Adsorptive removal of phosphorus from aqueous solutions using natural and modified coal solid wastes Open Access

Eutrophication is a global environmental concern that seriously threats aquatic ecosystems, and results in loss of biodiversity and large economic dissipations (Zhou et al. 2022). It occurs as a consequence of massive phosphorus (P) discharge into natural water resources through various anthropogenic activities including excessive use of agricultural compounds such as fertilizers and pesticides, paint industries, detergents, and various municipal wastes (Goscianska et al. 2018; Park et al. 2021). A maximum permissible range of 0.5–1 mg P L⁻¹ has been established by the Environmental Protection Agency (EPA) in wastewater (Park et al. 2021). Hence, in the past few decades, efforts have been significantly made to control this essential element regarding water quality standards.

Several physicochemical and biological techniques have been used to remove and recover P from wastewater and industrial effluents including electrochemical precipitation, filtration, ion exchange, and microbial degradation processes (Di Capua et al. 2022). Among them, adsorption has been developed as a promising water treatment strategy following its major advantages such as low cost and low energy requirements, simplicity, non-toxicity, and reversibility (Abdellaoui et al. 2021). Consequently, a large number of synthetic materials have been used for P adsorption including carbonized sludge (Zhang et al. 2018), calcium meta-silicate minerals (Obradović et al. 2017), iron-impregnated biochar (Lee et al. 2018), and layered double hydroxide-loaded biochar (Bolbol et al. 2019).

Recently, a new comprehensive approach has been established based on the circular economy concept on the valorization of agricultural and industrial waste by-products by employing them as novel composite materials for environmental decontamination, particularly water resources restoration (Xu et al. 2022). In this context, various solid waste materials have been examined for P adsorption such as different biomass-derived biochars (Jung et al. 2015; Ngatia et al. 2017), zirconium-loaded orange waste (Biswas et al. 2008), and concrete powder (Liu et al. 2020). Most of them have represented considerable P removal performances regarding their favorable physicochemical and structural characteristics including high porosity, surface area, and mineral composition, as well as their cost-effectiveness and environmental friendship.

Coal fly ash is one of the most important industrial by-products worldwide, which originates from the coal combustion process and is excreted in large quantities in the environment as a waste material, causing many environmental problems such as soil and water contamination, as well as waste management issues (Usman et al. 2022). From the sustainable management perspective, several utilizing plans have been suggested for these waste materials such as using them in cement industry (Singh et al. 2019), rubber production (Ren & Sancaktar 2019), and other engineering or agricultural applications (Ahmaruzzaman 2010). In addition, effective recapture of P from wastewater can be considered an important alternative approach to utilize coal fly ash following its great potential to act as a P adsorbent.

Coal solid wastes have recently received much attention by environmental researchers to be used in P removal reactions from aquatic environments. In this context, some efforts have been made to enhance their P removal efficiencies through different modification procedures such as pre-treatment with different metal ions, and impregnation in acidic or alkaline solutions (Wang et al. 2016a; Xu et al. 2022). However, limited research data regarding enhancement of their adsorption capacity through simultaneous application of physical and chemical modification methods are available. Since the adsorption functionality of a material depends largely on its size, shape, and surface morphology, its smaller size results in higher chemical reactivity. It can be achieved through size reduction of the coal wastes to nanoscales, creating new waste-based nano-adsorbents. On the other hand, chemical surface modification of these nano-adsorbents through embedding high P affinity metals such as Fe can increase their active surface adsorption sites and improve their P removal performance. Reusing coal solid wastes to remove phosphorus from water would be a novel scheme that synergistically supports waste management and environmental remediation issues. Hence, the present study was conducted to investigate the P removal efficiency of the coal solid wastes as an industrial adsorbent in the raw, nano-size, and Fe-modified forms in comparison with similar forms of zeolite as a well-known natural P adsorbent. The adsorbents were prepared, modified, and characterized using SEM-EDS, FTIR, and XRD instrumental analyses and their P adsorption performances were evaluated under batch conditions as affected by solution P concentration, pH, and contact time. Furthermore, the plausible mechanisms involved in the P adsorption processes were explored and discussed in detail regarding the obtained data.

The coal solid wastes used in this study were collected from the Zarand Coal Washing Plant in the Kerman province, southeast Iran, and the natural zeolite particles were obtained from the Sartakht Natural Zeolite Mine located at the Semnan region, north-central of Iran. All chemical reagents used in this study were of analytical grade (purity > 99%) provided from Merck, Germany, including potassium dihydrogen phosphate (KH₂PO₄), potassium antimony tartrate (K(SbO)·C₄H₄O₆·0.5H₂O), ammonium heptamolybdate tetrahydrate ((NH₄)₆Mo₇O₂₄), hydrochloric acid (HCl), sulfuric acid (H₂SO₄), sodium chloride (NaCl), sodium hydroxide (NaOH), iron chloride hexahydrate (FeCl₃·6H₂O), and ascorbic acid. Distilled water (DW) was used in the preparation of all chemical solutions except for P stock solution (1,000 mg L⁻¹) which was prepared by dissolving an appropriate amount of KH₂PO₄ in 0.01 M NaCl solution as background electrolyte. Desired P concentrations were then diluted from the prepared stock solution.

The adsorbents used in this study included the raw coal waste particles (CPs), coal waste nanoparticles (CNPs), FeCl₃-modified coal waste nanoparticles (MCNPs), raw natural zeolite particles (ZPs), zeolite nanoparticles (ZNPs), and FeCl₃-modified zeolite nanoparticles (MZNPs). The collected raw materials were powdered to pass through a 270-mesh size sieve, washed with DW several times to remove their impurities, and oven-dried at 70 °C for 24 h. The CNP and ZNP were prepared through physical modification (nano-size reduction) of the powdered particles using a Ball-mill for 10 h. The chemical surface modification of the prepared nanoparticles was carried out following Wang et al. (2016b). First, the CNP and ZNP particles were impregnated with 2 M NaOH and 1 M HCl, respectively, for 24 h, and after drying, they were added to 500 mL DW and the suspension pH was adjusted around 13. Then, a 0.5 M FeCl₃ solution was added drop-wise into the stirring suspension until the pH value decreased to 5, and the suspension was allowed to precipitate for 24 h. Finally, the suspension was centrifuged and the remaining residues were oven-dried at 105 °C for 3 h, ground and sieved by a 270-mesh size sieve.

A MIRA3-TESCAN scanning electron microscope (SEM) equipped with an energy-dispersive spectroscopy detector (EDS) was used to investigate the surface morphology and elemental composition of the prepared adsorbents, respectively. The crystallographic structure of the adsorbents was analyzed over 2θ range of 10–60° using a Philips X'pert Pro MPD model X-ray diffractometer (XRD) with Cu Kα radiation, running at 40 kV and 30 mA. The surface functional groups and chemical bonds were identified through infrared spectra (400–4,000 cm⁻¹) using a Bruker Tensor 27 Fourier transform infrared spectrometer (FTIR).

The pH_pzc of the prepared adsorbents was determined following the method described by Feizi & Jalali (2016). Briefly, 10 mL of 0.01 M NaCl was placed into a series of 50 mL centrifuge tubes and their initial pH values were adjusted in the range of 2–12 using 0.1 M HCl and NaOH solutions. Then, about 0.1 g of each adsorbent was added to the prepared solutions and the obtained suspensions were shaken for 48 h at room temperature. Finally, the equilibrium pH of each suspension was recorded and plotted across the initial pH value. The pH_pzc for each adsorbent was obtained from the intersection of the plotted curve with the blank sample line.

In these relations, Q_t and Q_max (mg g⁻¹) refer to the P adsorption capacity at times t and equilibrium, respectively, k₁ (min⁻¹) and k₂ (mg g⁻¹ min⁻¹) represent the rate constant of pseudo-first-order and pseudo-second-order models, respectively, and kp (mg g⁻¹ min^−0.5) and C (mg g⁻¹) refer to the rate constant and the intercept of the intra-particle diffusion model from the origin, respectively.

In these relations, Q (mg g⁻¹) and C_e (mg L⁻¹) represent the amount of adsorbed P per unit mass of adsorbent and the equilibrium P concentration in the solution, respectively. The Q_max (mg g⁻¹) and K_l (L mg⁻¹) are the Langmuir constants referring to maximum sorption capacity and the affinity of P ions to the sorption sites, respectively, and the K_F (mg g⁻¹), (L mg⁻¹)^1/n, and n (g L⁻¹) are the Freundlich constants referring to the sorption capacity and intensity, respectively. β is the activity coefficient related to the mean adsorption energy per mole, R is the universal gas constant (8.314 J mol⁻¹ K⁻¹), and T is the absolute temperature (K).

The correlation coefficients (R²) and standard errors of estimate (SEE) values were used to determine the conformity between the experimental data and model-predicted values in both kinetic and isothermal studies.

Figure 1(b) illustrates the FTIR spectra of the studied adsorbents. The broad absorption bands between 3,400 and 3,500 cm⁻¹ in all adsorbent's spectra are signatures of –OH stretching vibration, demonstrating the presence of water in the adsorbents structure or between their pores. The absorption band around 2,925 cm⁻¹ in the MCNP spectrum refers to the asymmetric stretching vibration of a methylene group (–C–H), which appeared after modification of CNP with FeCl₃. Also, the 1,447 cm⁻¹ peak which has intensified after FeCl₃ treatment was attributed to the bending vibration of the methylene group. In addition, the 1,614, 1,037, and 797 cm⁻¹ absorption bands were found in the adsorbents FTIR spectra, referring to the O–H, Si–O, and Si–O–Si groups, respectively (Fazlzadeh et al. 2017). The Al(Si)–O–(Si)Al asymmetric stretching vibration which has intensified in the MCNP spectrum, indicated its zeolite-like surface chemistry as a consequence of alkaline conditions (NaOH treatment) before FeCl₃ modification (Kobayashi et al. 2020). Moreover, the 1,632 cm⁻¹ band in the MZNP spectrum was related to the presence of the Fe–OH group in the zeolite surface after treatment with FeCl₃ (Tandon et al. 2013).

Studying the P speciation in solution as affected by pH provides useful information regarding the mechanisms involved in the P adsorption process. In the pH ranges between 2 and 6, the anion is the predominant P species in the solution, which mainly participates in the P adsorption process through OH⁻ exchange mechanism, resulting in an inner-sphere complex formation on the adsorbent surface (Krishnan & Haridas 2008; Feizi & Jalali 2016). When increasing the pH toward 10, the secondary orthophosphate (⁠⁠) prevails in the solution and its higher adsorption free energy rather than the ⁠, leads to a decrease in P adsorption (Chubar et al. 2005). The pH_pzc refers to the pH value that the net surface charge of an adsorbent is zero, and the anion exchange capacity equals the cation exchange capacity (Sparks 2003). This means that in pH ranges lower than the pH_pzc, the adsorbent has a net positive surface charge and it has more affinity to adsorb anions from the solution electrostatically. Accordingly, in pH ranges higher than the pH_pzc, the adsorbent prefers the cations to adsorb. The higher pH_pzc values imply that the charge balance occurs at higher pH values, and therefore, the adsorbent has a positive surface charge over a wider pH range (Feizi & Jalali 2016). Figure 3(c) and 3(d) displays the pH_pzc determination curves for the studied adsorbents. As can be seen, the pH_pzc of the CP, CNP, and MCNP are approximately coincided with each other around pH = 8, indicating that the physical and chemical modifications of coal waste particles did not affect their pH_pzc. However, in the case of ZP, ZNP, and MZNP, the pH_pzc points have been increased after modifications, so their pH_pzc were obtained as 3.8, 5.8, and 6.5, respectively. Overall, the optimum pH range for P adsorption by the studied adsorbents was lower than their pH_pzc and a decreasing trend was found in their P adsorption capacity toward pH_pzc. This suggests that the P adsorption process by the studied adsorbents has an electrostatic nature.

The FeCl₃ modification showed higher performance in the case of coal-based adsorbents rather than the natural zeolite nanoparticles. As can be deduced from Table 1, the modification of CNP with FeCl₃ has enhanced its P adsorption capacity 10-fold (from 3.23 to 30.48 mg g⁻¹), while in the case of ZNP, a 2.5-fold increase is observed (from 2.84 to 6.99 mg g⁻¹). Krishnan & Haridas (2008) reported a five to six times increase in P adsorption capacity (Q_max = 70.92 mg g⁻¹) by the Fe-modified coir pith rather than the native one, and related it to the cation bridge effect of Fe cations on the adsorbent surface. In another study on the native and FeCl₃-modified plant residues, Feizi & Jalali (2016) reported an approximately 2.5-fold increase in P adsorption capacity after modification of the adsorbents with Fe. They obtained Q_max values of 2.8, 4.3, 4, and 3.7 mg g⁻¹ for native sunflower, potato, canola, and walnut shell residues, while these values were increased to 6.6, 9, 8.4, and 9 mg g⁻¹, after Fe modification, respectively. Wang et al. (2012) also studied the P removal from an aqueous solution using two series of activated carbons modified by Fe (II) and Fe (III) and reported 14.12 and 8.73 mg g⁻¹Q_max values, respectively. Overall, it can be concluded from the findings of the present study that the P adsorption capacity found for MCNP was comparable to other adsorbents used in literature, making this industrial waste capable, cost-effective, and eco-friendly P adsorbent (Table 2).

The basis of the intra-particle diffusion kinetic model is to determine the diffusion rate of the adsorbate toward the adsorbent at the liquid–solid interface, which can be investigated by plotting the Q_t values against the square root of the contact time (t^0.5) (Weber & Morris 1963). In this context, a linear diagram crossing the origin of the coordinates implies the intra-particle diffusion mechanism as the main rate-limiting stage in the adsorption process. The P adsorption process generally occurs in three fundamental steps including film diffusion, intra-particle diffusion, and adsorption by the active surface sites. First, the P anions in bulk solution move to the film surrounding the adsorbent following the diffusion gradient and diffuse along it toward the adsorbent surface. Then the adsorption process gradually continues through an intra-particle mass transfer of P anions into the macro pores of the adsorbent, and finally, the physical or chemical adsorption reaction occurs depending on the surface reactivity (Wang et al. 2012). The final equilibrium occurs as a result of reduction in P concentration in solution, diffusion of P into micro-pores of the adsorbent, and electrostatic repulsion of P from the adsorbent surface.

The feasibility to improve phosphorus adsorption capacity of coal solid wastes was investigated through physical (nano-size reduction) and chemical (FeCl₃) modification methods. Characterization analyses including XRD, SEM-EDS, and FTIR confirmed the successful embedding of Fe (III) particles on the surfaces of the prepared adsorbents. The P adsorption capacities of the coal-based adsorbents were enhanced after both modification methods by approximately 9-fold for the physical (0.36–3.23 mg g⁻¹) and 85-fold for the chemical (0.36–30.48 mg g⁻¹) treatments, respectively. Whereas, the zeolite adsorbents showed a 3.5-fold (0.80–2.84 mg g⁻¹) and 9-fold (0.80–6.99 mg g⁻¹) increase in their adsorption capacity at the same modification conditions. The Langmuir equation described the isothermal data well, with a 4-fold maximum P adsorption capacity for MCNP (30.48 mg g⁻¹) rather than the MZNP (6.99 mg g⁻¹). These values demonstrated the higher potential of the coal waste materials to be used as P adsorbents rather than the zeolite ones. The kinetic data were explained by the pseudo-second-order model, implying the chemisorption mechanism involved in the P adsorption process. Besides the surface complexation, ligand exchange and electrostatic attraction were also found as effective mechanisms contributed to the P adsorption process. It can be concluded from the findings of the present study that coal waste materials are great choices to be used as invaluable P adsorbents regarding their desirable adsorption properties, suggesting a win-win paradigm supporting the sustainable management of waste materials as well as environmental protection disciplines.

The authors declare that no funds, grants, or other support were received during the preparation of this manuscript.

All the co-authors read and approved the final manuscript. S.H. contributed to the materials preparation, chemical analysis, and data collection. M.H.-M. contributed to the conceptualization, design, methodology, data analysis, and writing the first draft. H.H. and M.H.F. contributed to the review and editing.

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

The authors declare there is no conflict.