Complexation and Separation of Trivalent Actinides and Lanthanides by a Novel DGA Derived from Macrocyclic Crown Ether: Synthesis, Extraction, and Spectroscopic and Density Functional Theory Studies.

A DGA-arm-grafted macrocyclic aza-crown ether ligand (Cr6DGA) was synthesized, and its solvent extraction behavior toward trivalent americium and europium in nitric acid medium was studied. The effects of various parameters such as the contact time, temperature, concentration of the extractant, and acidity on the extraction by Cr6DGA were investigated. It was found that in 3 mol/L HNO3, the SFEu/Am value was about 2. The complexation energies calculated by DFT showed that the Eu(III) complexes were more stable than the corresponding Am(III) complexes in gas, aqueous, and organic phases. Furthermore, the coordination study showed that the metal/ligand ratio of the extracted species was 1:2 by mass spectrometry (MS) analysis. The time-resolved laser-induced fluorescence spectra (TRLFS) further proved that the extracted species contained one water molecule, and so the composition of the extracted complexes may be [EuL2NO3(H2O)]2+ or [EuL2(NO3)2(H2O)]+. Finally, DFT calculations revealed that [EuL2(NO3)2(H2O)]+ is a more stable species and the binding energy of Eu(III) with the DGA unit is lower than that with the crown unit.

Introduction

Since the first use of nuclear energy, its development has spanned more than half a century. At present, nuclear energy has become synonymous with “economic, green, and efficient” energy. With the continuous increase of population and the rapid development of the economy, the energy demand is increasing rapidly and therefore more attention needs to be paid to the safe use of nuclear energy.

For the safe development and effective utilization of nuclear energy, the disposal of high-level radioactive waste is crucial. In particular, the long-lived fission products and actinides urgently need to be remediated because of their strong radioactivity, strong toxicity, and high heat release rate. Among them, the separation of trivalent actinides and lanthanides has become one of the most challenging topics due to their very similar physicochemical properties.[1,2]

In recent years, a series of important ligands have been applied to the partitioning of lanthanides/actinides. These ligands include organic phosphorus compounds (for the TALSPEAK process),[3] diglycolamide (DGA),[4] malonamide (for the DIAMEX process),[5] and N-heterocyclic ligands (for the SANEX process).[6] Among them, it was found that the DGA extractant rapidly developed due to its advantages of stable chemical properties, radiation resistance, flammability, and absence of secondary pollution. Several DGA extractants were tested on a large scale and “thermal engineering operation” was performed.[7] In particular, N,N,N,N′-tetraoctyl diglycolamide (TODGA) was found to be the optimal ligand in terms of solubility and extraction capacity,[8] and it was certified that the stoichiometry of the extraction complexes depended on the nature of diluents.[9] In polar diluents, such as 1,2-dichloroethane, the dominant chemical species was the metal/TODGA in 1:2 ratio, while in nonpolar diluents, the 1:3 or 1:4 species was the predominant complex.[10] Obviously, the complexation of several TODGA molecules with a trivalent metal ion is not favorable to the entropy change during extraction.[11] In nonpolar diluents, such as n-dodecane, three to four TODGA molecules are involved in the formation of reverse micelles, the core of which shows an affinity for trivalent lanthanide and actinides in a dimensionally selective manner.[12] It can be predicted that the preattachment of three to four DGA molecules to a molecular platform will improve the affinity of the ligand to trivalent lanthanide and actinide ions, and this ligand may demonstrate a higher ability to extract trivalent lanthanide and actinide ions than TODGA. Therefore, some researchers have developed a series of multiple-DGA-functionalized ligands by attaching multiple DGA arms to a central platform to make the extraction process more thermodynamically favorable and independent of diluents, such as 2,2′-((9-methyl-9-(11-octyl-6,10-dioxo-2,8-dioxa-5,11-diazanonadecyl)-2,16-dioxo-7,11-dioxa-3,15-diazaheptadecane-1,17-diyl)bis(oxy))bis(N,N-dioctylacetamide) (T-DGA), 2,2′,2″-((((nitrilotris(ethane-2,1-diyl))tris(azanediyl))tris(2-oxoethane-2,1-diyl))tris(oxy))tris(N,N-dioctylacetamide) (TREN-DGA), and 2,2′,2″-(((((2,4,6-triethylbenzene-1,3,5-triyl)tris(ethane-2,1-diyl))tris(azanediyl))tris(2-oxoethane-2,1-diyl))tris(oxy))tris(N,N-dioctylacetamide) (Bz-T-DGA), and the M/L ratio of the extracted species by these ligands is 1:2, and only one DGA unit in a multiple-DGA extractant molecule participates in the coordination and extraction of metal ions.[13−15] Recently, 2,2′,2″-(((1,4,7-triazonane-1,4,7-triyl)tris(2-oxoethane-2,1-diyl))tris(oxy))tris(N,N-dioctylacetamide) (T9C3ODGA, Figure ) was designed by connecting three DGA arms to the macrocyclic aza-crown ether to reduce the steric hindrance of the ligand and make full use of the advantages of multiple-DGA-functionalized structures, and the distribution ratios of Eu(III)/Am(III) by T9C3ODGA were significantly higher than those of T-DGA and TREN-DGA.[16] Although the protonation of macrocyclic aza-crown ether occurs under strong acid conditions, the extraction experiments showed that the distribution ratio and the selectivity for Eu(III)/Am(III) are not significantly affected.[17] Subsequently, 2,2′,2″,2‴-(((1,4,7,10-tetraazacyclododecane-1,4,7,10-tetrayl)-tetrakis(2-oxoethane-2,1-diyl))tetrakis(oxy))tetrakis(N,N-dioctylacetamide) (T12C4ODGA, Figure ) has been reported. Compared with T9C3ODGA, the ability of this ligand to extract trivalent lanthanide/actinide metal ions was significantly improved (n-dodecane, DAm > 500, DEu > 103).[18] According to the experimental and theoretical results, the ligands T9C3ODGA and T12C4ODGA formed the same type of extraction complex (M/L = 1:1 and 1:2), and the 1:2 species was the optimal structure during the extraction process. It was found that the coordination modes of the two ligands were almost identical due to their similar structures. It needs to be noted that only one DGA arm in one ligand molecule participates in the coordination with one trivalent metal ion. In other words, during the extraction process, most of the DGA arms of these multiple-DGA ligands do not directly participate in the coordination of trivalent metal ions, which seems to be inconsistent with the atom economy principle of green chemistry.[19] Diglycolamide-functionalized calix[4]arenes (C4DGA) can significantly improve the selectivity of the metal Am3+ compared with UO22+, while the selectivity of the Eu ion is slightly decreased. Using this extractant, it is possible to quantitatively extract trivalent lanthanide and actinide ions (distributive ratio > 300) with little or no extraction of UO22+ and Sr2+.[4]arenes showing unusual complexation of actinide ions in room temperature ionic liquids: Role of ligand structure, radiolytic stability, emission pectroscopy, and thermodynamic studies. Inorg. Chem.. 2013 ">20]

Figure 1

Chemical structures of some multiple-DGA-functionalized ligands used to extract lanthanides and actinides.

In recent years, ionic liquids have been widely used in the extraction and separation of trivalent lanthanides and actinides.[21−28] Under the same conditions, the ionic liquid can affect the extraction ability and separation factor of the extractant on the trivalent lanthanide and actinide ions. Previous studies on T-DGA ligands showed that, under the same conditions, the DAm value was high in molecular diluents (DAm = 11.1) but significantly lower in room-temperature ionic liquids (DAm = 0.08). In an ionic liquid medium, TREN-DGA has a strong ability to extract trivalent lanthanide and actinide metal ions.[14] The solvent extraction studies suggested a unique selectivity reversal in the extraction of trivalent actinides versus trivalent lanthanides, which was observed when extraction studies in an ionic liquid was performed. Therefore, it is necessary to develop novel DGA extractants with a higher extraction efficiency for trivalent metal ions.

To investigate the effects of macrocyclic crown ether compounds on the extraction and separation of trivalent lanthanides/actinides, some extractants and synergistic agents derived from crown ethers were developed. For example, combined with 4-benzoyl-3-phenyl-5-isoxazolone (HPBI), the synergistic extraction D values of 15-crown-5 (15C5), benzo-15-crown-5 (B15C5, Figure S1), 1-benzyl-1-aza-15-crown-5 (BA15C5, Figure S1), and 2-[(1-aza-15-crown-5)-1-ylmethyl)]-4-(phenyldiazenyl)-naphthalen-1-ol (PDN1A15C5, Figure S1) for Eu(III) ions are 1.0, 0.5, 2.5, and 6.3 at pH 2.4, respectively.[29] Specifically, the La(III), Pr(III), and Eu(III) ions are extracted as Ln(PBI)3·2S complexes (S = PDN1A15C5), while the light lanthanoid(III) ions of the 4f-series (La–Gd) form the extraction complex as Ln(PBI)3·S species (S = B15C5 and 15C5). It should be noted that the addition of crown ether increases the extraction efficiency and the separation factors between the heavier Ln(III) and light Ln(III) are higher than those found for their extraction with the extractant alone. Further modification of the macrocyclic crown ether, N,N′-bis[(6-carboxy-2-pyridyl)methyl]-1,10-diaza-18-crown-6 (H2BP18C6, Figure ), showed that it has good separation performance for trivalent lanthanide ions (La–Gd), and the extraction distribution ratios are distributed within the range of 5.0 × 10–3–2 × 102 (conditions: 0.05 mol/L bis-2-ethyl(hexyl)phosphoric acid in o-xylene/0.001 mol/L H2BP18C6 in 1 mol/L NaNO3, 0.05 mol/L lactate at pH 3).[30] The above results can be explained based on a comparison of the following parameters: the size of the crown cavity, the diameter of cations, and the species of the donor groups in the crown ether host. Macrocyclic crown ethers have potential research prospects in the coordination and separation of Ln(III) and An(III).

Figure 2

Chemical structure of the crown ether-derived ligand H2BP18C6 used to separate lanthanides and actinides.

Based on the above research background, we designed a single DGA-arm-grafted aza-18-crown-6 macrocyclic ligand (Cr6DGA, Scheme ), reducing the number of DGA arms, to explore its extraction performance for trivalent lanthanide/actinide elements. In this paper, we first synthesized and characterized Cr6DGA and systematically studied its extraction behavior and the coordination mode with the trivalent metal ions. The possible extraction mechanism of Ln(III)/An(III) by this macrocyclic crown ether was experimentally and theoretically studied.

Scheme 1

Synthesis Route to the Cr6DGA Ligand

Results and Discussion

Preparation of the Ligand

The ligand, Cr6DGA, was prepared in a 25% yield according to Scheme . Note that the reaction conditions need to be strictly controlled, and an appropriate eluent and elution order are very important for purification of the crude product of Cr6DGA by column chromatography. After two consecutive silica gel column purifications (300–400 mesh, CH2Cl2/MeOH (v:v) = 250:2; CH2Cl2/MeOH (v:v) = 250:1), pure Cr6DGA was obtained as a pale yellow solid.

Solvent Extraction

First, the extraction equilibrium time of Am(III) and Eu(III) was investigated. It was found that the extraction equilibrium (Figure a) was reached within 10 min. Additionally, the extended contact time had also no adverse effect on extraction efficiency. Thus, to ensure that the measured samples had reached equilibrium, all the experiment time of extraction were carried out with the balanced contact time of 1 h. According to the acidity extraction results in Figure b, it was found that as the concentration of nitric acid increased from 1.0 × 10–1 to 3.0 mol/L, the extraction distribution ratios of Am(III) and Eu(III) were increased from 0.3 to 9 and 0.5 to 18 in the 2.0 × 10–2 mol/L Cr6DGA/cyclohexanone system, respectively. Increasing the concentration of nitric acid caused the increase of NO3– concentration, and the homo-ionic effect of NO3– promoted the extraction efficiency of metal ions. This trend is similar to the concentration of nitric acid extraction curves of DGA and multiple-DGA-functionalized ligands. While in this concentration range of nitric acid, the crown ether substructure reduced the extraction efficiency of metal ions due to protonation.[17] This suggested that the binding site of the metal ion to the Cr6DGA may be mainly in the amide pocket, whereas the crown ether unit is most likely not involved in the coordination.

Figure 3

(a) Effect of contact time on the extraction of Am3+ and Eu3+. Organic phase: 1.0 × 10–2 mol/L Cr6DGA in cyclohexanone. Aqueous phase: 3.0 mol/L HNO3. (b) Effect of the concentration of HNO3 on the extraction of Am3+ and Eu3+. Organic phase: 2.0 × 10–2 mol/L Cr6DGA in cyclohexanone. (c) Effect of the concentration of Cr6DGA (in mmol/L) on the extraction of Am3+ and Eu3+. Organic phase: Cr6DGA in cyclohexanone. Aqueous phase: 3.0 × 10 mol/L HNO3. (d) Effect of the concentration of NaNO3 on the extraction of Am3+ and Eu3+. Organic phase: 2.0 × 10–2 mol/L Cr6DGA in cyclohexanone. Temperature: 25 ± 0.5 °C.

The influence of the ligand concentration on the extraction of Am(III) and Eu(III) is shown in Figure c; the logarithmic curve of DM and the concentration of the ligand has a straight slope of 1.76 and 1.82, indicating that Cr6DGA with Am(III) and Eu(III) mainly exist in a 2:1 extraction species, which is very similar to other DGAs.[15,16]

The effect of a salting-out agent on the extraction efficiency is shown in Figure d, which is similar to that of the TOGDA ligand. In the range of 0.5–3.0 mol/L, the extraction efficiency of Cr6DGA for metal ions is positively correlated with the concentration of NaNO3. The results suggested that nitrate ions would promote the extraction of metal ions by Cr6DGA. The logarithmic curve of DM and the concentration of NO3– ions had a straight slope of 2.2 and 1.9, indicating that the extraction species mainly existed in the 2:1 form. However, it is unknown whether NO3– ions participate in the inner coordination of Am(III) and Eu(III) ions, which will be further discussed below.

The distribution ratios of various DGA ligands to Eu(III) and Am(III) ions in the 3 mol/L HNO3 solution are summarized in Table . As can be seen from Table , Cr6DGA has a high selectivity to Eu(III), DEu is higher than DAm, and the SFEu/Am value is about 2. However, under the same conditions, TODGA has a slightly higher selectivity for Am. It is worth noting that the selectivity trend of Cr6DGA is consistent with that of other DGA extractants, which all have better selectivity to Eu. The details of the extraction mechanism of these trivalent metal ions by Cr6DGA will be discussed further below.

Table 1

Extraction Behavior of Am(III)/Eu(III) in the 3 mol/L HNO3 Solution Using Different DGA Ligands at a Concentration of 2.0 × 10–2 mol/L (Temperature: 25 ± 0.5 °C)

metal ion	DCr6DGAa	DTODGAa	DT-DGAb,(16)	DTREN-DGAb,(16)	DBz-T-DGAb,(16)	DT9C3ODGAc,(18)	DT12C4ODGAc,(18)
Am(III)	9	105	11	0.4	235	71	255
Eu(III)	18	93	99	1	390	300	>103

Organic phase: cyclohexanone.

1.0 × 10–3 mol/L solutions in a 10% IDA + 90% n-dodecane mixture.

1.0 × 10–3 mol/L solutions in 5% IDA/n-dodecane medium.

ESI-MS Analysis

ESI-MS is particularly suitable for the mass determination of complexes because it can directly analyze nonvolatile macromolecules in the liquid phase.[31] Methanol solutions with Eu(NO3)3 concentrations of 5.0 × 10–3, 1.0 × 10–2, and 1.5 × 10–2 mol/L were stirred with 5.0 × 10–3 mol/L Cr6DGA for 1.0 h. In Figure , the peaks at m/z 709.8961 and 494.6093 were attributed to the complex species [Eu(Cr6DGA)2NO3]2+ and [Eu(Cr6DGA)2(NO3)2 + 2H]3+, respectively, which were consistent with the isotopic simulated cluster peaks. ESI-MS analysis showed that the binding ratio of Eu(III) to Cr6DGA was 1:2 and the extracted complex contained the nitrate ions, which were consistent with the results of extraction experiments.

Figure 4

ESI-MS analysis of Eu(III) complexes with the Cr6DGA ligand in methanol ([Eu(NO3)3] = 5.0 × 10–3 mol/L; [ligand] = 5.0 × 10–3 mol/L): (a) [EuL2NO3]2+ and (b) [EuL2(NO3)2 + 2H]3+. Temperature: 25 ± 0.5 °C.

Luminescence Spectral Analysis

Solvent extraction results showed that Cr6DGA and Eu(III) formed a 2:1 complex in the cyclohexanone diluent, but it was unclear whether there were water molecules in the inner layer of the Eu ion. The emission intensities for the 5D0 → 7F1 (λ = 594 nm) and 5D0 → 7F2 (λ = 619 nm) transitions are shown in Figure .[32,33] The complexation constants for spectral titration were fitted using the HyperSpec program,[34] and the complexation constants of ML and ML2 types of complexes were 4.91 and 9.88, respectively, which were consistent with the results of extraction experiments and ESI-MS analysis. However, in the case of T9C3ODGA, the stability constants of ML and ML2 types of complexes were 4.75 and 9.12, respectively.[16]

Figure 5

Fluorescence spectroscopic titrations of Cr6DGA with Eu(NO3)3 in acetonitrile (conditions: [Eu(NO3)3] = 4.0 × 10–3 mol/L, volume = 2.0 mL; [Cr6DGA] = 4.0 × 10–2 mol/L; temperature = 25 ± 0.5 °C).

TRLFS Analysis

Time-resolved laser-induced fluorescence spectra can provide information about the first coordination sphere of Eu(III) by the efficient energy transfer from the excited states of the metal ions to the ligand. There is a linear correlation between the fluorescence lifetime and the number of water molecules (eq ). The fluorescence lifetime τ (in ms) can be determined from the curve fitting with single exponential functions, and the values are listed in Table . The fluorescence decay curves with Cr6DGA/Eu(III) = 1:1 in methanol is shown in Figure . It can be seen that when the hydration number decreases from 1.25 to 0.3, the molar ratio of metal/ligand decreases from 1 to 0.25, indicating that the ligand gradually replaces the water molecule entering the central Eu(III) ion.

Table 2

Lifetime and Number of Water Molecules of Cr6DGA with Eu(NO3)3 in Methanol at Different M/L Molar Ratios (Initial Conditions: [Eu(NO3)3] = 2.5 × 10–3 mol/L, Volume = 1.5 mL; [Cr6DGA] = 4.0 × 10–2 mol/L; Temperature = 25 ± 0.5 °C)

M/L ratio	1	0.5	0.25
τ (ms)	0.57	0.78	1.16
N (H2O)	1.25	0.75	0.3

Figure 6

Fluorescence decay curve of Eu(NO3)3 with Cr6DGA in methanol solution (initial conditions: [Eu(NO3)3] = 2.5 × 10–3 mol/L, volume = 1.5 mL; [Cr6DGA] = 4.0 × 10–2 mol/L; temperature = 25 ± 0.5 °C).

Generally, for most DGA-derived ligands, the internal coordination of a metal ion usually does not involve water molecules.[35] However, for the ligand in this paper, there is one water molecule in the inner coordination layer of the metal ion. We speculate that for the 9-coordinated Eu(III) complex, its internal coordination layer contains two ligands and one water molecule, and there may also be 1–3 NO3– ions involved in the inner layer coordination. Based on the results of the above experiments, it can be inferred that the composition of the complex structure may be [Eu(Cr6DGA)2(NO3)(H2O)]2+ and [Eu(Cr6DGA)2(NO3)2(H2O)]+.

DFT Calculations

Although we have confirmed the composition of the metal–ligand complexes, the specific binding site of the ligand and metal ion is unclear since both the oxygen atom of the crown ether and the amide oxygen atom may bind to the metal ion. To identify the binding sites of the ligand with metal ions, we carried out DFT calculations to optimize the structure of these two possible complexes as shown in Figure S5 and found that the binding energy to the DGA unit was 9.9 kcal/mol lower than that at the “crown ether” unit (−11.5:–11.1, water/cyclohexanone). Correspondingly, in the follow-up research, we will focus on the binding mode of metal ions with the DGA unit for further discussion.

To elucidate the binding differences between the coordination of the ligand with Am(III) and Eu(III), the structures of the 1:2 species of [ML′2(NO3)(H2O)]2+ and [ML′2(NO3)2(H2O)]+ (M = Am and Eu) complexes (Figure ) were optimized; these structures were based on the results of the extraction study, spectrophotometric studies, and MS analysis. The relevant structural parameters of Eu(III) and Am(III) complexes are shown in Table , i.e., bond distance (d) and the Mayer bond order (MBO). The metal–ligand distances (M–O(CO) and M–O(ether)) and metal–oxygen (M–O(nitrate) and M–O(water)) bond lengths are provided as average values. The Eu–O(CO) bond distances were in agreement with the experimental values in Eu(TODGA)33+ (2.401 Å, EXAFS) complexes[36] and [Eu(TEDGA)3]3+ (2.389 Å, crystal structure) complexes.[37] Additionally, the Am–O(CO) bond lengths were also well consistent with those in the Am(TMOGA)3(ClO4)3 (2.459 Å) crystal structure.[38] Though the calculated M–O(ether) bond distances were longer than the experimental one by about 0.1 Å, this is quite understandable. The previous computational values reported the bond lengths between the Eu(III) and the ether O atom of TODGA within 2.56–2.65 Å, which was in line with the calculated bond lengths.[39−41] It should be noted that the M–O(CO) bond lengths in [ML′2(NO3)(H2O)]2+ complexes are shorter than the M–O(CO) bond lengths in [ML′2(NO3)2(H2O)]+ complexes. Accordingly, the bond orders of M–O(CO) in [ML′2(NO3)(H2O)]2+ complexes are higher than those of M–O(CO) in [ML′2(NO3)2(H2O)]+ complexes.

Figure 7

Optimized structures of metal complexes [ML′2(NO3)(H2O)]2+ and [ML′2(NO3)2(H2O)]+ (M = Am (a, b), Eu (c, d)) in the gas phase at the B3LYP/def2-SVP/RECP level.

Table 3

Calculated M–O Distances (dM–X, in Å) and Mayer Bond Order (MBO) in [ML′2(NO3)(H2O)]2+ and [ML′2(NO3)2(H2O)]+ (M = Eu and Am) Complexes in the Gas Phase

complexes	atom	d (Å)	MBO
[EuL′2(NO3)(H2O)]2+	O(CO)	2.398 (2.389 exp.)a	0.319
	O(ether)	2.574 (2.489 exp.)a	0.041
	O(H2O)	2.515	0.290
	O(nitrate)	2.447	0.406
[EuL′2(NO3)2(H2O)]+	O(CO)	2.418	0.306
	O(ether)	2.621	0.036
	O(H2O)	2.403	0.388
	O(nitrate)	2.428	0.347
[AmL′2(NO3)(H2O)]2+	O(CO)	2.445 (2.459 exp.)b	0.259
	O(ether)	2.619 (2.519 exp.)b	0.083
	O(H2O)	2.563	0.255
	O(nitrate)	2.491	0.374
[AmL′2(NO3)2(H2O)]+	O(CO)	2.464	0.256
	O(ether)	2.658	0.076
	O(H2O)	2.441	0.368
	O(nitrate)	2.481	0.312

[Eu(TEDGA)3][Ln(NO3)6] crystal structure (ref (36)).

Am(TMOGA)3(ClO4)3 crystal structure (ref (37)).

The Mulliken’s charges of Am(III) and Eu(III) ions are significantly smaller in comparison with the formal “+3” oxidation state. This implies a strong ligand to metal atom charge transfer (Table S1). In the metal complexes, the net charge of Eu(III) is smaller than that of Am(III), which means that the charge transfer from the ligand to Eu(III) is greater than that of Am(III). In addition, the Mulliken’s spin population of the metal (ρM) in the metal complexes obtained in the gas phase is also provided (Table S1) because it was shown that ρM is an important parameter for evaluating the bonding properties of the f-block complexes. The spin population values (ρM) of the metal atom of the discussed complexes were obtained by the Multiwfn code using Mulliken’s[42,43] and Löwdin’s[44] methods. When the deviation between ρM and 6.00 is greater, the bonding interaction between the ligand and the metal tends to be more covalent. Comparing the spin population values (ρM) between Am and Eu complexes, it was found that the Eu(III) complexes had a greater ρM value than the Am complexes by both Mulliken’s and Löwdin’s methods.

To understand the extraction mechanism further, the possible complexation reactions of the ligand and their complexation products [ML′2(NO3)(H2O)]2+ and [ML′2(NO3)2(H2O)]+ (M = Eu and Am) were investigated. The liquid extraction system is a very complicated process. For the process of liquid–liquid extraction in the HNO3 medium, eq can be used to describe the simplified extraction processes of Am(III) and Eu(III) with the studied Cr6DGA ligand.

The single-point energy calculations of the optimized Eu(III) and Am(III) complexes were performed in the aqueous phase and the organic (cyclohexanone) phase at the B3LYP/def2-SVP/RECP//def-TZVP level. The changes in the complex formation energy (ΔEcf) of reaction in the gas, aqueous, and organic phases were calculated and are given in Table . Obviously, the ΔEcf values for the reaction with [ML′2(NO3)2(H2O)]+ complexes turn out to be more negative than those of [M(L′)2(NO3)(H2O)]2+ complexes in all gas, aqueous, and organic phases, which suggests that the extraction metal complexes are more likely to form [ML′2(NO3)2(H2O)]+ species during the extraction process. As expected, the Eu(III) complexes are more stable than the corresponding Am(III) complexes in the gas phase (Δ(ΔEcf)Am/Eu = +4.1 and +3.7 kcal/mol for [ML′2(NO3)(H2O)]2+ and [ML′2(NO3)2(H2O)]+, respectively). By contrast, when the solvent effect of water (PCM) is taken into account, the differences between Am and Eu complexes decrease to +1.0 and +0.7 kcal/mol, respectively. When changing to cyclohexanone as a solvent, this trend is similar. The positive Δ(ΔEcf)Am/Eu values (+1.2 and +0.9 kcal/mol for [M(L′)2(NO3)(H2O)]2+ and [ML′2(NO3)2(H2O)]+, respectively) indicate that the Cr6DGA ligand has better selectivity for Eu(III) over Am(III). These results are consistent with the extraction results.

Table 4

Complex Formation Energies, ΔEcf (kcal/mol), and the Energy Differences, Δ(ΔEcf)Am/Eu, of the Formation of Eu(III) and Am(III) Complexes in the Gas Phase, Water, and Cyclohexanone, Calculated at the B3LYP/def2-SVP/RECP//def-TZVP Level

complexes	metal ion	phase	ΔEcf	Δ(ΔEcf)Am/Eu
[ML′2(NO3)(H2O)]2+	Eu3+	gas	–334.4
		water	–26.1
		cyclohexanone	–42.3
	Am3+	gas	–330.3	4.1
		water	–25.1	1.0
		cyclohexanone	–41.1	1.2
[ML′2(NO3)2(H2O)]+	Eu3+	gas	–469.3
		water	–34.1
		cyclohexanone	–56.6
	Am3+	gas	–465.6	3.7
		water	–33.4	0.7
		cyclohexanone	–55.7	0.9

Conclusions

In this paper, a novel single DGA-arm-grafted macrocyclic crown ether ligand was synthesized to investigate its extraction behavior for the Eu(III)/Am(III) cations. It was found that Cr6DGA was more selective than TODGA for the trivalent lanthanide ions in a 3 mol/L HNO3 solution, which may be explained by the introduction of the crown ether structure in the ligand. The coordination behavior of Cr6DGA with Eu(III) and Am(III) was identified by both slope analysis and luminescence spectroscopy, suggesting an ML2 complex with an inner-sphere water molecule. ESI-MS analysis indicated that the extraction complexes may exist in the form of [Eu(Cr6DGA)2NO3(H2O)]2+ or [Eu(Cr6DGA)2(NO3)2(H2O)]+. DFT calculations revealed that the binding ability of Eu(III) to the DGA unit of Cr6DGA is stronger than that of the crown ether unit, and [ML′2(NO3)2(H2O)]+ is more energy favorable. The calculation results also show that the Eu(III) complexes are more stable than the corresponding Am(III) complexes in gas, aqueous, and organic phases, which is in good agreement with the selectivity of the extraction results. These results can help us to understand the extraction mechanism of Am(III) and Eu(III) ions by the DGA-type ligands and provide some insights for the design of novel DGA-derived ligands for extraction of lanthanides and actinides. However, the role of the crown ether unit during the extraction process is quite complicate, which might be explained from the comparison of the cavity size of crown ethers and the diameter of cations, and the types of donor groups in the crown ether host. Since the specific function of the crown ether structure is not clear here, our group will continue to systematically study the selectivity of crown ethers and their derivatives to trivalent lanthanide and actinide metal ions.

Experimental Section

Materials and Instruments

The reagents mentioned were purchased from Aladdin (Shanghai Aladdin Biochemical Technology Co. Ltd, China) and were of analytical or chromatographic grade. CH2Cl2 and THF were distilled from CaH2. The reactions were conducted under N2 unless otherwise stated. 241Am and 152,154Eu, provided by the China Institute of Atomic Energy, were used as radioactive tracers in solvent extraction.

The NMR spectra were recorded on a Varian Inova NMR spectrometer (Bruker Inc., Switzerland). Electrospray ionization mass spectrometry (ESI-MS) data were obtained by an LCMS-IT-TOF spectrometer (Shimadzu, Japan). The concentration of Eu was determined using an inductively coupled plasma optical emission spectrometer (ICP-OES, Optima8000, PerkinElmer).

The activities of 241Am and 152,154Eu were measured by a NaI (TI) scintillation counter. Luminescence spectroscopic titration was obtained on a fluorescence spectrophotometer (F96pro, China). The Transient Fluorescence spectra was measured on a Fluorolog-3 spectrofluorometer (JobinYvon, Horiba).

Synthesis of Cr6DGA

Compounds 1 and 2 were synthesized according to the reported methods.[45,46] The synthesis of Cr6DGA was as follows: compound 1 (0.1 g, 0.56 mmol) was dissolved in 25 mL of anhydrous CH2Cl2. Then, compound 2 (0.36 g, 0.56 mmol), chlorotripyrrolidinyl hexafluorophosphate (0.24 g, 0.56 mmol), and diisopropylethyl amine (DIEA, 0.16 g, 1.25 mmol) were added in one portion. The mixture was stirred at room temperature (RT) in a nitrogen atmosphere for 4 days. Subsequently, it was washed three times with 10 mL of 5% hydrochloric acid. The crude product was collected by removing the organic solvent under vacuum and was purified by a silica gel column (300–400 mesh; CH2Cl2/MeOH (v:v) = 250:2; CH2Cl2/MeOH (v:v) = 250:1) twice. Finally, the desired product (Cr6DGA) was obtained as a pale-yellow solid with a Rf value of 0.5 (CH2Cl2/MeOH (v:v) = 10:1) and a yield of 25%. m.p.: 39–41 °C. HRMS: m/z calcd for Cr6DGA: 603.4579 [M + H]+; found: 603.4568. 1H NMR (400 MHz, CDCl3): δ = 4.48 (s, 4H), 3.88–3.46 (m, 24H), 3.23 (t, 2H), 3.09 (t, 2H), 1.50 (bs, 4H), 1.40–1.15 (m, 20H), 0.88 (t, 6H). 13C NMR (100 MHz, CDCl3): δ = 14.0, 22.94, 26.72, 27.29, 27.47, 28.78, 29.22, 29.38, 29.47, 29.57, 31.80, 31.92, 46.23, 47.07, 68.44, 76.70, 77.02, 77.33, 168.77.

The organic phases were prepared in cyclohexanone. For a single metal ion solution, it was prepared by dissolving metal ions in nitric acid solutions of different concentrations. The effects of various parameters on the extraction distribution ratio, including the contact time, nitric acid concentration (1.0 × 10–1–3.0 mol/L HNO3), extractant concentration (4.0 × 10–3–2.0 × 10–2 mol/L), and NaNO3 concentration (0.5–3.0 mol/L), were evaluated using radioactive tracers of the metal ion. The selectivity experiment was performed with 2.0 × 10–2 mol/L extractant in the presence of 152,154Eu and 241Am in various concentrations of HNO3 solution. The experiments were carried out using the following standard protocol: first, the organic phase was pre-equilibrated with HNO3 solution at least three times in a phase ratio of 1:1 at 25.0 ± 0.5 °C. Second, the aqueous phases were spiked with trace amounts of 241Am and 152,154Eu in nitric acid solution, and equal volumes of two phases were mixed thoroughly for 1 h in a thermostated water bath (25 ± 0.5 °C). After the phase separation by centrifugation, the count rates of 241Am and 152,154Eu were determined by a NaI (TI) scintillation counter. DM is defined as the ratio between the radioisotope activity or the element concentration in the organic and aqueous phases. The selectivity for 152,154Eu over 241Am is represented by the separation factor, SFEu/Am, which is defined as the ratio of the distribution ratios DEu/DAm. All of the extraction experiments were conducted twice under the same conditions, and the reported values are the average of these values.

LCMS-IT-TOF (Shimadzu, Japan) combines QIT (ion trap) and TOF (time-of-flight) technology, which can effectively introduce ions into QIT and spray to capture ion TOF at the same time. LCMS-IT-TOF can greatly assist in the identification of target compounds using high-speed/high-precision MSn (n ≥ 10). In the high-resolution mode, the detection range of the instrument is 50–5000 m/z. Methanol solutions with Cr6DGA concentrations of 5.0 × 10–3, 1.0 × 10–2, and 1.5 × 10–2 mol/L were stirred with 5 × 10–3 mol/L Eu(NO3)3 for 1.0 h, and the Eu(III) complexes were analyzed.

Luminescence Spectroscopic Titration

Luminescence spectroscopic titration was realized by changing the ratio of metal/ligand in the acetonitrile medium. Acetonitrile is a suitable solvent because of the solubility of both Eu(NO3)3 and Cr6DGA. The fluorescence titration was performed in a 10 mm quartz cell at 25 ± 0.5 °C. The initial concentrations of Cr6DGA and Eu(NO3)3 were 4.0 × 10–2 and 4.0 × 10–3 mol/L, respectively, controlling the titration ratio of metal/ligand from 0.2:1 to 4:1. The emission spectrum of Eu(III) was obtained in the range of 575–630 nm with an excitation wavelength of 395 nm and a bandwidth of 10 nm. The stability constants of the complexes were calculated by the HyperSpec program.[34]

Time-resolved laser-induced fluorescence spectra (TRLFS) were measured on a Fluorolog-3 spectrofluorometer with a SpectraLED (390 nm, S-390) as the excitation source and a picosecond photon detection module (PPD-850) as the detector. The TRLFS analysis test was carried out in a 10 mm quartz cell, and a methanol solution with Eu(III) at concentrations of 2.5 × 10–3, 5.0 × 10–3, and 1 × 10–2 mol/L was mixed with a methanol solution of Cr6DGA (1 × 10–3 mol/L) with a phase ratio of 1:1 at 25 °C.

All geometries were optimized with the hybrid B3LYP[47,48] functional implemented in Gaussian 09D software.[49] For Eu and Am metal atoms, relativistic effects were taken into consideration with the quasi-relativistic effective core potential (RECP). The corresponding ECP60MWB-SEG[50−52] and ECP28MWB-SEG[53,54] valence basis sets were used to describe Am and Eu atoms, respectively. The electronic configuration of Eu(III) and Am(III) in their septet state was chosen as the ground state of their complexes.[55−57] For all other light atoms (C, H, O, and N), the def2-SVP basis set[58,59] was employed for optimization. Frequency calculations were performed to ensure that the obtained stationary points were the minima on the potential energy surface. In summary, all of the geometries were optimized at the B3LYP/def2-SVP/RECP level in the gas phase. Many reported studies stated that the structures of the optimized complexes in aqueous and organic solutions do not change obviously in comparison with those in the gas phase.[60,61] Therefore, solvation energies in the aqueous and organic phases were estimated by single-point calculations with the def-TZVP[62] basis set for light atoms and the same basis set for Am and Eu with the corresponding ECP based on the structures optimized in the gas phase. The self-consistent reaction field (SCRF) polarizable continuum model (PCM)[63] was used to take into account the solvent effect of water and cyclohexanone (dielectric constants of 78.4 for water and 15.6 for cyclohexanone) according to the experimental report. The two solvents considered in this work, water and cyclohexanone, were used in experimental studies. To reduce the calculation cost, all alkyl chains connected with the amide nitrogen atom of the ligand were replaced by a methyl group, and this simplified ligand is represented as L′. We believe that this approximation is acceptable, because the length of the alkyl chain in DGA has no decisive effect on the separation behavior of Am(III) and Eu(III).[64] The Mayer bond order (MBO), Mulliken atomic charges, and Löwdin’s and Mulliken’s spin populations (ρM) of the metals in complexes were determined using Multiwfn software.[65] Visualizations were created with CYLview software.[66]