The FT-IR spectra of HPAMAM-ECH and HPAMAM-EGDE were showed in Figure 1A and Figure 1B, respectively. It can be observed that the absorption peaks around 3284 and 1640 cm⁻¹ in HPAMAM, which correspond to the N-H stretching vibration of amino and C=O stretching vibration of amide (Oliveira et al., 2021), respectively. After crosslinking with ECH, the stretching vibration peak of C=O were weakened and the C-O-C stretching vibration appeared at 1073 cm⁻¹ (Abdulhameed et al., 2019), which indicated the successful crosslinking between HPAMAM and ECH. The same phenomenon was observed in samples of HPAMAM-EGDE, the new peaks appeared around 1080 cm⁻¹ were assigned to the C-O-C stretching vibration in EGDE. In addition, the peak at 940 cm⁻¹ attributed to the symmetric stretching vibration absorption of epoxy functional group (Wang et al., 2017) was present in HPAMAM-EGDE but absent in all HPAMAM-ECH samples. This can be explained by the fact that a part of the epoxy groups in EGDE were not involved in the crosslinking reaction.

The surface morphologies (Figure 1C) of the materials were determined by SEM via choosing HPAMAM-ECH-3 and HPAMAM-EGDE-3 as models. The surface of the material crosslinked with epichlorohydrin was relatively smooth, and the particles were irregularly shaped and tightly arranged. However, the surface of the material crosslinked with ethylene glycol diglycidyl ether showed a wrinkled structure with a row of grooves and the particle size was relatively large. The possible reason was that the molecular chain of ethylene glycol diglycidyl ether was longer than that of epichlorohydrin, and the spatial structure formed in the cross-linking process was looser.

Specific surface area and pore size structure are important factors affecting the adsorption properties of materials. The BET results (Table 2) showed that the samples crosslinked with EGDE had relatively larger specific surface area and pore diameter than those crosslinked with ECH, which relying on nice stretch and flexibility of long-chain EGDE.

XPS results of wide-scan spectra of HPAMAM-ECH samples and HPAMAM-EGDE samples were shown in Figure 2. In the figure of wide scan XPS spectra (Figure 2A), three peaks appeared at 532.1, 399.3 and 284.7 eV in all samples, which can be assigned to O_1s, N_1s and C_1s, respectively. What’s more, the ratios of O to N in both HPAMAM-ECH and HPAMAM-EGDE materials increased with the increase of the amount of crosslinking agent and were all greater than that in HPAMAM (Table 3). The high-resolution N_1s spectra (Figure 2B) of HPAMAM curve fitted to two peak components with bonding energies at 399.6 eV for the amide group, and at 398.7 eV for the amino group. After crosslinking, the bond energies of the two functional groups increased. O_1s (Figure 2C) in HPAMMA contained only one peak which belonged to amide group, while cross-linked HPAMAM contained three peaks. Three peaks of O_1s in HPAMAM-ECH samples: 532.9, 532.1 and 531.5 eV, corresponding to C-O-C group, -OH group and O=C-N group, respectively (Li et al., 2018). In the HPAMAM-EGDE sample, the bond energies corresponding to these three peaks were 532.4, 531.8 and 530.9 eV. The binding energy of C_1s (Figure 2D) in cross-linked HPAMAM also shifted to the left compared with HPAMMA. The binding energies of C_1s that appeared at about 288, 287 and 285 eV indicated there were three types of C in the crosslinking HPAMAM structure, which could be attributed to O=C-N group, C-N or C-O group and C-C group, respectively (Lin et al., 2018). The appearance of C-O group and -OH group was indicative of the presence of epoxy crosslinking agent in the materials, which was consistent with the results of IR spectroscopy analysis.

The saturated adsorption capacities of HPAMAM-ECH and HPAMAM-EGDE for Hg(II) were presented in Figure 3. For HPAMAM-ECH obtained using different quantity of ECH, the adsorption capacities for Hg(II) fell in order of HPAMAM-ECH-3 > HPAMAM-ECH-4 > HPAMAM-ECH-5 > HPAMAM-ECH-2, while in the case of HPAMA-EGDE, the order was HPAMAM-EGDE-4 > HPAMAM-EGDE-3 > HPAMAM-EGDE-5 > HPAMAM-EGDE-2. This phenomenon can be explained as the structure and morphology of the material changed with the increase of the amount of crosslinking agent, resulting in the increase of the adsorption capacity. However, with the further increase of the amount of crosslinking agent, the number of amino groups decreased, resulting in the decrease of the adsorption capacity.

The pH value of the solution is an important factor affecting the adsorption capacity of the adsorbent because it has a great influence on the state of the existence of metal ions and the state of the functional group of the adsorbent. Figure 4 showed that the adsorption capacities of the HPAMAM-ECH samples and HPAMAM-EGDE samples for Hg(II) responded differently to changing pH value. As the solution pH was varied from 1.0 to 6.0, the adsorbed quantity increased first and attained the maximum at pH 5.0 both for HPAMAM-ECH samples and HPAMAM-EGDE samples. At low pH, NH₂ on the surface of adsorbents were positively charged owing to protonation, thus the adsorption was impeded by a large repulsive forces between the protonated amino group Hg(II). Moreover, the existence of a large amount of H⁺ in the solution may compete adsorption with Hg(II) (Zhang et al., 2021). The deprotonation of the NH₂ on adsorbents can effectively coordinate with Hg(II), resulting in increased uptake of Hg(II) as pH increased from 1.0 to 5.0. Above pH 5.0, the adsorption capacities decreased with the increasing of pH. This could attributed to the formation of hydroxyl species of mercury such as soluble Hg(OH)⁺ (Zeng et al., 2021).

Figure 5 showed the adsorption kinetic for Hg(II). The equilibrium kinetic adsorption capacity also followed the order of HPAMAM-ECH-3 > HPAMAM-ECH-4 > HPAMAM-ECH-5 > HPAMAM-ECH-2 for samples of HPAMAM-ECH and HPAMAM-EGDE-4 > HPAMAM-EGDE-3 > HPAMAM-EGDE-5 > HPAMAM-EGDE-2 for samples of HPAMAM-EGDE. For HPAMAM-ECH materials, the adsorption was swift at the inital 2 h and then tapered off until equilibrium was reached at about 10 h. Similarly, the adsorption within the first hour was rapid for HPAMAM-EGDE materials, which contributed most to the equilibrium adsorption and reached equilibrium at 10 h. The superiority of adsorption rate in the initial stage of HPAMAM-EGDE samples may come from the grooves of material surface and comparatively large surface area. The kinetics data were fitted into the pseudo-first-order and pseudo-second-order models (Niu et al., 2014): l n ( q e   −   q t )   =   l n q e   −   k 1 t   (3) t q t = 1 k 2 q e 2 + t q e (4) Where q _e and q _t (mmol/g) are the adsorption capacity for Hg(II) at equilibrium and time t (h), and k ₁ and k ₂ are the rate constant of pseudo-first-order (h⁻¹) and pseudo-second-order (g⋅mmol⁻¹h⁻¹) adsorption, respectively. Figure 5 and Table 4 gave fitting results and the kinetic parameters. The adsorption kinetic for the two materials fitted Pseudo-second-order model well as indicated by the higher correlation coefficients (R2 2), and the theoretical adsorption capacity was closer to the experimental value. Therefore, the adsorption kinetics of Hg(II) onto the adsorbents was better described by pseudo-second-order model. That is, the Hg(II) adsorbed rate determining step on HPAMMA-ECH and HPAMAM-EGDE may be dominant by chemisorption but not mass transfer in solution (Qu et al., 2018).

The adsorption isotherm of HPAMAM-ECH and HPAMAM-EGDE for Hg(II) were studied to depict the interactive behavior between adsorbent and adsorbate and the results were shown in Figure 6. With the increase of initial metal ion concentration, the adsorption increased accordingly due to the presence of higher driving force under high concentration. It was the high driving force that promoted the diffusion of metal ions and the adsorption capacity increased accordingly. When the initial metal ion concentration increased to a certain extent, the adsorption capacity would no longer increase on account of saturation of adsorbents. It can be also found that temperature had a remarkable effect on the adsorption capacity. It was worth mentioning that at lower initial concentrations (50 mg/L), HPAMAM-ECH-3 can completely adsorb Hg(II). The favorable of high temperature can be ascribed to the nature of the adsorption was endothermic.

The Langmuir and Freundlich equations were adopted to fit the experimental data. The Langmuir model assumes that adsorption occurs on a single layer of uniform surface with no interaction between the adsorbents, whereas the Freundlich model assumes heterogeneous adsorption with multiple molecular layers. The nonlinear expressions of the Langmuir and Freundlich models were as follows (Saadi et al., 2015): C e q e = C e q m a x + 1 q m a x K L (5) l n q e =   l n K F   + l n C e n (6) Where q _e (mmol/g) is the equilibrium adsorption capacity; C _e (mmol/L) is the Hg(II) concentration at adsorption equilibrium; q _max (mmol/g) is the maximum adsorption capacity; K _L is the Langmuir adsorption constant; and K _F and n are the Freundlich constants related to adsorption capacity and adsorption strength, respectively.

Figure 6 showed the fitting of the adsorption isotherms of HPAMAM-ECH-3 and HPAMAM-EGDE-4 at different temperatures to the Langmuir and Freundlich. For HPAMA-ECH-3 material, since the equilibrium concentration at low initial concentration was 0, the following five points were selected for fitting in order to ensure accuracy. From the correlation coefficients in Table 5, it can be concluded that the experiment data fitted Langmuir equation better than Freundlich equation for all cases of HPAMAM-ECH, revealing the adsorption of Hg(II) adsorption on HPAMAM-ECH obeyed the Langmuir adsorption isotherm. This implied that the adsorption of Hg(II) on HPAMAM-ECH followed the mechanism of monolayer adsorption (chemisorption) (Saadi et al., 2015). The difference was that for HPAMAM-EGDE samples, both Freundlich and Langmuir model had large correlation coefficients (R ² > 0.96). This phenomenon indicated that monolayer adsorption and multi-layer adsorption occurred simultaneously in the adsorption process of HPAMAM-EGDE materials. The multi-layer adsorption of HPAMAM-EGDE samples may be related to the wrinkled structure on the surface of the material and the internal cavity structure.

Table 6 showed the comparison of the maximum adsorption capacity of the adsorbents prepared in this paper with those reported in other literatures. The adsorption capacity of HPAMAM-ECH samples and HPAMAM-EGDE samples exhibited competitive adsorption performance for Hg(II) than most adsorbents due to high density nitrogen and oxygen ligands. This was evidence that crosslinked HPAMAM adsorbents prepared in this way can be potentially used for capture of Hg(II) from wastewater.

A series of binary ions coexistent systems, Hg(II)-Cu(II), Hg(II)-Co(II), Hg(II)-Cd(II), Hg(II)-Zn(II) and Hg(II)-Ni(II), were investigated to evaluate the adsorption selectivity of HPAMAM-ECH and HPAMAM-EGDE for Hg(II) and the results were listed in Table 7. It was obvious that HPAMAM-ECH and HPAMAM-EGDE exhibited excellent adsorption selectivity for Hg(II) from the binary metal ion systems, remarkably, the selective coefficients were infinite for Hg(II) in the presence of Co(II), Cd(II) and Zn(II). Therefore, the absorbents prepared in this work can probably be used for selective extracting and separating Hg(II) from aqueous system containing multiple metal ions.

The regeneration rate and adsorption capacity of the HPAMAM-ECH-3 and HPAMAM-EGDE-4 after three elution-regeneration experiments were shown in the Figure 7. The regeneration rate (Figure 7A) of HPAMA-ECH-3 and HPAMA-EGDE-4 remained above 90% after three cycles. The adsorption capacity (Figure 7B) was still superior, which was 1.31 mmol/g for HPAMAM-ECH-3 and 1.25 mmol/g for HPAMAM-EGDE-4, respectively. Therefore, we had reason to believe that the crosslinked HPAMAM adsorbents thus prepared possessed great potential land economic benefit due to satisfactory adsorption capacity and stable regeneration performance.

According to the data in Table 2, the amino and amide functional groups contents of HPAMAM-ECH-3 and HPAMAM-EGDE-4 were determined to be 7.39 and 6.13 (mmol/g). Thus, the coordination ratio of (NH₂ +CONH) to Hg(II) of can be calculated to be 2.03:1 (HPAMAM-ECH-3) and 1.40:1 (HPAMAM-EGDE-4). It can be found that the functional group utilization rate of HPAMAM-ECH sample was greater than that of HPAMAM-EGDE sample. This may be because the steric hindrance of the samples crosslinked by ECH was larger, hindering the entry of metal ions, while the samples crosslinked by long-chain EGDE were relatively loose. At the same time, the HPAMAM-EGDE sample had a relatively large specific surface area, and there were wrinkles and cavities on the surface, which may lead to the simultaneous chemical adsorption and physical adsorption, and improved the utilization rate of its functional groups.

The XPS survey spectra of HPAMAM-ECH-3 after adsorption as well as the N_1s, O_1s and C_1s high resolution spectra were used to gain further insight into the mechanism by comparing the changes of elements and binding energies before and after adsorption. The absorption peaks of Hg_4f and Hg_4d can be clearly seen in the Figure 8A, and the absorption peaks of Hg_4p can also be found, which verified the successful uptake of Hg(II). It can be observed from the high-resolution XPS spectra of Hg_4f (Figure 8B) that two absorption peaks located near 99.8 and 103.6 eV respectively corresponded to Hg_{4f 5/2} and Hg_{4f 7/2}. While, the main peak of Hg_{4f 7/2} contained two peaks of Hg²⁺ and Hg⁰, which were located at 99.6 and 100.5 eV respectively, indicating that Hg(II) may have undergone a redox reaction (Xu et al., 2021). Correspondingly, Figure 8C showed that after adsorption of Hg(II), the binding energy shifted from 399.3 to 399.8 eV for –NH₂ and from 401.0 to 401.5 eV for CONH due to the electrostatic interaction, and the peak areas were also increased. In addition, there was also a small peak near 406.2 eV in the spectrum, which belonged to the binding energy peak of N element in -NO₃. For O_1s (Figure 8D), the peak of CONH moved from 531.5 to 531.3 eV, while that of C-O-C group moved from 532.9 to 533.3 eV. After adsorption of Hg(II), the high-resolution C_1s spectra (Figure 8E) almost didn’t change, indicating the existence form of C element remained the same. From the XPS analysis, we can conclude that the N in the amino group and amide group, and O in the C-O-C and amide group in the materials were involved in Hg(II) adsorption and the main adsorption mechanism for Hg(II) was complexation.