Separation of terbium as a first step towards high purity terbium-161 for medical applications

Abstract

Terbium-161 is a medical radiolanthanide that has a beta decay energy and half-life similar to that of lutetium-177, which makes it a promising alternative for therapeutic purposes. The production route using an enriched gadolinium-160 target necessitates the purification of terbium-161 from the untransmuted target material as well as from its stable decay product, dysprosium-161. The separation of neighbouring lanthanides is challenging due to their similar chemical properties and prominent trivalent oxidation states. In this work, the aim is to change the oxidation state of terbium, resulting in the altering of chemical properties that ease the intragroup separation. To this end, a novel separation method is investigated, involving the electrochemical oxidation of terbium (3+) to terbium (4+) followed by anion exchange chromatography. The electrolysis conditions are set to the highest achievable conversion rate, followed by a dilution step during which the pH and electrolyte concentration are slightly lowered to obtain conditions that are compatible with the separation method. XAS analysis is done to characterize the carbonato complex of both oxidation states and to further elucidate the separation mechanism. The results show that the separation approach of combining electrochemical oxidation with anion exchange chromatography is promising for the purification of ¹⁶¹Tb for medical use.

---

Introduction

Terbium is called a ‘Swiss army knife’ for nuclear medicine,¹ since it has four different radioisotopes that are of potential interest to medical applications: ¹⁴⁹Tb (t_1/2 = 4.12 h, E_α = 3.967 MeV (16.7%)), ¹⁵²Tb (t_1/2 = 17.5 h, E_β⁺,av = 1.14 MeV (20.3%)), ¹⁵⁵Tb (t_1/2 = 5.32 d, E_γ = 86.55 keV (32%) and 105.3 keV (25%)) and ¹⁶¹Tb (t_1/2 = 6.96 d, E_β⁻,av = 0.154 MeV (100%)). These isotopes have complementary decay characteristics. ¹⁶¹Tb is suitable for therapeutic use due to the emission of low-energy β⁻ particles, conversion and Auger electrons.^{2 161}Tb has a half-life of 6.98 days, and emits low-energy β⁻-particles with a maximum energy of 0.59 MeV, which is very similar to clinically used ¹⁷⁷Lu (t_1/2: 6.65 d, E^max_β: 0.49 MeV).³

There are many in vitro and in vivo studies, comparing the therapeutic properties with a well-established ¹⁷⁷Lu counterpart.^3–7 Recently imaging of cancer patients has been successfully performed after administration of ¹⁶¹Tb-DOTATOC.⁸ Due to the comparable results and the fact that it is well-tolerated by the patients, the ¹⁶¹Tb-DOTATOC conjugate is considered to be a suitable alternative for ¹⁷⁷Lu-DOTATOC. The efficacy of ¹⁶¹Tb in targeted therapy and its potential for personalized treatment approaches, make it a great candidate for medical applications.

¹⁶¹Tb is commonly produced through the non-carrier-added reactor production route, which involves neutron irradiation (n, γ) of an enriched ¹⁶⁰Gd target. The neutron capture of ¹⁶⁰Gd produces short-lived ¹⁶¹Gd, which decays through β⁻-emission into ¹⁶¹Tb. When a highly enriched ¹⁶⁰Gd target material is used, radionuclidically pure ¹⁶¹Tb can be obtained at the end of the irradiation.^{9 161}Tb further decays to the stable ¹⁶¹Dy, which can interfere with radiolabeling due to their chemical resemblance. As a result, in addition to isolating ¹⁶¹Tb from ¹⁶⁰Gd, ¹⁶¹Tb and ¹⁶¹Dy must also be separated to preserve the high specific activity. The existing methods^3,10,11 only deal with the separation of micro-amounts of ¹⁶¹Tb from a macro-amount of ¹⁶¹Gd target material.

Lanthanides typically exist in the trivalent oxidation state, which results in a similar coordination chemistry for all the lanthanides. This gives a flexibility to select the radiolanthanide with the most suitable half-life, decay mode and decay energy for either diagnosis or therapy, from a range of possible radiolanthanides, that are all compatible with the same carrier molecule.^9,12–16 On the other hand, separation of lanthanides becomes challenging, due to these similar chemical properties. The current method to obtain non-carrier-added (nca) ¹⁶¹Tb is using cation exchange and extraction chromatography with α-hydroxy-isobutyric acid (α-HIBA) as the complexing agent.^3,11 However, these methods have only limited efficiency since the process requires long separation times and multiple separation steps. Another strategy involves substantially altering one of the lanthanide ions' chemical characteristics by changing its oxidation state, which creates possibilities for more effective separation methods.^17–19

Terbium (E⁰ = +3.1 V vs. SHE) is, besides cerium, one of the lanthanides that can be oxidized to its tetravalent state, but there is only limited information available in the literature on the oxidation of Tb³⁺.^20–22 In our previous work, a detailed parametric study was performed on electrochemical oxidation of Tb³⁺ in various aqueous media.²³ Although the applied potential could be lowered to +0.9 V vs. Ag/AgCl in periodate medium, the most stable Tb⁴⁺ complex was obtained in concentrated carbonate medium. Therefore, an aqueous carbonate medium seems to be the most promising compared to the other media for developing a separation process.

Although there are many examples on rare earth separation,^24–30 there are only a few examples of separation processes in general using carbonate medium. These examples consist of extraction of lanthanides and transplutonium elements from alkaline and carbonate media can be achieved mainly via liquid–liquid extraction,^31–36 using various extractants such as quaternary ammonium salts,^37–39 primary amines⁴⁰ or alkylpyrocatechols.⁴¹ There are also just a few examples of using extraction chromatography in carbonate media.^42,43 Thanks to its simplicity, ease of operation, speed, efficiency and the ease to be automated, solid phase extraction is used for the separation process.

In this study, a strategy for the separation of terbium from gadolinium is investigated based on the change of the valence state of terbium. An electrochemical oxidation is performed in highly concentrated carbonate medium, followed by a column separation using Dowex 1-X8 anion exchange resin. Prior to the separation, a dilution step is applied to lower the carbonate concentration. It is established in the work of Sherry et al.⁴² that at lower carbonate concentrations, higher distribution coefficients can be obtained. The Tb⁴⁺ complex was found to remain stable after dilution, which is advantageous for an anion exchange separation process. X-ray absorption spectroscopy measurements were performed to validate the change in the oxidation state as well as to gain information on the structure of the complexes.

Experimental

Chemicals

TbCl₃·6H₂O (99.9%), GdCl₃·xH₂O (99.9%) and DyCl₃·6H₂O (99.9%) were obtained from Strem Chemicals, Inc. (Newburyport, USA). K₂CO₃ (anhydrous, ≥98%) and KOH pellets (≥85%) were purchased from VWR International B. V. (Leuven, Belgium). Dowex 1-X8 (chloride form) 200–400 mesh ion exchange resin was obtained from Thermo Fisher Scientific (Merelbeke, Belgium). All chemicals were used as received without any further purification. All aqueous solutions were prepared with ultra-pure water (18.2 MΩ cm at 25 °C).

Production of ¹⁶¹Tb

[¹⁶¹Tb]TbCl₃ was produced by following an adapted literature procedure.¹¹ Highly enriched ¹⁶⁰Gd₂O₃ (1.0 mg, 98.2%, Isoflex USA) was loaded as a nitrate salt into a quartz ampoule, sealed inside an aluminum capsule. The irradiation was performed in the BR2 Reactor at the Belgian Nuclear Research Centre (SCK CEN) at a thermal neutron flux of 3.0 × 10¹⁴ n cm⁻² s⁻¹. After 10 days of irradiation, the target was removed from the reactor and cooled for 5 days. The irradiated material was then dissolved in trace-metal grade water and the ¹⁶¹Tb was purified using high-pressure ion chromatography (HPIC, Shimadzu), with a strong cation exchange column (ϕ: 6 mm, l: 50 mm, Shodex IC R-621) by elution with α-hydroxyisobutyric acid. The collected fractions containing [¹⁶¹Tb] were further concentrated by loading them onto an LN3 column (ϕ: 2.1 mm, l: 30 mm, TrisKem International) and eluted with 0.05 mol L⁻¹ hydrochloric acid. The isolated solutions of [¹⁶¹Tb]TbCl₃ had a radionuclidical purity of 99.998%. An activity of about 2.5 MBq ¹⁶¹Tb was added to a solution containing 4 mol L⁻¹ K₂CO₃ and 0.07 mol L⁻¹ TbCl₃·6H₂O. The pH was then increased to 14 by adding KOH. This step is crucial for the electrolysis to be successful as well as for the stability of the Tb⁴⁺ complex in carbonate medium.^20,23 The final volume was set to 10 mL to ensure the compatibility with the electrolysis cell volume.

Electrochemical oxidation of terbium

The electrochemical oxidation of terbium was performed in an undivided three-electrode voltammetry cell (BASI MF-1052 VC-2). A platinum-mesh electrode was used as working electrode, and a platinum wire (ϕ: 0.5 mm, l: 65 mm) served as auxiliary electrode and BASi MF-2056 Ag/AgCl in a 3 mol L⁻¹ KCl solution as reference electrode. A Metrohm Autolab PGSTAT302N potentiostat, operated by Nova 2.1.5 software was used in potentiostatic mode to apply constant a potential of +1.3 V vs. Ag/AgCl for 3 hours.

UV-Vis spectroscopy

During the electrolysis, a colour change was observed, starting from the electrode surface, and dissipating into the solution, from colourless to dark brown, indicating the Tb⁴⁺ complex formation. This change in the oxidation state was characterized with UV-Vis spectroscopy, by a broad charge transfer band with an absorption peak centred at 365 nm.^20,23

Absorption spectra were recorded in the 230–1000 nm range using a Shimadzu 1900i spectrophotometer, operated by LabSolutions UV-Vis software. The light source changeover occurred at 340.8 nm. A quartz cuvette with 1 mm path length was used. Spectra were collected at a fast-scanning rate of 300 nm min⁻¹ with steps of 0.5 nm. Baseline correction was done with the blank solutions of carbonate solutions without terbium.

Extraction chromatography

Experiments were conducted in a column separation setup using Triskem 2 mL bed volume cartridges with 1 cm inner diameter and a total length of 2.5 cm. An ISMATEC IPC 8-channel peristaltic pump was used to control the flow rate of the mobile phase. The Dowex 1-X8 resin in chloride form was packed into the column/cartridge as a slurry to ensure uniform packing, and to avoid cracks or holes that might end up in channel formation, hence impact the separation capability and column performance. The packed column was preconditioned with 4 mol L⁻¹ K₂CO₃ solution, and thus converted to carbonate form. The feed solution was prepared by diluting the initial carbonate concentration from 4 mol L⁻¹ to 1 mol L⁻¹ (pH ≈ 12), and thus the Tb³⁺ concentration decreases to 0.018 mol L⁻¹. 0.4 mL of the feed solution was loaded onto the column with a flow rate of 0.3 mL min⁻¹ by use of a peristaltic pump. Afterwards, the column was first eluted with water for the elution of Tb⁴⁺. The mobile phase was switched to 2 mol L⁻¹ K₂CO₃ after collecting 9 fractions of 0.5 mL. The fractions were measured by gamma spectrometry for ¹⁶¹Tb. All experiments were performed at room temperature.

Gamma spectrometry

Gamma spectrometric analysis of ¹⁶¹Tb (74.57 keV) and ¹⁷⁵Yb (396.33 keV) was done using a Canberra High Purity Germanium detector (model: GC2520) with a DSA-1000 Multi-Channel Analyzer (Canberra Semiconductors N. V., Olen, Belgium). Genie 200 gamma analysis software was used (Mirion Technologies). The relevant peaks showed at least 10 000 counts with the uncertainty of less than 1%, used for the quantitative analysis.

X-ray absorption spectroscopy

XAS spectra were collected at room temperature at the XAFS Beamline in Elettra Synchrotron Trieste (Italy).⁴⁴ The storage ring was operating at 2 GeV with an optimal current around 300 mA. The energy of the beam was tuned by using a fixed exit monochromator equipped with a pair of Si(111) crystals. An internal reference of cobalt foil is used for the energy calibration. Data were recorded at ambient pressure at Tb L₃-edge in dual transmission and fluorescence mode due to the low absorption of the material. A sample holder with Kapton® windows and a Teflon™ spacer with a thickness of 0.1 mm were used as sample holders. The spectra were recorded by scanning through the terbium L₃ absorption edge (ca. 7514 eV) in energy interval of 7305–7715 eV. XAS primary data analysis was performed with the Athena program. The ab initio simulation of the XANES spectra was performed using the FDMNES software.⁴⁵ The Tb L₃-edge was calculated in the photoelectron energy range −10 < E < 80 eV with respect to the Fermi energy level using the Muffin-Tin approximation. The Hedin–Lundqvist complex potential⁴⁶ was used to calculate the excited states. The absorption cross-section was calculated within the dipolar approximation. Clusters of 5 Å built around each non-equivalent absorbing atom were considered.

Results and discussion

Stability of Tb⁴⁺ complex

Dowex 1-X8 strong anion exchange resin was used in its carbonate form. There are multiple examples of this resin being used with alkaline media, specifically carbonate/bicarbonate systems.^{42,43,47–50} The resin shows good physical and chemical stability in the pH range of 0–14.⁵¹ The resin was prepared with washing 4 mol L⁻¹ K₂CO₃ and 0.08 mol L⁻¹ KOH mixture before introducing the feed solution. Thus, the resin was expected to convert from Cl⁻ form to CO₃²⁻ form rather than OH⁻ form due to the high concentration of carbonate ions in the solution. The selectivity towards Tb⁴⁺ complex rather than the Tb³⁺ complex, so that the former is retained on the column, could be explained by the difference in the total charge of these complexes. As the first eluent is water, the concentration of carbonate would drastically drop and preferentially elute the Tb³⁺ complex. The Tb⁴⁺ complex is then eluted by switching back to the carbonate solution, where the carbonate ions might be exchanging with the Tb⁴⁺ anionic complex.

Although the stability of Tb⁴⁺ in carbonate medium was found to be high enough to follow up with a separation process, the possibility of back-reduction due to the resin itself can result in lower separation factors. To investigate this effect, UV-Vis absorption spectra were collected after elution of different volumes of feed solution with water and measuring the related Tb⁴⁺ fractions. The UV-Vis spectra shown in Fig. 1, are for the feed solution after the electrolysis, and for the first fractions collected after 0.5 mL and 6 mL feed solution were introduced to the column respectively. In Fig. 2, it can be observed that the absorbance value only changes by 0.01 at 365 nm for the feed and fraction collected from 6 mL of feed, showing that the back reduction due to the column material can be considered as negligible. When a small volume of feed solution (0.5 mL) was introduced, due to the dilution effect by the eluent, the absorbance obviously decreased.

Fig. 1 UV-Vis absorption spectra of the fractions collected at different feed volumes of 0.07 mol L⁻¹ Tb⁴⁺ in 4 mol L⁻¹ K₂CO₃ solution at pH = 14. Electrolysis was performed at constant potential +1.3 V vs. Ag/AgCl for 3 h. The formation of Tb(IV) was followed by the absorbance peak at 365 nm. Blank carbonate solution was used as the reference sample for baseline correction. The cell path length was 1 mm.

Fig. 2 UV–Vis absorption spectra of the undiluted 0.07 mol L⁻¹ Tb⁴⁺ in 4 mol L⁻¹ K₂CO₃ solution at pH = 14, 1 : 1 (v/v) dilution and 1 : 2 (v/v) dilution. Electrolysis was performed on the undiluted sample at constant potential +1.3 V vs. Ag/AgCl during 30 min. Blank carbonate solution was used as the reference sample for baseline correction. The path length of the quartz cuvette was 1 mm.

In the studies of Sherry et al., carbonate and bicarbonate complexes of lanthanides were investigated in a K₂CO₃ concentration range of 0.3 to 3 mol L⁻¹.^42,43 These authors also showed that distribution coefficients of the rare earth carbonate complexes were drastically dropping with increasing carbonate concentration.^42,43 Therefore, to optimize the separation process, a dilution step was added after the electrolysis. The electrolysed solution was diluted from 4 mol L⁻¹ to 2 mol L⁻¹ and 1.3 mol L⁻¹ K₂CO₃. To check the stability of Tb⁴⁺ complex after the dilution step, UV-Vis spectra were recorded at various time intervals (ESI Fig. S1 and S2†). Absorbance was measured immediately after the electrolysis and any necessary dilutions are prepared swiftly in order to measure both diluted and undiluted samples in the same time frame. As can be seen in Fig. 2, the absorbance values, for the undiluted, 1 : 1 and 1 : 2 diluted solutions are 0.127, 0.061 and 0.038 at 365 nm respectively. These values agree well with the dilution factors and these results show that the Tb⁴⁺ complex remains stable under dilution, and the influence of back-reduction is negligible (ESI Fig. S1 and S2†).

In the proposed separation method (Fig. 3), Tb³⁺ and Tb⁴⁺ complexes show different affinities to the column, based on the difference in the charge of these complexes. Because of this, a selective separation can be achieved. Misumi et al. described the separation mechanism for tetravalent metal ion complexes in 0.2–0.6 mol L⁻¹ carbonate medium as follows:⁴⁸

M⁴⁺ + xCO₃²⁻ ↔ (MCO₃)_x^4−2x (1)

(MCO₃)_x^4−2x + (y/2)R₂CO₃ ↔ R_yM(CO₃)_x + (y/2)CO₃²⁻ (2)

Fig. 3 Schematic of the proposed purification method using Dowex 1-X8 resin packed column.

Based on these equations, y was equal to 2x − 4 to maintain the electrical neutrality of the resin. This was based on the assumption that the tetravalent metal ion forms a single complex with the carbonate ion, neglecting the presence of hydroxide ions, especially at higher concentrations of carbonate. In the study of Varlashkin et al.,²¹ several stoichiometries of solid state Tb⁴⁺ carbonato–hydroxo complexes were determined based on IR spectra, and it was assumed that the Tb⁴⁺ solutions could contain complexes similar to these containing a mixture of carbonate and hydroxide groups. Similarly, in the works of Sherry et al.,^42,43 rare earth carbonate systems are investigated by anion exchange column chromatography where the slope of plotted logarithmic functions of molality of the ligand and distribution coefficients, the ligand number of these complexes was predicted at lower carbonate concentrations. With increasing carbonate concentrations, the presence of hydroxide groups in the complex was assumed here as well. However, to the best of our knowledge, no further studies on Tb³⁺ and Tb⁴⁺ complexes in highly concentrated carbonate solutions have been reported.

Column separation using Dowex 1X-8 resin

In Fig. 4, the elution curve of a Tb³⁺ solution spiked with ¹⁶¹Tb that was not treated by electrolysis is shown. With water as the mobile phase, the Tb³⁺ carbonate complex was completely retained on the column. After the mobile phase was switched to 2 mol L⁻¹ K₂CO₃, Tb³⁺ is completely eluted in the collected fractions. The radioactivity of the column material was checked after all fractions were collected, showing no presence of Tb¹⁶¹ tracers. In addition, when the activity of 0.4 mL of the feed solution (178.73 ± 2.54 kBq) is compared to the total activity collected in the fractions (175.46 ± 2.47 kBq), values are in good agreement. When the remainder of the feed solution was electrolysed during 3 h, at constant potential of +1.3 V vs. Ag/AgCl, an immediate colour change was observed. In a second experiment, the same volume (0.4 mL) of feed solution was introduced onto the column. The elution curve of Tb⁴⁺ solution can be seen in Fig. 5. In this elution curve, it is observed that the Tb⁴⁺ carbonate complex is not retained by the anion exchange resin, in contrast to the Tb³⁺ carbonate complex. In our previous study,²³ an in-depth qualitative analysis was performed on the electrochemical oxidation of Tb³⁺ to Tb⁴⁺, but a quantitative electrochemical analysis could not be performed due to the limitation of the high formal reduction potential (E⁰ = +3.1 V vs. Ag/AgCl). In the current study, with the column chromatography experiments, full separation of Tb⁴⁺ and Tb³⁺ is achieved. An estimation of the conversion of Tb³⁺ to Tb⁴⁺ by electrochemical reduction can be done based on the assumption that back-oxidation is negligible during the dilution and separation, which was shown by the UV-Vis absorption spectroscopy results. Based on this, only ca. 11.2% of the mixture consisted of Tb⁴⁺ whereas majority (88.8%) was still Tb³⁺ after 2 hours of electrolysis.

Fig. 4 Elution curve of tracer spiked Tb³⁺ solution using a Dowex 1-X8 resin filled cartridge (ϕ: 1 cm, h: 2.5 cm, BV: 2 mL). Feed: 0.4 mL of a solution containing 0.018 mol L⁻¹ Tb³⁺ in 1 mol L⁻¹ K₂CO₃ solution (pH ≈ 12). The vertical dashed line denotes the change of mobile phase from water to 2 mol L⁻¹ K₂CO₃. Flow rate: 0.3 mL min⁻¹.

Fig. 5 Elution curve of tracer spiked and electrolysed Tb⁴⁺ solution using a Dowex 1-X8 resin filled cartridge (ϕ: 1 cm, h: 2.5 cm, BV: 2 mL). Feed: 0.4 mL of a solution containing 0.018 mol L⁻¹ Tb³⁺/Tb⁴⁺ in 1 mol L⁻¹ K₂CO₃ solution (pH ≈ 12). The vertical dashed line denotes the change of mobile phase from water to 2 mol L⁻¹ K₂CO₃. Flow rate: 0.3 mL min⁻¹.

As a next step, electrolysis was performed in a mixture of lanthanides gadolinium, terbium and dysprosium, with a concentration of each of 0.05 mol L⁻¹ in 4 mol L⁻¹ K₂CO₃ at pH = 14. After 2 hours of electrolysis at +1.3 V vs. Ag/AgCl, UV-Vis absorption spectra were recorded, as can be seen in Fig. 6. The spectra – recorded before and after the electrolysis – showed that, Tb³⁺ can be successfully oxidized in presence of other lanthanides without any inhibition. The concentration was chosen based on the solubility limit of the lanthanides in the carbonate solution, in which increasing the lanthanide concentration to more than 0.15 mol L⁻¹ resulted in precipitation. In the previous work, it was established that when terbium concentrations lower than 0.04 mol L⁻¹ were used, oxidation was no longer observed.²³ Furthermore, maximum Tb³⁺ to Tb⁴⁺ conversion was not achieved at 3 hours of electrolysis for terbium concentrations higher than 0.15 mol L⁻¹. This was explained by the strong complexation hindering the oxidation reaction at these concentrations.²³ Thus, in these tests equimolar lanthanide mixtures were used with the total concentration of 0.15 mol L⁻¹. After the electrolysis, the same protocol was followed for the separation process, except the feed volume was decreased to 0.2 mL instead of 0.4 mL due to the higher lanthanide concentration. In addition to ¹⁶¹Tb, ≈200 ± 18.2 kBq ¹⁷⁵Yb was also added. While ¹⁶¹Tb tracers are expected to oxidize to tetravalent state, ¹⁶⁵Yb tracers are added in the mixture as a representative for the trivalent lanthanides gadolinium and dysprosium. ¹⁶⁵Yb is deemed more of a convenience for the gamma spectrometry due to its availability and relatively short half-life (4.16 d). The elution curves for ¹⁷⁵Yb and ¹⁶¹Tb is shown in Fig. 7. The trivalent lanthanides, Gd, Dy and Tb were retained on the column as can be seen with ¹⁷⁵Yb and ¹⁶¹Tb elution curves, which is similar to the results containing only terbium. There were no trivalent lanthanides detected in the first elutions with water as mobile phase, whereas a brown coloured solution was collected in the first fractions, corresponding to Tb⁴⁺. When calculating the ratios of Tb³⁺ to Tb⁴⁺ in the feed solution, the integration of the peak areas of the elution curves were used, assuming that there was no back-reduction occurring during the separation process and the contribution of first elution peak was purely originating from Tb⁴⁺ complex whereas the second elution peak was only composed of Tb³⁺ complex. The amount of Tb⁴⁺ in the mixture was calculated to be ca. 8.6%, after 2 hours of electrolysis. This difference in the amount of Tb⁴⁺ can be explained by the difference in the electrolysis time. Therefore, Tb³⁺ was successfully oxidized to Tb⁴⁺ and stabilized in the aforementioned experimental conditions, as well as Tb⁴⁺ was separated from the other trivalent lanthanides.

Fig. 6 UV–Vis absorption spectra of 0.05 mol L⁻¹ Tb³⁺, 0.05 mol L⁻¹ Gd³⁺ and 0.05 mol L⁻¹ Dy³⁺ in 4 mol L⁻¹ K₂CO₃ at pH = 14, before and after the electrolysis. Electrolysis was performed at constant potential +1.3 V vs. Ag/AgCl for 2 h. The formation of Tb⁴⁺ was followed by the absorbance peak at 365 nm. Blank carbonate solution was used as the reference sample for baseline correction. The path length was 1 mm.

Fig. 7 Elution curves of tracer spiked electrolyzed lanthanide mixture solution (0.05 mol L⁻¹ Tb³⁺, 0.05 mol L⁻¹ Gd³⁺ and 0.05 mol L⁻¹ Dy³⁺ in 4 mol L⁻¹ K₂CO₃) solution before (left) and after (right) electrolysis at +1.3 V vs. Ag/AgCl, using a Dowex 1-X8 resin filled cartridge (ϕ: 1 cm, h: 2.5 cm, BV: 2 mL). Feed: 0.2 mL of a solution containing 0.018 mol L⁻¹ Ln³⁺ in 1 mol L⁻¹ K₂CO₃ solution (pH ≈ 12). The vertical dashed line denotes the change of mobile phase from water to 2 mol L⁻¹ K₂CO₃. Flow rate: 0.3 mL min⁻¹.

In order for this method to be a viable route for therapeutic radiolanthanide preparation, we should briefly evaluate its compatibility with exisiting radiolabelling strategies. Thanks to the similar properties with clinically-used ¹⁷⁷Lu, there has been many studies^3,5,7,11 on the therapeutic potential of ¹⁶¹Tb. As a result, similar approaches of radiolabelling has been applied and compared to ¹⁷⁷Lu. As the advantage of radiolanthanides, since the most stable oxidation state is trivalent, similar chemistry for radiolabelling could be applied. The most common chelators that are used for terbium isotopes are DOTA derivatives, whereas DTPA is also considered as alternative. The radiolabelling requires high temperature and a pH range of 4–5.5, where trivalent terbium is complexed with these chelators.¹⁶ Adjusting the pH of the eluted Tb⁴⁺ fractions to a value within the appropriate labelling window would thermodynamically favor reduction to Tb³⁺, thus making it even more analogues to the common labelling strategies used for ¹⁷⁷Lu.

XANES measurements

To explore the oxidation states of terbium ions during the course of electrolysis, XANES experiments have been carried out at the Tb L₃-edge. With the starting solution of 0.07 Tb³⁺ mol L⁻¹ Tb in 4 mol L⁻¹ K₂CO₃ at pH = 14, electrolysis was performed and recorded every 30 minutes. The normalized spectra at various electrolysis times can be seen in Fig. 8. The white lines of Tb-L₃ were observed with a maximum at 7519 eV, whereas a shoulder appeared at between 7525–7530 eV, which is indication of Tb⁴⁺ (ESI Fig. S3†). As the electrolysis time increases, the decrease in the white line could explain the conversion of Tb³⁺ to Tb⁴⁺, and the final spectrum containing a mixture of both valence states. In order to quantify the conversion, linear combination fitting (LCF) was performed, using initial Tb³⁺ solution without electrolysis performed (0.07 Tb³⁺ mol L⁻¹ Tb in 4 mol L⁻¹ K₂CO₃ at pH = 14) and Tb₄O₇ as the reference materials (ESI Fig. S4†). The final fitting using these references resembles the shape of the spectrum recorded after 24 hours of electrolysis. The conversion was calculated as 15% based on the fitting, which was in line with the results from the extraction chromatography experiments. On the other hand, even after 24 hours of electrolysis, the oxidation of Tb³⁺ to Tb⁴⁺ was only partial, meaning that after the oxidation the solution consists of a mixture of Tb³⁺ and Tb⁴⁺ complexes. Therefore, the uncertainty on the quantification based on LCF was accepted as high. The EXAFS spectrum of this mixture couldn't be analyzed due to the noise resulting from low terbium concentrations.

Fig. 8 Normalized XANES spectra of the oxidation of 0.07 mol L⁻¹ Tb³⁺ in a 4 mol L⁻¹ K₂CO₃ solution as a function of time (15 min–120 min).

Ab initio simulation of L₃ XANES spectra was carried out by considering the structures reported by Michiba et al.⁵² (ICSD 422736) and Dixey and Saines⁵³ (ICSD 7183) to gain more insight into the electronic and local structural features of terbium (Fig. 9). As can be seen in Fig. 10, there are five carbonato and two hydroxo ligands coordinated to Tb in the structure reported by Dixey and Saines,⁵³ (ICSD 7183) where only one carbonato is a bidentate ligand and the others are monodentate. Similarly, in the structure that was reported by Michiba et al.,⁵² the lanthanides were coordinated with five carbonato ligands (one bidentate and four monodentate), whereas the only difference was a third hydroxo ligand present in the structure. Although both structures contain Tb³⁺ surrounded by carbonates and hydroxides, the different local environment around terbium determines differences in the simulated L-edge. For ICSD 7183, terbium has an overall coordination number of 8, with bond valence sum of 2.85 (considering a bond search factor of 0.87). At the same time, the ICSD 422736 structure features terbium in a nine-fold coordination environment with average bond valence sum of 3.06 (bond search factor 0.87). Thus, the electronic distribution for the two structures is different to a small extent, with terbium in ICSD 422736 having a slightly higher effective charge. This can also be verified in the small shift of the edge toward higher energy in the simulated ICSD 422736 spectrum when compared to the respective one for ICSD 7183 (Fig. 9).

Fig. 9 Simulated L₃-edge spectra for reference structures ICSD-7183 and ICSD-422736 and comparison with experimental XANES traces of 0 h and 24 h electrolysis.

Fig. 10 The coordination environment of terbium in the Tb(OH)CO₃ crystal (ICSD 7183), as published by Dixey and Saines et al.,⁴⁸ visualized as a ball and stick model.

Interestingly, the experimental XANES trace recorded after 24 h of electrolysis displays more similarities with ICSD 422736, while the experimental spectrum at time zero (0 h) is closer to ICSD 7183. This might be related to a partially higher oxidation state of Tb at 24 h, as discussed above, as well as a different coordination environment, possibly a nine-fold coordination for the Tb⁴⁺/Tb³⁺ complex after 24 hours of electrolysis, vs. an eight-fold coordination for the Tb³⁺ complex.

Conclusions

¹⁶¹Tb tracer spiked solutions were successfully electrolyzed, and further diluted without affecting the stability of Tb⁴⁺. The separation of trivalent and tetravalent terbium was achieved by using Dowex 1-X8 anion exchange resin for the first time to the best of our knowledge. Tb⁴⁺ had no interaction with the column and eluted in the initial fractions, whereas the anionic carbonate complex of Tb³⁺ was retained on the column until the eluent is switched from water to concentrated carbonate solution. Full separation was achieved without any back-reduction of Tb⁴⁺ to Tb³⁺, proven by UV-Vis absorption spectroscopy results. Based on these results, an estimation of the electrolysis yield was done, which was calculated as ≈11.2% after 3 h of electrolysis at a constant potential at +1.3 V vs. Ag/AgCl. The separation method was proven also successfully in the presence of other lanthanides in the feed solution. Similar results were obtained when the same protocol was applied with an equimolar mixture of Gd³⁺, Tb³⁺ and Dy³⁺. Therefore, this separation method was found to be applicable for intergroup lanthanide separation. Finally, X-ray absorption spectroscopy results showed some minor differences in the structure of the Tb³⁺ and Tb⁴⁺ carbonate complexes, where different coordination numbers were obtained based on the simulations. As a result, the separation method looks promising towards the purification of ¹⁶¹Tb for medical use.

Data availability

The data supporting this article have been included as part of the ESI.†

Author contributions

M. Ö. A.: conceptualization, methodology, analysis, investigation, visualization, writing – original draft, writing – review & editing. A. M.: XAS analysis, writing – original draft, writing – review & editing. B. G.: conceptualization, methodology, analysis, investigation, writing – review & editing. K. V. H.: conceptualization, methodology, analysis, investigation, writing – review & editing. G. D.: XAS experiments, writing – original draft, writing – review & editing. G. A.: XAS experiments, writing – original draft, writing – review & editing. K. B.: conceptualization, methodology, analysis, investigation, writing – review & editing. T. C.: conceptualization, methodology, analysis, investigation, writing – review & editing.

Conflicts of interest

There are no conflicts to declare.

Acknowledgements

The research is funded by Research Foundation – Flanders (FWO) – SBO project Terbium Isotopes for Medical Applications in Flanders (Tb – IRMA – V, grant number S005019N). A. M. acknowledges Research Foundation – Flanders (FWO) for funding (fellowship no. 1228622N). The beamline team at Elettra Sincrotrone Trieste are acknowledged for granting the possibility to conduct XAS studies (with proposal number 20210288). M. Ö. A. is grateful to Robin Aerts, Wout Van Craenendonck and Dr Peter Zsabka for the support in XAS and gamma spectrometry measurements. Stijn Rameakers and Dr Michiel Van de Voorde are acknowledged for the production of radiotracers used in this work.

Notes and references

1. CERN Courier, Terbium: a new ‘Swiss army knife’ for nuclear medicine, https://cerncourier.com/a/terbium-a-new-swiss-army-knife-for-nuclear-medicine/, accessed 16 March 2020.
2. 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.
3. S. Lehenberger, C. Barkhausen, S. Cohrs, E. Fischer, J. Grünberg, A. Hohn, U. Köster, R. Schibli, A. Türler and K. Zhernosekov, Nucl. Med. Biol., 2011, 38, 917–924.
4. A. Al-Ibraheem and A. M. Scott, Nucl. Med. Mol. Imaging, 2023, 57, 168–171.
5. J. Grünberg, D. Lindenblatt, H. Dorrer, S. Cohrs, K. Zhernosekov, U. Köster, A. Türler, E. Fischer and R. Schibli, Eur. J. Nucl. Med. Mol. Imaging, 2014, 41, 1907–1915.
6. C. Müller, J. Reber, S. Haller, H. Dorrer, P. Bernhardt, K. Zhernosekov, A. Türler and R. Schibli, Eur. J. Nucl. Med. Mol. Imaging, 2014, 41, 476–485.
7. 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.
8. R. P. Baum, A. Singh, H. R. Kulkarni, P. Bernhardt, T. Rydén, C. Schuchardt, N. Gracheva, P. V Grundler, U. Köster, D. Müller, M. Pröhl, J. R. Zeevaart, R. Schibli, N. P. Van Der Meulen and C. Müller, J. Nucl. Med., 2021, 62, 1391–1397.
9. M. Van de Voorde, K. Van Hecke, T. Cardinaels and K. Binnemans, Coord. Chem. Rev., 2019, 382, 103–125.
10. E. J. Monroy-Guzman and F. Salinas, J. Mex. Chem. Soc., 2015, 59, 143–150.
11. N. Gracheva, C. Müller, Z. Talip, S. Heinitz, U. Köster, J. R. Zeevaart, A. Vögele, R. Schibli and N. P. Van Der Meulen, EJNMMI Radiopharm. Chem., 2019, 4, 12.
12. F. Rösch, Radiochim. Acta, 2007, 95, 303–311.
13. R. D. Teo, J. Termini and H. B. Gray, J. Med. Chem., 2016, 59, 6012–6024.
14. Z. Talip, C. Favaretto, S. Geistlich and N. P. Van Der Meulen, Molecules, 2020, 25, 966.
15. K. Mishiro, H. Hanaoka, A. Yamaguchi and K. Ogawa, Coord. Chem. Rev., 2019, 383, 104–131.
16. T. I. Kostelnik and C. Orvig, Chem. Rev., 2019, 119, 902–956.
17. J. M. Schwantes, R. Sudowe, H. Nitsche and D. C. Hoffman, J. Radioanal. Nucl. Chem., 2008, 276, 543–548.
18. S. A. Sayed, K. A. Rabie and I. E. Salama, Sep. Purif. Technol., 2005, 46, 145–154.
19. M. Van de Voorde, K. Van Hecke, K. Binnemans and T. Cardinaels, Sep. Purif. Technol., 2020, 232, 115939.
20. D. E. Hobart, K. Samhoun, J. P. Young, V. E. Norvell, G. Mamantov and J. R. Peterson, Inorg. Nucl. Chem. Lett., 1980, 16, 321–328.
21. P. G. Varlashkin, G. M. Begun and J. R. Peterson, J. Less-Common Met., 1985, 109, 123–134.
22. S. Qiang, W. Zhijian and G. Chongli, Solvent Extr. Ion Exch., 1995, 13, 275–288.
23. M. O. Arman, B. Geboes, K. Van Hecke, K. Binnemans and T. Cardinaels, J. Appl. Electrochem., 2022, 52, 583–593.
24. M. Gras, N. Papaiconomou, E. Chainet, F. Tedjar and I. Billard, Sep. Purif. Technol., 2017, 178, 169–177.
25. M. Atanassova, Solvent Extr. Ion Exch., 2009, 27, 159–171.
26. M. Krättli, T. Müller-Späth, N. Ulmer, G. Ströhlein and M. Morbidelli, Ind. Eng. Chem. Res., 2013, 52, 8880–8886.
27. K. L. Nash, Handb. Phys. Chem. Rare Earths, 1994, 18, 197–238.
28. K. Shimojo, N. Aoyagi, T. Saito, H. Okamura, F. Kubota, M. Goto and H. Naganawa, Anal. Sci., 2014, 30, 263–269.
29. N. Sivaraman, R. Kumar, S. Subramaniam and P. R. Vasudeva Rao, J. Radioanal. Nucl. Chem., 2002, 252, 491–495.
30. B. Gupta, P. Malik and A. Deep, Solvent Extr. Ion Exch., 2007, 21, 239–258.
31. V. V Smirnov, M. D. Karavan, M. V Logunov, I. G. Tananaev and B. F. Myasoedov, Radiochemistry, 2018, 60, 404–419.
32. E. V. Kuzovkina, E. A. Lavrinovich, A. P. Novikov, E. S. Stepanova, M. D. Karavan and I. V. Smirnov, J. Radioanal. Nucl. Chem., 2017, 311, 1983–1989.
33. I. V. Smirnov, E. S. Stepanova, N. M. Ivenskaya, M. D. Karavan, S. R. Zaripov, S. R. Kleshnina, S. E. Solovieva and I. S. Antipin, J. Radioanal. Nucl. Chem., 2017, 317, 1257–1265.
34. Z. K. Karalova, T. V Bukina, E. A. Devirts, Z. Z. Agaev and B. F. Myasoedov, Radiokhimiya, 1987, 29, 767–772.
35. Z. K. Karalova, E. A. Lavrinovich, B. F. Myasoedov, L. A. Fedorov and S. A. Sokolovskii, Radiokhimiya, 1988, 31, 38–45.
36. E. V Kuzovkina, E. A. Lavrinovich, A. P. Novikov, E. S. Stepanova, M. D. Karavan and I. V Smirnov, J. Radioanal. Nucl. Chem., 2018, 315, 639–642.
37. A. V. Boyarintsev, G. V. Kostikova, S. I. Stepanov, A. A. Shoustikov, A. M. Chekmarev and A. Y. Tsivadze, Solvent Extr. Ion Exch., 2021, 39, 745–763.
38. W. E. Clifford, E. P. Bullwinkel, L. A. McClaine and P. Noble, J. Am. Chem. Soc., 1958, 80, 2959–2961.
39. A. V. Boyarintsev, S. A. Perevalov, S. I. Stepanov and A. M. Chekmarev, J. Radioanal. Nucl. Chem., 2020, 327, 385–393.
40. Z. K. Karalova, B. F. Myasoedov, T. V Bukina and L. M. Rodionova, Radiokhimiya, 1985, 27, 47–49.
41. Z. K. Karalova, B. F. Myasoedov, L. M. Rodionova and S. A. Kuznetsova, Radiokhimiya, 1981, 25, 187–191.
42. H. S. Sherry and J. A. Marinsky, Inorg. Chem., 1964, 3, 330–335.
43. H. S. Sherry and J. A. Marinsky, Inorg. Chem., 1963, 2, 957–959.
44. A. Di Cicco, G. Aquilanti, M. Minicucci, E. Principi, N. Novello, A. Cognigni and L. Olivi, J. Phys.: Conf. Ser., 2009, 190, 012043.
45. Y. Joly, Phys. Rev. B: Condens. Matter Mater. Phys., 2001, 63, 125120.
46. L. Hedin, B. I. Lundqvist and S. Lundqvist, Solid State Commun., 1971, 9, 537–541.
47. N. Saito and T. Sekine, Nature, 1957, 180, 753.
48. S. Misumi and T. Taketatsu, Bull. Chem. Soc. Jpn., 1958, 32, 876–879.
49. T. Taketatsu, Bull. Chem. Soc. Jpn., 1963, 36, 549–553.
50. T. Taketatsu, Talanta, 1963, 10, 1077–1082.
51. R. S. Gärtner and G. J. Witkamp, Sep. Sci. Technol., 2005, 40, 2391–2410.
52. K. Michiba, T. Tahara, I. Nakai, R. Miyawaki and S. Matsubara, Z. Kristallogr., 2011, 226, 518–530.
53. R. J. C. Dixey and P. J. Saines, Inorg. Chem., 2018, 57, 12543–12551.

---

Footnote

† Electronic supplementary information (ESI) available. See DOI: https://doi.org/10.1039/d4ra02694b

---