La₂(CO₃)₃ was obtained from Guangfu Institute of Fine Chemicals, China. All other chemicals used (KH₂PO₄, Na₂SiO₄·9H₂O, NaCl, NaNO₃, Na₂CO₃, and Na₂SO₄) were of analytical grade. The measurements of pH were carried out using a PHS-3C pH-meter (Dapu instrumentation Corp., Ltd. Shanghai, China). All glassware used in experiments was carefully cleaned and rinsed with deionized water. The samples of natural organic matter (HA and FA) were collected from the soils of Jiufeng Mountain (Beijing) (Lin et al., 2011).

La₂(CO₃)₃ was equilibrated with a suitable amount of PO₄ ³⁻ solution (10–100 mg L⁻¹) using magnetic stirring for 20 h. Once the adsorbent was recovered by centrifugation, the concentration of PO₄ ³⁻ in the supernatant was measured using the ammonium molybdate blue method. The adsorption amount of PO₄ ³⁻ was calculated based on the difference between the balance and total amount. The influence of temperature was evaluated by setting the concentration of PO₄ ³⁻ from 10 to 40 mg L⁻¹ and applying a temperature which ranged from 303.15 to 323.15 K. To determine the concentration effect, the weight from 20 to 100 mg was added to 50 ml of PO₄ ³⁻ solution (200 mg L⁻¹), and then the suspension was agitated on a shaker for 24 h.

To study the influence of coexisting inorganic anions (Cl⁻, SO₄ ²⁻, NO₃ ⁻, CO₃ ²⁻, and SiO₃ ²⁻), 40 mg of La₂(CO₃)₃ was mixed with PO₄ ³⁻ solution (100 mg L⁻¹) and various competing ions. The effects of natural organic matter including HA and FA were also investigated, by combining 40 mg of La₂(CO₃)₃ with 100 mg L⁻¹ of PO₄ ³⁻ solution and HA and FA at the concentrations of 10, 30, or 50 mg L⁻¹. The mixture was shaken for 24 h, and the suspension was filtered through a 0.45-μm fiber membrane. The concentration of PO₄ ³⁻ in the solution was measured using the ammonium molybdate blue method.

The pH_zpc of La₂(CO₃)₃ was estimated according to the △pH method (Kinniburgh et al., 1975; Zhang et al., 2014), for which 50 mg of La₂(CO₃)₃ or PO₄ ³⁻-saturated La₂(CO₃)₃ was mixed with 50 ml of NaNO₃ (0.01 mol L⁻¹). These mixtures were shaken at room temperature for 20 h on an automatic shaker and then adjusted to various values of pH by additions of NaOH or HNO₃. After 60 min of equilibrium, pH was measured and defined as pH_(initial). Then, 1 g of NaNO₃ was added to each suspension. After shaking for 1 hour, the pH_(final) was measured, and the change in pH (△pH) was calculated as follows: pH_(final)-pH_(initial).

A JEOL JSM-6700F scanning electron microscope (SEM) was used to measure surface morphology. The energy dispersive spectrometer (EDS, Oxford X-MAX-20) associated with the SEM system and FTIR (Nicolet 6,700, United States) was utilized to examine chemical compounds on the surface and shape of the adsorbent before and after adsorption. Powder X-ray diffraction (XRD, Bruker D8 Advance, Germany) was also used to characterize the adsorbent. The chemical composition of La₂(CO₃)₃ after adsorption was determined by XPS (ESCALAB250 Thermo-VG Scientific, United States). The release of CO₃ ²⁻ and HCO₃ ⁻ from La₂(CO₃)_3, during adsorption was detected by acid–base titration.

The FTIR spectrums of La₂(CO₃)₃ prior to and after adsorption of PO₄ ³⁻ are shown in Figure 1. Compared with the spectrum of La₂(CO₃)₃, some vibration peaks corresponding to CO₃ ²⁻ at 1,420, 878, and 713 cm⁻¹ almost disappeared, while other peaks at 1,054, 616, and 542 cm⁻¹ were observed after the adsorption of PO₄ ³⁻. Furthermore, the peak at 1,054 cm⁻¹ was assigned to the asymmetric stretching vibration of P-O of the PO₄ ³⁻ group, and the peaks at 616 and 542 cm⁻¹ were assigned to the bending vibration of O-P-O (Li et al., 2014; Wang et al., 2016), which indicated that the adsorption mechanism of PO₄ ³⁻ on the surfaces of La₂(CO₃)₃ included a ligand exchange process.

The typical SEM images of La₂(CO₃)₃ prior to and after adsorption of PO₄ ³⁻ are shown in Figure 2. La₂(CO₃)₃ exhibited a more regular surface and better particle dispersion after the adsorption of PO₄ ³⁻. The adsorption of PO₄ ³⁻ on La₂(CO₃)₃ was confirmed by energy-dispersive spectroscopy analysis. As shown in Figure 2D, the characteristic peaks of P appeared in the spectra of La₂(CO₃)₃ after the adsorption of PO₄ ³⁻, which indicated that PO₄ ³⁻ was successfully adsorbed on the surfaces of La₂(CO₃)₃. Meanwhile, the characteristic peaks of C decreased significantly after the adsorption of PO₄ ³⁻ on La₂(CO₃)₃, evidencing that CO₃ ²⁻ was mostly replaced by PO₄ ³⁻ on the surfaces of La₂(CO₃)₃ (Huang et al., 2007).

The analysis of the powder X-ray diffractograms of the La₂(CO₃)₃ prior to and after adsorption of PO₄ ³⁻ was illustrated in Fig. S1, in which the XRD standard diffraction card was also presented. La₂(CO₃)₃ used as an adsorbent in this study can be indexed as La₂(CO₃)₃·8H₂O (JCPS card NO.25-1,400) from the XRD patterns. The XRD pattern of La₂(CO₃)₃ after the adsorption of PO₄ ³⁻ was closely matched to LaPO₄·0.5H₂O (JCPS card NO.46-1,439). These results demonstrated that a new substance was generated after PO₄ ³⁻ being adsorbed on the surfaces of La₂(CO₃)₃.

The pH can affect not only charges on the surfaces of La₂(CO₃)₃ but also dissociation and solubility of the adsorbent, which would influence the adsorption of PO₄ ³⁻ on La₂(CO₃)₃ (Yang et al., 2013). The effect of solution pH in the range of 1.0–8.0 is shown in Figure 3A. The adsorption amount increased sharply when the pH changed from 1.0 to 3.0 and then decreased slowly with an increase in pH. The maximum adsorption amount of PO₄ ³⁻ was obtained at pH 3.0 (101.6 mg g⁻¹), which is 15.6-fold greater than that at pH 1.1. It means that La₂(CO₃)₃ was an efficient adsorbent for the removal of PO₄ ³⁻ under acidic conditions. This is similar to the results of other research studies. Many researchers have studied the influence of the pH value on PO₄ ³⁻ adsorption capacity and found that the lanthanide adsorbent has a higher PO₄ ³⁻ removal efficiency only at a lower pH value (Lu et al., 2021). There have been many studies on the removal of PO₄ ³⁻ by various materials, including binary and ternary compounds, for e.g., ZrO₂, red-mud, La–Cu, Fe–Zr, Fe–Mn, and Fe–Al–Mn (Lǚ et al., 2013; Zhang et al., 2009; Liu et al., 2008; Huang et al., 2008; Zhao et al., 2014; Long et al., 2011). The adsorption amount of PO₄ ³⁻ on these materials was different and significantly affected by the solution pH (Table 1). The adsorption amount of binary materials (La–Cu) and ternary materials (Fe–Al–Mn) was significantly higher than that of other materials, and the adsorption capacity of La₂(CO₃)₃ was much greater than that of other materials. Meanwhile, the composition and synthesis method of La₂(CO₃)₃ was simple and better when applied to removal of PO₄ ³⁻.

The isoelectric points of La₂(CO₃)₃ prior to and after adsorption were 2.1 and 6.5, respectively (Figure 3B), indicated that PO₄ ³⁻ neutralized the positive charge on the surface of La₂(CO₃)₃ and caused an increase in the isoelectric point. These findings revealed that in the surface of La₂(CO₃)₃, there was a change from carbonate to PO₄ ³⁻ when the pH values were 2.0–6.0. The existence of PO₄ ³⁻ species mainly in the form of H₂PO₄ ⁻ is explained because the adsorbent gathered more positive charges on the surface under acidic conditions, which caused strong adsorption of PO₄ ^3- on the surface of La₂(CO₃)₃ by electrostatic attraction. Meanwhile, La³⁺ and CO₃ ²⁻ in La₂(CO₃)₃ were dissociated in weak acid solutions and could be replaced by H₂PO₄ ⁻ under acidic conditions (Haghseresht et al., 2009; Xie et al., 2014). When the pH was greater than 6.0, the adsorption amount of PO₄ ³⁻ on La₂(CO₃)₃ continue to decrease presumably due to the competition for the adsorption sites between PO₄ ³⁻ and other coexisting anions such as CO₃ ²⁻ or OH⁻.

The effect of dosage on the adsorption efficiency of PO₄ ³⁻ is shown in Figure 4. The adsorption efficiency varied from 38.4 to 94.6% for the range of concentrations of 20–80 mg 50 ml⁻¹, which suggested that the adsorption efficiency of PO₄ ³⁻ by La₂(CO₃)₃ was directly proportional to the dose of La₂(CO₃)₃. This is because there were more adsorption sites available for PO₄ ^3- as the dosage increased. When the concentration of La₂(CO₃)₃ was higher than 80 mg 50 ml⁻¹, it had a negligible effect on the adsorption efficiency of PO₄ ³⁻.

Figure 5A demonstrates the effect of time on the adsorption amount of PO₄ ³⁻ on the surfaces of La₂(CO₃)₃. The adsorption amount of PO₄ ³⁻ increased rapidly during the first 50 min, probably due to a fast exchange of CO₃ ²⁻ and PO₄ ³⁻ on the surface of La₂(CO₃)₃ and occupancy of the sites by PO₄ ³⁻. After that, the adsorption amount of PO₄ ³ increased slowly over time. This implied mass transfer of CO₃ ²⁻ and PO₄ ³⁻ and subsequent exchange, predominant during the adsorption process. There was no significant change in the adsorption amount of PO₄ ³⁻ on the surface of La₂(CO₃)₃ after 24 h.

A quantitative approach to determine adsorption is feasible using a kinetic model. The equation for pseudo-first-order kinetic was introduced by Lagergren (Eq. 1), which is used for the prediction of the physisorption of the adsorbate onto the adsorbent in a given system. ln [ q e − q t ] = ln ⁡ q e − k t , (1) where q_e is the amount of the adsorbate at equilibrium (mg g⁻¹), q_t is the amount of the adsorbate (mg g⁻¹) at time t (min), and K (min⁻¹) is the rate constant for the pseudo-first-order sorption.

This equation can also be expressed by the following alternative equation: q t = q e ( 1 − exp − k t ) . (2)

The kinetic curve of pseudo-first-order corresponding to the adsorption of PO₄ ³⁻ on La₂(CO₃)₃ is shown in Figure 5B. The q _e from the nonlinear optimization was 32.08 mg g⁻¹, the rate constant of the pseudo-first-order reaction was 6.3 × 10⁻³, and the correlation coefficient (r) of the formula was 0.95. These results indicated the existence of a reversible interaction between PO₄ ³⁻ and La₂(CO₃)₃.

Figure 6 displays the adsorption amount of PO₄ ³⁻ on La₂(CO₃)₃ at 303, 313, and 323 K. The adsorption amount of PO₄ ³⁻ increased gradually with the increase of temperature, significantly affecting the adsorption of PO₄ ³⁻ on the surfaces of La₂(CO₃)₃, since more carbonate ions could be dissociated from the surfaces of the adsorbent. The exchange rate of ions between PO₄ ³⁻ and CO₃ ²⁻ also increased, which could accelerate the adsorption reaction. The increase of the adsorption amount indicated that the adsorption of PO₄ ³⁻ on the surfaces of La₂(CO₃)₃ was an endothermic process (Mezenner and Bensmaili, 2009).

Inorganic ions such as Cl⁻, SO₄ ²⁻, NO₃ ⁻, CO₃ ²⁻, and SiO₃ ²⁻ are ubiquitous in environmental water. The influence of such anions on the adsorption of PO₄ ³⁻ is shown in Figure 7A. Compared with the control, all anions had a negative effect on the adsorption efficiency of PO₄ ³⁻. In particular, the coexistence of CO₃ ²⁻ and SiO₃ ²⁻ reduced the adsorption efficiency of PO₄ ³⁻ from 54.1 to 14.2%–9.5 and 7.6%, respectively, when the concentration of CO₃ ²⁻ and SiO₃ ²⁻ increased from 1.0 to 10 mmol L⁻¹. The previous studies have also examined the effect of ions on adsorption efficiency of phosphorus of other materials, such as La-porous carbon composites (Koilraj and Sasaki, 2017). The result was consistent with this study, and the adsorption capacity was disturbed by 20 mM CO₃ ²⁻ but not reduced in the presence 20 mM Cl⁻ and 20 mM SO₄ ²⁻. To further study the mechanism involved, the change of pH was measured after adsorption in the presence of CO₃ ²⁻ and SiO₃ ²⁻. The pH rose when the concentration was increased for CO₃ ²⁻ (pH = 7.1–10.2) and SiO₃ ²⁻ (pH = 7.7–11.9). This then entails a strong interfering effect on the adsorption of PO₄ ³⁻ due to CO₃ ²⁻ and SiO₃ ²⁻. On the one hand, the rose of pH would cause a decrease of positive charges on the surfaces of La₂(CO₃)₃, weakening the electrostatic attraction between PO₄ ³⁻ and La₂(CO₃)₃. On the other hand, the increase of anions could lead to stronger competitive adsorption with PO₄ ³⁻, resulting in a decrease of adsorption amount of PO₄ ³⁻.

HA and FA are the most important components in the natural organic matter, being ubiquitous in the aquatic environment. They are complex mixtures of heterogenous compounds with a negative charge, originated from the decomposition of plant and animal residues (Valencia et al., 2012; Mcintyre and Guéguen, 2013). The existence of HA and FA in environmental waters may cause interference in the adsorption of PO₄ ³⁻ on La₂(CO₃)₃ through a competitive process. As shown in Figure 7B, with no coexisting ions in solution as a blank controller (no co-ion), both HA and FA had a weaker effect, as interference on the adsorption of PO₄ ³⁻, than that observed for the anion. This is because the molecular weight of natural organic matter was larger than that of anion and requires a longer adsorption path. HA and FA were less competitive to PO₄ ³⁻ than inorganic anion, and FA, with a small molecular weight, had more influence on the adsorption than HA. Since HA and FA can be combined with PO₄ ³⁻ in aqueous solution, the complex of La₂(CO₃)₃/HA (or FA) can still be combined with PO₄ ³⁻, so the change in the adsorption amount of PO₄ ³⁻ was not perceivable in the presence of HA and FA.

As shown in Figure 8, the adsorption amount of PO₄ ³⁻ increased with the concentration of PO₄ ³⁻, reaching a maximum adsorption amount at 150 mg L⁻¹. When PO₄ ³⁻ concentration exceeded 150 mg L⁻¹, the adsorption amount decreased. The corresponding total molar volume of bicarbonate in the solution after adsorption is also shown in Figure 8, which rose when the concentration of PO₄ ³⁻ increased and attained equilibrium at variable concentration. In addition, the change of the adsorption amount of PO₄ ³⁻ (the initial concentration of PO₄ ³⁻ was 100–200 mg L⁻¹) could be due to the coexistence of anions such as OH⁻. Therefore, PO₄ ³⁻ in the solution would react with La₂(CO₃)₃, and HCO₃ ⁻ would be released.

The XPS spectra of La₂(CO₃)₃ after the adsorption of PO₄ ³⁻ is shown in Fig. S2(a). The content of C, O, La, and P at the surface of La₂(CO₃)₃ after adsorption were 13.7, 57.4, 18.0, and 10.8%, respectively. Fig. S2(b) and Fig. S2(c) show the fitted spectras corresponding to La_3d and P_2p, respectively. The peak at 835.10 eV corresponded to La_3d5/2 of LaPO₄ (Jorgensen et al., 2002). The binding energy of P 2p was 133.89 and 132.89 eV, which contributed to LaPO₄ (Ivanova et al., 1996) and HPO₄ ²⁻ (Kurmaev et al., 1996), respectively. The XPS spectrum indicated that the adsorption of PO₄ ³⁻ on the surfaces of La₂(CO₃)₃ might be the result of ion exchange between PO₄ ³⁻ and carbonate, according to the aforementioned results. Based on a previous study, the adsorption process of PO₄ ³⁻ on the surfaces of La₂(CO₃)₃ could be mainly explained in terms of ion exchange (chemisorptions). A tentative adsorption mechanism is presented in Fig. S3, when HPO₄ ²⁻ reacted with La₂(CO₃)₃, and HCO₃ ²⁻ was released in the solution.