1. Introduction

Rare earth elements (REEs) refer to a group of 17 elements, comprising the 15 lanthanides along with scandium and yttrium. These elements are widely recognized for their exceptional physicochemical properties, including outstanding optical,¹⁾ electrical,²⁾ and magnetic characteristics.³⁾ Due to these unique attributes, REEs have become indispensable in a broad range of industries, including renewable energy,⁴⁾ military technology, aerospace engineering, electronics, and advanced manufacturing.⁵⁾ As the global demand for high-performance materials continues to grow, REEs are increasingly regarded as strategic resources, making their sustainable supply and efficient utilization a critical concern.⁶⁾

However, the extraction, refining, and subsequent treatment of REE-containing ores generate substantial amounts of wastewater.⁷⁾ This wastewater is typically characterized by low REE concentrations and the presence of diverse and complex ion species, making the efficient separation and recovery of REEs highly challenging.⁸⁾ The inefficiency of conventional recovery methods not only leads to significant resource loss but also raises serious environmental concerns due to the potential release of toxic heavy metals and other pollutants into aquatic ecosystems.

To address these challenges, a variety of separation techniques have been developed, including ion exchange, chemical precipitation, electrochemical approaches, solvent extraction, and adsorption.⁹⁾ Among these, adsorption has garnered considerable attention due to its cost-effectiveness, ease of operation, and compatibility with large-scale industrial applications. Unlike solvent extraction and electrochemical methods, which often involve high energy consumption and the use of hazardous chemicals, adsorption-based techniques offer a more environmentally friendly approach to REE recovery. Nevertheless, existing adsorption strategies still face significant limitations, particularly in terms of selectivity, adsorption capacity, and material recyclability. These drawbacks undermine the urgent need for the development of novel adsorbents with enhanced performance.¹⁰⁾

Porous adsorbents, such as aerogels,¹¹⁾ cryogels,¹²⁾ foams,¹³⁾ and xerogels,¹⁴⁾ have emerged as promising candidates for REE separation due to their large surface area, tunable porosity, and high adsorption efficiency. Surface modifications can further enhance the selectivity of these materials, enabling them to target specific REEs over competing species.¹⁵⁾ However, a major challenge in the fabrication of these porous materials is the difficulty in the precise control of pore structure, which is largely dictated by preparation techniques such as freeze-drying and supercritical drying.¹⁶⁾ These conventional methods often result in inconsistent pore architectures, affecting the overall performance of the adsorbents.

High internal phase emulsions (HIPEs) offer a versatile and innovative strategy for the fabrication of highly porous materials with well-defined structures. HIPEs are emulsions in which the dispersed phase exceeds 74% of the total volume, which can lead to the formation of interconnected porous networks upon polymerization or gelation.^17,18) The unique morphology of HIPE-templated materials arises from the accumulation of large oil droplets, with smaller droplets filling the interstitial voids, thereby generating a highly porous structure with exceptional adsorption capabilities. This templating approach provides precise control over pore size distribution, mechanical stability, and surface chemistry, making HIPE-derived porous materials highly attractive for REEs adsorption.¹⁹⁾

In addition to high porosity, selective adsorption is a crucial requirement for efficient REE recovery. Selectivity is largely determined by the presence of functional groups that interact specifically with REE ions. Functional groups such as hydroxyl (-OH), carboxyl (-COOH), and phosphate (-PO₄^3-) groups have been widely explored for their ability to coordinate with REEs through strong chelation interactions.²⁰⁾ Among these, carboxyl groups are particularly attractive due to their ease of functionalization, strong binding affinity for REEs, and compatibility with a wide range of polymeric materials, especially biopolymers.²¹⁾

Poly(acrylic acid) (PAA) has emerged as a promising candidate for REE adsorption due to its abundance of carboxyl (-COOH) groups, which serve as highly effective coordination sites for REE binding.²²⁾ However, despite its strong affinity for REEs, PAA-based adsorbents often suffer from limited adsorption efficiency and insufficient structural stability, necessitating further optimization. One potential strategy to enhance the adsorption performance of PAA-based materials is the incorporation of citric acid (CA) as a functional modifier. CA, with its multiple carboxyl functional groups, can significantly increase the density of active binding sites and improve the overall binding affinity for REEs. By introducing CA into PAA-functionalized porous hydrogels, the adsorption selectivity and capacity can be greatly enhanced, making these materials more effective for practical applications in REE recovery.

To achieve the goal of designing absorbents with high adsorption capacity, high selectivity, and good reusability, the HIPEs-templating method was integrated with the PAA-CA crosslinking approach. To test their feasibility, the adsorption properties were studied for light and heavy REEs, namely, Pr³⁺ and Er³⁺. Variables such as pH, contact time, initial concentration, and temperature, were investigated to optimize the adsorption process. Additionally, adsorption kinetics and thermodynamics were analyzed to understand the adsorption mechanism. The effect of competing ions on adsorption and the reusability of the hydrogel absorbents were also evaluated.

2. Materials and Methods

2.1 Materials

Acrylic acid (AA), CA, N, N'-methylenebisacrylamide (MBA), xanthan, scandium nitrate hydrate (Sc(NO₃)₃·xH₂O, 99.99%), thulium nitrate pentahydrate (Tm(NO₃)₃·5H₂O, 99.9%), lutetium nitrate hexahydrate (Lu(NO₃)₃·6H₂O), gadolinium nitrate hexahydrate (Gd(NO₃)₃·6H₂O), copper nitrate trihydrate (Cu(NO₃)₂·3H₂O, 99.99%), zinc chloride (ZnCl₂, 99.95%), magnesium chloride hexahydrate (MgCl₂·6H₂O, 99.99%), and 3Å molecular sieve were purchased from Aladdin Biochemical Technology Co., Ltd. (China). Gelatin (gel strength ~100 g Bloom), p-xylene (PX), ammonium persulfate (APS), tetramethylethylenediamine (TEMED), yttrium nitrate hexahydrate (Y(NO₃)₃·6H₂O, 99.5%), lanthanum nitrate hexahydrate praseodymium (La(NO₃)₃·6H₂O, 99%), cerium nitrate hexahydrate (Ce(NO₃)₃·6H₂O, 99.99%), nitrate hexahydrate (Pr(NO₃)₃·6H₂O, 99%), neodymium nitrate hexahydrate (Nd(NO₃)₃·6H₂O, 99%), samarium nitrate hexahydrate (Sm(NO₃)₃·6H₂O, 99.9%), europium nitrate hexahydrate (Eu(NO₃)₃·6H₂O, 99.99%), dysprosium nitrate hexahydrate (Dy(NO₃)₃·6H₂O, 99.9%), holmium nitrate pentahydrate (Ho(NO₃)₃·5H₂O, 99.9%), erbium nitrate hexahydrate (Er(NO₃)₃·6H₂O, 99.99%), ytterbium nitrate pentahydrate (Yb(NO₃)₃·5H₂O, 99.9%), aluminum nitrate nonahydrate (Al(NO₃)₃·9H₂O, 99%), and iron nitrate nonahydrate (Fe(NO₃)₃·9H₂O, 99.9%) was purchased from Macklin Biochemical Co., Ltd. (China). Sodium hydroxide (NaOH, 96%), hydrogen chloride (HCl, 36–38%), potassium chloride (KCl, 99.5%), and anhydrous calcium chloride (CaCl₂, 96%) was purchased from Sinopharm Chemical Reagent Co., Ltd. (China). All chemicals and reagents used in this study were analytical reagent grade. Ultrapure water was used throughout the experiments.

2.2 Preparation of PAA-CA/PAA hydrogels

The synthetic route of the PAA-CA hydrogel adsorbent is shown in Fig. 1. As seen in Fig. 1, an emulsion-templating method was used to achieve controllability for a porous absorbent. AA and CA were crosslinked by APS and MBA, and added to the continuous phase with the presence of gelatin and xanthan. Stable emulsions were obtained after homogenization. A porous absorbent, PAA-CA was finally achieved by curing. The reference absorbent, PAA was prepared in the same methods without the addition of CA. Gelatin (0.1 g) and xanthan (0.05 g) were added to 10 mL of water and heated at 40°C while stirring until fully dissolved. MBA (0.15 g), AA (1.5 mL), and CA (0.5 g) were then added and stirred for 1 h at room temperature. The mixture was homogenized (homogenizer LC-SFJ-10) at 10,000 rpm while adding PX (30 mL) dropwise until the formation of emulsions. APS and TEMED were added and mixed for 3 min. The emulsion was transferred to a sealed container and cured at 60°C for 24 h in an oven. PX and residual monomers were removed by Soxhlet extraction with ethanol for 24 h. The material was sliced, immersed in a 5 wt% NaOH ethanol-water solution (V_ethanol/V_water = 7/3) for 24 h, dehydrated with industrial alcohol, embedded in 3Å zeolite molecular sieves, soaked in anhydrous ethanol for 24 h, and dried under vacuum at 60°C for 12 h. The reference sample of PAA hydrogels were prepared in similar methods at the same MBA and AA dosage.

Fig. 1.

Synthetic route of the PAA-CA hydrogel adsorbent.

2.3 Characterization

The morphologies of samples were characterized using scanning electron microscopy (SEM, ZEISS Gemini SEM 300). The content of metal ions was determined by an inductively coupled plasma emission spectrometer (ICP-OES, PerkinElmer Optima 5300 DV). Fourier transform infrared (FTIR) spectroscopy spectra were collected on a Bruker Tensor 27 FT-IR spectrometer using the KBr tableting method. The X-ray photoelectron spectroscopy (XPS) was recorded by an AXIS SUPRA⁺ instrument (Kratos Analytical, UK) with an Al Kα X-ray source.

2.4 Adsorption

The stock solution with an initial concentration of 1,000 mg/L were prepared for the adsorption of Pr³⁺ (praseodymium) and Er³⁺(erbium) REEs. The stock solutions were then directly diluted to obtain the adsorption working solution. Additionally, a binary mixed solution containing Er³⁺/K⁺, Mg²⁺, Ca²⁺, Zn²⁺, Al³⁺, and Fe³⁺ was prepared and the ion initial concentration was 50 mg/L. All experiments were conducted in 20 mL glass bottles by contacting 10 mL of the metal ion solution with 10 mg of hydrogel absorbents in a thermostatic shaker set at a constant speed of 150 rpm and a temperature of 298 K for a specified duration. The reaction mixtures were then filtered using 0.45 µm hydrophilic microporous filter membranes (polyethersulfone) and analyzed for metal ion concentrations using ICP-OES. The adsorption performance of PAA/PAA-CA for both single rare earth ions and binary component mixtures was evaluated. The optimal pH, initial concentration, and contact time for the maximum removal of Pr³⁺ and Er³⁺ were determined. The initial pH was adjusted by the dropwise addition of 0.1 mol/L HCl or NaOH solution, and pH values ranging from 1 to 7 were tested to prevent metal ion precipitation during adsorption. The water absorption swelling rate of 10 mg PAA/PAA-CA hydrogel was tested at different pH levels (1–9) to identify the optimal pH for the working solution. For the kinetic, isotherm, and thermodynamic experiments, the procedure remained the same, except that the time (1 to 480 min), concentration (100 to 1,000 mg/L), and temperature (298–318 K) were varied.

The amount of Pr³⁺, Er³⁺ or metal ions adsorbed by PAA/PAA-CA hydrogels was calculated from the following Eq. [1].

where Q is the adsorption capacity for ions at any time (mg/g), C₀ and C_e is the initial and equilibrium concentrations of rare earth ions or metal ions in solution (mg/L), V is the volume of solution (mL), m is the mass of hydrogels adsorbent (mg).

The competitive adsorption of K⁺, Mg²⁺, Ca²⁺, Cu²⁺, Zn²⁺, Al³⁺, and Fe³⁺ for Er³⁺ was examined. The distribution ratio was used to assess the selectivity of the PAA and PAA-CA (K_d), and the selectivity factor (k_Er/M) of Er³⁺ in comparison to other metal ions was determined using the following Eqs. [2] and [3].

where C₀, C_e represent the initial and final metal ion concentrations (mg/L), m represents the polymer mass (g), and V represents the solution volume (L). K_d is the ratio of Er³⁺ to other metal ions in milliliters per kilogram (mL/g). PAA and PAA-CA both have a selectivity coefficient of k.

The swelling ratio was calculated according to the following Eq. [4].

where S is the swelling ratio (g/g), M_s is the mass of the absorbent after swelling (g), M_d is the mass of the absorbent after drying (g).

2.5 Reusability

The reusability of the porous PAA/PAA-CA hydrogels adsorbent was investigated by ten consecutive adsorption-desorption cycles. When the adsorption equilibrium was reached, the adsorbent was separated and desorbed with 0.1 mol/L HCl solution. Then the adsorbent was regenerated in 0.1 mol/L NaOH solution for 30 min, washed with distilled water several times, dried at 60°C in a vacuum oven, and reused in the next adsorption experiments. There is no difference in gel content between PAA-CA and PAA hydrogels during preparation and circulation experiments.

3. Results and Discussion

3.1 Characterization and comparison of PAA and PAA-CA hydrogels

To confirm the formation of PAA-CA from the emulsion template, FTIR spectra of PAA-CA, PAA, and CA were recorded. As shown in Fig. 2, the characteristic absorption bands of CA at 3,496 cm^-1 (hydroxyl group), 1,748 and 1,705 cm^-1 (symmetric carboxyl groups), and 1,128 cm^-1 (stretching vibration of the tertiary alcohol hydroxyl group) were significantly weakened in PAA-CA. While, the infrared absorption peaks at 3,400 cm^-1 (hydroxyl group), 2,941 cm^-1 (asymmetric stretching vibration of methylene), and 1,574 and 1,421 cm^-1 (stretching vibrations of carbon-oxygen double and single bonds) were noticeably enhanced in PAA-CA, indicating the grafting of CA onto PAA.²³⁾ Additionally, PAA-CA exhibited a new broad absorption band at 1,228–1,007 cm^-1, assigning to the asymmetric stretching vibration of ester groups, further confirming the grafting between PAA and CA. The absence of a characteristic absorption band between 1,710 and 1,730 cm^-1 attributed to the C=O stretching of carboxyl groups, suggests that the carboxyl group of AA has been completely neutralized.

Fig. 2.

FTIR spectra of PAA-CA, PAA and CA.

Fig. 3 show the morphology and pore size distribution of PAA and PAA-CA hydrogels. PAA-CA has an increased number of smaller pores (average pore size of 8.87 ± 0.47 µm) than PAA (average pore size of 11.89 ± 0.53 µm). The increased number of smaller pores allows larger specific surface areas for adsorption, which may increase the adsorption capacity of REEs.²⁴⁾ Apart from the increased number of smaller pores, PAA-CA exhibits more uniform pore structure than PAA, which facilitates the structural integrity of absorbents both in dry (Fig. 3c) and wet (Fig. 3d, inset) conditions. The swelling ratios of absorbents were characterized at the pH range of 1–7. According to the results in Fig. 3d, the swelling ratios of PAA and PAA-CA increase with rising pH, which can be ascribed to the higher affinity of absorbents to H₂O molecules in less protonated conditions. In general, PAA has a bit higher swelling ratio than PAA-CA, possibly due to the relatively structural nonuniformity of PAA.

Fig. 3.

SEM images of the cross-section of PAA (a1, a2) and PAA-CA (b1, b2). (a2) and (b2): SEM images at a higher magnification, and corresponding pore size distributions. (c): Photo of PAA-CA absorbents. (d): Swelling ratio of PAA and PAA-CA at the pH values range of 1–7. (d): Insert a fully swelled PAA-CA hydrogel at pH = 7.

The absorbents shown in Fig. 3 were prepared at an oil phase volume fraction of 75%, which is HIPEs. In order to confirm the effect of surface area on adsorption capacity, PAA-CA absorbents with less specific surface area were also prepared by the same method at lower oil phase volume fractions for comparison. As shown in Fig. 4, the adsorption capacities of PAA-CA at the oil phase volume fractions of 50, 67, and 75% of Er³⁺ at room temperature are 243.4, 314.3, and 362 mg/g, respectively. Thus, the absorbents used in subsequent experiments were prepared with HIPEs (i.e. oil phase volume fraction of 75%).

Fig. 4.

Er³⁺ adsorption capacity of PAA-CA prepared at different oil phase volume fractions. Adsorption conditions: C₀ of 400 mg/L, time of 10 h, adsorption dosage of 10 mg/10 mL, pH = 5, and room temperature.

3.2 REEs enrichment by absorbents

3.2.1 Effect of pH value

To evaluate the adsorption capacity of REEs, PAA and PAA-CA were subjected to the adsorption tests of two REEs, including Pr³⁺ and Er³⁺, which belong to the light and heavy REEs, respectively. The effect of pH on the adsorption capacity of Pr³⁺and Er³⁺ by PAA and PAA-CA is presented in Fig. 5. The adsorption capacities of both hydrogels are similar at different pHs. Specifically, the adsorption capacity of Pr³⁺and Er³⁺ increases significantly as the pH increases from 1 to 3 and then stabilizes within the pH range of 3 to 7. This adsorption behavior is closely related to the interaction between the carboxyl groups (-COOH) of PAA and PAA-CA and the metal ions. In highly acidic conditions (pH from 1 to 3), most of the carboxylate ions (-COO⁻) are protonated into carboxyl groups (-COOH), which decreases the number of available active sites to bind metal ions.²⁵⁾ Additionally, the excess H⁺ ions in the solution compete with Pr³⁺ and Er³⁺ for the ion-exchange sites on the adsorbent surfaces, further reducing the adsorption capacity.²¹⁾ As the pH increases from 3 to 7, the carboxyl groups deprotonate, rendering more carboxylate ions (-COO⁻) that can bind with REEs through electrostatic interactions and coordination. In addition, Er³⁺ consistently shows a higher adsorption capacity than Pr³⁺ in the pH range from 2 to 7, which can be ascribed to the influence of ion radii and coordination preferences. Typically, heavy REE ions exhibit a slightly smaller size than light REE ions which render their diffusion into the absorbents’ networks. The reduced atomic radius of heavy REEs tends to have a decreased coordination number, facilitating the bulk diffusion of heavy REEs onto chelating sites of absorbents. Although PAA-CA has larger surface area and more carboxyl groups than PAA, no apparent increase in the adsorption capacity can be observed in PAA-CA, which is possibly a result of the restricted diffusion of REEs into the hydrogel networks.

Fig. 5.

Adsorption capacity of Pr³⁺ and Er³⁺ by PAA (a) and PAA-CA (b) at pH = 1–7. Adsorption conditions: C₀ of 400 mg/L, adsorbent dosage of 10 mg/10 mL, time of 0–8 h, and temperature of 25°C.

3.2.2 Effect of initial REE concentration

Due to their strong complexing abilities, hydrogel-based adsorbents exhibit a strong ability to adsorb metal ions.²⁶⁾ To accurately determine the maximum adsorption capacity and facilitate comparisons with other reported adsorbents, a relatively wide concentration range (100 to 1,100 mg/L) was chosen to investigate the effect of the initial concentration (Fig. 6). Both hydrogel absorbents exhibit similar maximum adsorption capacities for Pr³⁺ and Er³⁺. At concentrations below 400 mg/L, the adsorption capacity increases rapidly due to the abundant adsorption sites. From 400 to 600 mg/L, the adsorption capacity rises slowly until equilibrium at concentrations above 600 mg/L, which is ascribed to the limited adsorption sites of hydrogel absorbents at higher concentrations. Comparing Fig. 6a and b, one can see that the hydrogel absorbents reach adsorption capacity for Pr³⁺ at a bit lower concentration (ca. 500 mg/L) than that of Pr³⁺ (600 mg/L). As a result of the larger ion radii of Pr³⁺, absorbents for Pr³⁺ saturate at lower initial concentrations and adsorption capacities.

Fig. 6.

Adsorption capacity of Pr³⁺(a) and Er³⁺(b) at different initial concentrations. The inset displays the Langmuir isotherm model fitting results. Adsorption conditions: adsorbent dosage of 10 mg/10 mL, pH = 5, time of 0–8 h, and temperature of 25°C.

In adsorption studies, the Langmuir and Freundlich isotherm models are commonly used to analyze the adsorption mechanism. The Langmuir model assumes monolayer adsorption on localized sites without any interaction between adsorbed molecules, while the Freundlich model is used to describe adsorption on heterogeneous surfaces with varying energy distributions.²⁷⁾ The formula of these models are expressed the following Eqs. [5] and [6].

Langmuir isotherm model:

Freundlich isotherm model:

where C_e (mg/L) and Q_e (mg/g) represent the equilibrium concentration and adsorption capacity at equilibrium for a given initial concentration. Q_m (mg/g) is the maximum adsorption capacity, while K_L (L/mg) and K_F (mg1−1/n L^1/n/g) are the constants determined by linear regression of the experimental data for the Langmuir and Freundlich models, respectively. The constant n (dimensionless) describes the adsorption intensity. The estimated model parameters along with their correlation coefficients (R²) are presented in Table S1.

The adsorption capacity derived from the Langmuir isotherm model closely matched the experimental data, and the higher correlation coefficient (R_L² > 0.998) suggests that the Langmuir model is well-suited for describing the adsorption behavior. Note that the correlation coefficients of the Freundlich model are R_F² < 0.795. The key characteristic of the Langmuir isotherm can be represented by the dimensionless separation factor R_L (R_L = 1/(1 + K_LC₀), C₀ as the initial concentration of Pr³⁺ and Er³⁺ in mg/L, and K_L as the Langmuir constant. These factors are used to evaluate the preference of the adsorption process. Since the calculated R_L values are all in the range of 0 to 1 (Table S2), the adsorption of Pr³⁺ and Er³⁺ onto the hydrogel absorbents is preferred. Additionally, the decrease of R_L values with increasing initial concentration suggests that higher concentrations of Pr³⁺ and Er³⁺ enhance the adsorption process.^25,28) This indicates that Pr³⁺ and Er³⁺ are adsorbed as a monolayer on the adsorbent surface.

3.2.3 Adsorption kinetics and thermodynamics

Contact time is a crucial parameter for evaluating the adsorption kinetics of an adsorbent. Fig. 7 shows the variation in the adsorption capacity of Pr³⁺ and Er³⁺ on PAA and PAA-CA as a function of contact time at three different temperatures, including 298, 308, and 318 K. The results show that the adsorption capacity of PAA reaches 90% within 240 min, while the adsorption capacity of PAA-CA reaches 90% in just 120 min. The adsorption capacities for Pr³⁺ and Er³⁺ increase with rising temperature, indicating that the adsorption processes are endothermic. This suggests that higher temperatures enhance the interaction between the metal ions and the polymeric adsorption sites, possibly by providing more kinetic energy to overcome activation energy barriers associated with ion binding.²⁹⁾ The influence of temperature is more evident on the adsorption capacity of PAA than that of PAA-CA, possibly due to the higher flexibility of hydrogel structure under increased temperature. To analyze the controlling factors of the adsorption process, the pseudo-first-order and pseudo-second-order kinetic models were applied,³⁰⁾ the Eqs. [7] and [8] are given below:

Fig. 7.

Adsorption capacity of Pr³⁺and Er³⁺ by PAA (a, b) and PAA-CA (c, d) as a function of contact time at 298, 308, and 318 K. Adsorption conditions: C₀ of 400 mg/L, adsorbent dosage of 10 mg/10 mL, pH = 5, and time of 0–8 h.

where Q_e and Q_t are the adsorbed amount of REEs (mg/g) at equilibrium and the time t, respectively. K₁ and K₂ are the constants of the pseudo-first-order equation (min^-1) and pseudo-second-order equation (g/mg·min), respectively. The related parameters and correlation coefficients R² can be obtained by regressing the experimental data (Table S3). Both PAA and PAA-CA hydrogels have higher correlation coefficients of the pseudo-second-order model, indicating chemisorption was involved in the process, wherein ion exchange or complexation may play a dominant role.

To evaluate the thermodynamic parameters for the adsorption of Pr³⁺ and Er³⁺ by PAA and PAA-CA, following Eqs. [9] and [10] were utilized:

where K_T represents the thermodynamic equilibrium constant (L/g), R is the universal gas constant (8.314 J/mol/K), T is the temperature (K), and ΔG°, ΔS°, ΔH° denote the Gibbs free energy (kJ/mol), entropy (J/mol/K), and enthalpy (kJ/mol), respectively. These thermodynamic parameters can be calculated using Eqs. [9] and [10] and are listed in Table 1. The negative values of ΔG° for Pr³⁺ and Er³⁺ indicate that the adsorption process is spontaneous. The positive ΔH° values suggest that the process is endothermic, and the magnitude of ΔH° (ΔH° > 0 kJ/mol) implies that chemisorption occurs. Additionally, the positive ΔS° values indicate an increase in disorder, likely due to the exchange of metal ions with more mobile ions in the exchanger during adsorption.

3.2.4 Selectivity

REEs are often present alongside other common metal ions in environmental and mining wastes, the high selectivity of adsorbents toward rare earth ions is critical for practical applications. As both PAA and PAA-CA have higher adsorption capacity for Er³⁺, the selectivity for Er³⁺ of absorbents was studied only in the binary mixtures of Mⁿ⁺/Er³⁺. Mⁿ⁺ refers to common metal ions including K⁺, Mg²⁺, Ca²⁺, Cu²⁺, Zn²⁺, Al³⁺, and Fe³⁺. As illustrated in Fig. 8, PAA and PAA-CA exhibit a consistently higher adsorption capacity for Er³⁺ than other metal ions. According to Eqs. [9] and [10], one can obtain the selectivity factors in Table 2. Both PAA and PAA-CA have the highest selectivity in the binary mixtures of K⁺/Er³⁺, followed by Fe³⁺/Er³⁺ and M²⁺/Er³⁺, but exhibit little selectivity for Al³⁺/Er³⁺. This selectivity follows the general principle of selective absorbance for metal ions - the higher the valence of metal ions, the easier the absorbance. For comparison, PAA-CA exhibits evidently higher selectivity for all binary mixtures than PAA, suggesting its potential as a more effective adsorbent for selective separation of REEs.

Fig. 8.

Adsorption capacity of PAA (a) and PAA-CA (b) in Mⁿ⁺/Er³⁺ binary metal ion mixtures. Adsorption conditions: C₀ of 50 mg/L, adsorbent dosage of 10 mg/10 mL, pH = 5, time of 240 min, and room temperature.

3.2.5 Reusability

The reusability of an adsorbent is a crucial factor in determining its practical application. To assess the reusability of PAA and PAA-CA, ten consecutive adsorption-desorption cycles were conducted. The relationship between adsorption capacity and cycle number is illustrated in Fig. 9. It was observed that both adsorbents maintained high and stable adsorption capacities across all 10 cycles, which is not commonly seen in other biobased absorbents. This suggests that the adsorption of REEs onto PAA and PAA-CA is reversible in the presence of H⁺, facilitated by ion exchange. Besides, the maintained recyclability of absorbents can be ascribed to their hydrogel state which endures the shape recovery of absorbents in wet conditions. These hydrogel absorbents consistently demonstrated a higher adsorption capacity of Er³⁺ than Pr³⁺. These findings indicate that both PAA and PAA-CA exhibit excellent reusability for the adsorption of REEs.

Fig. 9.

Adsorbed amount of Pr³⁺ and Er³⁺ on PAA (a) and PAA-CA (b) as a function of adsorption-desorption cycle. Adsorption conditions: C₀ of 400 mg/L, adsorbent dosage of 10 mg/10 mL, pH = 5, time of 240 min, and room temperature.

3.3 Adsorption mechanism

To analyze the adsorption mechanism, the FTIR (Fig. 10) and XPS (Figs. 11 and 12) spectra of PAA and PAA-CA before and after the adsorption of Pr³⁺ and Er³⁺ were recorded. The FTIR spectra underline the changes in the functional groups before and after adsorption. It is obvious that the absorption band of the hybrid hydrogel adsorbent at 3,424 cm^-1 shifts to lower wavenumber (3,385 cm^-1) after adsorption of Pr³⁺ and Er³⁺, which is attributed to the interaction -OH groups with Pr³⁺ or Er³⁺. The asymmetric stretching vibration of -COO⁻ at 1,575 cm^-1 shifts to 1,545 cm^-1 after the adsorption of Pr³⁺ and Er³⁺. The absorption band at 1,419 cm^-1 derived from the symmetric stretching vibration of carboxylate reduces and the bending vibration C-H at 1,457 cm^-1 mergences with the neighouring adsorption peak.^26,31) The results above listed indicate that the carboxylate groups chelated with Pr³⁺ and Er³⁺ during the adsorption process. PAA-CA exhibits a similar C=O stretching band, but at different in intensity and position compared to PAA. Notably, the absorption bands around 1,417 and 1,451 cm^-1 show significant shifts following ion adsorption, indicating enhanced interaction between the modified polymer and the metal ions. These shifts in the characteristic absorption bands demonstrate that metal ion coordination primarily occurs through the carboxyl groups.³²⁾

Fig. 10.

FTIR spectra of PAA (a) and PAA-CA (b) before and after adsorption of Pr³⁺ and Er³⁺.

The XPS survey spectra of PAA and PAA-CA after the adsorption of Pr³⁺ and Er³⁺ are presented in Figs. 11 and 12. It is apparent in Fig. 11 that after adsorption, new peaks related to REEs at 217 and 169 eV appear, respectively, which confirms the successful adsorption by PAA and PAA-CA. The binding energy shifts of C1s and O1s spectra (Fig. 12) indicate that the oxygen atoms in the carboxyl groups serve as the primary adsorption sites for the REEs. After the adsorption of Pr³⁺ and Er³⁺ by PAA and PAA-CA materials, the binding energy of the C=O bond shifts to a higher energy level, indicating a decrease in electron cloud density. The larger shift of C=O in PAA-CA than in PAA can be correlated with the higher selectivity of PAA-CA than PAA. Besides, this shift can be attributed to the high charge density of Pr³⁺ and Er³⁺, which strongly coordinate with the carboxyl groups of polyacrylic acid, forming stable complexes. This coordination enhances the polarity of the C=O bond, redistributing its electron density and increasing its binding energy. Additionally, the interaction between rare earth ions and carboxyl groups withdraws electrons from the carboxyl oxygen, leading to a reduction in electron density, a shortened C=O bond length, and a higher binding energy. Furthermore, prior to adsorption, some carboxyl groups may have formed hydrogen bonds, which could have lowered the initial C=O binding energy. Upon adsorption of Pr³⁺ and Er³⁺, the deprotonation of carboxyl groups (COOH → COO⁻) further enhances the electron, resulting in a stronger C=O bond and a further increase in binding energy. The shifts in the N1s spectra of C-NH₂ groups indicate that C-NH₂ groups also participated in REEs adsorption.³³⁾

Fig. 11.

XPS spectra of PAA (a) and PAA-CA (b) before and after adsorption of Pr³⁺ and Er³⁺.

Fig. 12.

XPS analysis of PAA and PAA-CA before and after the adsorption of Pr³⁺ and Er³⁺.

4. Conclusions

Porous PAA and PAA-CA hydrogels were successfully synthesized using a gelatin-based HIPE-templating method. Their adsorption properties were tested both for heavy and light REEs. The high adsorption capacity of 400 mg/g is achieved both in PAA and PAA-CA in the pH value range of 3–7. The high adsorption capacity of PAA and PAA-CA hydrogel can be ascribed to the large surface area resulting from the HIPE template, and the carboxyl groups on material surfaces. The initial concentration of REEs Pr³⁺ and Er³⁺ have an important influence on the adsorption capacity of both PAA and PAA-CA hydrogels. Adsorption kinetics followed a pseudo-second-order model, and the process is spontaneous and endothermic. PAA-CA hydrogel reaches the adsorption capacity faster than PAA, indicating the efficiency of PAA-CA. PAA-CA hydrogel shows a higher affinity of rare earth ions than common metal ions, particularly for Er³⁺. Additionally, the prepared hydrogel absorbents maintained high adsorption capacity over 10 adsorption-desorption cycles, demonstrating excellent reusability. The selective adsorption performance of PAA-CA hydrogel is better than that of PAA hydrogel. The porous PAA-CA hydrogel is a promising candidate for the selective recovery of REEs from aqueous solutions with stable reproducibility.