Preparation of 93Mo solution using proton irradiated Nb

93Mo is an important long-lived radionuclide in nuclear waste, and is required to be measured during the characterization of decommissioning waste. However, no commercial 93Mo solution is available to be calibrated and used as standard in the analysis of nuclear waste. This work presents a method for separation of 93Mo from Nb metal used in cyclotron as a target holder and irradiated with protons for long time. The separation of 93Mo from Nb matrix was implemented by combination of precipitation and chromatographic separation. The Nb matrix was first removed by precipitating oxides-hydroxides of Nb (e.g. Nb2O5) and then by Fe(OH)3 co-precipitation; Mo in the solution was purified using an alumina (Al2O3) column. A decontamination factor of ca. 105 was achieved for Nb. A pure carrier-free 93Mo solution was successfully prepared, and the 93Mo purity was verified by liquid scintillation spectrometry.


Introduction
Mo is produced by neutron activation of stable 92 Mo (14.65% of abundance) within the structural materials of nuclear reactors. With progress in nuclear decommissioning activities, determination of 93 Mo in nuclear wastes (especially decommissioning wastes) is required in the waste characterization and management. However, 93 Mo is still not well characterized and investigated; its half-life [(4.0 ± 0.8) × 10 3 years [1]] is reported with a relatively high (20%) uncertainty. 93 Mo decays by electron capture, and does not emit any γ-line that can be easily measured by γ-spectrometry. It is mostly detected by X-ray spectrometry through its X-rays of 16.521 keV, 16.615 keV, 18.607 keV and 18.623 keV (which are the K α2 , K α1 , K β3 and K β1 lines of decay product Nb, respectively) [2][3][4][5]. However, 93m Nb also emits these X-rays, in the same ratio as 93 Mo. Therefore, complete separation of 93 Mo from 93m Nb before measurement of 93 Mo is of paramount importance [5,6]. The Mo/Nb ratios in the Mo fraction separated from zircaloy hulls from a nuclear reactor were reported to be 100-1000 [6].
By detection of its Auger-electrons, 93 Mo can also be measured by liquid scintillation counting (LSC). The major advantages of LSC over X-ray spectrometry for measurement of 93 Mo are better availability, higher detection efficiency and lack of self-absorption. Nevertheless, the measurement of 93 Mo using LSC is scare; only very few reported LS spectra of 93 Mo are available [7,8].
In addition, a standard solution of 93 Mo is not commercially available, even a 93 Mo solution is not accessible. Therefore, the preparation of a high purity 93 Mo solution is an important issue in the determination and investigation of 93 Mo. This work aims to separate 93 Mo from a proton irradiated 93 Nb (natural Nb) target.
The similar chemical properties of Mo and Nb makes the separation of trace amount of 93 Mo from the proton irradiated Nb bulk a major challenge. Only a few methods were reported for separation of 93 Mo from niobium matrix. Sadeghi [9] reported a solvent extraction procedure for separation of Mo from NbO 2 target solution, and obtained a Mo purity of > 99.9%. Das [10] reported a method for separation of Mo from Nb, Zr and Y using solvent extraction; but this method is mainly dedicated to separate large excess of Mo from trace amounts of Nb, Zr and Y.
Both Mo and Nb readily form anionic species, which behave usually similarly on anion exchange chromatography, and are not retained on cation exchange resin [11][12][13][14]. Their distribution coefficients on anion exchange resin differ significantly only in 40% HF media [15] due to their different behavior in the formation of anion with fluoride. However, it was reported that the recovery of Mo was unsatisfying using a method based on this feature [16]. Shimada reported two methods to separate Mo from -among others -Nb using TEVA resin [5,17]. The major drawback of these methods is the unsatisfying decontamination of Nb in the Mo fraction.
Butement [18] separated Nb from Mo by co-precipitating Nb on Fe(OH) 3 . After adding Fe 3+ to the solution, he adjusted pH to 10 with NH 3 ·H 2 O. With this method, more than 90% of Nb was precipitated, and less than 0.5% Mo was observed in the precipitate. As several ions are co-precipitated on Fe(OH) 3 , similar methods are often used to purify Mo [19] (but no explanation was found in the literature). Bombard [2] applied such a method to separate Mo from Nb in 4 M NH 3 ·H 2 O medium. Although some (≈ 10%) losses of Mo were observed, the separation was successful, as a decontamination factor (DF) of 1.98 × 10 4 was reported for Nb after two precipitation-filtration-separation cycles. Puech [20] also used this method and supposedly reported a very similar DF (1.8 × 10 4 ) [21].
Alumina (Al 2 O 3 ) is well known to retain Mo (both as MoO 4 2− and Mo 7 O 24 6− ) in diluted HNO 3 (pH = 3.0 ± 0.5) [22,23], which is widely used in production of 99 Mo/ 99m Tcgenerators. However, no data about the behavior of Nb on alumina column, or about the behavior of Mo or Nb on alumina column in HF medium have been reported.
For separation of Mo from irradiated Nb, a high DF is needed to get rid of the large amount of Nb matrix (including 93m Nb, 94 Nb and 94m Nb), whereas no reported method meets this requirement. This work aims to develop an effective method for separation of Mo from Nb in order to prepare a pure 93 Mo solution from irradiated Nb target.

Reagents and materials
Analytical grade reagents were used.
Ca. 35 mg irradiated Nb was taken from a Nb target holder of a GE Healthcare PETtrace 800 cyclotron, which has been used and exposed to proton (ca. 14 MeV) for a long time; so 93 Mo was expected to be produced by the reaction of 93 Nb(p, n) 93 Mo.
Single element standard solutions (10 μg/mL) of Mo and Nb were purchased from LabKings, and used as carrier in the chemical separation as well as calibration standards for ICP-AES measurement. 99.8% Nb powder and pieces of 99.95% Mo wire (both purchased from Alfa Aesar) were used in the process of method development.
Alumina column was prepared by loading 2 g of neutral activated alumina (≤ 100 mesh, Brockmann), which was soaked in water, in a standard TrisKem column (7 mm diameter × 6 cm length). The active length of column was 50 mm. The prepared column was rinsed with 20 mL distilled water (to remove the finest particles) and conditioned with 10 mL 0.05 M HF (or another appropriate concentration, corresponding to the loading solution).

Model experiments (using stable Mo and Nb metals)
Typically 25-50 mg Nb and ca. 16 μg Mo were dissolved in the mixture of 1 mL 40% HF and 1 mL 68% HNO 3 . Solution was evaporated to dryness, the residue was dissolved with 0.5 mL 40% HF, which was evaporated to dryness again. 5 mL 6 M HF was added to the residue and the beaker was warmed slightly, until the residue was completely dissolved.
Distribution of Mo between alumina and HF acid was studied in both batch and column models. In the batch experiments, 1 g alumina was added to 10 mL HF solution of different concentration with the same amounts of Mo (16 μg). Suspensions were shaken for 2 min and kept for overnight (> 12 h). The suspension was centrifuged to separate the supernatant. Mo concentrations in the supernatants were measured.
In the chromatographic experiments, Mo and Nb were prepared in 0.02 M, 0.05 M, 0.1 M and 0.5 M HF solutions, which were loaded onto conditioned alumina columns. After rinsing the columns with 5 mL HF (whose concentrations were the same as those of the sample solutions) and then with 10 mL distilled water, Mo was eluted using 20 mL 1 M NH 3 ·H 2 O.

Radiochemical separation of Mo from Nb
Ca. 15 mg irradiated Nb was transferred to a Teflon beaker, and dissolved in the mixture of 1 mL 40% HF and 1 mL 68% HNO 3 . 30 μg Mo carrier (3 mL standard solution) was added, and the solution was evaporated to dryness. After adding 0.5 mL 40% HF, it was evaporated to dryness again. 5 mL 6 M HF was added, and the beaker was covered and warmed slightly until the residue was completely dissolved. The solution was transferred into a 50 mL plastic centrifuge tube and diluted with distilled water until 25 mL.
10 mL 25% NH 3 ·H 2 O was added to form white Nb 2 O 5 precipitate. The sample was shaken for 5 min, and then centrifuged (4 min, 3000 rpm). The supernatant was transferred into another tube and 5 mL solution of FeCl 3 (16 mg/ mL Fe) was added to form brown Fe(OH) 3 precipitate. The supernatant obtained after shaking and centrifugation was transferred into a Teflon beaker, evaporated to dryness 3 times after adding a mixture of ca. 1 mL 36% HCl and ca. 1 mL 68% HNO 3 , then 3 times after adding ca. 1 mL 40% HF. It was dissolved in 5 mL 0.1 M HF, and 5 mL distilled water was added.
The solution was loaded onto a conditioned alumina column, followed by rinsing with 10 mL 0.05 M HF and then with 10 mL distilled water. Mo was eluted with 10 mL 6.7 M NH 3 ·H 2 O, the eluate was evaporated gently to dryness, and the residue was dissolved in 5 mL distilled water. A 200 μL aliquot was taken and diluted with 3 mL 3% HNO 3 for the measurement of stable Mo and Nb by ICP-AES to calculate chemical recovery of Mo and DF of Nb. 20 mL Ultima Gold LLT cocktail was added to the remaining Mo fraction for the LSC measurement of 93 Mo. A schematic diagram of the separation procedure is presented in Fig. 1.
The other half of the irradiated Nb (ca. 15 mg) was processed following the same procedure, but no carrier (stable Mo) was added. As HF damages the glass pieces of the instrument, samples with high (> 0.5 M) concentration of HF were diluted using 3% HNO 3 . To solutions containing less (≤ 0.5 M) HF, 0.5 M H 3 BO 3 was added in stoichiometric quantity to complex the excess HF [24]. It has to be taken into account that boric acid increases the number of false signals in the 201.512 nm and 203.846 nm lines for Mo measurement. Components of chemicals used in the separation procedure also interfere with some of the lines, such as Fe with the 201.942 nm line of Nb or Al with the 281.615 nm line of Mo [25].

Measurements
LSC spectra were acquired using a Wallac Quantulus 1220 ultra low level liquid scintillation spectrometer, and the spectra analysis was implemented using EasyView software.

Removal of niobium by precipitation of Nb 2 O 5
In the model experiments, almost all initial Mo (≥ 98%) and (9 ± 6)% of initial Nb was found in the supernatant during the Nb 2 O 5 precipitation step. (91 ± 6)% of Nb was precipitated, which is firmly better, than that reported by Espartero [26] with only 50-60% being precipitated, and 46-92% reported by Osváth [27] using a similar procedure. These results indicate that precipitation of Nb 2 O 5 can be used as a pre-concentration step for removal of most of Nb, but the DF (≈ 10) is insufficient.

Removal of niobium by precipitation of Fe(OH) 3
The preliminary experiments showed that precipitation of Fe(OH) 3 in water (adding 25% NH 3 ·H 2 O dropwise until pH ≈ 8) caused both Nb and Mo being co-precipitated, and only 11-15% of Mo was found in the supernatant. On the contrary, when a much bigger volume of 25% NH 3 ·H 2 O was added and Fe(OH) 3 Fig. 1 Flowchart of the separation procedure evaporating it repeatedly after adding a mixture of 36% HCl and 68% HNO 3 , then evaporating it repeatedly after adding 40% HF, and finally re-dissolving the residue in diluted HF solution. This probably is a consequence of the chemical form of Mo: a possible ammine complex has to be decomposed; but further investigation is needed to clearly explain the phenomenon.

Loading Mo onto alumina column
The batch experiments (Table 1) show that Mo was quantitatively adsorbed on the alumina (> 99%) when HF concentration was lower than 0.05 M. When the HF concentration was increased to 0.1-0.5 M, some of Mo occurred in the supernatant.
The column experiments ( Table 2) show that both Mo and Nb were well retained on alumina, when the loading solution was prepared in 0.02-0.1 M HF (Nb was also well retained in 0.5 M HF).
It was also observed that most of Mo (> 95%) could be adsorbed on the alumina column when the solution was prepared in diluted HNO 3 (< 1 M). However, some small fractions of Mo (up to 7%) occurred sometimes in the effluent. Importantly, various fractions of Nb (up to 51%) also occurred in the effluent, indicating that Nb could not be well retained on the alumina column.
No leakage of Mo or Nb was observed when column was rinsed with water. However, both Mo and Nb (retained from HF) have leaked when column was rinsed with 0.1-1 M HNO 3 . Therefore 0.05 M HF was finally selected as loading and rinsing solution.

Eluting Mo from alumina column
For elution of Mo, 1 M NH 3 ·H 2 O is typically used [29,30]; but highest recovery (≥ 80%) was achieved using concentrated (25% or 13.4 M) NH 3 ·H 2 O [31]. Our results (Table 3) show that the more concentrated the NH 3 ·H 2 O eluent, the more efficient the elution of Mo. However, total recoveries of Mo are within 86-99% (being typically over 92%). Concentrations of Nb in the eluate were below detection limit in all the Mo fractions, indicating that the DF of Nb is higher than 10 4 .

Performance of the complete separation procedure
Two simulated solutions with high Nb and low Mo concentration were treated using the developed procedure. The results ( Table 4) show that more than 75% of Mo can be separated, and DFs of Nb are higher than 3x10 5 . The separation of 93 Mo from the irradiated Nb target holder sample shows a relatively low recovery of only 48%. The slightly lower DF of ≥ 8.9 × 10 4 for Nb is estimated, this is attributed to the detection limit of Nb in the eluate of Mo.

Measurement using LSC
LSC spectrum of the separated 93 Mo is presented in Fig. 2. Peak of 93 Mo was found in channels 25-325. This is in good agreement with the reported LSC spectra of 93 Mo [7,8]. Spectral quench parameter (SQP) of spectrum was 828. According to Roos [32], this is corresponding to (52 ± 5)% detection efficiency, which gives an activity of (26.4 ± 2.6) Bq.

Conclusions
A fast and reliable procedure was developed to separate Mo from irradiated Nb. The method is based on the combination of precipitation of Nb 2 O 5 , co-precipitation with Fe(OH) 3 as well as alumina column separation. Purity of separated Mo was checked by ICP-AES; Mo recovery was     Section, Hevesy Laboratory, DTU Nutech for their kind help during the experiments. The first author is particularly grateful for the assistance given by the staff of Hevesy Laboratory, DTU Nutech.
Open Access This article is distributed under the terms of the Creative Commons Attribution 4.0 International License (http://creat iveco mmons .org/licen ses/by/4.0/), which permits unrestricted use, distribution, and reproduction in any medium, provided you give appropriate credit to the original author(s) and the source, provide a link to the Creative Commons license, and indicate if changes were made.