Uranium Extraction from Salt Water Using Formo-Phenolic Resin Containing Amido-β-phosphonic Acid Chelating Moiety

Abstract

Three chelating monomers with an amido-β-phosphonic acid moiety in varying α-position groups (hydrogen, methyl, phenyl) were synthesized and incorporated into formo-phenolic type resin, resulting in six adsorbents. Characterization showed an increase in chelating monomer content with hydrophobicity. The materials exhibited similar maximum adsorption capacities (~170 mg/g) for uranium, with adsorption kinetics that varied with the α-groups and monomer percentage. The resins demonstrated good selectivity for uranium, and maintained significant adsorption capacities in synthetic seawater. They could be reused without loss of capacity, with the uranium recoverable in sulfuric acid solution.

1. Introduction

Humanity is currently dependent on fossil fuels—the main cause of global warming—for energy and transport. The current trend is to promote a transition to an all-electric economic model, but this requires a switch to renewable or clean energy sources. Nuclear energy has long been seen as an alternative to polluting fossil fuels. According to the IAEA (International Atomic Energy Agency) Red Book 2022 edition, the demand for uranium could remain stable in Europe and the USA, but will increase by up to 80% worldwide by 2040—from 60,114 t/y in 2020 to 108,272 t/y [1].

Uranium is not yet listed as a critical material, but its cost and environmental impact are changing, and its supply needs to be considered in terms of production costs. A good complement to uranium production from mines is extraction from seawater or even brine from desalination plants [2,3,4,5]. For example, the current production of fresh water from desalination plants in Spain (2 million m³/day) represents approximately 5 tons per year of untapped local uranium, whose extraction has a very low environmental impact. The development of an efficient and sustainable process is currently a challenge, and several research teams are working on it. In practice, the easiest way to extract metal ions from large volumes of seawater (or brine) is solid–liquid extraction using an adsorbent [6].

The main concern in the case of uranium is that seawater contains macro-concentrations of sodium, magnesium, calcium, potassium and many other metals in lower concentrations but only 3.3 µg/L of uranium (~6 µg/L in brine). Under these conditions, uranium is mainly found as highly stable Ca₂UO₂(CO₃)₃ complexes [7,8,9].

Therefore, uranium extraction requires the development of new adsorbent materials with high affinity and especially important selectivity for uranium among competing cations. In addition to these properties, the adsorbent must exhibit a large adsorption capacity, a fast kinetic, a simple back-extraction process, and all of this with a cheap and simple preparation process. The materials that have been developed over the last few decades are diverse: porous carbon, metal–organic frameworks, covalent organic frameworks, biomaterials and porous polymers [4]. Typically, these adsorbents contain a functional group with an affinity for uranyl ions, mainly amidoxime derivatives, but also polydopamine, calix [6] arene, succinic acid or quinoline [10,11,12,13,14,15,16,17,18,19,20]. The main issue is that adsorbents show low adsorption capacity in simulated seawater conditions—with macro-concentrations of competing cations.

Recently a family of formo-phenolic resins containing a biscathecholamide group as a chelating moiety has demonstrated high adsorption capacity (Q_max up to 400 mg/g) and an excellent selectivity (SF_U/M > 200, where M = Mg, Ca, Sr, K) [21]. These resins had better performance after activation with NaOH to produce a deprotonated form of the phenolic function –ONa. To avoid the activation step and enhance the selectivity toward sodium cations, a new family of phenolic resin was designed based on a functional group previously developed for uranium liquid–liquid extraction [22]. Amido-β-phosphonate compounds have shown good affinity for uranium, with selectivity tuned by modifying (i) the amide function (primary, secondary or tertiary), (ii) the α-position (branching with alkyl group) and (iii) the phosphonate group (di/mono alkyl ester or acid).

Although there are a few examples of uranium sorbents incorporating phosphonic acid groups [23,24], interestingly, the development of a solid material containing an amido-β-phosphonate functional group for metal adsorption has not been previously investigated. This study describes the synthesis of three amido-β-phosphonic compounds with a phenolic moiety on the amide side and different alkyl groups on the α-position (Figure 1). Phosphonic acid and secondary amide were selected to provide hydrophilic properties to the chelating moieties. These compounds were incorporated into a formo-resorcinol resin matrix at different ratios, and their uranium extraction properties were investigated.

2. Experimental Part

2.1. Materials and Methods

All chemicals used in this study were analytically pure (Sigma–Aldrich (Saint Quentin Fallavier, France) or Thermo Fisher (Illkirch, France)) and were utilized without further purification. Anhydrous solvents were obtained from Pure Solv (Innovative Technology, China). Thin layer chromatography (TLC) was performed on silica plate 60F254 adsorbed onto alumina sheet (Merck TLC Silica Gel 60 F254, Burlington, MA, USA). Purification of compounds was carried out using BUCHI Reveleris X2 with BUCHI (Villebon sur Yvette, France) or Interchim (Montluçon, France) columns (silica or C18).

Characterization of all synthesized compounds was conducted using nuclear magnetic resonance (NMR) and mass spectrometry techniques (Figures S1–S33). ¹H and ¹³C NMR spectra were recorded on a Brucker Advance 400 MHz instrument, with chemical shifts reported in parts per million (ppm) using deuterated solvents as internal references. Electrospray ionization mass spectrometry (ESI-MS) was performed on a Flexar SQ 300 MS instrument (PerkinElmer, Waltham, MA, USA).

The synthesized resins were characterized through various techniques, including solid-state NMR, Fourier transform infrared spectroscopy (FT-IR), thermogravimetric analysis (TGA), and electron microscopy. Mechanical grinding of the samples was conducted at 25 Hz for 15 min using a Retsch mixer mill MM 200 with a zirconium ball. Solid-state magic angle spinning (MAS) ¹³C NMR spectra were recorded with a Varian VNMRS 300 solid spectrometer at a rotation speed of 12 KHz using 3.2 mm outer diameter rotors.

Attenuated total reflection (ATR) FTIR was performed on a Perkin Elmer 100 spectrometer equipped with an ATR crystal, covering a working range of 4000–615 cm⁻¹ with a resolution of 4 cm⁻¹.

Thermal analyses were conducted using a TGA/DSC 2 STARe system from Mettler Toledo, with a heating rate of 10 °C/min from 25 to 950 °C under air or nitrogen atmospheres.

Environmental scanning electron microscopy (ESEM) and X-ray energy dispersive spectroscopy (X-EDS) analyses were performed using a Bruker AXS X-Flash 5010 detector (Billerica, MA, USA) coupled with a FEI QUANTA 200 ESEM FEG model.

2.2. Preparation of Amido-β-phosphonic Monomers

All compounds were synthesized using commercially available starting materials. The final monomers were developed through novel synthetic protocols. Detailed synthesis procedures are provided in the “Compounds Synthesis” section, with additional experimental details available in the Supplementary Information (SI) file (Synthesis and characterization).

2.3. General Procedure for the Synthesis of Chelating Resin

The synthesis of the chelating resin involved dissolving the respective ratios of amido-β-phosphonic monomers (TyrAMeP, TyrAMe(Me)P and TyrAMe(Ph)P), resorcinol and sodium hydroxide (NaOH) pellets (2 eq.) in H₂O (100 eq.). A 37% formaldehyde solution (Equation (3)) was then added and the reaction mixture, which was stirred for 16 h. Following this step, the reaction mixture was transferred to a flask with a wide neck and a flat bottom and cured in a ventilated oven at 130 °C for 72 h. The resulting solid and robust resin was recovered, ground using a ball mill, and subsequently washed with 1 mol/L HCl and water, or 1 mol/L NaOH and water, depending on whether a protonated (-OH) or deprotonated (-ONa) form was desired. Finally, the washed resins were dried in a ventilated oven at 80 °C for 24 h, dispersed by ball milling, and stored for extraction experiments. A resin without functional groups R100 was also prepared using only resorcinol as the monomer under the same conditions and following the published procedures [25,26].

2.4. Batch Experiments

The ability of thermosetting resins to extract uranium and competing metals was evaluated through batch extraction experiments conducted using various aqueous solutions. In these experiments, a volume (V) to mass (m) ratio was fixed to V/m = 1, from 4 to 40 mg of resin and 4 mL to 40 mL of solution containing metal cations. The mixture was stirred at a controlled temperature (20 to 60 °C) for 16 to 120 h. Various aqueous solutions were used with a molar ratio fixed at around [CO₃]/[U] = 5.4:

(i)

Solution 1: a solution simulating seawater doped with 190 mg/L of uranium (8.0 × 10⁻⁴ mol/L), 259 mg/L of carbonates (4.3 × 10⁻³ M), and 2000 mg/L of sodium (8.7 × 10⁻² M) at pH = 8.2 ± 0.1;

(ii)

Solution 2: a solution simulating seawater doped with 1 to 800 mg/L of uranium with a molar ratio fixed at around [Na]/[U] = 120 at pH = 8.2 ± 0.1;

(iii)

Solution 3: a solution simulating seawater with salinity at around 36 g/L (24.53 g NaCl, MgCl₂ 5.20 g/L, Na₂SO₄ 4.09 g/L, CaCl₂ 1.16 g/L, KCl 0.695 g/L, NaHCO₃ 0.201 g/L, KBr 0.101 g/L) doped with 95 mg/L of uranium (4.0 × 10⁻⁴ mol/L), 96 mg/L of copper (1.5 × 10⁻³ mol/L), 97 mg/L of zinc (1.5 × 10⁻³ mol/L) and 39 mg/L of vanadium (7.7 × 10⁻⁴ mol/L) at pH = 5.9 ± 0.1;

(iv)

Solution 4: a solution simulating seawater with salinity at around 36 g/L (24.53 g NaCl, MgCl₂ 5.20 g/L, Na₂SO₄ 4.09 g/L, CaCl₂ 1.16 g/L, KCl 0.695 g/L, NaHCO₃ 0.201 g/L, KBr 0.101 g/L) doped with 105 mg/L of uranium (4.4 × 10⁻⁴ mol/L) at pH = 8.2 ± 0.1.

After the contact period, the mixture was centrifuged, and the supernatant was filtered through a 0.22 µm cellulose acetate membrane. The filtered solution was then diluted with 1% HNO₃ to achieve a suitable concentration for analysis.

The concentrations of various cations present in the filtrate were analyzed using inductively coupled plasma atomic emission spectrometry (ICP-OES) with an ICAP 7000 series spectrometer (Thermo Scientific). Wavelengths were carefully selected to avoid spectral interference among the elements.

-

The cation uptake capacity Q_ads (mg/g) was calculated using the equation:

-

$Q_{ads}=(C_i-C_f)\times {\displaystyle \frac{V}{m}}$, where C_i is the initial concentration of the metal ion in solution, C_f is the residual metal ion concentration, and V/m is the ratio of the solution volume by the resin’s mass;

-

The adsorption efficiency E (%) was calculated according to the following equation: $E={\displaystyle \frac{C_i-C_f}{C_i}}\times 100$;

-

The distribution coefficient, denoted K_D and expressed in mL/g, represents the ratio between the quantity of cation uptake by the resin and the quantity of cation remaining in solution after extraction, and was determined by the following formula:

$K_{\mathrm{D}}={\displaystyle \frac{C_{\mathrm{i}}-C_{\mathrm{f}}}{C_{\mathrm{f}}}}\times {\displaystyle \frac{V}{m}}\times 1000$;

-

The separation factor FS_U/M, with M the metal competitor, represents the ratio between the K_D of uranium and the K_D of the other metal. FS_U/M allows for quantification of the selectivity of a resin to extract uranium rather than other metals and was calculated by the following formula: ${FS}_{\mathrm{U}/\mathrm{M}}={\displaystyle \frac{K_{{\mathrm{D}}_{\mathrm{U}}}}{K_{{\mathrm{D}}_{\mathrm{M}}}}}$;

-

The back extraction or stripping efficiency S (%) is defined by the following equation: $S={\displaystyle \frac{Q_{\mathrm{e}}-Q_{\mathrm{f}}}{Q_{\mathrm{e}}}}\times 100$, where Q_e is the concentration of the metal ion loaded into the polymer at equilibrium and Q_f is the residual metal ion concentration in the polymer after the release.

Kinetics data were fitted using pseudo-first-order and pseudo-second-order kinetic models [27,28]. The pseudo-first-order model, proposed by Lagergren, is based on the solid sorption capacity and describes the adsorption process as diffusion-controlled, assuming that physisorption limits the adsorption rate of the particles onto the adsorbent. This suggests that, within the studied concentration range, a monolayer of the adsorbate forms on the adsorbent surface [29,30]. This model was employed to describe the sorption kinetics [31,32,33]. The pseudo-first-order rate is given as follows:

$$\log \left(Q_{\mathrm{e}}-Q_{\mathrm{t}}\right)=\log Q_{\mathrm{e}}-{\displaystyle \frac{K_1}{2.303}}\times t$$

(1)

where Q_e is the equilibrium metal ion concentration in solid-phase (mg/g), Q_t is the equilibrium metal ion concentration in the solid phase at time t (mg/g), K₁ is the pseudo-first-order equilibrium rate constant (min⁻¹) and t is the time (min).

The pseudo-second-order model is also based on solid-phase adsorption capacity but is controlled by sorption/desorption processes and diffusion toward chelating sites, considering chemisorption as the rate-limiting mechanism of the process [34,35,36]. The pseudo-second-order rate is given as follows:

$${\displaystyle \frac{t}{Q_{\mathrm{t}}}}={\displaystyle \frac{1}{Q_{\mathrm{e}}}}\times t+{\displaystyle \frac{1}{K_2{Q}_{\mathrm{e}}^2}}$$

(2)

where Q_e is the equilibrium metal ion concentration in the solid phase (mg/g), Q_t is the equilibrium metal ion concentration in the solid phase at time t (mg/g), K₂ is the pseudo-first-order equilibrium rate constant (min⁻¹) and t is the time (min).

Langmuir isotherm parameters [37,38] were estimated by the following equation:

$${\displaystyle \frac{C_{\mathrm{e}}}{Q_{\mathrm{e}}}}={\displaystyle \frac{1}{K_{\mathrm{L}}{Q}_{\mathrm{m}\mathrm{a}\mathrm{x}}}}+{\displaystyle \frac{1}{Q_{\mathrm{m}\mathrm{a}\mathrm{x}}}}\times C_{\mathrm{e}}$$

(3)

where C_e is the concentration of the metal ion in solution at equilibrium, Q_e is the amount of metal ion adsorbed at equilibrium, Q_max is the maximum amount of metal ion adsorbed at equilibrium and K_L is the Langmuir constant. The maximum sorption capacity can be estimated from this isotherm when a plateau is obtained [39].

Freundlich isotherm [40] parameters were estimated by the following equation:

$$\log Q_{\mathrm{e}}=\log K_{\mathrm{F}}+{\displaystyle \frac{1}{n}}\times \log C_{\mathrm{e}}$$

(4)

where C_e is the concentration of the metal ion in solution at equilibrium, Q_e is the amount of metal ion adsorbed at equilibrium, and K_F and n are Freundlich parameters.

2.5. Measure of Amido-β-phosphonic Ratio in Resin

To quantify the ratio of amido-β-phosphonic monomer within the formo-resorcinol resin matrix, a known quantity of resin was mineralized in concentrated HClO₄ at 180 °C for 2 h. The resulting soluble samples were then analyzed using ³¹P NMR spectroscopy (160 MHz, H₂O/HClO₄ + D₂O, 298 K). A solution of diethyl-4-aminobenzylphosphonate in D₂O was used as an internal standard (δ = 28.0 ppm–4.2 µmol/mL). The concentration of amido-β-phosphonic monomer in the resin was calculated by integrating the NMR signals to determine the phosphorus concentration in the mineralized samples (Figures S34 and S35).

3. Results and Discussion

3.1. Synthesis of Amido-β-phosphonic Monomers

To incorporate amido-β-phosphonic chelating monomers into the resin matrix, the compounds must contain a phenolic moiety. Therefore, TyrAMeP (3a) and TyrAMe(Me)P (3b) were synthesized in three steps from tyramine, with reasonable overall yields of 49% and 28%, respectively (Scheme 1). Initially, the primary amine was functionalized via the Schotten–Baumann reaction using chloroacetyl chloride or bromopropionyl bromide, respectively, to produce halogen-amide intermediates (1a,b) with yields of 60% and 43%, respectively. In a second step, the halogen atom was substituted by a phosphonate group through a Michaelis–Arbuzov reaction at 160 °C with tributylphosphite (P(OBu)₃), giving the phosphonate (2a,b). Finally, the phosphonate esters were quantitatively hydrolyzed using bromotrimethylsilane (TMSBr), yielding the desired amido-β-phosphonic monomers (3a,b).

The last monomer, TyrAMe(Ph)P (3c), with a phenyl group in the α-position, was synthesized using alcohol-protected tyramine (Scheme 2), specifically 4-methoxyphenethylamine, to minimize the formation of by-products. The synthesis pathway remained consistent with previous methods: the primary amine underwent a Schotten–Baumann reaction with 2-chloro-2-phenylacetyl chloride yielding the 2-chloro-N-(4-methoxyphenethyl)-2-phenylacetamide (1c), followed by a Michaelis–Arbuzov reaction with triethylphosphite (P(OEt)₃), leading to phosphonate 2c with a 59% yield over two steps. The hydrolysis of phosphonate esters and ether cleavage were conducted in a single step using boron tribromide (BBr₃), resulting in the amido-β-phosphonic compound TyrAMe(Ph)P (3c) with a 15% yield after purification on a C18 column. The relatively low yield in the final step could be improved by reducing the reaction time and promptly purifying the product upon completion.

Full characterization of the compounds, including ¹H, ¹³C, ³¹P NMR and HR-MS, is presented in the Supporting Information (Figures S1–S33).

3.2. Chelating Thermosetting Resin Preparation

The synthesis of formo-phenolic resin derivatives was achieved by substituting phenol, classified as CMR 2, with resorcinol. Formaldehyde, classified as CMR 1B, was retained to simplify the pre-polymerization step, although it should be replaced by terephthalaldehyde as demonstrated in a recent publication [41]. The pre-polymers were prepared using 0.34 or 0.66 equivalents of amido-β-phosphonic monomer, 0.66 or 0.34 equivalents of resorcinol, 3 equivalents of formaldehyde, and 2 equivalents of NaOH. The mixture was stirred in 100 equivalents of H₂O for 16 h, then cured at 130 °C for 72 h (Scheme 3). A total of 11 resins were prepared and their names, phenolic monomer ratios and forms are summarized in

Table 1. Molar ratio of amido-β-phosphonic monomers/resorcinol and form (OH or ONa) of the synthesized resins.

Resin	Chelating Monomer	Chelating Monomer Ratio	Resorcinol Ratio	Resin Form
R100	-	-	1	-OH/-ONa
TyrAMeP34	TyrAMeP	0.34	0.66	-OH/-ONa
TyrAMe(Me)P34	TyrAMe(Me)P	0.34	0.66	-OH/-ONa
TyrAMe(Ph)P34	TyrAMe(Ph)P	0.34	0.66	-OH/-ONa
TyrAMeP66	TyrAMeP	0.66	0.34	-OH
TyrAMe(Me)P66	TyrAMe(Me)P	0.66	0.34	-OH
TyrAMe(Ph)P66	TyrAMe(Ph)P	0.66	0.34	-OH

. The resins containing 0.66 equivalents of chelating monomer were not converted to the ONa form due to excessive polymer solubilization.

3.3. Characterization of Thermosetting Resins

Solid-state ¹³C and ³¹P NMR, FT-IR, TGA and SEM images were recorded and are presented in the Supporting Information (Figures S36–S42). The amount of amido-β-phosphonic contained in the material was determined by ³¹P NMR integration after mineralization of the samples in concentrated HClO₄ (

Table 2. Theoretical and experimental percentage of amido-β-phosphonic monomers in final matrix resins.

Resin	%theoretical	%measured	R
TyrAMeP34		29	H
TyrAMe(Me)P34	34	34	Me
TyrAMe(Ph)P34		37	Ph
TyrAMeP66		41	H
TyrAMe(Me)P66	66	49	Me
TyrAMe(Ph)P66		53	Ph

).

The measured values are lower than expected because (i) the weight of formaldehyde was not included in the theoretical value and (ii) some polymers were lost during the washing step; the remaining materials consisted only of insoluble polymers. An effect of the α-group was observed, with increasing hydrophobicity (H < Me < Ph) leading to a higher incorporation of monomers in the final resin.

3.4. Studies of Uranium Extraction Properties

The adsorption capacity was first evaluated as a function of contact time to determine the kinetic order of the adsorption reaction (Figure 2). Regardless of the percentage of monomer incorporated in the resins, the monomer TyrAMe(Ph)P resulted in a fast adsorption capacity up to 140–160 mg/g at 500 min.

Conversely, the resins containing TyrAMeP and TyrAMe(Me)P exhibited slower kinetics with different behaviors depending on the amount of monomer. The P34 resins displayed significantly faster kinetics than the P66 resins, which seem to be very slow—equilibria were probably not reached at 5760 min. Surprisingly, all the resins seemed to tend toward the same maximum adsorption capacity (around 160 mg/g) with different kinetics. The kinetic order of the adsorption process was modeled using Lagergren’s pseudo-first order (PFO) Equation (1) and pseudo-second order (PSO) Equation (2) (Figure S43, Table S1). Resins containing TyrAMe(Ph)P monomers fit well with the PSO model, suggesting an adsorption process driven by the coordination of uranium in the chelating sites. For the other chelating resins, the kinetic order seemed to depend on the amount of chelating monomers. In the case of TyrAMeP and TyrAMe(Me)P resins with 34% chelating monomer, the models are not very well correlated for either the PFO or the PSO model, with correlations of around 98%. It is therefore possible to consider either of these two models. TyrAMeP and TyrAMe(Me)P resins with 66% chelating monomers fit the PFO model better, suggesting that the adsorption rate of particles onto an adsorbent is primarily limited by physisorption. This phenomenon could be explained by the possibility of TyrAMeP66 and TyrAMe(Me)P66 to form—at the beginning of adsorption—L₂(UO₂) complexes, while TyrAMe(Ph)P66 could not form them due to steric hindrance.

Considering the rather slow kinetics of four out of six resins, experiments were performed to investigate the influence of the initial uranium concentration and temperature at 20 h contact time, although equilibrium was not fully reached.

Material adsorption models—Langmuir or Freundlich—were studied by measuring the adsorption capacity as a function of the initial uranium concentration (Figure 3). The adsorption capacity increased along with [U]₀ to a maximum value Q_max, but many resins exhibited a decrease in adsorption capacity at a high uranium concentration, possibly due to saturation of the contact area with uranyl or to the high ionic strength of the solution (sodium, carbonate and uranyl), which disrupts coordination.

However, it is clear that the incorporation of the amido-β-phosphonate moiety results in a higher adsorption capacity compared to R100 resins. The extraction and adsorption capacity values for [U] = 70 and 165 mg/L are detailed in the Supporting Information (Figure S44). Activation of R100 resin under the ONa form allows more uranyl adsorption (Q_max from 10 to 46 mg/g) while—as expected by incorporating amido-β-phosphonate into the resins—the ONa forms did not show better adsorption. The isotherm points were fitted with the two most common models—Langmuir and Freundlich (Figure 4).

The three P34 and TyrAMe(Ph)P66 resins fit well with the Langmuir model (even considering that Q_ads decreases at a high uranium concentration). This result means that the process is a single layer adsorption where all chelating sites are equal and do not interact with each other. Concerning TyrAMeP66 and TyrAMe(Me)P66, it is difficult to select a model because the measurements are far from the equilibrium state. In order to avoid hasty hypotheses, an isotherm was carried out with TyrAMeP66 at 40 °C with a contact time of 48 h (Figure 4d). This resin finally reached a similar adsorption capacity as the others (Q_max around 180 mg/g) and also showed a decrease at high uranium concentrations. Under these conditions, the points fit well with a Langmuir model. The effect of temperature on the adsorption was evaluated for the six resins (Figure 5).

For the TyrAMe(Ph)P34/P66—which has fast kinetics—increasing the temperature from 22 to 60 °C led to a slight effect on adsorption capacity within 16 h of contact time. The effect was steeper for TyrAMeP34 and TyrAMe(Me)P34, with adsorption capacities increasing from 120 mg/g at 22 °C to 180 mg/g at 50 °C.

The most significant effect was for TyrAMeP66 and TyrAMe(Me)P66, with a large gap in adsorption capacity between 22 °C and 60 °C. Finally, the temperature only increased the kinetics of adsorption by increasing the extraction capacity at a constant contact time, but it did not affect the maximum adsorption capacity of the materials, which tended to be around 170 mg/g for all of the resins.

In real media, zinc, copper and vanadium can be found in seawater at similar concentrations to uranium. Therefore, the selectivity of the chelating resins toward uranium in the presence of these competitors was evaluated (

Table 3. Extraction (%) and selectivity factors toward U^VI (SF_U/M, M = Zn, Cu, V^V) for the six studied resins. Conditions: Solution 3 synthetic seawater (36 g/L), [U] = 93 mg/L, [Zn] = 97 mg/L, [Cu] = 96 mg/L, [V] = 39 mg/L, pH = 5.9, V/m = 1 (4/4) at 35 ± 2 °C for 120 h.

	EU (%)	EZn (%)	ECu (%)	EV (%)	SFU/Zn	SFU/Cu	SFU/V
TyrAMeP34	94	17	55	85	76	13	3
TyrAMe(Me)P34	92	15	53	83	62	10	2
TyrAMe(Ph)P34	95	13	62	88	125	11	3
TyrAMeP66	98	18	58	96	295	46	3
TyrAMe(Me)P66	99	19	62	97	359	51	3
TyrAMe(Ph)P66	98	16	21	94	273	32	3

). The experiments were carried out in synthetic seawater with 36.88 g of salt per liter, but the pH was slightly reduced (pH = 5.9) to avoid the precipitation of metal hydroxides. Under these conditions, the resins extracted more than 90% of the uranium (>84 mg/g) and also some of the other metals. Vanadium, which is a known competitor, was extracted at over 80% (>31 mg/g), but always less than uranium—resulting in a selectivity SF_U/V = 2–3. The resins showed a poor affinity for zinc with extractions around 15% (≈15 mg/g) and separation factors between 62 to 359, and a slightly higher affinity for copper with extractions around 60% (≈58 mg/g) and separation factors between 10 and 51. Two significant observations were made: the total mass of metal adsorbed by the materials was around 195 mg/g (in agreement with previous experiments) and the resins containing 66% of chelating monomers offered better separation factors.

In order to be as close as possible to real seawater conditions, which are limited by laboratory analysis techniques, experiments were carried out with synthetic seawater (Solution 4, salt content 36 g/L at pH = 8.2) doped with 100 mg/L uranium. As reported in the literature, a significant decrease in efficiency was observed, but the synthesized materials still showed an interesting adsorption capacity ranging from 13 to 43 mg/g after 72 h at 25 °C (Figures S45 and S46). The α-methyl and α-phenyl resins were then subjected to reusability tests consisting of three cycles of extraction and re-extraction. The experiments were carried out at a higher temperature in synthetic seawater (40 °C) and the materials showed a high adsorption capacity between 70 and 90 mg/g (Figure 6).

Back extraction was carried out in sulfuric acid solution (1 mol/L) for 22 h at 40 °C. The first back extractions were not complete (35 to 70%), but they increased over 80% for the second and third back extractions. Note that the first back extraction was significantly higher for TyrAMe(Me)P than for TyrAMe(Ph)P, indicating the higher affinity of this moiety for uranium coordination.

4. Conclusions

The synthesis of three chelating monomers with an amido-β-phosphonic acid moiety for cation coordination and a phenol group for polymerization is reported. Three different precursors were used to obtain compounds in which the α-position carbon carries hydrogen, methyl, or phenyl groups. These monomers were successfully incorporated into formo-phenolic type resin using formaldehyde and resorcinol as reactive agents, resulting in six different adsorbents (34% and 66% monomers). The solid materials were fully characterized by IR, solid-state NMR, TGA, and SEM, and the amount of chelating monomer in the resin was determined by ³¹P-NMR analysis, showing an increase with the hydrophobicity of the monomer. The properties of the obtained materials were investigated by kinetics, isotherms, and variable temperature studies with solutions containing uranium, carbonate, and excess sodium. The maximum adsorption capacities were similar, around 170 mg/g, regardless of the nature of the chelating monomer and its amount in the matrix. However, the adsorption kinetics were influenced by α-groups (Ph > H ≈ Me) and monomer percentage (34% > 66%). As initially expected, these resins do not require alkaline activation, with OH being more efficient than ONa. Selectivity was also estimated, showing good results against zinc (SF_U/Zn = 62–359) and copper (SF_U/Cu = 10–51), but was limited against vanadium (SF_U/V = 2–3). Groups in the α-position slightly affected the selectivity, with the largest effect coming from the amount of chelating monomer; more amido-β-phosphonic moiety led to better selectivity. To approximate target media such as seawater or brine, extraction capacities were evaluated in synthetic seawater containing 36.88 g/L salt and doped with uranium. The materials maintained significant adsorption capacities between 13 and 46 mg/g under mild conditions (25 °C, 72 h) without extracting monovalent (Na, K) and divalent (Mg, Ca) cations, which are in large excess in these challenging media. Cyclic extraction/back-extraction tests (40 °C, 24 h) showed that the resins can be reused without loss of capacity (≈70 mg/g) and the uranium can be recovered in sulfuric acid solution (1 M).