Comparing the adsorption isotherms and kinetics of phosphate adsorption on various waste shells as adsorbent

Eutrophication is a main environmental issue globally and has received increased attention in recent decades (Kun et al. 2019). Phosphorus (P) is an important nutrient for the growth of biological organisms and algae (Abdul Salim et al. 2021a, 2021b). The release of high concentration of P into water bodies is a major cause of eutrophication (Abdul Salim et al. 2020). The excess P in water bodies promotes plant growth, reduces water quality, causes algal blooms, and kills aquatic life. The water bodies can also lose their vital functions, resulting in negative consequences to the environment and human health. The three common sources of P in water are polyphosphate, organically bound phosphate, and orthophosphate. Generally, orthophosphates are the most common P compounds found in wastewater (Jibing et al. 2011). The European Union (EU) and the United States Environmental Protection Agency (USEPA) have stated that uncontrolled of P release into surface waters may provoke legislation issues. The EU allows a discharge limit of 2 mg L⁻¹ per 100,000 population equivalents while the USEPA allows a total phosphorus (TP) effluent limit of less than 0.8 mg L⁻¹ (Abdul Salim et al. 2018).

Chemical precipitation, biological treatment, membrane separation, and adsorption are some of the phosphorus removal techniques that have been developed in the past few years to treat the excessive amount of P (Huan et al. 2021). Among these removal techniques, adsorption is frequently regarded as an excellent and promising approach for removing P from polluted water due to its high efficiency, ease of operation, and economic value (Jung et al. 2021; Payel & Animesh 2021). Some of adsorbents come from waste shells with potential to remove P from water due to calcium carbonate contained in the material. The abundance of waste shell produced in domestic usage can caused environmental problems (Tran et al. 2021). The use of waste shell, which represents wasted wealth, can give an excellent initiative to sustainability of environment cycle. Marsh clamshell (MCS), waste mussel shell (WMS), and eggshell (ES) are natural adsorbent materials that can be used to remove PO₄³⁻ from wastewater because all of them are waste materials from food processing industries that are normally dumped in foreshores (Nguyen et al. 2020; Aravind & Amalanathan 2021). The goal of this study was to investigate the feasibility of using MCS, WMS, and ES as adsorbents to remove PO₄³⁻ from a synthetic solution (KH₂PO₄). The effects of contact time (t) and various particle sizes on the removal of PO₄³⁻ were studied. To investigate the adsorption mechanism, Freundlich and Langmuir isotherm models were used in this study.

MCS, WMS, and ES were used as the adsorbents in this study. The samples were randomly collected from different locations in Malaysia. For the preparation of raw adsorbent, which is MCS, the sample was washed with tap water several times and rinsed using deionised water. The cleaned sample was dried in open air under sunshine for several hours before drying it in an oven at 30 °C for 2 days. The sample was then crushed using mortar and grinder before sieving it to sizes of 1.18, 0.60, 0.30, 0.15, and 0.075 mm. The sieved samples were then weighed and packaged into 2 g per package for each size. For the preparation of calcined adsorbents, which are WMS and ES, the samples were prepared in the same method as MCS. After the sieving process, the samples were calcined in a furnace at 800 °C for 2 h before being weighed and packaged (Nguyen et al. 2020).

PO₄³⁻ stock solution (100 ppm) was prepared by dissolving 0.1433 g potassium dihydrogen phosphate (KH₂PO₄) in 1 L deionised water and the pH range is 6.6–7.4 with average of pH 7.1. To achieve the required concentration of 10 ppm, the stock solution was diluted with deionised water and then analysed by using an HACH DR6000 UV–vis spectrophotometer to ensure that it can be adopted for the batch experiment (Abdul Salim et al. 2020).

HACH DR6000 UV–vis spectrophotometer was used to measure PO₄³⁻ concentration in the samples using molybdate and amino acid reagents. The surface morphology of MCS, WMS, and ES were observed by using a scanning electron microscope (SEM) (EM-30AX Plus, COXEM, Daejeon, Korea). The chemical compositions of the adsorbents were identified by using energy-dispersive X-ray fluorescence (EDXRF) spectrometer, while BRUKER D2 Phaser benchtop X-ray diffraction (XRD) was used to perform a detailed study of the crystal phase composition of the adsorbents. The functional group of the raw MCS, calcined WMS, and calcined ES were determined by using Fourier transform infrared (FTIR) spectroscopy (Perkin Elmer Spectrum Two FTIR Spectrometer, USA).

Adsorption kinetics is useful because it explains the potential rate-controlling step of the adsorption process as well as the true meaning of their observed rate coefficients (Azizian 2004; Liu & Shen 2008). Two types of adsorption kinetic models were used to explain the mechanism of the adsorption process in this study, which are pseudo-first-order (PFO) and pseudo-second-order (PSO) models (Table 1). The PFO equation was used to describe the sorption kinetics and can be written as shown in Table 1 (Ho & McKay 1998; Qiu et al. 2009). The PSO kinetic model describes the relationship between the adsorbent's adsorption capacity and time (Edet 2020). The PSO equation can be written as shown in Table 1 (Ho 2006).

Two adsorption isotherm models are commonly used to explain the adsorption isotherm, namely Freundlich and Langmuir models. According to the Freundlich model, the adsorbates from a heterogeneous surface are formed on the adsorbent's surface through multilayer sorption of different adsorption energies (Nodeh et al. 2017). In contrast, the Langmuir model explains monolayer sorption on distinct localised adsorption sites. It assumes uniform energies of monolayer sorption onto the sorbent surface and indicates no adsorbate transmigration in the plane of the surface (Singh et al. 2018). The linear form of the Freundlich isotherm model can be written as shown in Table 1 (Efome et al. 2018).

According to the Freundlich isotherm, a plot of ln q_e against ln C_e should yield a straight-line intercept at K_F with a slope of 1/n (Table 1). The affinity of the adsorbate–adsorbent is indicated by the adsorption coefficient K_F. The exponent n is related to the adsorbent surface's energetic heterogeneity and determines whether the curve is favourable or unfavourable (Worch 2012). The linear form of the Langmuir isotherm model can be written as shown in Table 1 (Li et al. 2013). K_L and q_max values represent the Langmuir isotherm constant and maximum adsorption capacity, respectively.

EDXRF, SEM, XRD, and FTIR analysis showed the characterisation of raw MCS, calcined WMS, and calcined ES. Figure 1 shows the surface morphology of the shells at 10000× magnification obtained from SEM. Figure 1(a) shows the SEM image of raw MCS that has a compact texture and low porosity (Abdul Salim et al. 2021a, 2021b). The SEM micrograph of calcined WMS is shown in Figure 1(b), while Figure 1(c) shows the SEM image of calcined ES. The calcined shells have small pores at the shell surface.

The elemental compositions of raw MCS, calcined WMS, and calcined ES were examined by EDXRF (EM-30AX Plus, COXEM, Daejeon, Korea). Table 2 shows the elemental composition of raw MCS, calcined WMS, and calcined ES. The major elements in all three shells were Ca and O, in which 27.96% Ca and 55.33% O were measured in raw MCS, 59.74% Ca and 38.92% O in calcined WMS, and lastly, 37.61% Ca and 53.92% O in calcined ES. Normally, materials that contain Ca have the potential to remove PO₄³⁻ from an aqueous solution (Malihe et al. 2019). So all three shell varieties have the potential to adsorb PO₄³⁻ from an aqueous solution.

The XRD patterns (refer Table 2) of raw MCS, calcined WMS, and calcined ES were determined by using BRUKER D2 Phaser benchtop XRD instrument. The XRD pattern (refer Figure 2(a)) for raw MCS sample indicates that the major components are aragonite (CaCO₃) and calcium carbonate (CaCO₃). Next, the XRD pattern of calcined WMS (Figure 2(b)) indicates that lime (CaO) is a major component of calcined WMS. Portlandite (HCaO₂), iron (Fe), oxygen (O₂), magnesium (Mg), and strontium cobalt manganese oxide are also present in the XRD pattern of calcined WMS. The EDX pattern of calcined ES (Figure 2(c)) shows that the major component is lime (CaO). The EDX pattern also shows the presence of other components such as portlandite (H₂CaO₂), periclase (MgO), oxygen (O₂), sodium peroxide (Na₂O₂), and magnesium (Mg). Other researchers have reported that lime (CaO) and aragonite (CaCO₃) have a good phosphate adsorption property (Xie et al. 2017).

The functional groups of raw MCS, calcined WMS, and calcined ES were determined by using FTIR spectroscopy (Perkin Elmer Spectrum Two FTIR Spectrometer, USA). The FTIR spectra of raw MCS, calcined WMS, and calcined ES before and after PO₄³⁻ adsorption are listed in Table 3. The difference in frequencies for (a) raw MCS was in the range of −3.37 to 0.11 cm⁻¹, (b) calcined WMS was in range of −2.58 to 0.51 cm⁻¹, and (c) calcined ES were the range of 1.45–6.9 cm⁻¹. The position and shape of the PO₄³⁻ stretching band in the FTIR spectra of raw MCS, calcined WMS, and calcined ES are influenced by the nature and position of the surface functional groups (Merenda et al. 2016).

The frequencies of raw MCS before PO₄³⁻ adsorption were 1,082.78, 857.45, 712.68, and 699.82 cm⁻¹ and significantly changed to 1,082.89, 856.92, 712.77, and 699.93 cm⁻¹ after PO₄³⁻ adsorption from synthetic solution. The difference in the frequency spectrum of 0.0011 cm⁻¹ (1,082.78–1,082.89 cm⁻¹) was attributed to the CO stretching vibrations (Yuri & Anatoliy 2011). Adsorption of PO₄³⁻ from aqueous solution onto the surface of raw MCS caused a change on the C = C bending frequency spectrum by 0.09 cm⁻¹ (712.68–712.77 cm⁻¹) and 0.11 cm⁻¹ (699.82–699.93 cm⁻¹) due to the interaction of PO₄³⁻ and C = C functional groups. A change in the C–H bending occurred after PO₄³⁻ adsorption where a new peak was formed at 1,455.63 cm⁻¹; in contrast, the C–H bending was attributed to the 1,459.36 cm⁻¹ peak before PO₄³⁻ adsorption.

The FTIR spectra of calcined WMS as shown in Table 3 significantly changed from 3,641.13, 1,410.29, 873.61, and 713.1 cm⁻¹ before adsorption to 3,641.64, 1,407.71, 872.48, and 712.09 cm⁻¹ after PO₄³⁻ adsorption. The difference of 0.51 cm⁻¹ (3,641.13–3,641.64 cm⁻¹) between two spectra was due to the O–H stretching at the surface of calcined WMS being affected by the vibration of the asymmetric stretching mode (Rao 2001). Other than that, the frequency spectrum of S = O stretching had a difference of −2.58 cm⁻¹ (1,410.29–1,407.71 cm⁻¹). The frequencies of C–H bending and C = C bending also shifted by −1.13 cm⁻¹ (873.61–872.48 cm⁻¹) and −1.01 cm⁻¹ (713.1–712.09 cm⁻¹), respectively.

PO₄³⁻ adsorption onto the surface of calcined ES resulted in an increase by 5.65 cm⁻¹ (3,645.3–3,639.65 cm⁻¹) on the O–H stretching frequency spectrum, because the interaction between PO₄³⁻ and O–H functional group can affect the vibration asymmetric stretching mode (Gilbert 1999). The band near 1,418.12 cm⁻¹ was attributed to O–H bending. The peak at 1,058.85 cm⁻¹ corresponded to C–O stretching (Peter 2018). Because of the vibration, this can influence the stretching (Norman et al. 1990). A new peak was observed at 3,379.56 cm⁻¹ and was associated with the stretching of the N–H functional group (Salim et al. 2021).

Figure 3 shows the phosphate removal efficiency by (a) raw MCS, which reached 73, 17, 21, 37, and 62%, while (b) calcined WMS reached 96, 96, 95, 92, and 94% and (c) calcined ES reached 98, 97, 93, 99, and 96% after a contact time of 5,760 min for raw MCS and 95 min for calcined WMS and calcined ES. For raw MCS, the PO₄³⁻ removal efficiency constantly increased for the 30, 60, 120, 180, 300, and 420 min. The increment occurred until a constant value was reached at 420 min as shown in Figure 3(a). The fast adsorption of PO₄³⁻ onto raw MCS during the initial stage was due to the abundance of free active sites on the adsorbent's surface, where it is related to the different SEM analysis on Figure 1 (Leone & Vincenzo 2020). It shows the fastest adsorption on early stage but reducing interaction of PO₄³⁻ due to deficiency of surface area contact (Abdul Salim et al. 2020). Figure 3(b) shows the phosphate removal efficiency of calcined WMS. The efficiency increased drastically during the first 5 min and continued to reach equilibrium. The times of equilibrium were 10, 15, 40, and 50 min for the adsorption of phosphate. The rate of phosphate adsorption onto calcined WMS increased because the adsorbent contained many active sites on its surface, where; CaO is major components in adsorbents as mentioned in Table 2 (Suprakas et al. 2020). PO₄³⁻ removal efficiency of calcined ES also rapidly increased during the first 5, 10, 15, and 20 min. Then, it slowly increased until it reached equilibrium at 30 min (Figure 3(c)). The availability of many free active sites on the adsorbent surface was the factor of the rapid phosphate adsorption on calcined ES.

Tables 4 and 5 show that the correlation coefficient for raw MCS (R² = 0.9991) for the PSO model was higher than that for PFO model (R² > 0.8001). Based on the lower F_e value and the R² value being near to 1, the PSO model is more suitable than the PFO model to explain the adsorption. For calcined WMS, the best kinetic model is the PSO model, in which the value of R² (0.9999) is larger than that of the PFO model (R² = 0.3273), and it also has a lower F_e value. For the calcined ES, the more suitable kinetic model is also the PSO model (R² = 0.9997), which has a higher correlation coefficient than the PFO model (R² = 0.5314) and lower F_e value. The adsorption process in this study can be classified as chemisorption because of the exchange of adsorbent and aqueous solution (Boyd et al. 1947).

The experimental data were also fitted to the Langmuir and Freundlich models for the analysis of the adsorption isotherm of PO₄³⁻ onto raw MCS, calcined WMS, and calcined ES. All the values of the parameters calculated using these two isotherm models are shown in Table 6. The experimental data are better fitted to the Freundlich isotherm model compared to Langmuir isotherm model, in which the R² is close to 1. The R² for raw MCS in the Freundlich model is better (0.9491) compared to that of the Langmuir model. It is the same with calcined WMS and calcined ES, in that the correlation is fitted better with R² = 0.3965 and R² = 0.9377, respectively as shown in Figure 4.

The outcome shows that all three shells can adsorb PO₄³⁻ from the synthetic solution. This is because the affinity between calcium oxide present on the surfaces of raw MCS, calcined WMS, and calcined ES and phosphate ions is quite strong since both compounds have different charges, causing them to attract each other (Barry & James 2016). Figure 2 shows the PO₄³⁻ adsorption efficiency rates onto calcined WMS and calcined ES were better and only took 95 min to reach the equilibrium point, compared to the PO₄³⁻ adsorption onto raw MCS, which took a long period to reach the equilibrium point (Mengyuan et al. 2021). This is because the calcination process produces a change in calcium carbonate and altered the physical and chemical properties of the WMS and ES. Calcium carbonate changes to calcium oxide at 600–800 °C (Rafia et al. 2021). This is supported by Figure 1 that shows the difference between the porosity of the calcined WMS and calcines ES with the raw MCS. The raw MCS has a compact texture, while calcined WMS and calcined ES have small pores in their structures.

The feasibility of using MCS, WMS, and ES as an adsorbent to remove PO₄³⁻ from synthetic solution was investigated. The adsorption kinetics data of MCS, WMS, and ES fitted well with the PSO model with R² values of 0.9991, 0.9999, and 0.9997, respectively, indicating that the PO₄³⁻ adsorption on the various adsorbents is a chemisorption process. The adsorption isotherms data of MCS, WMS, and ES fitted well with the Freundlich model, the study shows the PO₄³⁻ adsorption is suitable for MCS and ES. This present study provided more information on the adsorption of phosphate from aqueous solution onto raw MCS, calcined WMS, and calcined ES. The application of waste material to adsorb the phosphate from aqueous solution shows the potential of a new adsorbent for use in real adsorption wastewater treatment technologies.

This research was supported by the Ministry of Higher Education (MOHE) through the Fundamental Research Grant Scheme (FRGS/1/2020/TK0/UTHM/02/22) or Vot No. K308. The authors would also like to thank the Neo Environment Technology (NET), Centre for Diploma Studies (CeDS), Research Management Centre, Universiti Tun Hussein Onn Malaysia for its support.

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