficient, sustainable and affordable adsorbent with a promising potential for eliminating Cr(VI) from tannery effluents.

All relevant data are included in the paper or its Supplementary Information.

The authors declare there is no conflict.