Synthesis, solid-state, solution, and theoretical characterization of an “in-cage” scandium-NOTA complex

Abstract

Developing chelators that strongly and selectively bind rare-earth elements (Sc, Y, La, and lanthanides) represents a longstanding fundamental challenge in inorganic chemistry. Solving these challenges is becoming more important because of increasing use of rare-earth elements in numerous technologies, ranging from paramagnets to luminescent materials. Within this context, we interrogated the complexation chemistry of the scandium(III) (Sc³⁺) trication with the hexadentate 1,4,7-triazacyclononane-1,4,7-triacetic acid (H₃NOTA) chelator. This H₃NOTA chelator is often regarded as an underperformer for complexing Sc³⁺. A common assumption is that metalation does not fully encapsulate Sc³⁺ within the NOTA³⁻ macrocycle, leaving Sc³⁺ on the periphery of the chelate and susceptible to demetalation. Herein, we developed a synthetic approach that contradicted those assumptions. We confirmed that our procedure forced Sc³⁺ into the NOTA³⁻ binding pocket by using single crystal X-ray diffraction to determine the Na[Sc(NOTA)(OOCCH₃)] structure. Density functional theory (DFT) and ⁴⁵Sc nuclear magnetic resonance (NMR) spectroscopy showed Sc³⁺ encapsulation was retained when the crystals were dissolved. Solution-phase and DFT studies revealed that [Sc(NOTA)(OOCCH₃)]¹⁻ could accommodate an additional H₂O capping ligand. Thermodynamic properties associated with the Sc-OOCCH₃ and Sc-H₂O capping ligand interactions demonstrated that these capping ligands occupied critical roles in stabilizing the [Sc(NOTA)] chelation complex.

---

Introduction

Recent advances in inorganic chemistry have supported development of a wide range of technologies based on rare-earth elements (Sc, Y, La, and lanthanides).^1–6 This application space spans from superconducting materials and paramagnets^7–9 to luminescent technologies^10–12 as well as radiotherapeutics and radiodiagnostics.^13–24 To advance the state-of-the-art in the aforementioned rare earth technologies, there is need to develop a better understanding of rare-earth element chemistry, in general. One area in need of attention is associated with rare-earth chelation. On the fundamental side, identifying chelators that bind strong and selective to Sc³⁺, Y³⁺, Lu³⁺, and the rest of the 4f-elements represents a longstanding challenge in inorganic chemistry. On the applied front, rare-earth chelation chemistry is important for purifying rare-earth elements in large-scale processing, remediating environmental contamination areas, and quantitatively analyzing for rare-earth elements in analytical samples.

Breakthroughs in the radiopharmaceutical field have ignited recent interest in complexing one particular rare-earth element, namely Sc³⁺. The ⁴⁴Sc and ⁴³Sc isotopes are promising agents for positron emission tomography (PET) imaging and ⁴⁷Sc is a potential therapeutic that emits low energy Auger electrons.^13,25–33 Pharmaceutically relevant scandium chelators need to bind Sc³⁺ rapidly, irreversibly, under mild conditions, and in high yield.³⁴ Most Sc³⁺ chelator design and metalation strategies are based on successful lanthanide chelation chemistry because Sc³⁺ shares many physical properties with lanthanide(III) cations. These characteristics include a highly stable +3 oxidation state, oxophilicity, and strong Lewis acidity. From this perspective, it is not surprising that many previous studies repurposed common lanthanide chelators for application with Sc³⁺. Some representative examples are provided in Chart 1; e.g. diethylenetriaminepentaacetic acid, H₅DTPA, 1,4,7,10-tetraazacyclododecane-1,4,7,10-tetraacetic acid, H₄DOTA, and AAZTA.^35–38

Chart 1 (a) Examples of some common chelates previously considered as Sc³⁺ binding agents; (b) an illustration of Sc³⁺ bound “in-cage”, with full chelation by the macrocycle, vs. “out-of-cage” with interactions between only the metal and pendant acetate arms.

Despite strong similarities between Sc³⁺ and the other rare-earth elements, scandium is a first-row transition metal, which leads to significant differences compared to the lanthanides. It has valence 3d-orbitals, not 4f- and 5d-orbitals. It is also much smaller than the 4f-elements and, consequently, more Lewis acidic. For example, the eight-coordinate ionic radius for Sc³⁺ (0.87 Å) is over 0.3 Å smaller than the largest lanthanide (La³⁺ at 1.18 Å) and 0.1 Å smaller than the smallest lanthanide (Lu³⁺ at 0.97 Å).³⁹ We propose that these distinctions should endow Sc³⁺ with chemical characteristics that can be tailored specifically to Sc³⁺ chelation and worry that treating Sc³⁺ as a small lanthanide is an oversimplification that limits innovation in Sc³⁺ chelator design. Within this context, we questioned why the small, hexadentate chelate H₃NOTA (1,4,7-triazacyclononane-1,4,7-triacetic acid) was not more routinely used in Sc³⁺ chemistry. Reviewing the literature suggested NOTA³⁻ often underperforms in Sc³⁺ binding, especially when compared to the larger DOTA⁴⁻ chelate.^35–38 This underperformance has been attributed to NOTA³⁻ binding Sc³⁺ in a labile fashion, which is correlated with NOTA³⁻ failing to encapsulated Sc³⁺ into its binding pocket where Sc³⁺ can interact with both the pendent acetate functional groups and the triaza-macrocyclic backbone (Chart 1). In the alternative and undesirable “out-of-cage” binding, Sc³⁺ only interacts with the acetate functional groups.^36,38,40,41 Only by modifying NOTA's acetate substituents (e.g. longer linkers or alternate metal binding functionality) has “in-cage”, stable and robust Sc³⁺ binding been consistently realized with the 1,4,7-triaazacylcononane backbone of NOTA.^34,42–44

Herein, we undertook a coordination chemistry study focused on better understanding Sc³⁺ complexation chemistry with NOTA³⁻. We synthesized an “in-cage” complex and used numerous characterization techniques to probe the interactions between Sc³⁺ and NOTA³⁻. For example, the solid-state structure of [Sc(NOTA)(OOCCH₃)]¹⁻ was determined, for the first time, using single crystal X-ray diffraction. Subsequently, we showed that “in-cage” Sc³⁺ binding by NOTA³⁻ was preserved in aqueous solution, using ⁴⁵Sc NMR spectroscopy. Interpretation of the solution phase data was guided by theoretical studies that, when combined with experimental results, provided insight into the behavior of the [Sc(NOTA)] complex in aqueous solutions. These studies also highlighted the critical role of capping ligand(s), defined here as ligands occupying vacant coordination sites after metalating a chelator. For example, we observed that the capping acetate and H₂O ligands were more than arbitrary ancillary binding agents that filled out Sc³⁺'s first coordination sphere. Instead, coordination of the capping ligands, specifically acetate, contributed substantially to the overall stability of the “in-cage” [Sc(NOTA)] chelation complex.

Results and discussion

Synthesis of Na[Sc(NOTA)(OOCCH₃)]

Sodium scandium(III) 1,4,7-triazanonane-1,4,7-triacetate κ²-acetate, Na[Sc(NOTA)(OOCCH₃)], was prepared from a reaction that fully encapsulated Sc³⁺ into 1,4,7-triazacyclononane-1,4,7-triacetate (NOTA³⁻). The resulting complex contained both Sc–O_NOTA and Sc–N_NOTA bonds (Scheme 1). Our experimental procedure was carried out under mild conditions. Metalation was achieved in aqueous media, under ambient atmosphere (air), using low temperatures, and in an aqueous solution that had been adjusted to a slightly acidic pH [acetate buffer, HOOCCH₃/NaOOCCH_3(aq); 0.1 M; pH of 5]. Under these conditions H₃NOTA·3 HCl (1 equivalent) was combined with anhydrous scandium(III) triflate, Sc(OTf)₃ (1 equivalent), and heated mildly (at 50 °C) for several hours to drive the reaction as far as possible toward complete complexation of Sc³⁺ by NOTA³⁻. This approach achieved high-yielding (∼70% isolated) and complete Sc³⁺ complexation with a 1 : 1 ligand to metal ratio.

Scheme 1 Synthesis of Na[Sc(NOTA)(OOCCH₃)] complex in acetate buffered solution.

After heating, a small amount of free Sc³⁺_(aq) (not complexed) was easily removed from the reaction mixture. This was achieved by precipitation of the uncomplexed Sc³⁺, which occurred rapidly when the pH was raised to 8 by addition of aqueous sodium hydroxide [NaOH_(aq), 1 M, K_sp for Sc(OH)₃ = 2.22 × 10⁻³¹].³⁹ The resultant fine, white precipitate was removed via filtration and the filtrate was collected. Subsequent removal of the volatiles in vacuo left a white residue that contained a mixture of leftover reagents (NaOOCCH₃, NaOH), byproducts (NaOTf and NaCl), and the Na[Sc(NOTA)(OOCCH₃)] target compound. Purification of the Sc³⁺ complex from these salts required extraction of the reaction residue with a solvent mixture of methanol and water (1 : 1) followed by crystallization at reduced temperature (10 °C, 1 week), with acetone added as an antisolvent. This crystallization method yielded plate-like, colorless single crystals.

Solid-state characterization with single crystal X-Ray diffraction

Analysis of the colorless plates by single crystal X-ray diffraction confirmed formation of Na[Sc(NOTA)(OOCCH₃)]. The crystallographic data were modeled in a C-centered, monoclinic crystal system in the C2/c space group. This model showed two [Sc(NOTA)(OOCCH₃)]¹⁻ anions within the asymmetric unit (Fig. 1) that had nearly identical structures (see ESI† for more discussion on this topic). They each contained a single Sc³⁺ cation bound by the hexadentate NOTA³⁻ chelate. There were three Sc–N_NOTA bonds associated with the cyclic aza-macrocycle and three Sc–O_NOTA bonds from the acetate arms. Although this binding motif is prototypical of complete metal binding by polyaza acetate chelators,^{38,42,45–47} it has not been previously observed for NOTA³⁻ binding of Sc³⁺. The three acetate arms and their corresponding nitrogen atoms (from the macrocycle) were eclipsed, rather than twisted.^48–51 Each [Sc(NOTA)] fragment was additionally capped by an acetate ligand (κ²-OOCCH₃¹⁻) bound through both acetate oxygen atoms. This ligand environment around Sc³⁺ provided a total coordination number of eight, a common aqueous coordination number for Sc³⁺. Constraints imposed by the alkyl linkages of the hexadentate chelator (NOTA³⁻) and the capping ligand (κ²-OOCCH₃¹⁻) prevented adaptation of an idealized 8-coordinate geometry for Sc³⁺ bound O and N atoms. Instead, the Sc³⁺ inner coordination sphere contained five O and three N atoms arranged somewhere between a bicapped trigonal antiprism (C_2v) and a dodecahedron (D_2d). These restrictions imparted low molecular symmetry. We identified only one pseudo-reflection plane, which contained the Sc³⁺ cation and the planar κ²-OOCCH₃¹⁻ ligand, and no rotational axes within each anionic fragment. Based on these findings, it seemed most appropriate to describe the [Sc(NOTA)(OOCCH₃)]¹⁻ anion fragment using the C_s point group.

Fig. 1 Single crystal X-ray structure of Na[Sc(NOTA)(OOCCH₃)]; Sc = orange, Na = aquamarine, N = blue, O = red, C = grey. Thermal ellipsoids are shown at 50% probability. Calculated H-atoms and disordered –CH₂– groups on the NOTA³⁻ backbone were omitted for clarity. Top – Asymmetric unit contents are shown, with two full [Sc(NOTA)(OOCCH₃)]¹⁻ units bridged by Na¹⁺ cations via carbonyl oxygens of the NOTA³⁻ ligand and OOCCH₃¹⁻. Bottom left – An overlay showing that the geometries for Sc1 are similar to Sc2 within the [Sc(NOTA)(OOCCH₃)]¹⁻ substructure. Bottom right – The geometry of Sc1 in [Sc(NOTA)(OOCCH₃)]¹⁻. Data is from solution 1.

Looking beyond the first few coordination spheres revealed that the [Sc(NOTA)(OOCCH₃)]¹⁻ anions were arranged in two-dimensional sheets that extend parallel to the bc-plane (Fig. 2). Within each sheet, [Sc(NOTA)(OOCCH₃)]¹⁻ anions were linked through a multifaceted Sc–O⋯Na⋯O=C–O–Sc network that sandwiched the Na¹⁺ cations between [Sc(NOTA)(OOCCH₃)]¹⁻ anions. This arrangement caused the boundary of each sheet along the bc-plane to consist of hydrophobic CH₂ groups from the macrocyclic NOTA-backbone (polyaza ring). Within the sandwich, the Na¹⁺ cations were bridged by NOTA³⁻ oxygen atoms that were not directly bound to Sc³⁺ (terminal carbonyl-like oxygens), as well as linkages involving both of the acetate oxygen atoms bound to Sc³⁺ in each anion fragment. Although the sum-total Na¹⁺ coordination number (for each Na¹⁺ atom) was six, the geometries around each Na¹⁺ cation were not regular nor did they fall into idealized six-coordinate geometries (i.e. not perfectly octahedral or trigonal anitprismatic).

Fig. 2 Extended structure of Na[Sc(NOTA)(OOCCH₃)] with sheets of [Sc(NOTA)(OOCCH₃)]_x¹⁻ linked through Na¹⁺ cations; Sc = orange, Na = aqua, N = blue, O = red, C = grey. The unit cell is shown and first-coordination sphere for Sc³⁺ and Na¹⁺ have been represented as orange and aqua polyhedra, respectively. Calculated H-atoms were omitted for clarity.

Many molecules of water (H₂O) filled the void spaces between consecutive Na[Sc(NOTA)(OOCCH₃)] layers. Exact numbers and positions of H₂O molecules were difficult to determine. Hence, we investigated two crystallographic solutions. In the first model, solution 1, the electron density within the void space was treated using the “squeeze” function in Olex2 refinement software (with Platon, running ShelX refinement package).^52–54 In the second model, solution 2, the residual void space electron density was modeled with discrete H₂O molecules. Interatomic distances and angles for Na[Sc(NOTA)(OOCCH₃)] from the two different models were essentially indistinguishable and largely fell within the uncertainty of the solution statistics. Solution 1 gave a slightly lower R-value compared to solution 2, R₁ = 0.0427 vs. 0.0608. Based on this metric, we only included bond distances and angles from solution 1 herein. Related to this topic is a subtle detail associated with the Na[Sc(NOTA)(OOCCH₃)] structure. Three Na¹⁺ cations co-crystalized alongside two Sc(NOTA)(OOCCH₃)¹⁻ anions. The system charge balances with one of the interstitial “water” molecules likely being a hydroxide (OH¹⁻). Given the quality of the data, no attempt was made to model the position of the OH¹⁻ or the accompanying and likely proton disorder.

Investigation of solution structure with ⁴⁵Sc NMR

A series of scandium-45 (⁴⁵Sc) nuclear magnetic resonance (NMR) measurements were made to determine if the solid-state [Sc(NOTA)(OOCCH₃)]¹⁻ structure described above persisted in solution. These studies exploited the unique accessibility of ⁴⁵Sc NMR spectroscopy. First, the nuclear properties of naturally occurring ⁴⁵Sc made it well suited for NMR spectroscopy: 100% abundant, nuclear spin of 7/2, quadrupole moment of −0.22b, and a magnetogyric ratio (γ_45-Sc) of 6.5081 × 10⁷ rad T⁻¹ s⁻¹ (0.3γ_1-H = γ_45-Sc).^55,56 Second, the ⁴⁵Sc NMR experiment could be conducted with common instrumentation because the ⁴⁵Sc NMR excitation frequency is very close to that of the more commonly probed ¹³C nuclide (24.97% vs. 25.14% frequency ratios, respectively). This property makes standard broadband probes that have low frequency channels for ¹³C data acquisition compatible with ⁴⁵Sc NMR experiments.

The ⁴⁵Sc NMR spectrum from single crystals of Na[Sc(NOTA)(OOCCH₃)] dissolved in deuterated water (D₂O) showed an intense resonance at 88.8 ppm (top, Fig. 3). This chemical shift was comparable to those reported previously from other examples of Sc³⁺ cations completely encapsulated by related chelating agents, e.g. Sc(DOTA) at 90 ppm, Sc(DTPA) at 79 ppm, Sc(DO3AP) at 100 ppm, and Sc(AAZTA) at 80 ppm.^38,41,42 This good agreement suggested that upon dissolution, the Sc³⁺ cation remained encapsulated by the NOTA³⁻ chelate, like the “in-cage” solid-state structure described above and depicted in Scheme 1. Consistent with this observation was the mismatch between our ⁴⁵Sc chemical shifts and those reported for the “out-of-cage” complexation of Sc³⁺ by polyaza acetic acid ligands. Of specific relevance were the observations made by Huclier-Markai and co-workers.³⁶ Those researchers observed [Sc(NOTA)] as an intermediate to “out-of-cage” complex, which had a diagnostic resonance shifted far up-field, near 20 ppm. That shift was similar, within 10 ppm, to Sc(oxalate)₄⁵⁻ and the related “out-of-cage” Sc(DO3AP).^38,41 Hence, the complexation method used by Huclier-Markai and co-workers led to “out-of-cage” Sc³⁺ binding, while the method used in the present work leads to “in-cage” Sc³⁺ binding.

Fig. 3 ⁴⁵Sc NMR spectra from single crystals of Na[Sc(NOTA)(OOCCH₃)] dissolved in D₂O (top), D₂O with H₂¹⁷O (middle), and DMSO-d₆ (bottom). Spectra were collected at 20 °C with an operating frequency of 97 MHz. For additional discussion of the DMSO-d₆ spectrum, see ESI.†

We attributed the resonance at 88.8 ppm to the [Sc(NOTA)(OOCCH₃)]¹⁻ compound. However, the small downfield shoulder at 99.8 ppm clearly indicated that there was an additional ⁴⁵Sc complex present in solution, one also encapsulated “in-cage” by the NOTA³⁻ chelate (as indicated by the chemical shift). These signals were successfully deconvoluted using Gaussian fitting (in MestreNova V.14.1), which showed the 88.8 to 99.8 ppm peak intensity ratio was about 1.5 to 1 at 20 °C. The ⁴⁵Sc spectrum also showed dependence on temperature (see Fig. S5–7† and discussion below). Increasing the temperature from 20 to 80 °C shifted the peak positions determined by Gaussian deconvolution slightly (Δ < 1 ppm). More significantly, increasing temperature also caused the peak at 99.8 ppm to decrease in intensity relative to the peak at 88.8 ppm. This behavior suggested the Sc³⁺ species associated with these two features were related by a dynamic exchange process in solution. We speculated that this exchange subtly altered the first coordination sphere of [Sc(NOTA)(OOCCH₃)]¹⁻ (responsible for the peak at 88.8 ppm). Given that both resonances were consistent with Sc³⁺ encapsulation by NOTA³⁻, this first coordination sphere change was reasonably associated with the capping ligand. Likely processes responsible for the two peaks were (1) slipping the acetate from bidentate κ²-OOCCH₃¹⁻ to monodentate κ¹-OOCCH₃¹⁻, (2) hydration to add Sc-(H₂O) bonding interactions (Fig. 4a), or (3) substitution of the capping acetate ligand with water (Fig. 4b). The following experiments and computational studies (vide infra) suggested that contributions from scenarios 1 and 2 were most likely responsible for the observed exchange behavior (Fig. 4a), while scenario 3 was significantly unfavorable (Fig. 4b).

Fig. 4 Two possible interpretations of the equilibrium process observed by ⁴⁵Sc NMR spectroscopy of Na[Sc(NOTA)(OOCCH₃)] crystals dissolved in D₂O. The aqueous environment gives rise to two species of varying concentrations in solution; (a) coordination of water without loss of the acetate ligand vs. (b) exchange of the capping acetate ligand for water. Note, our analyses suggest that process (a) is the primary exchange process accessible in aqueous solutions.

To more precisely characterize the origin of the two ⁴⁵Sc resonances, two additional experiments were carried out. First, the NMR solvent was changed from D₂O to DMSO-d₆. The ⁴⁵Sc NMR spectrum obtained from single crystals of Na[Sc(NOTA)(OOCCH₃)] dissolved in DMSO-d₆ showed a single broad feature at 99.47 ppm. We attributed this feature to the anhydrous [Sc(NOTA)(OOCCH₃)]¹⁻ complex (see ESI†). This observation reinforced our conclusion that dissolving the crystalline material generated a solution that contained a single ⁴⁵Sc species of the general formal [Sc(NOTA)(OOCCH₃)]¹⁻; however, we acknowledged that a DMSO adduct was also possible. The second ⁴⁵Sc NMR experiment resembled the original measurement shown in Fig. 3 in that Na[Sc(NOTA)(OOCCH₃)] crystals were dissolved in D₂O. However, this solution was spiked with ¹⁷O isotopically enriched H₂¹⁷O (10 μL at 90% enrichment). The quadrupolar ¹⁷O nucleus (spin 5/2) provided an opportunity to identify if Sc-(H₂O) bonds formed in solution. Close contact between the quadrupolar ¹⁷O and ⁴⁵Sc nuclei through Sc-(H₂O) interactions should enhance spin–spin relaxation, shorten the nuclear relaxation times (T₂) for interacting ¹⁷O and ⁴⁵Sc nuclei during the NMR experiment, and dramatically broaden the observed ⁴⁵Sc NMR resonance.^55,56 As demonstrated in Fig. 3, this prediction aligned well with the changes observed spectroscopically upon inclusion of H₂¹⁷O. The ⁴⁵Sc NMR spectrum in D₂O spiked with H₂¹⁷O showed a single resonance at 87.6 ppm. Meanwhile, the downfield shoulder at 99.8 ppm vanished into the baseline, likely owing to formation of Sc–¹⁷O_H2O bonds. Reasonable attempts to deconvolute the spectrum with two Gaussians failed, indicating that the remaining signal intensity at 87.6 ppm originated from a single ⁴⁵Sc species.

Based on the above experimental results (and the calculations described below), we concluded that the ⁴⁵Sc NMR spectrum showed a combination of two species that coexisted in H₂O solutions on the ⁴⁵Sc NMR time scale. Two possible scenarios for this exchange process were presented in Fig. 4. The major contributor to the spectrum at room temperature was the anhydrous [Sc(NOTA)(OOCCH₃)]¹⁻ that exhibited the ⁴⁵Sc resonance near 88 ppm. There was also a minor species that exhibited a ⁴⁵Sc resonance near 100 ppm, namely the hydrated [Sc(NOTA)(OOCCH₃)(H₂O)]¹⁻ (Fig. 4a). The data showed the Sc³⁺ cation in both [Sc(NOTA)(OOCCH₃)(H₂O)]¹⁻ and [Sc(NOTA)(OOCCH₃)]¹⁻ resided in the binding pocket of NOTA³⁻, akin to the solid-state structure (vide supra), as evident from the dramatic downfield shifts of both species relative to the Sc(NO₃)₃(H₂O)_x standard (0 ppm). The only observable difference between the two coordination complexes was the presence of an H₂O ligand in [Sc(NOTA)(OOCCH₃)(H₂O)]¹⁻ (Fig. 4a). The experimental data did not provide insight into the Sc-OOCCH₃ binding mode and ¹H and ¹³C NMR spectra did not generate additional insight. Hence, we have refrained from speculating on the degree of mono- vs. bidentate character associated with the Sc-OOCCH₃ interaction (in either species) and simply acknowledge that both are possible in solution.

Experimental approximation of H₂O coordination thermodynamics

Characterizing the solution-phase exchange process between [Sc(NOTA)(OOCCH₃)]¹⁻ (88.8 ppm, linewidth = 450 Hz) and [Sc(NOTA)(OOCCH₃)(H₂O)]¹⁻ (99.8 ppm, linewidth = 880 Hz) by variable temperature ⁴⁵Sc NMR spectroscopy (in D₂O) provided an estimate of the thermodynamic properties associated with the H₂O coordination process (note, chemical shifts dependence on temperature were reversible). This was achieved after assuming that these two ⁴⁵Sc species had similar quadrupolar relaxations, which seemed reasonable owing to similar chemical shifts, linewidths, and coordination environments. Hence, a Van't Hoff approximation for the thermodynamic parameters associated with the H₂O coordination process (Fig. 4a) was generated from the following equations:

Here, K_eq was the equilibrium constant, R was the ideal gas constant (1.987 × 10⁻³ kcal mol⁻¹ K⁻¹), T was temperature (Kelvin), and ΔH and ΔS were the enthalpy and entropy changes for the exchange process, respectively. Plotting the dependence of K_eq on showed a linear relationship over the temperature region probed (Fig. 5). A linear fit of the data gave a ΔH = −1.7 ± 0.2 kcal mol⁻¹ as described in eqn (1) and ΔS = −0.0064 ± 0.0005 kcal mol⁻¹ K⁻¹. As expected, these data suggested ΔS favored formation of reactants, [Sc(NOTA)(OOCCH₃)]¹⁻ and H₂O. This value was offset slightly by enthalpic stability that favored the [Sc(NOTA)(OOCCH₃)(H₂O)]¹⁻ product, which we attributed to formation of the Sc-H₂O bond. The entire process was close to thermoneutral at room temperature, with a calculated ΔG of +0.2 ± 0.2 kcal mol⁻¹ at T = 293.15 K. Overall, these values match the qualitative observation of more [Sc(NOTA)(OOCCH₃)]¹⁻ in solution compared to [Sc(NOTA)(OOCCH₃)(H₂O)]¹⁻ (roughly 1.5 : 1).

Fig. 5 Top: A Van't Hoff analysis of the equilibrium constant (K_eq) that relates conversion of [Sc(NOTA)(OOCCH₃)]¹⁻ to [Sc(NOTA)(OOCCH₃)(H₂O)]¹⁻. Metrics from this plot were used to solve eqn (1) and (2). Bottom: Normalized distribution of the minor [Sc(NOTA)(OOCCH₃)(H₂O)]¹⁻ and major [Sc(NOTA)(OOCCH₃)]¹⁻ peaks observed by ⁴⁵Sc NMR spectroscopy from Na[Sc(NOTA)(OOCCH₃)] crystals dissolved in D₂O as a function of temperature. Peak deconvolution was performed using the “peak fitting” tool in MestreNova software V 14.1. The total peak area, [Sc(NOTA)(OOCCH₃)]¹⁻ + [Sc(NOTA)(OOCCH₃)(H₂O)¹⁻], was normalized to 1.0 (100%), see ESI† for more details.

Calculated [Sc(NOTA)(OOCCH₃)]¹⁻ and [Sc(NOTA)(OOCCH₃)(H₂O)]¹⁻ coordination thermodynamics

To guide the interpretation of [Sc(NOTA)(OOCCH₃)]¹⁻ solution phase behavior, particularly in terms of interpreting the ⁴⁵Sc NMR results discussed above, density functional theory (DFT) calculations were performed. Emphasis was placed on assessing computationally determined vs. experimentally derived Gibbs free energy (ΔG) parameters associated with conversion of Na[Sc(NOTA)(OOCCH₃)] and H₂O to Na[Sc(NOTA)(OOCCH₃)(H₂O)]. Note, these calculations included the [Sc(NOTA)(OOCCH₃)]¹⁻ and the Na¹⁺ counter cation. Calculations were also made that accounted for potential coordination with DMSO. The calculated structure for Na[Sc(NOTA)(OOCCH₃)] agreed well with the experimental data (Table 1 and in the ESI†). For instance, the calculations showed stable “in-cage” Sc³⁺ binding by NOTA³⁻ with bidentate binding of OOCCH₃¹⁻. The calculated Sc–ligand distances were similar to those observed experimentally. One slight deviation was associated with the Sc–N_NOTA bond distances of the NOTA³⁻ macrocyclic backbone. Here, theoretically predicted Sc–N_NOTA distances were slightly longer than the single crystal X-ray diffraction data by about +1.4% (on average). Similar differences in M–N distances between DFT optimized structures and X-ray diffraction data have been observed previously, specifically with M(DOTP)⁵⁻ (M = Ac, Am, Cm, La)⁵⁷ and M(HOPO)¹⁻ (M = Am, Cm, Cf) complexes;⁵⁸ DOTP is 1,4,7,10-tetraazacyclododecane-1,4,7,10-tetra(methylene)phosphonic acid and HOPO is an octadentate hydroxypyridinone ligand. Usually, such discrepancies emerge when calculations focus on a single molecule and omit crystal packing effects from the extended solid, e.g. the presence of multiple counter cations surrounding each anion fragment (Fig. 2, see ESI†). It is also worth noting that the differences are relative to the static structure observed in the solid state, while the thermodynamic comparisons were made versus the solvated complex.

Table 1 Top: Selected bond distances (Å) and angles (°) from the Na[Sc(NOTA)(OOCCH₃)] crystal structure. Analogous bond distances or angles between the two [Sc(NOTA)(OOCCH₃)]¹⁻ fragments in the asymmetric unit are listed in the same row. Atomic numbering follows Fig. 1. Bottom: Comparative bond distances for Na[Sc(NOTA)(OOCCH₃)] from experimental X-ray crystallographic structure determination vs. DFT calculated structure

Experimental	Bond	Fragment 1 distance/angle (°/Å)	Bond	Fragment 2 distance/angle (°/Å)
Sc-NOTA	Sc1–N1	2.384(2)	Sc2–N4	2.376(2)
	Sc1–N2	2.374(2)	Sc2–N5	2.368(2)
	Sc1–N3	2.393(2)	Sc2–N6	2.436(2)
	Sc1–O1	2.116(1)	Sc2–O9	2.131(1)
	Sc1–O3	2.165(1)	Sc2–O11	2.158(1)
	Sc1–O5	2.155(1)	Sc2–O13	2.158(1)
Sc-OOCCH3	Sc1–O7	2.377(1)	Sc2–O15	2.328(1)
	Sc1–O8	2.213(1)	Sc2–O16	2.180(1)
κ2-(OOCCH3)	O7–C13–O8	118.89(17)	O15–C27–O16	118.75(17)
	C13–O7	1.258(2)	C27–O15	1.262(2)
	C13–O8	1.283(2)	C27–O16	1.278(2)

---

Theory	Bond	Average distance
		Experiment (avg., Å)	Theory (Å)
a Error reported as the standard deviation of the mean at 1σ.
Sc-NOTA	Sc–ONOTA	2.147 ± 0.019^a	2.136
	Sc–NNOTA	2.389 ± 0.025^a	2.423
Sc-OOCCH3	Sc–OOOCCH3 (short)	2.197 ± 0.023^a	2.183
	Sc–OOOCCH3 (long)	2.353 ± 0.035^a	2.357

---

With a reasonable and validated calculated structure for [Sc(NOTA)(OOCCH₃)]¹⁻ in hand, two categories of reactions involving water with Na[Sc(NOTA)(OOCCH₃)] were explored. We initially probed association of water to the complex to generate Na[Sc(NOTA)(OOCCH₃)(H₂O)] (Fig. 4a and S15b†) and then interrogated acetate substitution by water to form [Sc(NOTA)(H₂O)] or [Sc(NOTA)(H₂O)₂] (Fig. 4b and S15b†). Acetate substitution by water was calculated to be highly unfavorable based on the large and positive ΔG values calculated for one (+18.17 kcal mol⁻¹) and for two (+24.94 kcal mol⁻¹) water molecules (see ESI†). In contrast, the calculations showed that water association was thermodynamically favorable with a small, slightly negative ΔG of −0.94 kcal mol⁻¹. A similar value of −0.74 kcal mol⁻¹ was obtained if the calculations took into account a 1^st-shell-hydrated Na¹⁺ cation that stabilized [Sc(NOTA)(OOCCH₃)]¹⁻ complex, see ESI.† Both of these values were in exceptional agreement with the experimentally derived +0.2 ± 0.2 kcal mol⁻¹ at T = 293.15 K, especially when uncertainty between the experimental and theoretically derived values were considered. These metrics suggested that this reaction was close to thermoneutral at room temperature. The computed ΔH (−4.58 kcal mol⁻¹) and ΔS (−0.01 kcal mol⁻¹ K⁻¹) parameters were also in agreement with the values estimated from experiment (−1.7 ± 0.2 and −0.0064 ± 0.0005 kcal mol⁻¹ K⁻¹, respectively), suggesting that the reaction was enthalpically favored. The small (close-to-zero) Gibbs free energy suggested that the [Sc(NOTA)(OOCCH₃)]¹⁻ and [Sc(NOTA)(OOCCH₃)(H₂O)]¹⁻ structures can coexist in solution. These computational results were consistent with the mixture of two species observed experimentally. In addition to the water addition/water-acetate substitution reactions, we also considered DMSO–acetate substitution and DMSO addition reactions with Na[Sc(NOTA)(OOCCH₃)]. Our calculations show that they are less thermodynamically favorable (Fig. S17 and S18†), with significantly more positive ΔG values in the range of +7.81–17.46 kcal mol⁻¹ (eqn (S8)–(S10) in the ESI†) than the reaction of water association with Na[Sc(NOTA)(OOCCH₃)].

In the calculated water addition reaction, acetate slips from bidentate (κ²-OOCCH₃¹⁻) to monodentate (κ¹-OOCCH₃¹⁻). This modification accommodated water occupation without increasing the Sc³⁺ coordination number beyond eight (Fig. 4, 6 and S12, S15a†). The calculated Sc–O_H2O (O9) distance was 2.374 Å and compared reasonably well with other analogous Sc–O_H2O experimental values.⁴² It seemed likely that this calculated structure was further stabilized by hydrogen bonding between the coordinated H₂O ligand and the κ¹-OOCCH₃¹⁻ (Fig. 6).

Fig. 6 Association of a water molecule to Na[Sc(NOTA)(OOCCH₃)]. Initially (a) the acetate is bound to Sc in a bidentate fashion (κ₂-OOCCH₃¹⁻); upon water coordination (b) the acetate slips to a monodentate binding mode (κ₁-OOCCH₃¹⁻).

It was important to evaluate the challenge associated with computationally determining the absolute value of Gibbs free energy for the hydration reactions considered. This exercise put our calculated ΔG of −0.94 kcal mol⁻¹ (−0.74 kcal mol⁻¹ taking into account 1^st-shell-hydrated Na¹⁺ stabilizing [Sc(NOTA)(OOCCH₃)]¹⁻ complex) for the hydration reaction into perspective. To approximate hydrogen bonding interactions that occur in solution, we used clusters of water molecules based on previously determined global minima from the gas-phase calculations.⁵⁹ In this way, effects stemming from explicit waters of solvation were included, granted to a limited extent. The approach did not model dynamic hydrogen bonding interactions present in bulk water nor did it capture contributions from the outer sphere sodium cations. It did, however, provide a more reasonable alternative to ignoring bulk effects altogether (using a single water molecule on the reactant side), which is well documented to severely overestimate ΔG.^57,58,60,61 It is worth noting that the calculated ΔG values fluctuated with sizes and shapes of water clusters considered in the hydration process for water association (see eqn (S4)–(7) in the ESI†). These results illustrate the challenge in determining the absolute values of ΔG for hydration reactions with high accuracy.

The DFT calculations provided guidance for interpreting what reactions were responsible for the exchange process observed experimentally (Fig. 3 and 4). These results attributed the two species observed by ⁴⁵Sc NMR spectroscopy to a mixture of [Sc(NOTA)(OOCCH₃)]¹⁻ and [Sc(NOTA)(OOCCH₃)(H₂O)]¹⁻. The calculated thermodynamic parameters were on the same order of magnitude (within a few kcal mol⁻¹) as the experimentally determined values. This interpretation was also self-consistent with the H₂¹⁷O labeled ⁴⁵Sc NMR experiment and the ⁴⁵Sc NMR experiment carried out in DMSO-d₆. Alternative mixtures, e.g. [Sc(NOTA)(OOCCH₃)]¹⁻ and [Sc(NOTA)(H₂O)_n] (n = 1, 2), and [Sc(NOTA)(OOCCH₃)]¹⁻/[Sc(NOTA)(DMSO)_n] (n = 1, 2), were discarded based on the massively positive calculated ΔG values.

Outlook

We have established an alternate synthetic approach that fully encapsulated Sc³⁺ into the NOTA³⁻ binding pocket, providing the Na[Sc(NOTA)(OOCCH₃)] complex in 69% isolated yield. Conditions for Sc³⁺ complexation by NOTA³⁻ were identified that could be carried out under relatively mild conditions. The procedure successfully proceeded in aqueous media, under ambient atmosphere (air), in acetate buffer solutions [HOOCCH₃/NaOOCCH_3(aq); 0.1 M] that were slightly acidic (pH = 5), and at relatively low temperatures (50 °C). Characterization in the solid-state showed that Sc³⁺ was bound “in-cage” by NOTA³⁻. Achieving this “in-cage” binding required a gentle synthetic hand. For example, previous studies showed the ease at which “out-of-cage” binding prevailed. Only by heating Sc³⁺ with H₃NOTA, for longer reaction times, and in an acetic acid/acetate buffer solution did we generate an “in-cage” [Sc(NOTA)] complex.^36,38 Our [Sc(NOTA)] species was capped with a bidentate κ²-OOCCH₃¹⁻ ligand, which provided stability (see below) and gave Sc³⁺ a coordination number of eight. Reaching steric saturation in aqueous media with only eight ligands also represented a distinction for Sc³⁺ in comparison to the larger lanthanide(III) cations, whose aqueous coordination numbers often range from 9–11.

Solution phase ⁴⁵Sc NMR studies demonstrated that the “in-cage” binding of Sc³⁺ by NOTA³⁻ was preserved in aqueous solution. Computational results were consistent with this interpretation and suggested that in solution a mixture of [Sc(NOTA)(OOCCH₃)]¹⁻ and [Sc(NOTA)(OOCCH₃)(H₂O)]¹⁻ coexisted. These calculations were supported by ⁴⁵Sc NMR measurements made on Na[Sc(NOTA)(OOCCH₃)] dissolved in water with H₂¹⁷O. The presence of H₂¹⁷O dramatically enhanced relaxation of the ⁴⁵Sc resonance from [Sc(NOTA)(OOCCH₃)(H₂¹⁷O)]¹⁻, as a result of the close ⁴⁵Sc–¹⁷O_H2O interaction. We interpret our results as suggesting that [Sc(NOTA)(OOCCH₃)]¹⁻ was the dominant species in solution and that the [Sc(NOTA)(OOCCH₃)]¹⁻ to [Sc(NOTA)(OOCCH₃)(H₂O)]¹⁻ ratio was ∼1.5 to 1.

While characterizing the dynamic exchange behavior for conversion of [Sc(NOTA)(OOCCH₃)]¹⁻ to [Sc(NOTA)(OOCCH₃)(H₂O)]¹⁻, we identified the capping ligand(s) (OOCCH₃¹⁻ and H₂O) substantially influenced thermodynamic stability of [Sc(NOTA)(L)_x] (L = capping ligands). Both [Sc(NOTA)(OOCCH₃)]¹⁻ and [Sc(NOTA)(OOCCH₃)(H₂O)]¹⁻ were calculated to be markedly more stable (by a ΔG of ∼20 kcal mol⁻¹) than the simple hydrate, [Sc(NOTA)(H₂O)_n] (n = 1 or 2). These results highlight the importance of the capping ligands (L) in a coordination complex with the general formula M(chelator)(L)_x; M being a metal like Sc³⁺ and chelator representing a binding agent like NOTA³⁻. Our results suggest that the capping ligand is more than a simple ancillary ligand and can be used to influence M(chelator) stability.

Taken as a whole, the synthetic and computational assessment of “in-cage” Sc³⁺ binding by NOTA³⁻ provided insight into rare-earth complexation chemistry, particularly for Sc³⁺. The results highlighted how subtle variation for a given complexation method led to profoundly distinct outcomes, e.g. “in cage” and inert complexation vs. “out-of-cage” and labile binding. The data also piqued our interest in better defining the capping ligand's role in stabilizing rare-earth chelation complexes. Hence, current efforts are underway to further define Sc³⁺ speciation in aqueous media that contain a wider range of complexing agents and capping ligands. It is our aspiration that future studies carried out by us and others in the field will advance understanding of rare-earth complexation chemistry and contribute to developing the next generation of selective and strong binding rare-earth chelates. Success could impact rare-earth technologies broadly, and aid in solving 4f-element chelation changes.

Methods

General considerations

All reactions and manipulations were carried out in air at ambient pressures and temperatures. Anhydrous scandium(III) triflate [Sc(OTf)₃, Sigma-Aldrich, 99%], scandium(III) nitrate hexahydrate [Sc(NO₃)₃·6 H₂O, Sigma-Aldrich], 1,4,7-triazacyclononane-1,4,7-triacetic acid (H₃NOTA·3 HCl, Macrocyclics), sodium acetate (NaOOCCH₃, Fischer), and sodium hydroxide (NaOH, pellets, Fischer) were used as received without further purification. Deuterated water (D₂O, Sigma-Aldrich) and deuterated dimethylsulfoxide (DMSO-d₆, 99.9%, anhydrous, Sigma-Aldrich), used as NMR spectroscopy solvents, were also purchased and used without further purification. Oxygen-17 enriched water (H₂¹⁷O; estimated 90% ¹⁷O) was procured from legacy chemical inventories at LANL, and had poorly identified origins. Isotopic enrichment was determined experimentally prior to use. All pH measurements were made using Whatman pH paper (0–14 pH).

Instrumentation

NMR experiments were performed on a Bruker AVANCE™ 400 MHz solution spectrometer equipped with a 5 mm broadband tunable probe for X nuclei. Infrared spectra were collected on a commercial ThermoFisher Scientific Nicolet iS5 ATR-FTIR. High resolution mass spectrometry was performed on an Agilent 6210 LC-TOF (ESI, APCI, APPI).

Single crystal X-Ray diffraction

Data for Na[Sc(NOTA)(OOCCH₃)] were collected on a Bruker D8 Quest diffractometer configured with a CPAD Photon II™ area detector and MoKα (λ = 0.71073 Å) IμS 3.0 micro source™. The crystal was cooled to 100 K employing an Oxford Cryostream 800™ liquid nitrogen cryostat. A hemisphere of data was collected using omega scans and 1.00° frame widths. Data collection and initial indexing and cell refinement were handled using APEX 3 software.⁶² Frame integration, including Lorentz-polarization corrections, and final cell parameter calculations were carried out using SAINT+ software.⁶³ The data were corrected for absorption using redundant reflections and the SADABS program.⁶⁴ The structure was solved using Intrinsic Phasing and difference Fourier techniques. All hydrogen atom positions were idealized, and rode on the atom to which they were attached. The final refinement included anisotropic temperature factors on all non-hydrogen atoms. Structure solution, refinement, graphics, and creation of publication materials were performed using SHELXTL.⁵⁴ Hydrogen atom positions were idealized, and all non-hydrogen atoms were refined anisotropically. Cell indexing, data collection, integration, structure solution, and refinement were performed using Bruker and SHELXTL software. Additional details are included in the ESI.†

CCDC deposit number: 2072300 and 2076191.†

Synthesis of sodium scandium(III) 1,4,7-triazacyclononane-1,4,7-triacetate κ₂-acetate, Na[Sc(NOTA)(OOCCH₃)]

A scintillation vial (20 mL) was loaded with Sc(OTf)₃ (49 mg, 0.1 mmol), a Teflon stir bar, and an aqueous solution of NaOOCCH₃ (8 mL; 0.1 M in NaOOCCH₃; pH = 5; adjusted by addition of acetic acid). Note, all pH measurements were determined using pH paper. Separately, H₃NOTA·3 HCl (41 mg, 0.1 mmol) was dissolved in an aqueous solution of NaOOCCH₃ (3 mL; 0.1 M NaOOCCH₃; pH = 4). This colorless aqueous solution of H_3−nNOTAⁿ⁻ was added dropwise to the stirring Sc solution described above. Upon addition the reaction solution pH dropped and was subsequently adjusted to a pH of 5 by dropwise addition of NaOH (2 M). Because the homogeneity of this solution is quite sensitive to the NaOH_(aq) addition [i.e. rapid addition can result in Sc(OH)_xⁿ⁻ precipitate] the base was added slowly. The reaction solution was then heated to 50 °C and stirred overnight (∼12 h). The solution was cooled and aqueous NaOH (2 M) was added to adjust the pH to 8. This caused unreacted Sc³⁺ to precipitate as scandium(III) hydroxide [Sc(OH)_x(H₂O)_9−x^3−x] and slight opacity of the initially transparent solution. The precipitate represented a small percentage of dissolved Sc³⁺ and was easily removed by passing the solution through a filter stick packed with filter paper and Celite. Collecting the transparent and colorless filtrate into a clean scintillation vial and removing the solvent by rotary evaporation yielded a white residue. The residue was extracted with a mixture of H₂O and MeOH (3 mL, 1 : 1, v : v). The extracts were filtered through Celite into a glass V-vial (10 mL) and the filtrate was carefully layered with acetone (1 mL). This solution was then stored at 10 °C for 1 week. From the crystallization solution, colorless plate-like crystals formed (18 mg, 41% yield) that were suitable for characterization by single crystal X-ray diffraction. Addition of more acetone (1.5 mL) to the equilibrated solution produced a second crop of crystals (12 mg, 28%). After drying the crystals, the total crystalline yield for Na[Sc(NOTA)(OOCCH₃)] was 69%.

NMR: ¹H (D₂O, 20 °C, 400.13 MHz): δ 3.69 (br, s), 3.54 (br, s), 2.95 (br, m), 1.78 (s, OOCCH₃). ¹H (DMSO-d₆, 20 °C, 400.13 MHz): δ 3.50 (br, s), 3.39 (s), 3.12 (br, m), 2.91 (br, m), 1.83 (s). ⁴⁵Sc (D₂O, 20 °C, 97.198 MHz): δ 99.8 ppm (fitted peak width, 840 Hz), 88.8 ppm (fitted peak width, 440 Hz). ⁴⁵Sc (DMSO-d₆, 20 °C, 97.198 MHz): δ 99.47 ppm (fitted peak width, 3744 Hz).

IR (cm ⁻¹ ): 3370 (br, m); 3130 (s); 3040 (s), 2860 (br, shoulder); 1630 (br, shoulder), 1570 (s); 1400 (s). See ESI† for spectra and additional details.

HRMS [Sc(NOTA)]: (ToF, positive ion mode) ScN₃O₆C₁₂H₁₉ (M + H)⁺ 346.0827, observed 346.0841; ScNaN₃O₆C₁₂H₁₈ (M + Na)⁺ 368.0658, observed 368.0647.

Computational details

All complexes were optimized without any constraints using self-consistent reaction field approach based on the integral equation formalism of the polarized continuum model (PCM)^65,66 as implemented in Gaussian 16 software package (Version B.01).⁶⁷ Water and dimethylsulfoxide (DMSO) were used as solvents in the respective reactions. Harmonic frequency calculations were performed to confirm that the optimized structures were stationary points on the potential energy surface. The initial geometry of the Na[Sc(NOTA)(OOCCH₃)] complex was taken from the experimental XRD data. Energies and geometries were reported from results using the PBE0 ⁶⁸ hybrid DFT functional and def2-TZVP⁶⁹ triple zeta valence basis set. Dispersion corrections (D3)⁷⁰ were applied to account for possible intramolecular noncovalent interactions that can be important for the correct description of solvent-complex interaction. As reported previously for various transition-metal-based complexes,⁷¹ PBE0-D3 was found to be the best functional in the complete benchmark set relative to estimated CCSD(T)/CBS reference data, with a mean absolute deviation from the reference values of 1.1 kcal mol⁻¹. Wavefunctions of the studied species were found to be stable, indicating that the calculations converged to the ground electronic state. Water clusters containing 16–19 molecules were used to more accurately calculate the energy of water binding/exchange between bulk water and complexes. Using only a single water molecule would neglect the energy from hydrogen bonding in water clusters and lead to less accurate ΔG values. The geometries of these water clusters were taken from the previously established global minima gas-phase structures,⁵⁹ and were fully optimized in solvent medium using the PCM model in this study.

Conflicts of interest

The authors declare no competing financial interests.

Acknowledgements

We acknowledge the US Department of Energy, Office of Science, Office of Basic Energy Sciences, Heavy Element Chemistry program (2020LANLE372) and LANL's LDRD-DR project (20180005DR) for support. Los Alamos National Laboratory is operated by Triad National Security, LLC, for the National Nuclear Security Administration of US Department of Energy (contract no. 89233218CNA000001). We are grateful for postdoctoral support provided (in-part) by the Glenn T. Seaborg Institute (KEA), by the LANL Director's Fellowship Program (SMG), and by the J. Robert Oppenheimer Distinguished Postdoctoral Fellowship (IAP). I. A. P. acknowledges the computational resources at the ARCC HPC cluster at the University of Akron. We would like to thank Mr Saman Shafaie and the Northwestern University IMSERC for performing the HRMS analyses.

Notes and references

1. Z. Guo and P. J. Sadler, Adv. Inorg. Chem., 2000, 49, 183–306.
2. N. Farrell, Coord. Chem. Rev., 2002, 232, 1–4.
3. K. D. Mjos and C. Orvig, Chem. Rev., 2014, 114, 4540–4563.
4. J. Karges, ChemBioChem, 2020, 21, 3044–3046.
5. B. A. Corbin, A. C. Pollard, M. J. Allen and M. D. Pagel, Mol. Imaging Biol., 2019, 21, 193–199.
6. K. J. Franz and N. Metzler-Nolte, Chem. Rev., 2019, 119, 727–729.
7. A. J. Dos santos-García, M. Á. Alario-Franco and R. Sáez-Puche, Lanthanides: Superconducting Materials, in Encyclopedia of Inorganic and Bioinorganic Chemistry, ed. R. A. Scott, 2012,  DOI:10.1002/9781119951438.eibc2039.
8. J. E. McPeak, S. S. Eaton and G. R. Eaton, Electron paramagnetic resonance of lanthanides, Elsevier Inc., 1st edn, 2021, vol. 651.
9. D. Parker, E. A. Suturina, I. Kuprov and N. F. Chilton, Acc. Chem. Res., 2020, 53, 1520–1534.
10. J. C. G. Bünzli, Handb. Phys. Chem. Rare Earths, 2016, 50, 141–176.
11. D. Parker, J. D. Fradgley and K. L. Wong, Chem. Soc. Rev., 2021, 50, 8193–8213.
12. S. V. Eliseeva and J. C. G. Bünzli, Chem. Soc. Rev., 2010, 39, 189–227.
13. C. Müller, K. A. Domnanich, C. A. Umbricht and N. P. Van Der Meulen, Br. J. Radiol., 2018, 91, 1–13.
14. R. F. Cavaier, F. Haddad, T. Sounalet, T. Stora and I. Zahi, Phys. Procedia, 2017, 90, 157–163.
15. F. D. C. Guerra Liberal, A. A. S. Tavares and J. M. R. S. Tavares, Appl. Radiat. Isot., 2016, 110, 87–99.
16. S. J. Goldsmith, Semin. Nucl. Med., 2020, 50, 87–97.
17. C. Müller, K. Zhernosekov, U. Köster, K. Johnston, H. Dorrer, A. Hohn, N. T. Van Der Walt, A. Türler and R. Schibli, J. Nucl. Med., 2012, 53, 1951–1959.
18. R. P. Baum, A. Singh, M. Benešová, C. Vermeulen, S. Gnesin, U. Köster, K. Johnston, D. Müller, S. Senftleben, H. R. Kulkarni, A. Türler, R. Schibli, J. O. Prior, N. P. Van Der Meulen and C. Müller, Dalton Trans., 2017, 46, 14638–14646.
19. C. Müller, C. A. Umbricht, N. Gracheva, V. J. Tschan, G. Pellegrini, P. Bernhardt, J. R. Zeevaart, U. Köster, R. Schibli and N. P. van der Meulen, Eur. J. Nucl. Med. Mol. Imaging, 2019, 46, 1919–1930.
20. A. K. M. Rahman and A. Awal, J. Radioanal. Nucl. Chem., 2020, 323, 731–740.
21. C. Vermeulen, G. F. Steyn, F. Szelecsényi, Z. Kovács, K. Suzuki, K. Nagatsu, T. Fukumura, A. Hohn and T. N. Van Der Walt, Nucl. Instrum. Methods Phys. Res., Sect. B, 2012, 275, 24–32.
22. E. Hindie, P. Zanotti-Fregonara, M. A. Quinto, C. Morgat and C. Champion, J. Nucl. Med., 2016, 57, 759–764.
23. C. Champion, M. A. Quinto, C. Morgat, P. Zanotti-Fregonara and E. Hindié, Theranostics, 2016, 6, 1611–1618.
24. C. Müller, E. Fischer, M. Behe, U. Köster, H. Dorrer, J. Reber, S. Haller, S. Cohrs, A. Blanc, J. Grünberg, M. Bunka, K. Zhernosekov, N. van der Meulen, K. Johnston, A. Türler and R. Schibli, Nucl. Med. Biol., 2014, 41, e58–e65.
25. C. Müller, M. Bunka, J. Reber, C. Fischer, K. Zhernosekov, A. Türler and R. Schibli, J. Nucl. Med., 2013, 54, 2168–2174.
26. P. Santhanam, D. Taieb, L. Solnes, W. Marashdeh and P. W. Ladenson, Clin. Endocrinol., 2017, 86, 645–651.
27. S. Ferguson, H. S. Jans, M. Wuest, T. Riauka and F. Wuest, EJNMMI Phys., 2019, 6, 1–14.
28. G. A. Wiseman, K. Pacak, M. S. O'Dorisio, D. R. Neumann, A. D. Waxman, D. A. Mankoff, S. I. Heiba, A. N. Serafini, S. S. Tumeh, N. Khutoryansky and A. F. Jacobson, J. Nucl. Med., 2009, 50, 1448–1454.
29. E. Eppard, A. de la Fuente, M. Benešová, A. Khawar, R. A. Bundschuh, F. C. Gärtner, B. Kreppel, K. Kopka, M. Essler and F. Rösch, Theranostics, 2017, 7, 4359–4369.
30. S. M. Qaim, B. Scholten and B. Neumaier, J. Radioanal. Nucl. Chem., 2018, 318, 1493–1509.
31. J. H. Turner, Br. J. Radiol., 2018, 91, 1–9.
32. C. S. Loveless, L. L. Radford, S. J. Ferran, S. L. Queern, M. R. Shepherd and S. E. Lapi, EJNMMI Res., 2019, 9, 1–10.
33. B. A. Vaughn, A. J. Koller and E. Boros, in Methods in Enzymology, Elsevier Inc., 1st edn, 2021, vol. 651, pp. 343–371.
34. B. A. Vaughn, S. H. Ahn, E. Aluicio-Sarduy, J. Devaraj, A. P. Olson, J. Engle and E. Boros, Chem. Sci., 2020, 11, 333–342.
35. M. Połosak, A. Piotrowska, S. Krajewski and A. Bilewicz, J. Radioanal. Nucl. Chem., 2013, 295, 1867–1872.
36. S. Huclier-Markai, A. Sabatie, S. Ribet, V. Kubíček, M. Paris, C. Vidaud, P. Hermann and C. S. Cutler, Radiochim. Acta, 2011, 99, 653–662.
37. A. Majkowska-Pilip and A. Bilewicz, J. Inorg. Biochem., 2011, 105, 313–320.
38. M. Pniok, V. Kubíček, J. Havlíčková, J. Kotek, A. Sabatie-Gogová, J. Plutnar, S. Huclier-Markai and P. Hermann, Chem. – Eur. J., 2014, 20, 7944–7955.
39. M. L. Williams, Occup. Environ. Med., 1996, 53, 504.
40. S. Huclier-Markai, R. Kerdjoudj, C. Alliot, A. C. Bonraisin, N. Michel, F. Haddad and J. Barbet, Nucl. Med. Biol., 2014, 41, e36–e43.
41. R. Kerdjoudj, M. Pniok, C. Alliot, V. Kubíček, J. Havlíčková, F. Rösch, P. Hermann and S. Huclier-Markai, Dalton Trans., 2016, 45, 1398–1409.
42. G. Nagy, D. Szikra, G. Trencsényi, A. Fekete, I. Garai, A. M. Giani, R. Negri, N. Masciocchi, A. Maiocchi, F. Uggeri, I. Tóth, S. Aime, G. B. Giovenzana and Z. Baranyai, Angew. Chem., Int. Ed., 2017, 56, 2118–2122.
43. J. Sinnes, J. Nagel and F. Rösch, EJNMMI Radiopharm. Chem., 2019, 4, 1–10.
44. B. A. Vaughn, A. J. Koller, Z. Chen, S. H. Ahn, C. S. Loveless, S. J. Cingoranelli, Y. Yang, A. Cirri, C. J. Johnson, S. E. Lapi, K. W. Chapman and E. Boros, Bioconjugate Chem., 2021, 32, 1232–1241.
45. S. Aime, A. Barge, F. Benetollo, G. Bombieri, M. Botta and F. Uggeri, Inorg. Chem., 1997, 36, 4287–4289.
46. F. Benetollo, G. Bombieri, L. Calabi, S. Aime and M. Botta, Inorg. Chem., 2003, 42, 148–157.
47. P. Vojtíšek, P. Cígler, J. Kotek, J. Rudovský, P. Hermann and I. Lukeš, Inorg. Chem., 2005, 44, 5591–5599.
48. F. Benetollo, G. Bombieri, L. Calabi, S. Aime and M. Botta, Inorg. Chem., 2003, 42, 148–157.
49. C. Kumas, W. S. Fernando, P. Zhao, M. Regueiro-Figueroa, G. E. Kiefer, A. F. Martins, C. Platas-Iglesias and A. D. Sherry, Inorg. Chem., 2016, 55, 9297–9305.
50. K. J. Miller, A. A. Saherwala, B. C. Webber, Y. Wu, A. D. Sherry and M. Woods, Inorg. Chem., 2010, 49, 8662–8664.
51. J. Blahut, P. Hermann, Z. Tošner and C. Platas-Iglesias, Phys. Chem. Chem. Phys., 2017, 19, 26662–26671.
52. A. L. Spek, Acta Crystallogr., Sect. C: Struct. Chem., 2015, 71, 9–18.
53. O. V. Dolomanov, L. J. Bourhis, R. J. Gildea, J. A. K. Howard and H. Puschmann, J. Appl. Crystallogr., 2009, 42, 339–341.
54. G. M. Sheldrick, SHELXTL, University of Gottingen, Germany, 2016.
55. J. Mason, Multinuclear NMR, Plenum Press, New York, 6th edn, 1987.
56. A. Abragam, The Principles of Nuclear Magnetism, Clarendon Press, 1961.
57. B. W. Stein, A. Morgenstern, E. R. Batista, E. R. Birnbaum, S. E. Bone, S. K. Cary, M. G. Ferrier, K. D. John, J. L. Pacheco, S. A. Kozimor, V. Mocko, B. L. Scott and P. Yang, J. Am. Chem. Soc., 2019, 141, 19404–19414.
58. M. P. Kelley, G. J. P. Deblonde, J. Su, C. H. Booth, R. J. Abergel, E. R. Batista and P. Yang, Inorg. Chem., 2018, 57, 5352–5363.
59. J. T. Su, X. Xu and W. A. Goddard III, J. Phys. Chem. A, 2004, 108(47), 10518–10526.
60. M. P. Kelley, J. Su, M. Urban, M. Luckey, E. R. Batista, P. Yang and J. C. Shafer, J. Am. Chem. Soc., 2017, 139, 9901–9908.
61. X. Cao, D. Heidelberg, J. Ciupka and M. Dolg, Inorg. Chem., 2010, 49, 10307–10315.
62. APEX3, Bruker, AXS Inc., Madison, Wisconsin, 2019.
63. SAINT+, Bruker, AXS Inc., Madison, Wisconsin, 2013.
64. L. Krause, R. Herbst-Irmer, G. M. Sheldrick and D. Stalke, J. Appl. Crystallogr., 2015, 48, 3–10.
65. B. Mennucci, E. Cancès and J. Tomasi, J. Phys. Chem. B, 1997, 101, 10506–10517.
66. E. Cancès and B. Mennucci, J. Math. Chem., 1998, 23, 309–326.
67. M. J. Frisch, G. W. Trucks, H. B. Schlegel, G. E. Scuseria, M. A. Robb, J. R. Cheeseman, G. Scalmani, V. Barone, G. A. Petersson, H. Nakatsuji, X. Li, M. Caricato, A. V. Marenich, J. Bloino, B. G. Janesko, R. Gomperts, B. Mennucci, H. P. Hratchian, J. V. Ortiz, A. F. Izmaylov, J. L. Sonnenberg, D. Williams-Young, F. Ding, F. Lipparini, F. Egidi, J. Goings, B. Peng, A. Petrone, T. Henderson, D. Ranasinghe, V. G. Zakrzewski, J. Gao, N. Rega, G. Zheng, W. Liang, M. Hada, M. Ehara, K. Toyota, R. Fukuda, J. Hasegawa, M. Ishida, T. Nakajima, Y. Honda, O. Kitao, H. Nakai, T. Vreven, K. Throssell, J. A. Montgomery Jr., J. E. Peralta, F. Ogliaro, M. J. Bearpark, J. J. Heyd, E. N. Brothers, K. N. Kudin, V. N. Staroverov, T. A. Keith, R. Kobayashi, J. Normand, K. Raghavachari, A. P. Rendell, J. C. Burant, S. S. Iyengar, J. Tomasi, M. Cossi, J. M. Millam, M. Klene, C. Adamo, R. Cammi, J. W. Ochterski, R. L. Martin, K. Morokuma, O. Farkas, J. B. Foresman and D. J. Fox, Gaussian, Inc., Wallingford CT, 2016.
68. C. Adamo and V. Barone, J. Chem. Phys., 1999, 110, 6158–6170.
69. F. Weigend and R. Ahlrichs, Phys. Chem. Chem. Phys., 2005, 7, 3297–3305.
70. S. Grimme, J. Antony, S. Ehrlich and H. Krieg, J. Chem. Phys., 2010, 132, 132.
71. M. Steinmetz and S. Grimme, ChemistryOpen, 2013, 2, 115–124.

---

Footnote

† Electronic supplementary information (ESI) available: NMR spectra including fits for VT NMR; additional discussion of X-Ray diffraction data; additional information on the computed reactions and method details. CIF files for SOLUTION_1 of the solid-state single crystal structure. CCDC 2072300 and 2076191. For ESI and crystallographic data in CIF or other electronic format see DOI: https://doi.org/10.1039/d1dt03887g

---