Na@La-modified zeolite particles for simultaneous removal of ammonia nitrogen and phosphate from rejected water: performance and mechanism

Rejected water is a low C/N ratio, high ammonia nitrogen and phosphate wastewater produced in the process of excess sludge dewatering (Cheng et al. 2010). Presently, the traditional treatment method for rejected water is to transport it to the first wastewater tank in wastewater treatment plants (WWTPs) and mix it with the total influent of the process (Hu et al. 2015). Although it only accounts for about 5% of the wastewater processed by WWTPs, it can contribute 10–50% of the nitrogen load and 10–80% of the phosphorus load to the biochemical pond (Ivanov et al. 2009; Torà et al. 2014; Ren et al. 2015). If the treatment scale is small, there is a huge risk that the effluent quality will not satisfy the discharge standard of the WWTPs and even threaten their usual operation. In China, the production of rejected water was about 310 million m³/d in 2020 (Hu et al. 2015; Ge et al. 2019). Hence, it is important to find a rejected water treatment technique to protect the water.

Presently, various wastewater treatment technologies have been developed to remove different pollutants, such as biological treatment, advanced oxidation technology, and adsorption (Zhang et al. 2016a, 2016b, 2016c; Kumari et al. 2020; Iqbal et al. 2021). Due to the advantages of simple processing technology and high efficiency, adsorption has attracted most attention (Iqbal et al. 2019; Pham et al. 2019; Song et al. 2020). The development and application of low-cost materials with high adsorption performance have been a focus of investigation. Until now, various materials have been widely used to remove ammonia nitrogen and phosphorus from wastewater, including biochar, bentonite, ion exchange resins, graphene, magnetic cationic hydrogel and zeolite (Copetti et al. 2016; Mazloomi & Jalali 2016; Zhang et al. 2016a, 2016b, 2016c; Dong et al. 2017; Xu et al. 2018).

Among them, natural zeolite is a three-dimensional framework composed of silicon, aluminum and oxygen (Motsi et al. 2009). The charge is unbalanced in the aluminum oxygen tetrahedron, therefore the natural zeolite surface is negatively charged (Jiang et al. 2018). To maintain charge equilibrium, alkaline earth metal cations are used to compensate around the tetrahedron. These metal cations can combine with the skeleton structure through weak bonds and flow effortlessly, so ion exchange reactions are easy with in aqueous solutions (Montalvo et al. 2012).

Consequently, cheap and high cation exchange performance makes natural zeolite promising in ammonia nitrogen treatment (Huang et al. 2015). Moreover, the special framework structure of natural zeolite has a huge pore volume and specific surface area. Lots of adsorption sites can be provided for phosphate, hence it is frequently used to eliminate phosphate in wastewater (Alshameri et al. 2014). These functional features make zeolite a potential adsorbent for removing ammonia nitrogen and phosphate from rejected water.

However, the cation exchange capacity of natural zeolite is fairly low. Additionally, the aluminum oxygen tetrahedron in the framework structure is negatively charged, which leads to weak phosphate adsorption capacity because of electrostatic repulsion (Jiang et al. 2013; Wen et al. 2018). Therefore, natural zeolite needs to be modified to improve its ion exchange and phosphate adsorption capacity. Modification with inorganic salts and metal oxides are extensive and excellent methods to enhance the nitrogen and phosphorus adsorption performance of zeolite (Xu et al. 2019). The cations in salt solutions have been used to substitute the alkaline earth metal cations in the zeolite framework (Fu et al. 2020), strengthening the exchange capacity of the zeolite.

Na⁺ has a stronger in ion exchange capacity for ammonium compared with various cations, hence a NaCl modification technique has been used to improve the nitrogen removal capacity of zeolite (Fu et al. 2020).

Some studies have focused on the loading modification of natural zeolites with metal oxides (Guaya et al. 2015; Li et al. 2017; Maulana & Takahashi 2018). The unique properties of metal oxides remedy the low phosphate adsorption capacity of natural zeolites. Lanthanum is a rare earth element with high phosphate binding ability and its excellent selectivity has attracted many researchers’ attention (Chen et al. 2016; Zhang et al. 2016a, 2016b, 2016c; Xu et al. 2017). In addition, the huge pore volume and specific surface of zeolite provide a number of attachment sites for lanthanum, and lanthanum has shown excellent phosphate adsorption after loading on the zeolite (Xie et al. 2014a, 2014b; Pham et al. 2019; Song et al. 2020).

Zeolite modified with Na and La can achieve better adsorption performance. However, most reported studies used zeolite powder, which made the recovery and reuse of adsorption materials difficult. Moreover, most studies on zeolite concerned the removal of ammonia nitrogen or phosphate individually. There are few investigations into the simultaneous removal of ammonium and phosphate by zeolite. Therefore, it is essential to prepare a recyclable material base on zeolite for the simultaneous removal of ammonia nitrogen and phosphorus from rejected water.

As mentioned above, cheap natural zeolite was selected as the raw material to prepare a composite material (sodium- and lanthanum-modified zeolite particles; Na@La-MZP) for rejected water treatment. The objectives of this study were: (a) to investigate the removal efficiency of Na@La-MZP for ammonia nitrogen and phosphorus, and study the important influencing parameters; (b) to analyze the adsorption characteristics of Na@La-MZP adsorption characteristics by kinetics and isotherms; (c) to study the simultaneous removal mechanism of ammonia nitrogen and phosphorus from rejected water; (d) to evaluate the material's reuse performance, the dynamic treatment capacity of actual wastewater and its economic performance.

The natural clinoptilolite used in this study was purchased from Zhejiang Shenshi Mining Co., Ltd (China), and its chemical components are displayed on Table 1. Natural zeolites of 40–60 mesh particle size were used as the original materials. All chemicals in this study, such as NH₄Cl, KH₂PO₄, LaCl₃·6H₂O, NaCl and C₂H₆O, were analytical grade and obtained from National Pharmaceutical Group Co., Ltd (China). The -N and phosphate content in rejected water from Hubei Township Wastewater Treatment Plant is shown in Table 2. Therefore, simulated wastewater was prepared by dissolving NH₄Cl and KH₂PO₄ for use in this study.

The preparation process of Na@La-MZP was as follows (Xu et al. 2017): Firstly, 10 g natural zeolite was added to 1.71 mM NaCl solution at a solid-liquid ratio of 1:10 kg/L, and the mixture was shaken at 50 °C for 6 hours. The resulting material was washed multiple times with deionized water until neutral, and dried at 105 °C for 12 hours. Next, 1.25 g LaCl₃·6H₂O was dissolved in 50 mL of 20 wt% ethanol solution and stirred for 1 hour. After that, 5 g NaCl-modified zeolite was added to mixture and the pH was adjusted to 11 with 2 M NaOH solution. The mixture reacted in a water bath at 60 °C for 2 hours and then aged in an air-blast drying oven at 80 °C for 12 hours. Next, the zeolite particles were rinsed using deionized water until neutral, and then dried.

Batch experiments were carried out to study the simultaneous removal performance of -N and phosphate in rejected water by Na@La-MZP. To determine the appropriate adsorbent dosage, different dosages (2.5–20.0 g/L) of Na@La-MZP were added to the rejected water (⁠-N: 50 mg/L, phosphate: 60 mg/L) at initial pH = 6 and shaken for 12 hours. The effect of initial pH on -N and phosphate adsorption was examined using 1.25 g Na@La-MZP and 100 mL rejected water in the pH range of 2–12, and samples were analyzed at intervals. The pH of the wastewater was adjusted with 0.10 M HCl and NaOH solution. Common anions such as Na⁺, Ca²⁺, K⁺, Mg²⁺, Cl⁻, ⁠, ⁠, and F⁻ were evaluated to analyze their effect on the adsorption performance of Na@La-MZP under stable conditions (dosage: 12.50 g/L, initial pH: 6). All the above experiments were performed at 25 °C with a shaking speed of 160 rpm. At the end of the adsorption period, the supernatant was filtered through a 0.45 μm membrane and the residual concentration in solution was measured by spectrophotometry.

Adsorption kinetics experiments for -N and phosphate were performed. For -N experiments, 1.25 g of Na@La-MZP was added to 100 mL solution with 60 mg/L phosphate and 20–500 mg/L -N. Then the mixtures were stirred at initial pH = 6 and 25 °C, and measured at intervals. The adsorption kinetics of phosphate was similar to -N. Water samples containing 50 mg/L -N and 30–300 mg/L phosphate were prepared. The isothermal adsorption experiments were performed at 25 °C, 35 °C, and 45 °C. The equations and parameters of relevant models are detailed in Text S1 of the Supplementary Information (Chen et al. 2013; Xu et al. 2019).

The adsorbent-saturated Na@La-MZP was collected to investigate reuse performance. 2 g adsorbent-saturated Na@La-MZP was added to 50 mL 1.0 M NaOH solution, and the mixture was stirred at 25 °C for 2 hours. Then, the supernatant was removed and 50 mL NaOH and Na₂CO₃ mixed solution was added to the original flask and stirred at 50 °C for 2 hours. The desorbed zeolite was washed to neutral with deionized water and dried at 105 °C for 12 hours. The zeolite adsorption-desorption experiment was repeated five times.

A fixed-bed column experiment was carried out in the adsorption column of the up-flow, which was made of a plexiglass column with inner diameter of 4 cm and a height of 30 cm. Coarse-grained quartz sand was arranged between the filter plate and zeolite to prevent adsorbent loss, then 100 g Na@La-MZP was packed into the column. A peristaltic pump was used to control the inlet water flow rate (10 mL/min) and empty bed contact time (12.56 min). The rejected water was applied in the continuous flow systems and analyzed at intervals.

The concentration of -N and phosphate were determined by Nessler's and molybdenum blue reagent spectrophotometry, respectively. The pH value of the solution was measured with a precision pH meter (STARTER 3100). The concentration of metal cations in solution was determined by inductively coupled plasma-optical emission spectrometry (ICP-OES, PerkinElmer, USA). The specific surface area, pore volume and mesopore distribution of the zeolite were observed using a multi-channel specific surface area and pore size analyzer (TriStar II 3020, USA) at −196.15 °C, and samples were degassed at 200 °C for 4 hours before testing. A field emission scanning electron microscope (SEM, JEM-7500F, Japan) equipped with energy dispersive X-ray spectroscopy (EDS, X-maxN80, Japan) was employed to determine the morphology and elemental composition of zeolite. The crystal structure of the zeolite was recorded by X-ray diffraction (XRD, D8 Advance, Germany). The chemical groups contained in the zeolite were detected by Fourier transform infrared spectrometer (FTIR, Nexus, USA) at room temperature according to the transmission method. The crushed sample and KBr powder were mixed uniformly and compressed into tablets before the FTIR investigation. The surface element types and chemical valences of the zeolite were measured by X-ray photoelectron spectra (XPS, Escalab 250Xi, USA) with monochromatic Al-Kα X-rays as anode. The binding energy (EB) was calibrated using the C 1s peak at 284.8 eV, and XPS Peak 4.1 software was used to perform the peak fitting analysis of O 1s. The surface zero potential (pH_pzc) of the zeolite was analyzed by the pH drift method (Sarma & Mahiuddin 2014; Reck et al. 2018).

The surface morphological characteristics of natural zeolite and Na@La-MZP are shown in Figure 1. The surface of the natural zeolite was rough, with a lot of debris particles (Figure 1(a)), however the Na@La-MZP showed an obvious network structure (Figure 1(b)). Simultaneously, a large number of rod-shaped microcrystals was discovered in its surface, and these crystal structures were similar to La(OH)₃ (Dong et al. 2017). Compared with natural zeolite, the manufactured Na@La-MZP presented bigger specific surface and pore size (Supplementary Information, Table S1), which was beneficial for -N and phosphate removal (Pham et al. 2019). Additionally, the proportion of Na and La element increased from 0.56 to 0.74% and 0.00 to 40.80%, respectively (Figure 1(d)). This observation further confirmed that the adsorption properties of zeolite were enhanced.

The crystal structure of Na@La-MZP is shown in Figure 1(c). The main diffraction peaks of zeolite appeared at 11.22°, 22.45°, 25.67° and 26.65°, which matched the diffraction peaks of clinoptilolite (JCPDS 80-1557). Diffraction peaks of 16.62°, 27.85°, 39.46°, and 48.54° appeared in Na@La-MZP, which proved to be the characteristic peaks of La (OH)₃ (JCPDS 83-2034). Therefore, it can be concluded that the lanthanum compound on the zeolite surface was La(OH)₃, which was consistent with the SEM results.

Figure 1(d) depicts the pH_pzc of natural zeolite and Na@La-MZP. The pH_zpc of Na@La-MZP increased from 6.81 to 7.89 compared with natural zeolite. This result could be attributed to the increase in hydroxyl groups on the surface of the zeolite (Yu & Chen 2015). Simultaneously, the high pH_zpc implied that the Na@La-MZP was electropositive in a neutral solution, and phosphate adsorption performance was improved by electrostatic attraction.

To determine the optimal dosage of Na@La-MZP, the effect of dosage on the simultaneous removal of -N and phosphate was studied. In Figure 2, the removal efficiency of -N and phosphate were improved with the increase of Na@La-MZP dosage. It was observed that the removal efficiency of -N was stable when the dosage reached 7.50 g/L, while the removal efficiency of phosphate increased linearly until the dosage was increased to 12.50 g/L. In addition, it is worth noting that the adsorption amount of per unit mass of zeolite decreased with increasing dosage. This result could be attributed to the fact that available sites were superimposed during adsorbent aggregation (Xie et al. 2014a, 2014b). Thus the adsorption efficiency of Na@La-MZP was affected. Considering the economic and removal efficiencies, the adsorbent dosage of Na@La-MZP was selected as 12.50 g/L in subsequent experiments.

The solution pH is a critical factor in the adsorption process, because the physico-chemical properties of adsorbent and the type of adsorption are affected by different pH (Liu et al. 2013; Xie et al. 2014a, 2014b). As shown in Figure 3(a), the Na@La-MZP showed excellent adsorption performance for -N in the pH range of 4–8, and the removal efficiency reached the highest point, 92.61%, at pH 4 after 12 hours. However, the removal efficiency of -N dropped to 69.58 and 20.49% at pH 2 and 12, respectively. These results could be due to the smaller diameter H⁺ affecting the ion exchange performance of Na@La-MZP in a strongly acid solution (Vu et al. 2017). However, due to the reaction between and OH⁻ in a strongly alkaline solution, free NH₃ is formed. NH₃ is a neutral molecule and could not exchange with metal cations in the Na@La-MZP (Huang et al. 2010), so the removal efficiency of -N decreased.

The main chemical form of phosphate was in the pH range of 4–6, which made ligand exchange with OH⁻ on Na@La-MZP easier than the other types. Furthermore, the pH_zpc of the modified zeolite was 7.89. The positively charged La-(OH)²⁺ was formed through the deprotonation process of metal active groups on the Na@La-MZP surface (Liu et al. 2013), thus the ligand exchange process was promoted and the negatively charged was adsorbed by electrostatic attraction (Dong et al. 2017). Lanthanum on the Na@La-MZP surface changed to negatively charged La-O⁻ in an alkaline environment (Xie et al. 2014a, 2014b). In this environment, the electrostatic repulsion between Na@La-MZP and HPO₄²⁻ resulted in the reduction of ligand exchange. Therefore, maintaining the pH at 4–6 could mean a crucial improvement in the simultaneous removal of -N and phosphate in rejected water by Na@La-MZP.

The adsorption performance of Na@La-MZP was significantly affected by the various complex ions in rejected water. Therefore, the common metal cations of Na⁺, Ca²⁺, K⁺, Mg²⁺ and anions of Cl⁻, ⁠, ⁠, ⁠, F- were selected in this study, and the results are illustrated in Figure 4. The -N adsorption capacity of Na@La-MZP sharply reduced in the presence of K⁺. In contrast, the -N adsorption capacity was found to slightly decrease in the presence of Mg²⁺. This phenomenon was attributed to the exchange sequence and hydration radius of cations. Weatherley & Miladinovic (2004) reported the selective exchange order of zeolite was K⁺ > > Na⁺ > Ca²⁺ > Mg²⁺. Furthermore, the K⁺ ion has the smallest hydration radius compared to other cations (Rajic et al. 2011; Tansel 2012), hence it was easily adsorbed by the Na@La-MZP. This was why the competitive effect of K⁺ was greatest for -N removal.

The effect of coexisting anions on the phosphate adsorption performance of Na@La-MZP is shown in Figure 4(b). There was no obvious impact on phosphate adsorption in the presence of and ⁠. It is worth noting that the phosphate adsorption rate dropped significantly in the presence of Cl⁻, which resulted in a longer adsorption equilibrium time. However, phosphate removal efficiency significantly declined from 99.67 to 92.69% and 72.46% in the presence of and F⁻. The decrease of phosphate adsorption performance resulted in ligand exchange between deprotonated hydroxyl groups on the Na@La-MZP and coexisting ions (Shi et al. 2019). In addition, the affinity of Na@La-MZP to was higher than PO₄³⁻ because the K_SP value of La₂(CO₃)₃ (3.98 × 10⁻³⁴) was lower than LaPO₄ (3.7 × 10⁻²³) (He et al. 2016; Koilraj and Sasaki 2017).

Kinetic analysis was performed to investigate the adsorption of -N and phosphate from rejected water by Na@La-MZP. In Figure S2 (Supplementary Information), -N and phosphate were quickly adsorbed in the early stages, and then reached equilibrium slowly. Simultaneously, the adsorption capacity of Na@La-MZP was improved with the increase in their initial concentrations. The adsorption capacity reached maximum values of 15.83 mg/g and 8.98 mg/g when the initial concentrations of -N and phosphate were increased to 500 and 300 mg/L, respectively. The fitting results of pseudo-first-order, pseudo-second-order dynamic models and correlation parameter of models are shown in Table 3. The correlation coefficients (R² > 0.99) of the pseudo-second-order kinetics were higher than the pseudo-first-order kinetics for the adsorption of -N and phosphate on Na@La-MZP, and the saturation adsorption amounts were closer to actual data. This result implied that pseudo-second-order dynamic models were more suitable to describe the simultaneous adsorption process of -N and phosphate on Na@La-MZP, which suggested that the adsorption process was controlled by chemical forces (Zhu et al. 2012).

The intra-granular diffusion model was applied to clarify the diffusion mechanism of -N and phosphate on Na@La-MZP (Supplementary Information, Figure S3, Table S2). The relation diagram between q_t and t^1/2 showed that fitting lines of -N and phosphate did not pass through the origin, indicating that the adsorption rate was jointly controlled by liquid film diffusion and particle diffusion (Qiu et al. 2017). In addition, the intra-particle diffusion model showed two linear parts, indicating that the adsorption of -N and phosphate was a multistage adsorption processes onto Na@La-MZP (Shi et al. 2019). In the first stage, the adsorption sites on the Na@La-MZP surface were quickly occupied by -N and phosphate because of the adsorbent concentration difference and electrostatic attraction. When the surface adsorption sites reached saturation, -N and phosphate diffused into the interior of Na@La-MZP and combined with the interior sites. As the amounts of -N and phosphate in the interior of Na@La-MZP increased, the ion diffusion rate decreased with the increase of diffusion resistance, which was consistent with the fact that the rate constant k_d1 was obviously larger than k_d2. At the same time, the rate constant proved that intragranular diffusion process decided the adsorption rate-control of -N and phosphate (Huang et al. 2014).

Adsorption isotherms were used to describe the relationship between the adsorption capacity of the material and the residual concentration in solution (Özer & Dursun 2007). The experimental data were analyzed with Langmuir and Freundlich isotherm adsorption models, and the fitting results are presented in Figure 5 and Table S3 (Supplementary Information). According to the correlation coefficient (R²), two isothermal adsorption models could fit the adsorption process of Na@La-MZP for -N. But the R² value of Langmuir (>0.98) was higher than Freundlich (>0.93); it showed that -N combination sites were evenly distributed on the Na@La-MZP surface and that adsorption was a monolayer adsorption process (Wang et al. 2005). The maximum adsorption amount of -N calculated with the Langmuir model was 17.92 mg/g in 45 °C. It is worth noting that the Freundlich (R² > 0.98) could better describe adsorption behavior of phosphate on Na@La-MZP compared with Langmuir (R² > 0.91), indicating that phosphate adsorption was a multi-layer chemical adsorption process, and there were different types of chemical reactions on Na@La-MZP surface (Wang et al. 2005). In addition, the 1/n value in the Freundlich equation is frequently used to describe the adsorbent ability of adsorbents. In this study, values of 1/n were less than 0.5, which showed that phosphate was easily adsorbed (Lǚ et al. 2013). The maximum adsorption amount of phosphate was 9.53 mg/g. Isotherm analysis showed that the adsorption mechanism of Na@La-MZP for -N and phosphate were significantly different.

The zeolite structure contains many cavities and different size pores. The zeolite framework is negatively charged due to the unbalanced charge of the aluminum oxygen tetrahedron. In order to maintain the charge balance, alkaline earth metal cations are needed to compensate. Common metal cations are Na⁺, K^+, Ca²⁺, and Mg²⁺. These metal cations are combined with the framework structure through weak bonds in the cavities and pores inside the zeolite. Due to the weak binding force of metal ions with the zeolite framework, it shows great fluidity and ion exchange can occur with in rejected water. Therefore zeolite could effectively remove ammonia nitrogen. Table S4 (Supplementary Information) displays the concentration and proportion of cations in solution after adsorption equilibrium with the initial -N concentration of 50 mg/L. Clearly, the metal cation content in solution increased significantly in the adsorption equilibrium. This was because the weak binding force within the framework of metal cations in the zeolite and metal cations could be replaced into the solution by the ion exchange reaction with in rejected water. Specifically, the Na⁺ content increased from 0 to 44.39 mg/L and accounted for 89.78% of total cations, which was significantly higher than other metal cations. These results indicated that ion exchange between and Na⁺ on Na@La-MZP was the vital mechanism for -N removal in rejected water. It is worth noting that the total amount of cations in solution after adsorption equilibrium was slightly lower than the amount on Na@La-MZP, which might be attributed to few being adsorbed by electrostatic action.

The FTIR spectra of Na@La-MZP before and after adsorption are shown in Figure S4 (Supplementary Information). The absorption peaks near 1,043, 796 and 462 cm⁻¹ in the figure are attributed to the unique framework (Si/Al-O tetrahedron) vibration of zeolite. For example, the absorption peak near 1,043 cm⁻¹ is Si-O-Si asymmetric tensile vibration, the absorption peak near 796 cm⁻¹ is the symmetrical tensile vibration of Si-O-Si or Si-O-Al, and the absorption peak near 462 cm⁻¹ is the bending vibration of Si-O or Al-O. In addition to the zeolite skeleton vibration, there are also obvious O-H vibrations such as the O-H stretching vibration in the hydroxyl group near 3,612 and 3,468 cm⁻¹ and the O-H bending vibration in the bound water near 1,639 cm⁻¹. A new band at 1,401 cm⁻¹ was clearly observed on the adsorbed Na@La-MZP, which was attributed to N-H bending vibration of ⁠. Simultaneously, it was observed that the vibration frequency of Si-O-Si or Si-O-Al bands near 796 cm⁻¹ was weakened and shifted to low wave number. It was related to the reduction of Na⁺ content in the adsorbed Na@La-MZP, so that the spatial structure of material was affected (Król et al. 2012; Huang et al. 2014).

These results further demonstrated that the removal of -N was related to Na⁺ on Na@La-MZP. The bands around 3,612 and 3,468 cm⁻¹ were ascribed to the O-H stretching vibration in hydroxyl group (Yu & Chen 2015). The O-H band was significantly weakened after adsorption, which could be attributed to the ligand exchange between hydroxyl groups in the Na@La-MZP and phosphate. In addition, the band initially at 1,043 cm⁻¹ shifted to 1,047 cm⁻¹ and the vibration frequency was strengthened, which was assigned to the asymmetric tensile vibration of P-O (Li et al. 2009). At the same time, the band vibration frequency at 611 cm⁻¹ was significantly enhanced due to the bending vibration of O-P-O (Liu et al. 2013). Therefore, Na@La-MZP surface inner-sphere complexes were formed by ligand interaction.

The XPS characterization of material was used to further explore the removal mechanism of -N and phosphate by Na@La-MZP (Figure 6). A new peak of P 2p (132.34 eV) clearly appeared after adsorption (Figure 6(c)), and the N 1s (401.31 eV) peak was also enhanced (Figure 6(b)). These phenomena indicated that P and N were successfully adsorbed by Na@La-MZP. In detail, the binding energy of P 2p XPS spectrum was reduced by 1.66 eV compared with the standard P 2p (134.00 eV) of pure KH₂PO₄ (Zhang et al. 2016a, 2016b, 2016c), which confirmed the formation of new phosphate complexes on Na@La-MZP (Acelas et al. 2015; Wu et al. 2019). Two sets peaks of La 3d in the XPS spectrum were observed (Figure 6(e)). E_B[La 3d_5/2] = 834.38 eV and 838.08 eV and E_B[La 3d_3/2] = 851.28 eV and 854.88 eV before adsorption were consistent with the binding energy of La(OH)₃ (Sunding et al. 2011). The peak values of La 3d_5/2 (834.88 eV and 838.28 eV) and La 3d_3/2 (851.38 eV and 855.18 eV) moved in the direction of high binding energy. In addition, Δ (La 3d_5/2) decreased from 3.70 eV to 3.40 eV while Δ (La 3d_3/2) increased from 3.60 eV to 3.80 eV, which indicated that La-OP inner-sphere complexes were formed under strong interaction force on Na@La-MZP (Fang et al. 2018; Hao et al. 2019).

According to the different binding energy of different oxides, O 1s spectrum of Na@La-MZP was separated into three types: bound water (H₂O) corresponds to 532.09 eV, metal-bound hydroxyl (M-OH) corresponds to 531.07 eV and metal oxides (M-O) correspond to 530.11 eV (Figure 6(f)) (Islam et al. 2014; Yu & Chen 2015). The detailed analysis of O 1s before and after adsorption is summarized in Table S5 (Supplementary Information). Note that the relative content of M-OH in Na@La-MZP decreased from 60.05 to 49.42%, while M-O increased from 26.96 to 35.61%. These results indicated ligand exchange of O-H with O-P, and inner-sphere complex formation in La-O-P (Yu & Chen 2015; Zhang et al. 2016a, 2016b, 2016c; Shi et al. 2019).

In summary, the effective removal mechanism of -N could be attributed to ion exchange with Na⁺ in Na@La-MZP, and the removal mechanisms of phosphate were electrostatic adsorption, formation of La-O-P inner-sphere complexes and ligand exchange of hydroxyl groups.

Recycling performance is an important index to evaluate the performance of adsorbents. The -N and phosphate in the saturated modified zeolite were regenerated successively by two-step desorption. From Figure S5a (Supplementary Information), NaOH solution showed good desorption capacity for -N compared with phosphate. The -N desorption efficiency could reach 93.89% while phosphate was only 4.40% in 1.0 M NaOH solution. This difference in desorption might be attributed to the ion exchange reaction between Na⁺ and on the modified zeolite and PO₄³⁻ being less affected by NaOH. Therefore, NaOH-Na₂CO₃ mixed solution was used as regenerators in the second step of desorption to further improve the phosphate desorption efficiency of Na@La-MZP. The desorption efficiency is shown in Figure S5 (Supplementary Information).

The phosphate desorption rate gradually increased with the rise of NaOH concentration at a fixed molar ratio of NaOH and Na₂CO₃, while the -N desorption efficiency remained at about 4.00%. The phosphate desorption efficiency reached a maximum of 66.47% with 3 M NaOH mixed with 3 M NaCO₃. Thus, the concentration of NaOH and Na₂CO₃ mixed solution was selected as the desorption agent to perform the second step of the regeneration of saturated Na@La-MZP.

Adsorption-desorption cycle regeneration experiments were performed five times to evaluate the reusability of Na@La-MZP (Figure 7). Overall, the adsorption performance of Na@La-MZP decreased gradually with the increase of the regeneration times. The removal efficiency of -N and phosphate decreased significantly, and it decreased from 91.84 and 98.84% to 81.42 and 85.72% in second adsorption, respectively. After five cycles of adsorption-desorption, the removal rates of -N and phosphate decreased by 16.32 and 21.41%, respectively, compared with the initial Na@La-MZP. This phenomenon of adsorption capacity reduction was attributed to the leaching of Na/La compounds and presence of some irreversible adsorption sites such as chemical adsorption sites of phosphate. Nevertheless, the removal rate of -N and phosphate could remain above 75% after five cycles, which indicated that Na@La-MZP could be reused in rejected water treatment.

In order to prove the practical application potential of Na@La-MZP, a fixed-bed adsorption experiment was carried out with the rejected water of Hubei Township Sewage Treatment Plant. The water quality parameters of wastewater are listed in Table 2. The removal efficiency of Na@La-MZP with -N and phosphate in actual wastewater was slightly lower than that in simulated wastewater (Figure S6, Supplementary Information). In particular, the breakthrough time of -N and phosphate in actual wastewater was reduced by 3.11 hours and 4.53 hours, respectively. This was because there are various competitive ions and organic pollutants in actual wastewater alongside the target pollutants. These substances will occupy the active adsorption sites on the surface of the adsorbent, thereby influencing the adsorption performance of the material. Nevertheless, Na@La-MZP still showed excellent treatment capacity for actual wastewater from the perspective of actual treatment effect, which indicated that Na@La-MZP could be used as a potential adsorbent to simultaneously remove -N and phosphorus from rejected water.

Na@La-MZP after adsorption can be used in agriculture and as a raw material in fertilizers because these zeolite solids are loaded in nitrogen and phosphorus (Markou et al. 2014). In addition, its selling price can justify the expense of zeolite consumed during the process of adsorption. The cost of the process was mostly due to the brine used for the modification of the zeolite. Thus, ignoring the staffing and energy costs, an estimated cost for treating 1 L rejected water is dependent on the cost of 8.84 mM LaCl₃·6H₂O (market price 0.074 US$/kg). Consequently, the treatment cost of the proposed process can be calculated to be 0.22 US$/m³ rejected water.

In this study, a new type of Na@La-MZP adsorbent was successfully fabricated and used to simultaneously remove -N and phosphorus from rejected water. Na@La-MZP could remove 92.61% ammonia nitrogen (50 mg/L) and 99.01% phosphate (60 mg/L) at the optimal conditions of dosage 12.5 g/L, initial pH 6.0 and reaction time 12 hours, which made the effluent satisfy the discharge standard (GB 18918-2002) for municipal WWTPs in China. The Na@La-MZP exhibited excellent adsorption performance for -N and phosphorus at pH 4–6.

Coexisting cations inhibited the removal of ammonium in the order K⁺ > Na⁺> Ca²⁺ >Mg²⁺, and anions affected the adsorption of phosphate in the order F⁻ > > Cl⁻ > > ⁠. Pseudo-second-order kinetics could well describe the adsorption process, and the adsorption rate was mainly controlled by the intra-particle diffusion process. The effective removal mechanism of -N could be attributed to ion exchange with Na⁺ in Na@La-MZP and the removal mechanisms of phosphate were electrostatic adsorption, formation of La-O-P inner-sphere complexes and ligand exchange of hydroxyl groups.

The results of reuse and actual wastewater treatment experiments implied that Na@La-MZP as an adsorbent could simultaneously and efficiently remove -N and phosphate. Research into inexpensive modification technology will be a vital development direction in the future. We suggest strengthening the research on the regeneration of zeolite, which would help to enlarge the utilization period of zeolite. The development of economical synthetic zeolite is an essential research direction for achieving the sustainable utilization of zeolite.

This work was supported by National Natural Science Foundation of China project 51108360 and 51208397, Science and Technology Infrastructure Program in Hubei (Grant No.2015BCA304).

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