Tuning aminopolycarboxylate chelators for efficient complexation of trivalent actinides

Abstract

The complexation of trivalent lanthanides and minor actinides (Am³⁺, Cm³⁺, and Cf³⁺) by the acyclic aminopolycarboxylate chelators 6,6′-((ethane-1,2-diylbis–((carboxymethyl)azanediyl))bis–(methylene))dipicolinic acid (H₄octapa) and 6,6'-((((4-(1-(2-(2-(2-hydroxyethoxy)ethoxy)ethyl)-1H-1,2,3-triazol-4-yl)pyridine-2,6-diyl)bis–(methylene))bis–((carboxymethyl)azanediyl))bis–(methylene)) dipicolinic acid (H₄pypa-peg) were studied using potentiometry, spectroscopy, competitive complexation liquid–liquid extraction, and ab initio molecular dynamics simulations. Two studied reagents are strong multidentate chelators, well-suited for applications seeking radiometal coordination for in-vivo delivery and f-element isolation. The previously reported H₄octapa forms a compact coordination packet, while H₄pypa-peg is less sterically constrained due to the presence of central pyridine ring. The solubility of H₄octapa is limited in a non-complexing high ionic strength perchlorate media. However, the introduction of a polyethylene glycol group in H₄pypa-peg increased the solubility without influencing its ability to complex the lanthanides and minor actinides in solution.

Introduction

At the end of its service in a commercial nuclear reactor the irradiated nuclear fuel consists of 95% uranium dioxide, 1% plutonium dioxide, roughly 3.5% fission products, and ≈ 0.5% minor actinides (An; predominantly Np, Am, and Cm)¹. The majority of countries with established nuclear energy technologies manage used nuclear fuel as waste for a variety of socio-economic reasons. Globally, as our society continues to mature, the ethics of sustainability and environmental responsibility will continue to evolve, strengthening the argument for a more resourceful approach. In a closed nuclear fuel cycle option the actinides are reused in reactors to generate electric power². This approach enables increased utilization of uranium resources, and helps manage the long-term fate of radioactive byproducts through a reduction of bulk volume, radiotoxicity, and decay heat load inside a geologic repository³.

Actinide recovery/recycle/isolation techniques center on the development of efficient sequestering reagents, capable of selective coordination of the 5f elements in aqueous effluents containing nearly a third of the periodic table^4,5,6. A variety of metal ion complexants have been explored throughout the years, mainly targeting the rich redox chemistry of major actinides (U, Pu, and Np), and electronic properties of trivalent minor actinides (Am and Cm)⁷. The isolation of the 5f metals commonly employs solvent extraction methods, exploiting the chemical binding preferences to partition actinides into an immiscible liquid phase, and separate them away from unwanted metal ions. Efficient differentiation of trivalent minor actinides and trivalent lanthanides (Ln) is of particular importance as the 4f fission products are neutron poisons¹, but selective binding of the trivalent An is obstructed by similar solution chemistry of the trivalent Ln⁸. The f-elements are collectively hard Lewis acids, highly hydrated spherical cations in solutions with compact f-orbitals, embedded in the interior of an ion. Similar ionic radii, contracting across the 4f and 5f series, yield trivalent cations of nearly equivalent charge density (e.g., Am³⁺ and Pm³⁺). Accordingly, such ions are indistinguishable to hard Lewis bases (oxygen donor ligands). The selective binding of trivalent actinides is realized due to the enhanced spatial extension of the 5f orbitals, relative to 4f orbitals, inviting stronger interaction with reagents containing less electronegative atoms (e.g., N and S)⁸.

The European strategy for trivalent An/Ln differentiation employs heterocyclic nitrogen donor ligands^9,10. The two classes of such reagents, bis–triazinylpyridines, bis–triazinyl bipyridines, and bis–triazinyl phenanthrolines contain highly acidic amine binding sites, sustaining efficient An/Ln selectivity in mixtures containing molar quantities of nitric acid^9,10. Recently, a bis–triazolylpyridine chelator was identified as a suitable choice, accomplishing the An³⁺/Ln³⁺ task in moderately acidic aqueous mixtures (0.25 M nitric acid)¹¹. In the United States, acyclic aminopolycarboxylate N-donor ligands are utilized to facilitate An/Ln separations^2,12,13,14. As with the heterocyclic nitrogen donors the presence of amine moieties enhances the stability of actinide complexes. Efficient group differentiation is accomplished if a multidentate reagent like diethylenetriamine-N,N,N',N'',N''-pentaacetic acid (H₅dtpa) is employed^12,13. The structure of H₅dtpa is shown in Fig. 1. The weaker [Ln(dtpa)]²⁻ complexes dissociate, and the Lns are extracted into the non-aqueous environment. The stronger [An(dtpa)]²⁻ complexes remain in the aqueous layer. Although simple and eloquent in its original recipe, the aminopolycarboxylate (APC) approach offers a solution in buffered, mildly acidic (pH > 2) aqueous environment, where metal complexation by APCs is strong. The liquid–liquid distribution is also slow to attain equilibrium as low proton content inhibits the hydrogen ion catalyzed mechanism of metal complex dissociation^12,13.

Figure 1

Chemical structures of different aminopolycarboxylate ligands.

The operational pH window for the APC chemistry can be expanded to higher aqueous acidities (0.5 < p[H⁺] < 2.0) through structural modification of the chelator, targeting functional groups which increase the acidity of amine sites. Our recent structure–function relationship studies for aminopolycarboxylate complexants identified a 6-carboxypyridin-2-yl-methyl pendant arm, commonly known as N-2-methylpicolinate, as a promising substitutent to strengthen metal complexation in acidic media¹⁵. A replacement of a single acetate arm of ethylenediamine-N,N,N′,N′-tetraacetic acid, H₄edta, by a picolinate moiety yielded H₄edta-mpic (Fig. 1), which significantly increased the total ligand acidity (∑pK_a), while maintaining the capacity for An³⁺/Ln³⁺ differentiation. Further symmetrical acetate-for-picolinate exchange on the H₄edta molecular motif yields H₄octapa, an octadentate acyclic aminopolycarboxylate reagent, recently demonstrated as a versatile radiometal chelator for radiopharmaceutical uses (Fig. 1)^16,17. The N-2-methylpicolinate groups offer structural pre-organization to increase the stability of metal chelate and strong electron-inducing effects which increase the total acidity of the APC ligand. Accordingly, aqueous complexants such as H₄octapa may sustain An³⁺/Ln³⁺ differentiation in aqueous mixtures of increased acidity. For H₄edta-mpic, the β₁₀₁(Am)/β₁₀₁(Nd) ratio remained similar to that of H₄edta. This would indicate that, despite the introduction of additional N-donor atom, no appreciable enhancement in 4f/5f differentiation should be expected for H₄octapa upon the introduction of a second picolinate group.

An overall enhancement in ligand softness, which has been shown to play an important role in separation of trivalent An from Ln, can be attained through an introduction of additional nitrogen donor atoms into the APC structure. This study examines trivalent f-element complexation by a novel APC structure, H₄pypa-peg, where a pyridine ring is added to H₄octapa ligand as illustrated in Fig. 1. Both reagents are based on the ethylenediaminetetraacetatic acid structural motif, where, for H₄octapa, two N-2-methylpicolinate groups replace two acetate pendant arms, and, for H₄pypa-peg, a pyridine ring is also centrally added. The H₄pypa-peg also includes a polyethylene glycol moiety to improve solubility in aqueous electrolyte mixtures. The structural modifications enable the study of chelate stability induced by the rigidity of the binding pocket, total ligand acidity and the electron-inducing effects on the soft nitrogen donor atoms for reagents containing 2-methylpicolinate pendant arms. A rigid, pre-organized binding pocket of H₄octapa is compared with a more flexible H₄pypa-peg, where a central pyridine also introduces additional nitrogen donor to seek enhanced preference for actinide complexation.

Trivalent actinide coordination with H₄pypa-peg and H₄octapa was studied to determine whether the insertion of the pyridine functionality in H₄octapa structure yields a structural modification suitable for deployment in trivalent 4f/5f element separations. Potentiometric titrations were performed for both APC ligands to compare their protonation equilibria and interpret metal complexation results. The complexation of trivalent Ln and An (Am³⁺, Cm³⁺, and Cf³⁺) was characterized using potentiometry, spectrophotometry, and competitive solvent extraction and the interpretation of the acquired data was aided by trivalent f-element complexation studies for H₄edta, H₅dtpa, and triethylenetetraamine-N,N,N',N'',N''',N'''-hexaacetic acid (H₆ttha) in 2.0 M (Na⁺/H⁺)ClO₄ electrolyte medium.

Methods

Reagents

Caution: All radiological manipulations were performed in a HEPA filtered fume hood or negative-pressure glovebox approved for radiological work.

Preparation and standardization of Ln salts and electrolyte salts was reported previously^15,21,22. Trivalent ²⁴³Am and ²⁴⁸Cm were retrieved from Idaho National Laboratory (INL) stocks, and purified using a diglycolic acid (DGA) extraction chromatographic resin (Eichrom), as described previously¹⁵. The DGA resin was also used to recycle trivalent ²⁴⁹Cf (sourced as the chloride salt from the Isotope Development and Production for Research and Applications Program through the Radiochemical Engineering and Development Center at Oak Ridge National Laboratory (ORNL) from previous experimental work). The ²⁴³Am(NO₃)₃ and ²⁴⁸Cm(ClO₄)₃ working stocks were prepared in 0.01 M HNO₃ and HClO₄, respectively, and standardized using UV/Visible spectroscopy^15,23,24,25. Radiotracer mixtures of ¹⁵⁴Eu³⁺ (Eckert & Ziegler) and ²⁴³Am³⁺, ²⁴⁸Cm³⁺ and ²⁴⁹Cf³⁺ for solvent extraction competitive studies were further diluted from stocks to ensure 300–500 Bq per 5 µL content. Bis(2-ethylhexyl)phosphoric acid (97%, Millipore-Sigma, HDEHP) was purified using the copper salt precipitation method^26,27. Potentiometric titration indicated 99.7 ± 0.3% purity of the HDEHP extractant. Octane diluent (99 + %) was purchased from Millipore-Sigma and used without any further purification.

Ligand synthesis

H₄octapa

The ligand was synthesized using an experimental protocol reported by Platas-Iglesias et al.²⁸ It was recrystallized from water twice prior to thermodynamic studies. The purity of the complexant was > 99% as verified using ¹H NMR.

H₄pypa-peg

6,6'-((((4-(1-(2-(2-(2-hydroxyethoxy)ethoxy)ethyl)-1H-1,2,3-triazol-4-yl)pyridine-2,6-diyl)bis–(methylene))bis–((carboxymethyl)azanediyl))bis–(methylene)) dipicolinic acid, H₄pypa-peg, was prepared in five synthesis steps that are described in detail in the Supplemental Information. Final purity of product was > 99% as verified using ¹H NMR and complexometric titration with a well-characterized ²⁴³Am stock solution.

Potentiometric titrations

All potentiometric titrations were run using a Mettler Toledo T90 autotitrator equipped with an Orion Ross Semi-micro glass electrode with the junction stabilization solution changed to 5.0 M NaCl to prevent precipitation with perchlorate media. All titrations were performed under a hydrated nitrogen atmosphere. The forward titration measurements (acidic titrand, basic titrant) of H₄octapa in nitrate electrolyte medium were performed as described previously due to the limited solubility of this ligand in a 2 M (Na⁺/H⁺)ClO₄ medium¹⁵. Acid dissociation constants of H₄octapa in perchlorate electrolyte medium were also determined by titrating in reverse (basic titrand, acidic titrant) to alleviate the solubility concerns. Titrations of H₄pypa-peg were performed in reduced-volume glass titration cups (10 mL nominal) due to the limited supply of this ligand. All potentiometric titrations were run in triplicate. The choice of using p[H⁺] was explained in Heathman et al.¹⁵, and the calibration of the p[H⁺] scale utilized Gran analysis through titration of a strong acid with the base used in the potentiometric titrations^29,30,31. Potentiometric titration data was modeled and analyzed using Hyperquad2013 to determine acid dissociation and metal complexation stability constants^32,33.

Spectroscopic studies

Changes in the optical absorption characteristics of Nd³⁺ were monitored using a Cary 6000i UV/Vis–NIR Spectrometer (Agilent) using 1 cm semimicro quartz cuvettes (Starna) in double-beam mode, where the ⁴I_9/2 → ⁴G_5/2, ²G_7/2 transitions³⁴ were monitored between 560 and 605 nm with a 0.075 nm interval and 1.0 nm spectral bandwidth. Changes in the optical absorption characteristics of ²⁴³Am³⁺ and ²⁴⁸Cm³⁺ were monitored using a Flame-S-VIS–NIR-ES spectrometer coupled to a DH-2000-BAL dual-mode light source (Ocean Optics) with a single 1 cm semimicro quartz cuvette (Starna). The light source and spectrometer were optically coupled to the CUV sample (Ocean Optics) holder using 2 m fiberoptics (200 μm) and the parameters for acquisition were 0.369 nm interval and 1000 scans averaged with an integration time/scan 1.08 ms (total acquisition time ~ 1 s). The complexation of Am³⁺ with the ligands was monitored in the 495–525 nm spectral region, looking at the ⁷F₀′ → ⁵L₆′ transition^23,35. For Cm³⁺, ligand complexation was monitored in the 365–410 nm spectral region, looking at the ⁸S_7/2 → ²J_i (i = {15}, {13,9}, {11,17}) manifold transitions^24,25,36. Changes in aqueous p[H⁺] were monitored throughout each titration step using a Gran-calibrated Orion Ross Semi-micro glass electrode. Baseline correction was applied to all datasets and was typically a simple linear subtraction. However, in the case of the ²⁴⁸Cm/H₄pypa-peg spectra, the tail end of the ligand absorption overlapped the hypersensitive Cm peaks. In this case the background correction was applied using a 5th-order empirical polynomial. Finally, the spectroscopic data was analyzed using the HypSpec software package^33,37. Full details of the titrand and titrant solutions for each titration are included in the Supplemental Information.

Fluorescence measurements used a HORIBA Jobin Yvon IBH FluoroLog-3 fluorometer adapted for time-resolved studies. A submicrosecond xenon flash lamp (Jobin Yvon, 5000XeF) was used as a light source. The DAS 6 software (HORIBA Jobin Yvon IBH) was used for decay analysis and data fitting. Single and double exponential decay curves were used to model the observed luminescence lifetime data yielding χ² values ranging between 1.01 and 1.08.

Phase distribution measurements

Stability constants for the coordination of Eu³⁺, Am³⁺, Cm³⁺, and Cf³⁺ were also achieved by monitoring the suppression of liquid–liquid partitioning of the radiotracer due to the increasing presence of an aqueous complexant. Aqueous solutions were prepared containing sub-millimolar amounts of ligand while maintaining p[H⁺] at 1.7, 1.8, 1.9, 2.0, and 2.1. Ionic strength was maintained at 2.00 M using sodium perchlorate. Aqueous complexation was balanced with the appropriate choice of HDEHP concentration in octane to allow accurate quantification of metal distribution. Non-aqueous phases were pre-equilibrated three times with 2.00 M NaClO₄ at the corresponding p[H⁺]. Preliminary time-dependent studies indicated the phase transfer equilibrium was attained for all measurements with H₄octapa and H₄pypa-peg (Figures S2 and S3 in the SI). For the mixtures traced with radioisotopes, the activity was measured using gamma spectroscopy (ORTEC GEM50P4 coaxial HPGe detector, DSPEC gamma spectrometer) for ¹⁵⁴Eu (123.07 keV), ²⁴³Am (74.66 keV), and ²⁴⁹Cf (388.17 keV) and liquid scintillation counting (Perkin Elmer Tri-Carb 3180 TR/SL) for ²⁴⁸Cm. The ratio of radioisotope activity in the organic and aqueous phases defined the liquid–liquid distribution.

Computational methods and models

Density functional theory (DFT) calculations using Gaussian 16³⁸ have been performed to identify the most stable structures of the Eu³⁺ complexes with the H₄octapa and H₄pypa ligands containing two and one water molecules, respectively (see the SI section for details). Each complex was placed in a periodic cubic box of 15.226 Å (0.47 mol/L) and explicitly solvated with ≥ 85 water molecules. The initial configurations for AIMD simulations were generated by the MedeA Amorphous Cell Builder³⁹ and preequilibrated for AIMD using the PCFF + ⁴⁰ force field supported in MedeA-LAMMPS^39,41. These simulations were carried out in an NVT ensemble at a temperature of 300 K for 5 ns, in which only the solvent molecules were allowed to equilibrate.

After initial equilibration of the [Eu(octapa)(H₂O)₂]⁻ and [Eu(pypa)(H₂O)]⁻ complexes with explicit water molecules, first-principles molecular dynamics (MD) simulations based on density functional theory (DFT) were carried out in the Born–Oppenheimer approximation using the VASP software^42,43,44,45. The valence electronic states were expanded in a basis of plane waves. The core-valence interaction was described using the Projector Augmented Wave (PAW) approach^46,47. All ab initio MD (AIMD) calculations utilized a plane wave kinetic energy cutoff of 400 eV and the $\Gamma$ point approximation. The PBE GGA functional^48,49 was used to describe the exchange–correlation energy together with the DFT-D3 method of Grimme⁵⁰ for dispersion interactions. The Self-Consistent Field (SCF) convergence threshold was set to 10⁻⁴ eV in all the calculations. During the SCF solution, a Pulay scheme⁵¹, as implemented in VASP was used for charge density mixing. The trivalent Eu was modeled using the large core (LC) pseudopotential with six f electrons placed in the core. This is justified, because the Eu 4f orbitals typically have a negligible overlap with the frontier orbitals responsible for the Eu-ligand bonding⁵², whereas the placement of electrons in the valence shell can lead to spurious delocalization of localized f electrons. The employed method was shown to provide reliable Eu-EDTA bond distances in the solid state¹⁵, justifying it for predicting the solution structure of Eu³⁺ complexes.

For [Eu(octapa)(H₂O)₂]⁻ and [Eu(pypa)(H₂O)]⁻ complexes, the initial configurations thus obtained were subjected to short geometry optimization using VASP, followed by AIMD simulations with a timestep of 1.0 fs. The Nosé-Hoover thermostat was adopted to maintain a temperature of 300 K for the complex with [octapa]⁴⁻, but the simulation temperature for the complex with [pypa]⁴⁻ was raised to 328 K to speed up the equilibration process. Additional AIMD simulations were performed for the [pypa]⁴⁻ complex at 300 K using the CP2K software⁵³ to verify that the temperature variation has a minimal impact on the metal–ligand bond distances (Table S1). The–1 charge on the complex was compensated by a uniform background charge. We obtained ≥ 116 ps long trajectories, showing a transition of one of the water molecules from the inner to the outer shell at ~ 70 ps in both cases. Two windows of the last 40 ps for two different coordination geometries were chosen for structural analysis. Graphical visualization of the solvated structure was prepared using VMD⁵⁴ and the analysis of the trajectory was performed using GROMACS⁵⁵. The coordination number, CN(t), was defined by Eq. (1), as follows⁵⁶,

$$CN\left( t \right) = \mathop \sum \limits_{i = 1}^{Natom} \frac{{1 - \left( {\frac{{r_{i} \left( t \right)}}{{r^{\dag } }}} \right)^{12} }}{{1 - \left( {\frac{{r_{i} \left( t \right)}}{{r^{\dag } }}} \right)^{24} }}$$

(1)

where ${r}_{i}$ is the distance of the i-th atom (oxygen or nitrogen atom) from Eu³⁺ and r^† is the location the first minimum appearing after the first Eu³⁺–O or Eu³⁺–N radial distribution function peak, as shown in Figure S5.

Results and discussion

Synthesis of Complexant with Improved Aqueous Solubility

H₄pypa features a centrally located pyridine, increasing the donor group count by one compared to H₄octapa. However, compared to H₄octapa, which has only two pyridine units, H₄pypa's hydrophilicity is reduced due to the higher number of aromatic heterocycles in its structure. Nonetheless, the central pyridine serves as a further functionalization platform, as sucessfully demonstrated with the bioconjugation studies in nuclear medicine applications^18,19,20. A more practical approach was devised to functionalize H₄pypa to enhance its solubility in moderately acidic aqueous electrolyte mixtures (Fig. 2). The synthetic scheme incorporates a reactive site in the central pyridine fragment early on, enabling the Pd-catalyzed Sonogashira coupling reaction to introduce a further functionalizable alkyne moiety. Intermediate 2 is highly versatile as it can be utilized to append reactive or unreactive structural fragments for a variety of applications via Cu-catalyzed azide-alklyne cycloadition. To synthesize H₄pypa-peg, 2 was reacted with 2-[2-(2-azidoethoxy)ethoxy]ethanol, followed by acid-catalyzed hydrolysis of methyl and tert-butyl esters. This simple structural modification resulted in significantly improved aqueous solubility of H₄pypa-peg when compared to unfunctionalized H₄pypa.

Figure 2

Synthesis of new H₄pypa-peg complexant with improved aqueous solubility.

Acid dissociation constants

Changes in hydrogen ion concentration were monitored using a glass electrode for a solution of H₄octapa as either NaOH or HClO₄ were titrated into the solutions. Low solubility of H₄octapa in an aqueous sodium perchlorate electrolyte medium prompted initial ligand titrations in a 2.0 M (Na⁺/H⁺)NO₃ ionic background. The experimental potentiometric curve collected for a forward titration of H₄octapa in (Na⁺/H⁺)NO₃ is presented in Fig. 3A, alongside the distribution curves for the identified H_nL species. Following those studies, the acid/base conditions were reversed and octapa⁴⁻ was titrated in 2.0 M (Na⁺/H⁺)ClO₄ using HClO₄ (SI Figure S7). The potentiometric curves are well respresented by five proton dissociation equilibria. The calculated equilibrium constants correspond with the protonation reactions for two aliphatic amines, two pyridine nitrogens, and one carboxylate group with the protonation sequence described by Eq. (2). The acid dissociation constants (pK_a =  − log₁₀K_a) for H₄octapa are reported in Table 1 alongside previously reported values in the literature.

Figure 3

Potentiometric p[H⁺] trends collected for (A) H₄octapa protonation titration at T = 25.0 ± 0.1 °C and I = 2.00 ± 0.01 M (Na⁺/H⁺)NO₃, and (B) H₄pypa-peg protonation titration at T = 25.0 ± 0.1 °C and I = 2.00 ± 0.01 M (Na⁺/H⁺)ClO₄. (A) Titrand: V_initial = 25.012 mL, C_H4octapa = 4.875 mM, C_H+ = 0.043 M. Titrant: 0.397 M NaOH and 1.60 M NaNO₃. (○) Experimental p[H+], (red solid line) calculated p[H⁺], (black dashed line) L⁴⁻, (red dashed line) HL³⁻, (green dashed line) H₂L²⁻, (blue dashed line) H₃L⁻, (pink dashed line) H₄L_(Aq), (violet dashed line) H₅L⁺. (B) Titrand: V_initial = 10.067 mL, C_H4pypa-peg = 4.851 mM, C_H+ = 0.044 M. Titrant: 0.125 M NaOH in 1.875 M NaClO₄. (O) Experimental p[H⁺], (red solid line) calculated p[H⁺], (orange dashed line) L⁴⁻, (red dashed line) HL³⁻, (green dashed line) H₂L²⁻, (blue dashed line) H₃L⁻, (pink dashed line) H₄L_(aq), (violet dashed line) H₅L⁺, (black dashed line) H₆L²⁺.

Table 1 Acid dissociation constants determined for H₄octapa compared with those reported in literature^16,17.

The initial three equilibrium reactions, (H₈L⁴⁺, H₇L³⁺, and H₆L²⁺), occur in aqueous conditions too acidic for accurate quantification using a glass electrode. Those reactions correspond to the protonation reactions of carboxylate groups. The fourth acid dissociation equilibria (H₅L⁺) may be attributed to the weakest carboxylic acid proton. The remaining pK_a sequence describes the protonation reactions of amine sites, with the initial two (H₄L and H₃L⁻) consistent with acidities of nitrogen within the pyridinecarboxylate structure, and the latter two (H₂L²⁻ and HL³⁻) characteristic of the amine sites of the ethylenediamine backbone. Some differences between nitrate and perchlorate aqueous electrolyte media are noticeable, especially for protonation constants ascribed to the amine groups. Also, as expected, ionic strength influence on the acid dissociation constants is evident, when compared with acid dissociation constants reported by Kálmán et al.¹⁶ and Jaraquemada-Peláez et al.¹⁷ However, a strong electron-withdrawing influence of the N-2-methylpicolinate groups is evident, significantly decreasing the pK_a values of the amine nitrogens.

$${\text{H}}_{9 - a} {\text{L}}^{ + 5 - a} { \leftrightarrows }{\text{H}}_{8 - a} {\text{L}}^{ + 4 - a} + {\text{H}}^{ + } \;\;\;K_{a} = \frac{{\left[ {H^{ + } } \right] \cdot \left[ {{\text{H}}_{8 - a} {\text{L}}^{ + 4 - a} } \right]}}{{\left[ {{\text{H}}_{9 - a} {\text{L}}^{ + 5 - a} } \right]}}\;\;\;a = { 1} \ldots \;{8}$$

(2)

The experimental potentiometric titration curve and the distribution curves for the associated H_nL species for H₄pypa-peg are presented in Fig. 3B. The measured pK_a values for H₄pypa-peg in 2.0 M (Na⁺/H⁺)ClO₄ are summarized in Table 2. Fitting of the experimental curves showed convergence when six acid dissociation constants were included in the model. For comparison, the protonation equilibria of H₄edta were also studied in 2.0 M (Na⁺/H⁺)ClO₄. The pK_a values for H₄edta are listed in Table 2 and collected potentiometric titration curves (forward and reverse) are presented in SI Figure S8. The protonation constants for similar ligands reported in literature (H₄pypa¹⁸ and H₄py4pa⁵⁷) are also listed in Table 2 for comparison. For H₄pypa-peg coordination pocket the protonation of H₉L⁵⁺, H₇L³⁺, and H₆L²⁺ cannot be characterized using a glass electrode. The H₆L²⁺ species likely represents a protonated carboxylate group, while H₅L⁺ and H₄L may describe the pyridinecarboxylate nitrogens. The equilibrium involving H₃L⁻ likely describes the dissociation of the protonated pyridine nitrogen, while the two most basic species in this ligand (HL³⁻ and H₂L²⁻) are likely the bridging amines in the H₄pypa-peg backbone. Equation (3) describes the protonation sequence for H₄pypa-peg.

$${\text{H}}_{10 - a} {\text{L}}^{ + 6 - a} { \leftrightarrows }{\text{H}}_{9 - a} {\text{L}}^{ + 5 - a} + {\text{H}}^{ + } \;\;\;K_{a} = \frac{{\left[ {H^{ + } } \right] \cdot \left[ {{\text{H}}_{9 - a} {\text{L}}^{ + 5 - a} } \right]}}{{\left[ {{\text{H}}_{10 - a} {\text{L}}^{ + 6 - a} } \right]}}\;\;\;a = { 1} \ldots {9}$$

(3)

Table 2 Acid dissociation constants determined for H₄pypa-peg, H₄edta and other structurally related APC ligands previously reported^18,57.

Metal ion coordination by APC ligands in aqueous acidic mixtures is strongly dependent on the basicity of the dialkylamine sites. The sum of the acid dissociation constants for such amine sites, ΣpK_a(N), provides a relative comparison scale. This total dialkylamine basicity follows the H₄edta > H₄pypa-peg > H₄octapa trend indicating that ethylenediamine backbones of H₄pypa-peg and H₄octapa are more acidic, relative to H₄edta. The lowering of ΣpK_a(N) for both APC ligands may be attributed to the electron-withdrawing influence of the N-2-methylpicolinate groups, relative to N-acetate arms of H₄edta. The strong inducing effects are lessened through the introduction of the pyridine group in the H₄pypa-peg ligand.

Lanthanide Complexation Behavior

The coordination of La³⁺, Nd³⁺, Eu³⁺, Tb³⁺, Dy³⁺, Ho³⁺, and Lu³⁺ with H₄octapa and H₄pypa-peg was studied. Figures 4A, B show the potentiometric curves collected when Nd³⁺ was titrated with H₄octapa and H₄pypa-peg, respectively. Both p[H⁺] curves show the characteristic buffering region due to deprotonation of metal–ligand complexes when titrated with base. The experimental p[H⁺] trend is best represented when considering the presence of MHL_(aq) and ML^– complexes as generalized by Eqs. (4) and (5). The progressive formation of the metal complexes throughout the titration is illustrated by species distribution curves, also shown in Figs. 4A, B. The complex formation equilibria are described by the conditional stability constants, β₁₀₁ and β₁₁₁, listed in Table 3.

Figure 4

Potentiometric p[H⁺] trends collected for complexometric titrations between Nd³⁺ and (A) H₄octapa at T = 25.0 ± 0.1 °C and I = 2.00 ± 0.01 M (Na⁺/H⁺) NO₃ and (B) H₄pypa-peg at T = 25.0 ± 0.1 °C and I = 2.00 ± 0.01 M (Na⁺/H⁺)ClO₄. (A) Titrand: V_initial = 25.244 mL, C_H4octapa = 4.803 mM, C_H+ = 0.043 M, C_Nd3+ = 4.887 mM. Titrant: 0.397 M NaOH and 1.60 M NaNO₃. (○) Experimental p[H⁺], (red solid line) calculated p[H⁺], (orange dashed line) Nd³⁺, (green dashed line) NdHL_(aq), (red dashed line) NdL⁻. (B) Titrand: V_initial = 10.153 mL, C_H4pypa-peg = 4.700 mM, C_Nd3+  = 4.841 mM, C_H+  = 0.044 M. Titrant: 0.125 M NaOH in 1.875 M NaClO₄. (○) Experimental p[H⁺], (red solid line) calculated p[H⁺], (orange dashed line) Nd³⁺, (green dashed line) NdHL_(aq), (red dashed line) NdL⁻.

Table 3 Conditional stability constants for the formation of ML^- and MHL_(aq) complexes for selected lanthanides (La³⁺, Nd³⁺, Eu³⁺, Tb³⁺, Dy³⁺, Ho³⁺, and Lu³⁺) for H₄octapa, H₄pypa-peg, and other structurally similar ligands.

The coordination of Nd³⁺ by H₄octapa and H₄pypa-peg was also studied using spectroscopic methods. Figures 5A, B show the collected spectrophotometric signatures for H₄octapa and H₄pypa-peg, respectively.

$$M^{3 + } + L^{4 - } { \leftrightarrows }ML^{ - } \;\;\;\beta_{101} = \frac{{\left[ {ML^{ - } } \right]}}{{[M^{3 + } ]\left[ {L^{4 - } } \right]}}$$

(4)

$$M^{3 + } + H^{ + } + L^{4 - } { \leftrightarrows }MHL_{{\left( {aq} \right)}} \;\;\;\beta_{111} = \frac{{\left[ {MHL_{{\left( {aq} \right)}} } \right]}}{{[M^{3 + } \left] {\left[ {L^{4 - } } \right]} \right[H^{ + } ]}}$$

(5)

Figure 5

Changes in the optical absorption spectra for neodymium ion induced by (A) H₄octapa in 2.00 M (Na⁺/H⁺)NO₃ and (C) H₄pypa-peg in 2.00 M (Na⁺/H⁺)ClO₄ at T = 20 ± 1 °C. (A) Titrand conditions: V_initial = 0.809 mL, C_Nd3+= 7.62 mM, p[H⁺]_initial = 1.40. Titrant conditions: V_{titrant added} = 0.529 mL, C_Nd3+= 7.65 mM, C_H4octapa = 9.75 mM, p[H⁺] ≈ 6. (C) Titrand conditions: V_initial = 0.801 mL, C_Nd3+  = 9.817 mM, p[H⁺] = 1.61. Titrant: C_H4pypa-peg = 24.584 mM, C_Nd3+  = 9.913 mM, p[H⁺] = 3.59. The calculated molar absorptivities (ε_A)are shown in panels (B) for H₄octapa species and (D) for H₄pypa-peg species.

Optical absorption features of ⁴I_9/2 → ⁴G_5/2, ²G_7/2 transitions for Nd³⁺ are perturbed due to the introduction of increasing quantities of complexants, resulting in characteristic red-shifted changes. In Fig. 5A a broad absorption band with λ_max = 577 nm is distinctly different to the optical absorption of free Nd³⁺ metal in Fig. 5B. The differences can be attributed to the nitrate complexation of Nd³⁺ as the spectrophotometric titration with H₄octapa was performed in 2.0 M (Na⁺/H⁺)NO₃. Analysis of the spectral data collected for the titration with H₄octapa (Fig. 5C) suggests the presence of two uniquely absorbing metal species: a ${\text{NdNO}}_{3}^{{{2} + }}$ and NdL⁻, suggesting a quantitative displacement of the weak nitrate ligand by H₄octapa. The HypSpec interpretation of the Nd/pypa-peg titration points to the presence of three absorbing species: Nd³⁺, NdHL_(aq), and NdL⁻. The calculated molar absorptivities for those species are shown in Fig. 5D. The conditional stability constants obtained spectroscopically are listed in Table 3, showing good agreement with the stability constants determined using potentiometry. The complex stability constants for structurally similar APC ligands found in the literature are also presented (H₄pypa¹⁸ and H₄py4pa⁵⁷).

Relative to H₄edta, both H₄octapa and H₄pypa-peg form complexes of increased stability, as the substitution of two N-acetate pendant arms by N-2-methylpicolinates increases the denticity of the coordination pocket. The H₄octapa is a well-known octadentate chelator of trivalent f-elements, while, with the additional donor group, H₄pypa-peg has the potential to displace all water molecules from the hydration zone of trivalent f-elements. The luminescence lifetime measurements of Eu³⁺ in presence of H₄pypa-peg were performed to study the ligand-induced metal dehydration. The aqueous p[H⁺] = 1.54 and 4.70 were chosen for the measurements as conditional stability constants indicated high abundance of [Eu(Hpypa-peg)]_(aq) and $\left[ {{\text{Eu}}\left({{\text{pypa-peg}}} \right)} \right]_{{({\text{aq}})}}^{ - }$ species, respectively. An estimated 1:1 molar ratio of those metal complexes was also investigated at p[H⁺] = 1.99. The luminescence lifetime decay trends are shown in SI Figure S9, and the fitted decay constants (τ) and estimated inner-sphere hydration numbers (${\eta }_{\mathrm{H}2\mathrm{O}}$) for each complex are summarized in SI Table S2. The data suggests that at low p[H⁺], the [Eu(Hpypa-peg)]_(aq) complex is heptadentate with two inner-sphere waters, while at higher p[H⁺], the [Eu(pypa-peg)]⁻ complex is ocatadentate with one water molecule in the coordination sphere of Eu³⁺. The presence of single water molecule in the ML⁻ complex of H₄pypa-peg may indicate that not all donor atoms of this chelator participate in metal ion complexation. The inspection of the acquired conditional stability constants may help with identifying this donor group.

The stability of the ML⁻ complexes for light lanthanides with H₄pypa-peg shows remarkable similarity to H₄octapa. This supports the luminescence finding, pointing to the same binding denticity for both APC ligands. The coordination pockets are well suited to accept the bulkier, less charge dense 4f cations. A sharp contrast is observed for the complexation of heavier lanthanides (Tb³⁺ through Lu³⁺) with H₄octapa and H₄pypa-peg. After initially increasing, the stability constant trend for H₄octapa reaches a plateau and β₁₀₁ constants for the complexation of Eu³⁺, Tb³⁺, Dy³⁺, and Ho³⁺ with H₄octapa do not vary appreciably, and eventually decrease for Lu³⁺. This stability constant trend has been attributed to steric factors which increase when H₄octapa coordinates heavier trivalent lanthanides^16,17. Similar impact of steric hindrance on the stability constants for trivalent f-element complexation has been observed for large polydentate APC chelators such as H₅dtpa and H₆ttha^58,59. The cross-lanthanide stability constant trend increases monotonically for H₄pypa-peg, resembling H₄edta. This trend is also observed with the non-functionalized H₄pypa, showing a 1000-fold stability increase for ML⁻ complexes. Accordingly, the introduction of the central pyridine ring in H₄pypa-peg lowers the rigidity of the coordination pocket of H₄octapa. The MHL_(aq) complex for H₄pypa-peg forms more readily (higher abundance) at higher p[H⁺] relative to H₄octapa in accordance with its lower overall total ligand acidity.

Structural analysis of the EuL⁻ complexes from ab initio molecular dynamics simulations

In order to further understand the structural properties of metal complexes with H₄octapa and H₄pypa-peg, we simulated the [Eu(octapa)(H₂O)₂]⁻ and [Eu(pypa)(H₂O)]⁻ complex using AIMD. Figure 5 shows evolution of various bond distances and coordination numbers (Eq. (1)) over the course of the simulations. The data illustrate the interconversion dynamics between ten and nine-coordinate complexes, capturing the departure of one water molecule from the inner to the outer shell for both the Eu³⁺-octapa and Eu³⁺-pypa complexes. The snapshots of the Eu³⁺-octapa and Eu³⁺-pypa complexes in two coordination states (CN = 9 and 10) are shown in Figure S4. Structural correlations based on the radial distribution functions (RDFs) for [Eu(octapa)(H₂O)_n]⁻ (n = 1, 2) and [Eu(pypa)(H₂O)_n]⁻ (n = 0, 1) complexes are provided in Figure S5.

Table 4 provides a comparison of the Eu–O and Eu-N bond distances for Eu³⁺-octapa in the presence of either one or two inner-sphere water molecules and for Eu³⁺-pypa in the presence or absence of one inner-sphere water molecule. Analysis of the Eu–O and Eu–N bond lengths shows that the metal–ligand bond lengths are in most cases shortened in the nine-coordinate structure compared to those in the ten-coordinate structure. The stronger bond shortening going from CN = 10 to CN = 9 is observed for the acetate-COO⁻ groups that become largely indistinguishable from the pyridine-COO⁻ groups. At the same time, the bond distances to the amine N atoms remain substantially longer (by 0.10–0.18 Å) compared to pyridine N atoms, which can be attributed to higher donor ability and better preorganization of the latter. An exception to this is one terminal pyridine N₁ atom that exhibits a substantial bond elongation in the ten-coordinate pypa complex. Likewise, the amine N₂ atom adjacent to it shows the longest bond elongations and the largest bond fluctuation, providing flexibility to accommodate one inner-sphere water molecule.

Table 4 Bond distances (Å) in EuL⁻ complexes with H₄octapa and H₄pypa ligands from AIMD simulations. The last 40 ps segment for each coordination state was taken for structural analysis.

Based on luminescence and paramagnetic NMR measurements²⁸, the Ln³⁺-octapa complexes in solution were inferred to be nine-coordinate, with one water molecule completing the primary coordination sphere. Unlike coordination environments in solution, the X-ray studies in the solid state for La³⁺¹⁷ and Gd³⁺⁶⁰ showed a ten-coordinate geometry, where two water molecules reside in the primary shell. Differences between the solid- and solution-state interactions for lanthanide complexes that include crystal packing and outer-sphere solvent effects, respectively, are not uncommon, with the Eu³⁺-edta complex being another example^15,61, where the number of water molecules increases from 2 in solution to 3 in the solid state. There is a relative paucity of knowledge pertaining to solution phase coordination that is hard to unambiguously probe experimentally. Thus, our DFT-based AIMD simulation was used to shed light into to (i) the evolution of the crystalline-like complex in aqueous environment and (ii) corroborate the hydration numbers from the luminescence lifetimes.

Two unsuccessful departure attempts for one of the inner-sphere water molecules in [Eu(octapa)(H₂O)₂]⁻ were observed during the first 15 ps of the AIMD simulation. This transition finally succeeded at ~ 70 ps and the resulting [Eu(octapa)(H₂O)]⁻ complex remained stable during the rest of the simulation (Fig. 6). Thus, in bulk water, according to the present MD simulation results, the interaction of the second water with the metal ion center is less favorable than the water-water interaction, which agrees with the presence of one coordinating water molecule based on luminescence studies.

Figure 6

Time-dependent structural metrics for EuL⁻ complexes with H₄octapa and H₄pypa-peg ligands. (A) Fluctuation of the individual Eu³⁺–O bond distances, (B) oxygen coordination number of Eu³⁺ considering the Eu³⁺–O cutoff distance of 3.4 Å (based on the first minimum of the Eu³⁺–O RDF shown in Figure S5), (C) fluctuation of the individual Eu³⁺–N bond distances, and (D) nitrogen coordination number of Eu³⁺ considering the Eu³⁺–N cutoff distance of 3.5 Å (based on the first minimum of the Eu³⁺–N RDF shown in Figure S5). Structures of ten-coordinated complexes are shown in the inset. The middle pyridine nitrogen in pypa-peg is labeled as N_5-pyr.

An unsuccessful attempt for one of the pyridine-COO⁻ groups in [Eu(pypa)(H₂O)]⁻ to depart the inner coordination shell was observed during the first 5 ps of the AIMD run. At around 70 ps the inner sphere water molecule moved to the second shell, rendering the complex nine-coordinate. However, unlike the octapa complex, the CN plot (Fig. 6B) shows several attempts for another water molecule to approach the Eu³⁺ center as close as 3.5 Å. It is hypothesized that longer AIMD simulations could result in a fully coordinated complex. This conjecture is based on the following two arguments. First, in the EuL⁻ complex with pypa, a large portion of the space around Eu³⁺ remains unobstructed by the ligand, potentially allowing one water molecule in and out of the inner shell. Second, AIMD simulations using the CP2K software package at 298.15 K (Fig. 5) indicate that the potential energies for CN = 10 and CN = 9 states are highly overlapping, which are essentially the same within the statistical uncertainty. This suggests a possibility for a stronger stabilization of the CN = 10 at room temperature compared to the elevated temperature (T = 328 K) at which VASP simulations were run. Interestingly, the water remains fully coordinated in the [Eu(pypa)(H₂O)]⁻ complex in the AIMD simulations performed at T = 298.15 K, which passed the time threshold for water dissociated in the AIMD simulations performed at T = 328 K.

Finally, cluster calculations with implicit solvation were utilized to check which functional group is more likely to be protonated in the [Eu(Hpypa)(H₂O)] complex. The results suggested a strong energetics preference (by at least 5 kcal/mol) for protonating one of the carboxylic groups attached to the amine over the protonation of the terminal pyridine nitrogen and significantly larger energy penalty for protonating the amine nitrogen (by more than 10 kcal/mol). This is consistent with the pK_a values of the functional groups present in the ligands.

Actinide complexation behavior

Liquid–liquid partitioning of trivalent f-elements was monitored in presence of H₄octapa and H₄pypa-peg using a strong liquid cation exchanger, bis–(2-ethylhexyl)phosphoric acid, HDEHP. This competitive solvent extraction methodology was used to determine the conditional stability constants for complexation of Eu³⁺, Am³⁺, Cm³⁺, and Cf³⁺ with H₄octapa and H₄pypa-peg. The metals, present at radiotracer concentrations (>1 μM), distributed between an organic phase containing HDEHP in n-octane and an aqueous phase containing the complexant at varying concentrations and aqueous acidity. The interpretation of the competitive complexation liquid–liquid distribution data (SI Tables S3 through S5 for H₄octapa and S6 through S8 for H₄pypa-peg) considered two complex formation equilibria as described by Eqs. (4) and (5). All collected distribution data indicated the average ligand-to-metal stoichiometry of 1.00 ± 0.04 for H₄octapa and 1.00 ± 0.03 for H₄pypa-peg (see SI Figures S10A and S11A). The partitioning of a trivalent metal ion as facilitated by a dimerized HDEHP cation exchanger can be expressed by Eq. (6) as quantified by the extraction constant, K_ex (subscript org refers to organic-soluble species).

$$M^{3 + } + 3\left( {HDEHP} \right)_{2,org} { \leftrightarrows }M\left( {H\left( {DEHP} \right)_{2} } \right)_{3,org} + 3H^{ + } \;\;\;K_{ex} = \frac{{\left[ {M\left( {H\left( {DEHP} \right)_{2} } \right)_{3,org} } \right]\left[ {H^{ + } } \right]^{3} }}{{\left[ {M^{3 + } } \right]\left[ {\left( {HDEHP} \right)_{2,org} } \right]^{3} }}$$

(6)

The stoichiometry for the extraction of trivalent metal ion by HDEHP was verified at 2.85 ± 0.03 by slope analysis study using ²⁴⁹Cf³⁺ as representative ion for H₄octapa (SI Figure S10B) and 2.78 ± 0.13 for H₄pypa-peg (SI Figure S11B). The distribution of the metal in this competitive complexation liquid–liquid environment can be described by Eq. (7).

$$D_{M} = \left( {\left[ {M\left( {H\left( {DEHP} \right)_{2} } \right)_{3} } \right]_{org} } \right)/\left( {\left[ {M^{3 + } } \right] + \left[ {ML^{ - } } \right] + \left[ {\left( {MHL} \right)_{aq} } \right]} \right)$$

(7)

Substitution of equilibrium and extraction constants Eqs. (3)–(5) into Eq. (6) yields Eq. (8).

$$D_{M} = \left( {K_{ex} \left[ {\left( {HDEHP} \right)_{2,org} } \right]^{3} \left[ {H^{ + } } \right]^{3} } \right)/\left( {1 + \beta_{101} \left[ {L^{4 - } } \right] + \beta_{111} \left[ {L^{4 - } } \right]\left[ {H^{ + } } \right]} \right)$$

(8)

At constant HDEHP concentration and p[H⁺], when the distribution of the metal ion in absence of ligand is labelled as D⁰, the liquid–liquid partitioning of the metal will adhere to Eq. (9),

$$\frac{{D^{0} }}{{D_{M} }} - 1 = \beta^{app} \left[ {L^{4 - } } \right]$$

(9)

where β^app is the apparent stability constant as represented by β ^app = β₁₀₁ + β₁₁₁[H⁺]. Accordingly, when metal distribution is plotted as a function of free ligand concentration, the presence of ML⁻ and MHL_(aq) is revealed if the experimental data shows variance with aqueous p[H⁺].

The Figs. 7A for H₄octapa and Fig. 7C for H₄pypa-peg show representative Eu³⁺, Am³⁺, and Cf³⁺ sets of dependencies collected at constant aqueous acidities to determine β^app constants from the slopes of the error-weighted least-squares regression analyses. Equivalent dependencies were also collected at different p[H⁺] conditions to study variation of β^app trends with p[H⁺] as presented in Figs. 7B,D for H₄octapa and H₄pypa-peg, respectively. The distribution plots for Eu³⁺, Am³⁺, and Cf³⁺ for H₄pypa-peg clearly show variation with the changing aqueous acidity, which yielded good estimates for the conditional stability constants for the formation of ML⁻ and MHL_(aq) (Table 5). The collected trends for Cf³⁺ in the presence of H₄octapa vary to a lesser extent, enabling the determination of β₁₀₁ only. The p[H⁺] variation was not monitored for the coordination of ²⁴⁸Cm³⁺ with both APC ligands due to limited availability of this radioisotope. Accordingly, only β₁₀₁ values are reported in Table 5 for the complexation of Cm³⁺ with H₄octapa and H₄pypa-peg. Good overall agreement between the stability constants acquired using solvent extraction and potentiometry was observed for the complexation of Eu³⁺.

Figure 7

(A) β^app trends determined for the complexation of Cf ³⁺ (p[H⁺] 1.88), Am³⁺ (p[H⁺] 1.87) and Eu³⁺ (p[H⁺] 1.87) by octapa⁴⁻. Distribution ratio data is listed in Tables S3, S4, S5 in Supplemental Information. (B) p[H⁺]-dependence of β^app trends describing trivalent f-element coordination by H₄octapa. Symbol legend: circles represent distribution data at p[H⁺] 1.80 (Cf), 1.80 (Am), 1.78 (Eu), squares at p[H⁺] 1.89 (Cf), 1.88 (Am), 1.88 (Eu), diamonds at p[H⁺] 2.01 (Cf), 2.02 (Am), 1.99 (Eu). (C) β^app trends determined for the complexation of Eu³⁺ (p[H⁺] = 2.00), Am³⁺ (p[H⁺] = 2.00), and Cf ³⁺ (p[H⁺] = 1.98) by pypa-peg⁴⁻. Distribution ratio data is listed in Tables S6, S7, S8 in the Supplemental Information. (D) p[H⁺]-dependence of β^app trends describing trivalent f-element coordination by pypa-peg⁴⁻. Symbol legend: circles represent distribution data at p[H⁺] = 2.00 (Eu), 1.79 (Am), 1.78 (Cf); squares at p[H⁺] = 2.09 (Eu), 1.90 (Am), 1.88 (Cf); diamonds at p[H⁺] = 2.25 (Eu), 2.00 (Am), 1.98 (Cf).

Table 5 Conditional stability constants for trivalent actinides measured spectrophotometrically and through competitive metal complexation in liquid–liquid extractions.

To further validate the competitive solvent extraction methodology the complexation of Am³⁺ with H₄octapa and Cm³⁺ with H₄pypa-peg was studied spectrophotometrically. These titrations monitored the ligand-induced changes on the optical absorption spectra of free Am³⁺ at λ_max = 503.0 nm, (Fig. 8A) and Cm³⁺ at λ_max = 375.0, 380.6, 396.5 nm, (Fig. 8C). The presence of both ligands induced new red-shifted absorption signatures in each monitored titration. For H₄octapa, Fig. 8B shows two light absorbing metal species were observed, assigned to Am³⁺/AmNO₃²⁺ composite and [Am(octapa)]⁻ complex. Figure 8D shows the calculated molar absorptivity spectra for H₄pypa-peg/Cm³⁺ spectrophotometric titration, where the presence of free Cm³⁺ and the ML⁻ complex were also observed. The calculated conditional stability constants are listed in Table 5. Excellent agreement between the solvent extraction and spectrophotometry was attained for the β₁₀₁ values for the formation of [Am(octapa)]⁻ albeit the uncertainty associated with a constant measured using competitive liquid–liquid distribution is rather large and points out the shortcomings of this method. A variance between two analytical methods is larger for [Cm(pypa-peg)]⁻ possibly due to free ligand absorption interferences.

Figure 8

(A) Spectrophotometric titration of Am³⁺ with H₄octapa. Experiment was conducted at 20 ± 1 °C with I = 2.00 M (Na⁺/H⁺)NO₃. Titrand conditions: V_initial = 0.822 mL,C_Am3+ = 0.788 mM, p[H⁺]_initial = 1.34. Titrant conditions: V_{titrant added} = 0.655 mL, C_Am3+= 0.791 mM, C_H4octapa = 4.93 mM, p[H⁺] ≈ 7. (B) Calculated molar absorptivities (ε_A) of the two fitted species in solution, Am³⁺ and [Am(octapa)]⁻. (C) Spectrophotometric titration of Cm³⁺ with H₄pypa-peg. Experiment was conducted at 20 ± 1 °C with I = 2.00 M (Na⁺/H⁺)ClO₄. Titrand: V_initial = 0.809 mL, C_Cm3+  = 1.000 mM, p[H⁺] = 1.34. Titrant: C_H4pypa-peg = 15.098 mM, C_Cm3+ = 1.002 mM, p[H⁺] ≈ 5–6. (D) Calculated molar absorptivities (ε_A) of the two fitted species in solution, Cm³⁺ and [Cm(pypa-peg)]⁻.

As observed for the trivalent 4f elements, the insertion of a central pyridine ring in the H₄pypa-peg structure decreases the rigidity of the coordination pocket of H₄octapa. Similar observation is noted for the trivalent 5f elements as the stability constants follow the Am³⁺ < Cm³⁺ < Cf³⁺ trend. This increase is also preserved in the case of H₄octapa, which suggests that complexation of Cf³⁺ with this rigid ligand adheres to the trend followed by the light members of the f-elemental series. This increase is in accordance with the studies by H₅dtpa Brandau et al. and Leguay et al.^62,63 Accordingly, as in the case of bulky H₅dtpa, the steric factors are not manifested for the complexation of Cf³⁺ with H₄octapa.

The effect of a fifth nitrogen donor atom in H₄pypa-peg on the coordination of trivalent actinides can be evaluated through a comparison of metal complexation equilibria for trivalent 4f and 5f elements of similar charge density. Similar stabilities of metal complexes of Am³⁺ and analogous complexes with Nd³⁺ or Sm³⁺ are expected when complexing agents consist solely of oxygen donor atoms are expected⁸. The presence of donor atoms softer than oxygen, i.e. nitrogen or sulfur, is typically manifested by a steeper slope when the stabilities of 1:1 or 1:2 metal–ligand chelates of Am³⁺ versus Nd³⁺ and Sm³⁺ ions are co-related^8,59. Figure 9 shows this linear free energy relationship for APC reagents containing aminoacetate blocks only. For consistency, stability constants used to establish this trend have been determined in 2.0 M (Na⁺/H⁺)ClO₄^15,64,65,66. The stability constants for Am³⁺ and Nd³⁺ complexation with H₅dtpa and H₆ttha, were also examined in 2.0 M (Na⁺/H⁺)ClO₄ in a series of potentiometric and spectroscopic measurements to broaden the consistency of APC data presented in Fig. 9. For each APC reagent, acid dissociation constants (SI Table S9) and stability constants for coordination of Nd³⁺ and Am³⁺ (SI Table S10). The spectrophotometric titration plots are also provided for Nd³⁺ and Am³⁺ with H₄edta (SI Figures S12 and S13), with H₅dtpa (SI Figures S14 and S15) and with H₆ttha (SI Figures S16 and S17). The slope of 1.06 for the linear free energy relationship constructed in Fig. 9 indicates the stability of the AmL complex is approximately 6% higher, relative to NdL, for the considered APC ligands. Previous studies of APC chelators containing N-2-pyridylmethyl groups have found the Am³⁺/Nd³⁺ stability constant ratios are even higher, finding departures from this linear free energy trend^22,67,68. This is demonstrated for H₄dtta-pym, where a single N-acetate group of H₅dtpa was replaced by N-2-pyridylmethyl group²². This reagent does not adhere to the linear trend established in Fig. 9 and the upward deviation suggests that trivalent actinide binding strengthens with the addition of sp²-hybridized nitrogens²². In contrast, the presence of a pyridine ring in APC reagents containing 2-methylpicolinate functionality does not increase the preference for trivalent actinide binding as evidenced for H₄edta-mpic¹⁵. This may be explained by the strong electron-withdrawing influence imposed on the pyridine nitrogen, depleting its electron density and rendering it similar to sp³-hybridized amine groups. These observations are also preserved for H₄octapa, which shows no departure from the free energy relationship, and H₄pypa-peg where only a small increase in the stability constant ratio for the complexation of Am³⁺ and Nd³⁺ is found. Figure 9 highlights both reagents to emphasize this. The participation of the central pyridine nitrogen in metal binding may be inferred from the enhanced differentiation of Am³⁺/Nd³⁺ pair. However, a lower β₁₀₁ ratio, relative to H₄dtta-pym, suggests that strong electron-inducing forces impact the softness of the central pyridinyl nitrogen of H₄pypa-peg.

Figure 9

Am³⁺/Nd³⁺ stability constant ratio for H₄pypa-peg (●) and H₄octapa (□) as compared with those established for conventional APC reagents containing aminoacetate blocks only, and APC reagents containing N-2-pyridinylmethyl groups. The linear free energy relationship diagram was built by combining previously reported stability constants for Am³⁺ and Nd³⁺ complexation with ethylenediamine-N,N′-di(acetylglycine)-N,N'-diacetate, H₄eddag-dag³⁹, diethylenetriamine-N,N''-di(acetylglycine)-N,N',N''-triacetate, H₅dtta-dag⁴⁰, N-(hydroxyethyl)-diethylenetriamine-N,N′,N'',N''-tetracetate), H₄hedtta,⁴¹ and ethylenediamine-N-(2-methylpicolicate)-N,N',N'-triacetic acid, H₄edta-mpic¹⁰, and stability constants for Am³⁺ and Nd³⁺ complexation with ethylenediamine-N,N,N',N'-tetraacetic acid, H₄edta, diethylenetriamine-N,N,N',N'',N''-pentaacetic acid, H₅dtpa, and triethylenetetraamine-N,N,N',N'',N''',N'''-hexaacetic acid, H₆ttha, redetermined in 2.0 M (Na⁺/H⁺)ClO₄. For comparison, data for N-2-pyridylmethyl-diethylenetriamine-N,N′,N′′,N′′-tetraacetic acid, H₅dtta-pym ( ⊕)¹⁷, is also included as another APC reagent containing a pyridine ring.

Conclusions

The coordination studies of trivalent Ln and An by APC radiometal chelators H₄octapa and H₄pypa-peg show an intricate balance between the rigidity of the binding pocket, total ligand acidity, and the electron-inducing effects on the soft nitrogen donor atoms for reagents containing 2-methylpicolinate pendant arms. The 6-carboxypyridine-2-yl-methyl substituents increase the nitrogen acidity on the aminoacetate APC backbone, substantially lowering the operational pH window for efficient metal coordination. This structural modification comes at a cost of reduced rigidity of a binding pocket and aqueous solubility as demonstrated by the H₄octapa. Steric hindrance, observed in the coordination of trivalent f-elements by H₄octapa, reduces the chelate stability variance across the 4f series, similar to a bulky H₆ttha, and in contrast to the trends observed for conventional APC reagents H₄edta and H₅dtpa. The inhibited rotational flexibility of H₄octapa can be overcome with the addition of a central pyridine ring of H₄pypa-peg, which effectively reorganizes the binding pocket. The AIMD simulations show greater spatial availability for structural reorganization and exchange of water molecules for Eu-pypa-peg complex, relative to Eu-octapa. The observed linearly increasing trend of chelate stabilities across the 4f series is similar to that of H₄edta. The addition of a polyethylene glycol chain averts solubility challenges observed with H₄octapa. The presence of strongly electron-inducing 2-methylpicolinate moieties is also manifested in the observed differences in chelate stabilities of trivalent lanthanides and actinides. For H₄octapa, the enhanced actinide/lanthanide differentiation, as inhered from the β₁₀₁ ratio for the coordination of Am³⁺ and Nd³⁺, matches that observed for conventional APC reagents. H₄pypa-peg shows a small improvement in the preference for actinide binding, which may be assigned to the presence of pyridine.