Materials

Glycine (Gly) (C₂H₅NO₂, purity: 99%, CAS Number: 56-40-6), ZnCl₂ (purity: 98%, CAS Number: 7646-85-7), MnSO₄·H₂O (purity: 98%, CAS Number: 10034-96-5), FeSO₄ (purity: 90%, CAS Number: 10034-96-5), CaCl₂ (purity: 96%, CAS Number: 10043-52-4), NaCl (purity: 99%, CAS Number: 7647-14-5), phosphoric acid (H₃PO₄, purity: 85%, CAS Number: 7664-38-2), BeO (purity: 99%, CAS Number: 1304-56-9), Ca(OH)₂ (purity: 99%, CAS Number: 1305-62-0), were purchased from China Shanghai McLean Biochemistry Co., Ltd.

Preparation of CP@Glycine

The CP@Glycine was prepared by the following procedure. 1 g of glycine was dissolved in 25 mL of deionized water and then 1.5 g of calcium hydroxide was added to the mixture. After stirring at 65 ℃ for 2 hours, 1.5 g of phosphoric acid was poured uniformly and the mixture was then rapidly mixed. The obtained uniform solution was heated in an oven at 105 ℃ to dry and turn into a powder. The CP@Glycine powder prepared in this way was mostly used to remove Be(II) in U/Be solutions.

The preparation process of the adsorbent does not produce any toxic and harmful gases, and can efficiently remove beryllium from beryllium-containing wastewater.

Solution Preparation

According to previous studies [20], U/Be wastewater usually contains various elements such as Be(II), Zn(2), Mn(2), Ca(2), Fe(2), Na(1), and U(6). In the present investigation, U/Be solutions with low concentrations (5mg·L^-1 and 10 mg·L^-1) was appropriately prepared for batch experiments, adsorption kinetics, and thermodynamic studies, whereas mixed solutions containing Be, Zn, Mn, Ca, Fe, Na, and U (10mg·L^-1-50 mg·L^-1) were simulated for adsorption selectivity investigations.

Batch Experimental Analysis

The accuracy of results was demonstrated by the preparation of adsorbent, initial pH, amount of adsorbent, and adsorption of a simulated solution from U/Be solution. Three sets of parallel samples were utilized in batch experiments to verify the validity of the results. Batch test of $W_{{\text{H}}_{\text{3}}{\text{PO}}_{\text{4}}}/W_{\text{Ca}{\left(\text{OH}\right)}_{\text{2}}}/W_{\text{Glycine}}$ (wt:wt:wt) was conducted to investigate the effects of $W_{{\text{H}}_{\text{3}}{\text{PO}}_{\text{4}}}/W_{\text{Ca}{\left(\text{OH}\right)}_{\text{2}}}/W_{\text{Glycine}}=1:1:1$, 2:2:1, 3:3:1, 4:4:1, 5:5:1 on the resulting adsorption. 0.05 g of five various substances were placed in a 50 mL solution to form a mixture. The mixture was placed in a shaking table under reaction conditions of 25 ℃, 175 rpm, and 16 h. The mixture was then appropriately separated, and the concentration of Be(II) in the clarified solution was measured by an ultraviolet spectrophotometer (UV-2550, Shimadzu).

Adsorption Kinetics and Thermodynamics Studies

0.1 g of CP@Glycine was weighed and placed in 50 mL beryllium solutions (C₀=10, 60, 180, 200, 220, 250, 300, and 350mg·L^-1), then divided into 8 groups with three parallel samples in each group. The above solutions were placed in a constant temperature shaking table and then oscillated at 175 rpm at three temperatures (i.e., 15 ℃, 25 ℃ and 35 ℃) until arriving at the equilibrium. The adsorption kinetics of Be(II) were investigated with pseudo-first-order (PFO), pseudo-second-order (PSO), Elovich, and intra particle diffusion models. The adsorption kinetics and thermodynamic models refer to the equation section entitled "Equations" in the Supporting Information.

Adsorption Selectivity Studies

The mixed solutions containing Be, Zn, Mn, Ca, Fe, Na, and U (10mg·L^-1-50 mg·L^-1) were appropriately simulated for the adsorption selectivity studies. The selective adsorption performance of Be(II) from the solutions with CP@Glycine was evaluated using the dispersion coefficient, K_d (mL·g^-1). The dispersion coefficient (K_d) can be calculated as follows [21]: $K_{\text{d}}=\frac{\left(C_0-C_{\text{e}}\right)\times V}{C_{\text{e}}\times m}\times 1000$ (1) Where C_e (mg·L^-1) is the equilibrium concentration; C₀ (mg·L^-1) is the initial concentration of Be. m (g) is the mass of the adsorbent. V (mL) is the solution volume.

Characterization

Figure 1(a) demonstrates the SEM analysis diagram of CP@Gycine, from which it can be seen that CP@Glycine shows good agglomeration and irregular shape. Figure 1(b) shows the results of the TEM analysis. The original sample showed an irregular shape, and the average particle size of about 200 nm, CP@Gycine particle size greater than compared with separate calcium phosphate [22, 23] and amino acids. This result may be related to the presence of amino acids that alter the growth of calcium phosphate.

Fig. 1Full-screen View

(Color online) SEM analysis of CP@Glycine (a), TEM analysis of CP@Glycine (b), FT-IR analysis of CP@Glycine (c), XRD analysis of CP@Glycine (d)

Figure 1(c) illustrates the FT-IR analysis of CP@Glycine. The bending at 3400 cm^-1 -3500 cm^-1 was mainly attributed to O-H [24-26]. The vibration of the bands at 2900 cm^-1 and 1650 cm^-1 is essentially ascribed to the N-H bond [27, 28], whereas the vibration at 1060 cm^-1 is mainly attributed to the P-OH bond [29]. At 1155 and 1343 cm^-1, the characteristic band of in-plane bending of P-O-H has been presented [30]. The vibration of the bands at 890 cm^-1 and 610 cm^-1 is essentially attributed to the P-O bond [31].

Figure 1(d) shows the XRD analysis results of CP@Gycine, revealing multiple characteristic peaks. The presence of Ca₃(PO₄)₂ (PDF#48-0488) leads to the characteristic peak at 2θ=24.2°, 25.6°, 32.2°, while NH₄N₃ (PDF#09-0499) leads to the characteristic peak at 2θ=29.3° and 38.8°. The obtained results by the SEM, TEM, FT-IR, and XRD revealed that the CP@Glycine sample was successfully prepared.

Batch Experiments

The preparation process of adsorbent significantly affects the adsorption effect. The results of the batch experiments are illustrated in Fig. 2.

Fig. 2Full-screen View

(Color online) (a) Effect of sample preparation, (b) Influence of adsorbent dosage, (c) Effect of initial pH, (d) Zeta potential, (e) Adsorption and desorption experiments of CP@Glycine ((e) 5 mg/L, (f) 10 mg/L)

Five different ratios were tested for sample preparation ($W_{{\text{H}}_{\text{3}}{\text{PO}}_{\text{4}}}/W_{{\text{Ca(OH)}}_{\text{2}}}/W_{\text{Glycine}}=1:1:1$, 2:2:1, 3:3:1, 4:4:1, 5:5:1). The removal efficiency of Be(II) by CP@Glycine is positively correlated with the weight ratio of $W_{{\text{H}}_{\text{3}}{\text{PO}}_{\text{4}}}/W_{\text{Ca}{\left(\text{OH}\right)}_{\text{2}}}/W_{\text{Glycine}}$ and reached the maximum when the weight ratio was 3:3:1 (Fig. 2(a)). Therefore, the optimal weight ratio of $W_{{\text{H}}_{\text{3}}{\text{PO}}_{\text{4}}}/W_{\text{Ca}{\left(\text{OH}\right)}_{\text{2}}}/W_{\text{Glycine}}$ was 3:3:1.

The amount of absorbent shows the economic benefits of the absorbent material. Figure 2(b) illustrates the effect of adsorbent dosage. The graph presents that the optimal adsorption rate (99%) is achievable in the case of 2mg·L^-1 adsorbent. By this virtue, the adsorbent dosage in this experiment was set as 2mg·L^-1.

The initial pH value of the solution influences the chemical properties and zeta potential of the adsorption process [21]. The plotted results in Fig. 2(c) reveal that the adsorption rate of Be(II) is positively correlated with increasing pH. The optimum initial pH appears at 6 and the removal efficiency of beryllium reaches 99%. This is because Be(II) mainly exists in the form of ${\text{Be}}_4{\left(\text{OH}\right)}_4^{4+}$ for the case of pH>6, which hardly binds to phosphate and ammonia, thus affecting the adsorption efficiency [32]. This phenomenon is also reflected by the zeta site (PH_zpc). Figure 2(d) demonstrates that the PH_zpc of CP@Glycine is 2.25. When the pH is lower than PH_zpc, the adsorption efficiency is enhanced. The surface of CP@Glycine is positively charged, whereas Be(II) exists as a cation, leading to electrostatic repulsion leading to low adsorption rates [15]. As the pH increases, the deprotonation degree of CP@Glycine gradually decreases. When the surface of CP@Glycine is negatively charged and the phosphate and ammonia groups are likely to form stable complexes with Be(II) [17], the adsorption capacity could be remarkably enhanced, as explained by De [16].

The Be(II) desorption capacity and reusability of CP@Glycine were also studied. Figures 2(e-f) demonstrate that the removal efficiency of Be(II) decreases slightly, while the desorption rate of beryllium decreases during 5 adsorption-desorption cycles for both 5mg·L^-1 beryllium solutions and then remains constant. In the first adsorption and desorption cycle, the removal efficiency of Be(II) is 99% and decreases to 90% (in cycle 3). After 5 cycles, the absorption rate of Be(II) remains 85%. The results indicate that the desorption rates of Be(II) during 5 cycles are above 90%. Therefore, CP@Glycine has good reusability in the experimental range, which proves that CP@Glycine has economic benefits and acceptable recycling performance.

Equilibrium Modeling

Adsorption kinetics are able to effectively react to the required retention time of the adsorbent. Four kinetic models, namely pseudo-first-order kinetics, pseudo-second-order kinetics, Elovich, and intra-particle diffusion, were used to fit the data of CP@Glycine adsorption of beryllium. Figs. 3(a-b) illustrates the fitting of the adsorption kinetics. The results showed that Be(II) is rapidly absorbed in the initial stage of adsorption. As CP@Glycine occupies many adsorption sites in the early stages, Be(II) is removed in solution [15]. As is seen, the PSO model (R²=0.999) is more suitable for fitting the Be(II) removal process by CP@Glycine compared to the PFO model (R²=0.993), intraparticle diffusion (R²=0.921–0.940), and Elovich (R²=0.987–0.996). In addition, the Chi-Square parameter (χ²) of the fitted model was lower, and the estimated Q_e values were closer to the experimental data. This suggests that the removal process is regulated by chemisorption and physical adsorption at the same time [16].

Fig. 3Full-screen View

(Color online) Adsorption kinetics fitting (a-d), Adsorption isotherm fitting (e-i)

Adsorption isotherms were also examined based on the Langmuir model, Freundlich model, Redlich-Peterson model, and Temkin model. Figure 3(c) illustrates that Q_e and temperature exhibit a positive correlation with the initial concentration of beryllium. The fitting results of the adsorption isotherm models are demonstrated in Figs. 3(d-f), and the adsorption isotherm data are listed in Table S2. According to the relevant coefficient (R²), the Freundlich model (R²=0.961) is more suitable for fitting the Be(II) removal process with CP@Glycine for the tested temperature range than the Langmuir model (R²=0.830) at 35 ℃ [13]. At the temperature of 25 ℃, the Temkin model (R²=0.991) also seems suitable for fitting the adsorption process [32, 33]. By comparing the Chi-Square parameter (χ²) of the four equations, it can be seen that the χ² values of Langmuir and Temkin at 25 ℃ are 0.75049 and 0.54103, respectively, which is much lower than that of Freundlich. However, the χ² of Freundlich was lower than that of Langmuir and Temkin at 15 and 35 ℃. Therefore, the Redlich-Peterson model was introduced, which is a mixture of the Langmuir and Freundlich models [34]. Additionally, the experimental isotherm data fitted by the Redlich-Peterson model revealed that phosphoric acid adsorption on pyrolytic hydrocarbons does not follow the ideal monolayers or heterogeneous surface adsorption, but the mixed adsorption behavior of both. The maximum Q_e of CP@Glycine for beryllium is obtained as 66mgg^-1 at 35 ℃. Compared to other adsorbents, CP@Glycine exhibits a higher Q_e for beryllium (Table S3). In addition, the cooperative effect of P-O with N-H and O-H may enhance the removal of beryllium by CP@Glycine.

Adsorption thermodynamics of Be(II) with CP@Glycine was also determined. According to Table S4, during the tested temperature range, the calculated enthalpy change (△H⁰) is positive (> 0), indicating that the removal progress is associated with a negative heat response [28]. Gibbs free energy (△G⁰) < 0, indicating that the removal progress is autogenic [21]. Entropy change (△S⁰) > 0 signifies that the disorder between solid and liquid is increased in the removal progress [35]. To conclude, the removal progress is an autogenous negative heat process, which is consistent with the results of Fig. 3 and Table S4.

Adsorption Selectivity

The selective adsorption experiment of beryllium by CP@Glycine was explored. In U/Be solutions, multiple coexisting ions usually compete for active sites [36]. The separation behaviors of Be(II) with CP@Glycine from the mixed solutions with Zn, Mn, Ca, Fe, Na, and U were also carefully examined. Figure 4(a) illustrates the competitive adsorption of Be(II) from binary solutions that contain Be(II) and other different metal ions. The obtained results indicate that CP@Glycine exhibits a good adsorption rate (99%) of Be(II) from mixed U/Be solutions with various ion concentrations. The defined distribution coefficient (K_d) shows the adsorption selectivity of Be(II) with CP@Glycine. Figure 4(b) demonstrates the K_d of each competitor ion at the same concentration. The K_d value of Be(II) with CP@Glycine (K_d=2.53×10⁴ mg·L^-1) is much higher than the other six elements, which confirms that CP@Glycine has excellent selectivity of Be(II) over other impurities (i.e., Zn(2) (K_d=0.133×10⁴ mg·L^-1), Mn(2) (K_d=0.023×10⁴ mg·L^-1), Ca(2) (K_d=0.014×10⁴ mg·L^-1), Fe(2) (K_d=1.53×10⁴ mg·L^-1), Na(1) (K_d=0.013×10⁴ mg·L^-1), and U(6) (K_d=0.56×10⁴ mg·L^-1)). To investigate the adsorption selectivity of Be(II) over competitive ions, the influence of different concentrations of competitive ions on K_d of Be(II) was also determined. Figure 5 illustrates the effect of various ion concentrations (i.e., Fe(2), U(6), Zn(2), Mn(2), Na(1), and Ca(2)) on the K_d value of Be(II). Figure 5(a) demonstrates that the K_d value of Be(II) grows with increasing Fe(2) concentration because increasing pH may lead to precipitation of Fe(2) and co-precipitation with Be(II) [21]. Figures 5(b-f) illustrate that within the experimental range, the concentrations of coexisting ions (i.e., Zn, Mn, Ca, Na, and U) have little influence on the K_d value of Be(II). The orders of magnitude of the K_d value of Be(II) are as high as 2.53×10⁴ mg·L^-1, indicating the strong removal selectivity of CP@Glycine for Be(II) from U/Be solutions.

Fig. 4Full-screen View

(Color online) (a) Effect of the coexisting ion concentration on the adsorption rate of beryllium, (b) Effect of the single concentration of coexisting ions on K_d

Fig. 5Full-screen View

(Color online) Influence of different coexisting ions on K_d adsorbed by CP@Glycine at different concentrations

Adsorption Mechanism

The physical and chemical properties of CP@Glycine before and after adsorption (Used-CP@Glycine) were analyzed by different characterization to discover the adsorption mechanism of Be(II) with CP@Glycine. The surface morphology of CP@Glycine and Use-CP@Glycine is presented in Figs. 6(a-f). By comparing the morphology of CP@Glycine and Use-CP@Glycine, many flocs were detected on the surface after adsorption. The energy dispersive spectrum (EDS) analysis (Fig. 6(g)) of CP@Glycine and Use-CP@Glycine was further performed. Figure 6(h) illustrates that the peaks of O element are remarkably increased, indicating that a much amount of O accumulates on Used-CP@Glycine. However, the Be element peak is not appropriately detected because Be(II) is below the EDS detection line.

Fig. 6Full-screen View

(Color online) (a-c) SEM images of CP@Glycine, (d-f) Used-CP@Glycine, (g) EDS images of CP@Glycine, (h) Used-CP@Glycine, (i) FT-IR analysis, (j) XRD analysis of CP@Glycine and Used-CP@Glycine, (k) TG analysis of CP@Glycine, (l) TG analysis of Used-CP@Glycine

The interaction between Be(II) and the adsorbent molecular groups plays a vital role in the removal process [37]. Figure 6(i) illustrates the FT-IR analysis of CP@Glycine and Use-CP@Glycine. The O-H strength of Use-CP@Glycine increased significantly. CP@Glycine has a strong vibration at 2900 cm^-1 (N-H), and the vibration becomes weak after adsorption. The N-H bond shifts to 1650 cm^-1, which indicates the coordination change of the N-H bond during the reaction process [31] and the formation of a new substance with Be(II). The strength of the P-OH bond enhanced after the reaction in 1060 cm^-1, which reveals the increase of the -OH group radical on the CP@Glycine surface. These findings are in reasonable agreement with the EDS results. The removal band at 555 cm^-1 is essentially designated as the P-O bending from the ${\text{PO}}_2^{3-}$ group [38]. The results indicate that The P-O bond on CP@Glycine was shifted after adsorption, and a new vibration peak of the P-O bond at 690 cm^-1 appears, which proves that the P-O bond and Be(II) interact to form new products [39].

To characterize the chemical composition of the newly formed product after adsorption, X-ray diffraction (XRD) analysis (Fig. 6(j)) of CP@Glycine and Use-CP@Glycine was performed before and after adsorption. Figure 6(j) illustrates that new diffraction peaks appear in the XRD pattern of Use-CP@Glycine compared to the XRD pattern of CP@Glycine, which implies the formation of new products. According to standard diffraction card data, the diffraction peaks of the formed products correspond to BeNH₄PO₄ (PDF#31-0185) and Be(OH)₂ (PDF#39-1411). It can then be concluded that beryllium compounds are produced during the adsorption process and the products are mostly BeNH₄PO₄ and Be(OH)₂. This finding confirms that the separation of Be(II) is dominated by chemisorption and that both P-O bonds and N-H bonds of CP@Glycine involve the formation of BeNH₄PO₄ precipitates. Hence, it is thought that the central Be(II) ion formed a complex due to the presence of two distinct coordinate bonds of CP@Glycine to achieve a chelating effect.

To further confirm the composition of the surface precipitates after adsorption, thermogravimetric analysis (TG) was implemented and the corresponding results of CP@Glycine and Used-CP@Glycine are presented in Fig. 6(k-i). The plot associated with CP@Glycine exhibits multiple endothermic peaks, the first peak happens around 65 ℃ and the loss value is around 4%, which indicates the lack of water in the adsorption material [40]. The second peak occurs around 200 ℃ and the loss value was 6.8%, which was essentially related to the loss of ammonia in CP@Glycine [41]. The third endothermic peak occurs at about 400 ℃ with a loss value of 13.1%, indicating the dehydroxylation of CP@Glycine [42]. The endothermic loss peak of Used-CP@Glycine is estimated to be 12.76% at about 65 ℃-280 ℃, which indicates the loss of water and ammonia [41]. The thermogravimetric loss from 300 ℃ to 690 ℃ corresponds to the dehydroxylation process of Used-CP@Glycine with a loss of 7.84%. The third endothermic peak appears at the temperature of 700 ℃-800 ℃ and the loss rate is 2.34%, revealing the conversion process of pyrophosphate ion ${\text{P}}_{\text{2}}{\text{O}}_7^{4-}$ [42].

A previous study has shown that elemental valence changes occur during the adsorption process [31]. Herein, X-ray photoelectron spectroscopy (XPS) (see Fig. 7) has been utilized to analyze the elemental valence changes of CP@Glycine and Used-CP@Glycine C 1s, O 1s, N 1s, and P 2p were further analyzed. Figure 7(A) illustrates the entire XPS spectrum of CP@Glycine and Used-CP@Glycine. According to Fig. 7(B), Be(II) is detected in Used-CP@Glycine, indicating that beryllium is adsorbed to CP@Glycine. Figure 7(C) demonstrates that the O-C=O bond is converted into a C-O-P bond and a C-OH bond during the reaction, indicating that phosphate is involved in the reaction between Be(II) and CP@Glycine [43, 44]. Compared with CP@Glycine, the area of the $-{\text{NH}}_3^{+}$ group on Used-CP@Glycine exhibits an obvious growth (Fig. 7(D)), and the energy shifts. This appropriately interprets the increase of the $-{\text{NH}}_3^{+}$ group on the surface of Used-CP@Glycine [29] and the participation of such forms. Additionally, the performed analysis shows that the peak at 530.70 eV is considered the O 1s binding site of the phosphate group (Fig. 7(E)). The -OH group is added at 532.74 eV on Used-CP@Glycine [45]. Figure 7(F) demonstrates the analysis of P 2p, the peak value of P 2p decreases and shifts after adsorption, indicating the formation of complex [46]. The obtained XPS results confirm that the phosphate and amino acids, ${\text{PO}}_4^{3-}$, ${\text{NH}}_4^{+}$, and OH^- of the as-prepared CP@Glycine involve the adsorption of beryllium, which further validates the chelating-like interaction between Be(II) and CP@Glycine.

Fig. 7Full-screen View

(Color online) XPS analysis of CP@Glycine and Used-CP@Glycine

Adsorption Mechanism

In this paper, the interaction of Be (2) with a phosphate group and glycine in CP@Glycine has been investigated utilizing the quantum chemistry calculation method. The stable existence forms of glycine, ${\text{H}}_{\text{2}}{\text{PO}}_4^{-}$, ${\text{HPO}}_4^{2-}$, and ${\text{PO}}_4^{3-}$ in aqueous solution are shown in Fig. 8.

Fig. 8Full-screen View

(Color online) Optimized structures of Be(II) and possible ligands ((a) $\text{Be}{\left({\text{H}}^{2+}\text{O}\right)}_4^{2+}$, (b) glycine, (c) ${\text{H}}_2{\text{PO}}_4^{-}$, (d) ${\text{HPO}}_4^{2-}$, (e) ${\text{PO}}_4^{3-}$)

The optimized structure and binding energy of beryllium and different ligands are presented in Fig. 9.

Fig. 9Full-screen View

(Color online) Optimized structures and binding energies of Be²⁺ with different ligands: (a) Be²⁺-glycine complex, (b) ${\text{Be}}^{2+}-{\text{H}}^2{\text{PO}}_4^{-}$ complex, (c) ${\text{Be}}^2{}^{+}-{\text{HPO}}_4^{2-}$ complex, (d) ${\text{Be}}^{2+}-{\text{PO}}_4^{3-}$ complex, (e) Be²⁺- glycine complex, (f) ${\text{Be}}^{2+}-{\text{H}}^{\text{2}}{\text{PO}}_4^{-}$-glycine complex, (g) ${\text{Be}}^{\text{2+}}{\text{-HPO}}_4^{2-}$-glycine complex, (h) ${\text{Be}}^{\text{2}}{}^{+}-{\text{PO}}_4^{3-}$-glycine complex, (i) ${\text{H}}_{\text{2}}{\text{PO}}_4^{-}-{\text{Be}}^{2+}-{\text{H}}_{\text{2}}{\text{PO}}_4^{-}$ complex

Figure 9(a) are the optimized structures and binding energies of the possible Be(II) complexes formed with glycine. It is found that both the amino and carboxyl groups participate in the bonding with Be²⁺, the formed complex has the lowest binding energy -89.24 kcal·mol^-1, so this binding mode is the most thermodynamically stable [47]. Figure 9(b) shows the optimized structure and binding energy of Be²⁺ with ${\text{H}}_{\text{2}}{\text{PO}}_4^{-}$, which shows the binding energy of Be²⁺ and ${\text{H}}_{\text{2}}{\text{PO}}_4^{-}$ was -99.92 kcal·mol^-1. Figure 9(c) shows the optimized structure of Be²⁺ and ${\text{HPO}}_4^{2-}$, and the binding energy between Be²⁺ and ${\text{HPO}}_4^{2-}$ is -145.52kcal·mol^-1. In Fig. 9(d), the structure of Be²⁺ and ${\text{PO}}_4^{3-}$ is optimized and the binding energy is determined as -179.97kcal·mol^-1. Hence, it can be seen that the binding energy between Be²⁺ and ${\text{PO}}_4^{3-}$ is the lowest as compared with that of the possible Be(II) complexes formed with glycine, indicating that the structure of Be²⁺ and ${\text{PO}}_4^{3-}$ might be the possible thermodynamically stable structure. To further explore the possible chelating structures and binding energies of Be²⁺ with glycine and phosphate ligands, the different Be(II) complexes were optimized and the corresponding binding energies were obtained (Fig. 9(e-i)). Figure 9(e-i) show that among all the composite structures, Gly-Be-${\text{PO}}_4^{3-}$ (Fig. 9(h)) has the lowest binding energy, so it is the most thermodynamically stable in the optimized structure. Therefore, the localized electron density function diagram of the optimized structure is shown in Fig. 10.

Fig. 10Full-screen View

(Color online) Electron localization density function diagram of Be²⁺ combined with glycine and phosphate ligands

Figure 10 presents the Electron Localized Function (ELF) diagram of the electron density distribution among Be, glycine, and phosphate ligands. The cross-section of Fig. 10(a) is the plane where nitrogen/oxygen atoms and Be²⁺ in glycine are located, and the cross-section of Fig. 10(b) is the plane where two oxygen atoms and Be²⁺ directly interact with Be²⁺ in ${\text{PO}}_4^{3-}$ are located. It can be found that the ELF function value between glycine and Be²⁺ is close to 1, so it can be considered that there is a covalent bond between the glycine and Be²⁺, and the ELF function value between ${\text{PO}}_4^{3-}$ and Be²⁺ is about 0.8, so there is a covalent bond between the ${\text{PO}}_4^{3-}$ and Be²⁺. The overlap of electron density among Be, glycine, and phosphate ligands is the maximum for adsorption of Be at the N-O-P site, as shown in Fig. 10(a). This is because glycine and ${\text{PO}}_4^{3-}$ have a strong ability to provide electrons, and Be(II) accepts electrons and coordinates with functional groups on the surface of the adsorbent.This indicates that N,O, and P in glycine and phosphate groups provide coordination for Be.

According to the above characterization results, it is speculated that both the N-H bond and P-O bond of CP@Glycine form a stable metal-chelated complex with Be(II) during adsorption [18]. N, O, and P of CP@Glycine all contribute electrons to the coordination of Be²⁺. The hydroxyl and amino groups of CP@Glycine lead to more anions on the solid surface that produce an electrostatic attraction to positively charged beryllium ions [31]. Moreover, the O-H on the CP@Glycine combined with Be(II) to form a beryllium hydroxide precipitate [15]. In addition, amino acids have been reported to produce ammonia and phosphate due to the deamination process [47, 48]. These adducts may promote the formation of stable metal chelate complexes with Be(II). The quantum chemical calculation indicates higher stability of Gly-Be-${\text{PO}}_4^{3-}$ and ${\text{PO}}_4^{3-}$-Be-${\text{PO}}_4^{3-}$ than other structures. The higher stability is due to the bonding affinity of beryllium between the ${\text{PO}}_4^{3-}$ and glycine of CP@Glycine. Hence, it can be concluded that the introduction of phosphate and glycine will not only increase the number of active sites on the surface of CP@Glycine but also help in the formation of a stable beryllium complex [36]. To sum up, the selective removal of Be(II) with CP@Glycine is dominated by chemisorption. Further, the occurrence of adsorbents are precipitates of BeNH₄PO₄ and Be(OH)₂, which show a chelate-like effect during the adsorption process.