Study on an effective industrial waste-based adsorbent for the adsorptive removal of phosphorus from wastewater: equilibrium and kinetics studies

Phosphorus is an unrenewable resource which is essential for the growth of organisms in ecosystems. However, excess of bioavailable phosphorus lead to eutrophication in water bodies, resulting the bloom of aquatic plant and algal growth, and cause dissolved oxygen depletion.

Phosphates (P) may come from agricultural fertilizers and other sources such as treated and untreated sewage, which is present in organic matter and also in soaps and detergents (Kuroki et al. 2014). It is of great importance to remove phosphates from wastewater before being dispersed into environment (Xie et al. 2014a). Typical removal methods such as biological treatment (Lopez-Vazquez et al. 2008), chemical precipitation (Lu & Liu 2010), ion exchange (Chubar et al. 2005), membrane filtration (Kim et al. 2008) and adsorption (Kuroki et al. 2014) have been successfully applied. Among them, biological P removal is limited in low concentration, chemical precipitation has been hindered by its high chemical consumption and excessive production of chemical sludge (Fang et al. 2014). Ion exchange may lose its uptake capability after multiple regenerations, while membrane filtration is limited by fouling and it is economically unattractive (Awual & Jyo 2011). Adsorption is recommended as one of the most effective immobilization processes for treating P. Adsorptive removal of P is high efficiency and easy handling even at low P concentration. Various adsorbents such as bentonite (Kuroki et al. 2014), zeolite (Jiang et al. 2013), red mud (Zhao et al. 2011), lanthanum hydroxide (Xie et al. 2014b) and iron-related adsorbents (Chen et al. 2015) were successfully used in the adsorptive removal of phosphate. However, the high cost of these adsorbents is the main obstacle for their wide application.

Lithium silica fume (LSF) is a byproduct of silicon or alloys containing silicon industry. The main constituents of LSF are SiO₂ and Al₂O₃ (Siddique 2011), containing some trace element such as iron, magnesium, calcium, and alkali oxides. Silicon fume has remarkable physical and chemical properties such as high chemical stability, extremely small particle size, large surface area, and high electrical resistivity (Siddique 2011), which make it a potentially attractive adsorbent.

Lanthanum compounds are highly active in P adsorption, phosphate precipitation by lanthanum with molar ration of 1:1 formed lanthanum–phosphate complex, which is known to be the least soluble among the rare earth-phosphate salts and able to form when present in low concentrations and at low pH values = 4 (Diatloff et al. 1993; Haghseresht et al. 2009). Many researchers trying to develop efficient P adsorbent by doping lanthanum to various kinds of material, such as bentonite (Spears et al. 2013), zeolite (Xie et al. 2014a) and montmorillonite (Tian et al. 2009), red mud (Zhao et al. 2011), et al. The resulting adsorbents showed remarkable improvement in P removal after doping with lanthanum; however, these materials are not easily accessible and are not cheap.

The emphasis of this work is to study the feasibility of using LSF, an industrial waste, for the removal of phosphorus from aqueous solution. The surface structure of the materials was investigated by means of scanning electron microscope, an N₂ adsorption–desorption technique (Brunauer–Emmett–Teller, BET), and X-ray diffraction (XRD). The effects of temperature and pH on phosphate adsorption capability were investigated. Moreover, the phosphate adsorption equilibrium, kinetics and thermodynamics by these novel adsorbents were described for further understanding of the adsorption mechanism.

The LSF used in the present study was obtained from a Regenerated Resources Recycling Co., Ltd of Sichuan Province, China. The chemical compositions of the LSF consists primarily of SiO₂ (60 wt.%), Al₂O₃ (18.33 wt.%), SO₃ (6.08 wt.%) and CaO (4.28 wt.%). There are also small amounts of alkali metals (K₂O 0.73%, Li₂O 0.43%, Na₂O 0.36%), iron metal oxides (Fe₂O₃ 0.56%), etc. All the chemicals including LaCl₃•7H₂O, KH₂PO₄, NaOH, HCl and H₂SO₄ were of analytical grade, which were obtained from Kelong Chemical Co. (China) and used as received.

Raw LSF with particle size <75 μm was selected for subsequent modification. After dried in oven at temperature of 110 °C to constant weight, 10 g of LSF was dispersed into 50 mL 0.08 mol/L LaCl₃•7H₂O solution, the pH was adjusted to 11.0 by dropwise adding 1 M NaOH under vigorous stirring. Afterwards, the solution was kept for 24 h before separating the solids and liquid. The resulting solids were washed with deionized water until pH neutral and dried at 110 °C, and then cooled in a desiccator. The modified LSF with lanthanum (III) solution was symbol as LLSF.

The crystallinity and phase identification of LSFs were determined by XRD (X-Pert PRO MPD) with Cu Ka as the radiation source under 36KV with 2θ range of 10–80°. The scanning electron microscopy (SEM) measurement was carried out using SEM (JSM7500F from JEOL Company). The specific surface areas and pore volumes were analyzed by using the N₂ adsorption–desorption method (Micromeritics ASAP 2460). The amount of lanthanum determination is analysis by X-ray fluorescence spectrometry (XRF-1800, Japan). pH point of zero charge (pH_PZC) was measured to investigate the total surface charge of the adsorbents using the method described by Moharami and Jalali (Moharami & Jalali 2013).

A series of batch tests were conducted to investigate the phosphate adsorption performances of LSFs. Phosphate solution was prepared by dissolving anhydrous K₂HPO₄ in deionized water. Typically, different dosage of the as-prepared adsorbent was added into 100 mL of 1.50 mg P/L phosphate solution in sealed conical flasks. After being shaken for a certain time at room temperature (25 ± 1 °C) at 120 rpm, the suspensions were filtered and the concentration of phosphate was analyzed according to ammonium molybdate spectrometric method by UV–visible spectrophotometer (Jiang et al. 2013). The effect of pH on phosphate removal was obtained by adding 0.1 M NaOH or 0.1 M HCl solution to desired pH values ranging from 1 to 10 with 100 mL solution containing 1 mg/L phosphate. Adsorption kinetic studies were carried out as follows: 1.6 g of LSFs was added in 800 mL phosphate solution with different initial concentration of 1 mg/L, 10 mg/L and 50 mg/L, respectively. The suspension was continuous mixing by magnetic stirrers for 8 h at room temperature (25 ± 1 °C) with an initial solution pH of 7. Samples were taken at various times during the adsorption process. An adsorption isotherm study was carried out by shaking 0.2 g of LSFs in 100 mL KH₂PO₄ solutions with different initial concentrations (5, 10, 15, 25, 50, 75, and 100 mg/L) for 6 h under 25, 30, 35 and 45 °C, with initial solution pH of 7.

Desorption experiments conducted to explore the recovery of the phosphate were carried out with pre-sorbed P-saturated LLSF. For the desorption tests, the phosphate-containing LLSF were added into 100 mL 0.1–4.5 M NaOH solution. The mixture was stirred for 6 h and then separated from the NaOH solution.

in which R (8.314 J/(mol K)) is the gas constant, T (K) is the absolute temperature, K_d is the thermodynamic equilibrium constant of the adsorption process (Chen et al. 2011). ΔH⁰ and ΔS⁰ values were evaluated according to Van 't Hoff equation.

in which q_e and q_t are the amounts of P adsorbed (mg/g) at equilibrium and time t (h), respectively. k₁ and k₂ (g/mg h) are rate constants of pseudo-first-order adsorption (1/h), and second-order adsorption (g/mg h).

in which k_id is the constant of intra-particle diffusion rate (mg/(g h^1/2)) and C is proportional to the extent of the boundary layer thickness.

Table 1 shows the specific surface areas and total pore volumes of the LSF and LLSF. The N₂ adsorption/desorption isotherm showed a mesoporous pattern. As can be seen from Table 1, the raw LSF has a surface area of 14.3683 m²/g with an average diameter of 9.02 nm. After lanthanum (III) modification, the surface area of LLSF increased 69.8% and reached 20.4024 m²/g, the average diameter increased to 11.3 nm. An increment of total pore volume was also observed, while micropore volume decreased 3.2%, suggesting some amount of mesopores or macropore were formed on LSF surface after lanthanum (III) modification.

Thermodynamic calculations by MINEQL + showed that four species of phosphate exist in solution, which is H₃PO_4, H₂PO₄⁻, HPO₄²⁻ and PO₄³⁻ with different ratios according to pH (Chubar et al. 2005). Phosphate acid undergoes dissociation and H₂PO₄⁻ and HPO₄²⁻ are superior in solution form with a solution pH between 8 and 10 (Ye et al. 2006). Moreover, in this pH range, the LLSF surfaces carry more negative charges and thus would significantly repulse the negatively charged species in solution (Ye et al. 2006). As a result, the P adsorption amounts decreased.

In order to further study the kinetic mechanism that controls the phosphate adsorption, pseudo-first-order model and pseudo-second-order model (Figure 5(a)) were employed to fit the experimental data. The corresponding parameters are listed in Table 2. From Table 2, we could see pseudo-second-order was better fitted than pseudo-first-order model, which suggests chemisorption might be involved in the adsorption process. Similar results were found in lanthanum-loaded tourmaline (Li et al. 2015), betonies (Kuroki et al. 2014) and oak sawdust (Wang et al. 2015). The initial adsorption rate (h) derived from the adsorption kinetic data using different initial P concentration of 1 mg/L, 10 mg/L and 50 mg/L are 40.323 mg/(g.h), 121.951 mg/(g.h) and 178.571 mg/(g.h), respectively. It confirms that faster removal rate was achieved under higher P concentration.

To further understanding the diffusion mechanism, we studied intra-particle diffusion model in the adsorption process. As can be seen in Figure 5(b), the linear fit of intra-particle diffusion model presents three linear portions, which indicates multi-stage adsorption process was involved. The first sharper portion is caused by the external surface adsorption or instantaneous adsorption which is driven by the initial phosphate concentration differences (Huang et al. 2014). The second linear portion is the gradual adsorption stage where the intra-particle diffusion is the rate-limiting step. The third portion is attributable to the final equilibrium state in which the intra-particle diffusion slows down due to less residual phosphate remains in the solution. As can be seen from Table 2, k_id1, k_id2 and k_id3 increase with increasing initial phosphate concentration, which could be attributed to higher concentration, provides a higher gradient which drives P into the pores. At the same initial concentration, the value of k_id2 is greater than k_di3, suggesting that intra-particle diffusion controls the process of adsorption (Xu et al. 2009).

Adsorption experiments were conducted at 25, 30, 35 and 45 °C to investigate the effect of temperature, with initial P concentration of 5–100 mg/L, LLSF dosage of 2 g/L, pH = 7. Positive value of △H⁰ of 25.007 KJ/mol can be observed from Table 4, which indicates that the adsorption process is endothermic, and a chemical reaction such as ion-exchange is involved during the adsorption process (Silva et al. 2013). The negative values of ΔG⁰ at 25–45 °C indicates that the adsorption of phosphate onto LLSF is spontaneous. The values of ΔS° were positive, thus indicating a good affinity of phosphate ions towards LLSF. The increases in adsorption capacity of phosphate at higher temperatures may be caused by the enlargement of pore size and/or activation of the LLSF surface (Benyoucef & Amrani 2011).

As can be seen from Table 5, lanthanum hydroxide has the highest adsorption capacity towards phosphate; however, lanthanum hydroxide is a relatively expensive material for phosphorus removal. Granulated ferric hydroxide (GFH) can remove phosphate well, and GFH is less expensive; however, 96 h contact time is needed for GFH to reach equilibrium, and GFH is highly dissolvable in acidic solution (Genz et al. 2004). Our LLSF shows greater adsorption capacity than some other adsorbents.

This study shows LLSF is effective for the removal of phosphate from aqueous solution, Langmuir isotherm best fitted the equilibrium data, and the maximum adsorption capacity was 24.096 mg P/g. pH and desorption studies showed that both chemisorption and ion exchange are involved in the adsorption process.

Adsorption kinetic data were adequately fitted by the pseudo-second-order kinetic model and, although intra-particle diffusion was not the only rate controlling step, it still played a significant role in the adsorption mechanism. Thermodynamic studies demonstrate that the adsorption process was endothermic and spontaneous in nature.