Production and separation of 163Ho for nuclear physics experiments

This paper describes the production and chemical separation of the 163Ho isotope that will be used in several nuclear physics experiments aiming at measuring the neutrino mass as well as the neutron cross section of the 163Ho isotope. For this purpose, several batches of enriched 162Er have been irradiated at the Institut Laue-Langevin high flux reactor to finally produce 6 mg or 100 MBq of the desired 163Ho isotope. A portion of the Er/Ho mixture is then subjected to a sophisticated chemical separation involving ion exchange chromatography to isolate the Ho product from the Er target material. Before irradiation, a thorough analysis of the impurity content was performed and its implication on the produced nuclide inventory will be discussed.


Introduction
The radioactive isotope 163 Ho (t 1/2 = 4567 a [1,2]) has gained considerable attention within the physics community due to its very low Q-value for electron capture decay of only 2.8 keV [3,4]. This property and recent developments in high precision calorimetric measurements have facilitated several research projects devoted to measuring the mass of the neutrino. Among them, the HOLMES collaboration aims at implanting 3x10 5 Bq of isotopically pure 163 Ho a grid of 10 3 transition edge sensor micro-calorimeters to precisely measure the endpoint region of the energy spectrum of the 163 Ho electron capture decay [5]. With an envisaged statistical mass sensitivity around 1 eV, this measurement will provide an alternative technique to spectrometry to answer the long lasting question in physics about the neutrino mass [6].
The isotope 163 Ho is also an interesting nuclide in terms of nuclear physics research in the field of bound state beta decay. It has been experimentally observed that under fully ionized conditions (such as in stellar environments), the previously stable 163 Dy isotope becomes radioactive and decays to 163 Ho with a half-life of 47 d [7]. This circumstance significantly influences the s-process pathway in the A = 163 mass region, opening an additional branch a1111111111 a1111111111 a1111111111 a1111111111 a1111111111 towards isotopes accessible mainly via the p-process. Within this context, the neutron-capture cross section of the 163 Ho isotope is foreseen to be measured at the n_TOF CERN facility at thermal to stellar energies for the first time. While the Maxwellian average neutron capture cross-section (MACS) of 163 Ho has been measured in 1995 at stellar energies [8], a new measurement with several mg quantities of 163 Ho is envisaged to determine the neutron capture cross section in the eV to keV region typical for stellar environments [9].
In collaboration with the Institut Laue-Langevin (ILL), Grenoble, France, and the Paul Scherrer Institut (PSI), Villigen, Switzerland, the HOLMES project aims at producing roughly 100 MBq of 163 Ho by irradiation of enriched 162 Er material in a nuclear reactor. The production route 162 Er(n,γ) 163 Er followed by decay to 163 Ho is depicted in Fig 1 together with other neutron capture reactions. While neutron irradiation of 162 Er has been shown to be the most efficient way of producing the 163 Ho isotope [10,11], alternative routes via proton irradiation of nat Dy or enriched 164 Dy foils have also been reported in literature [1,12,13]. The HOLMES collaboration has chosen to pursue the former production method due to higher 163 Ho production yields [10] and better availability of irradiation resources, taking into account its unavoidable disadvantage.
The main drawback arising from neutron irradiation of enriched Er is the inevitable formation of 166m Ho (see Fig 1). This isotope has a half-life (t 1/2 = 1132 a [14]) comparable to that of 163 Ho and has a complex decay scheme, which drastically deteriorates calorimetric decay measurements of 163 Ho contaminated with 166m Ho [5]. Since the latter isotope cannot be separated from 163 Ho by chemical means, a mass separation of these two isotopes is mandatory before the implantation of Ho into a calorimetric measurement system [5]. The main production pathways of 166m Ho may be summarized as follows: 2. from the reaction 164 Er(n,γ) 165 Er(EC) towards 165 Ho and subsequent reaction 1; 3. from capture reactions of 164-x Dy(xn,γ) 165 Dy(β -) towards 165 Ho and subsequent reaction 1; 4. from neutron captures on 163 Ho to 164g+m Ho, that undergo either EC decay feeding reaction 3 or βdecay towards 164 Er feeding reaction 2.
The production rates of 166m Ho from 165 Ho and the isotopes 161 Dy, 162 Dy, 163 Dy and 164 Dy as function of their concentration relative to 164 Er for a 50 d irradiation with a thermal neutron flux of 10 15 n cm -2 s -1 are given in Fig 2. As can be seen, the dominant impurities transmuting into 166m Ho are 165 Ho and 164 Dy. While the main production pathway surely proceeds over the 164 Er(n,γ) reaction, an impurity content of 5% relative to 164 Er of 165 Ho or 164 Dy almost equally contributes to the 166m Ho production. Since the thermal neutron capture cross section of 166m Ho is in the order of 3 kb [15,16], its production and destruction rate will soon be in equilibrium at a neutron flux of 10 15 n cm -2 s -1 , resulting in a constant ratio of 165 Ho to 166m Ho after an irradiation time of 50 d (see also Fig 2 in [17]). Other production routes of 166m Ho, such as 169 Tm(n,α) 166m Ho or 166 Er(n,p) 166m Ho, are of minor importance due to the small reaction cross sections for these reactions [18,19] and the relatively well thermalized neutron flux during irradiation. For example, the fraction of epithermal and fast neutrons in the V4 beam tube of the ILL reactor is below 10%. Yet another production route of 165 Ho proceeds via double neutron capture on 163 Ho and largely depends on the involved thermal capture cross-sections of 163 Ho and 164g+m Ho. Up to date, only the former is known as it has been recently measured to be σ 163Ho = (156 + 23) b for the formation of 164g Ho and 164m Ho, respectively [11]. Neutron capture reactions on 163 Ho significantly influence the production rate and thus, the total amount of 163 Ho after neutron irradiation. In addition, capture reactions on 164g+m Ho yield again 165 Ho, a pathway that competes with both 164g Ho decay modes yielding 164 Er and 164 Dy (see hashed white lines in Fig 1). Thus, even for a chemically and isotopically pure 162 Er material, the production of 163 Ho in a high flux reactor is always accompanied by a number of parasitic reactions leading to 166m Ho.
In addition to 166m Ho, the irradiation will result in substantial amounts of 170 Tm and 171 Tm from capture reactions of 168 Er(n,γ) 169 Er(β -) 169 Tm(n,γ) 170 Tm and 170 Er(n,γ) 171 Er(β -) 171 Tm, respectively. Both of these isotopes are produced in the GBq range, representing the most significant hazard for handling the irradiated sample material. Moreover, any contamination of the final Ho product with kBq amounts of these isotopes will certainly deteriorate the sensitivity of 163 Ho neutrino mass measurements. Additionally, capture reactions on impurity 158 Dy will form 159 Dy (t 1/2 = 144 d). Thus, a sophisticated chemical separation of Ho not only from massive amounts of Er, but also from Tm and Dy has to be accomplished to assure a radiochemically clean product. A decrease in the production of 166m Ho is achieved by choosing high enrichment grades of the 162 Er material (with low 164 Er content) and substantially reducing the amount of impurities (Ho, Dy) present in the initial material [21]. In order to quantify the amount of 163 Ho and parasitic 166m Ho formed during a reactor irradiation, a careful characterization of the initial Er material including isotope composition and impurity content should thus be performed.
Ultimately, the recovery of irradiated and purified 162 Er is an additionally anticipated task. Due to the absence of any Er isotopes having half-lives exceeding 9.4 d ( 169 Er), the recycled 162 Er material is essentially free from any radioactivity after a cool-down period of 1 year and might be reused for new irradiations. This is especially desirable considering the price of enriched 162 Er (> 100 $/mg). Such recycled material will moreover be depleted in 167 Er, the strongest neutron absorber among all stable Er isotopes, which results in self-shielding and neutron flux depression during the first irradiation.
A detailed description of a possible separation procedure has been very recently given in [11]. The authors of this work suggest a pre-purification of the starting 162 Er material in order to mitigate impurities lighter than Er. This approach maximizes the isotopic purity of 163 Ho, but it is experimentally proven that the final Ho product will always result in a mixture of 163 Ho, 165 Ho and 166m Ho after irradiation. Apart from the described purification, the authors provide a very thorough analysis of the final 163 Ho material yielding 1.2x10 18 atoms of 163 Ho, 6.3x10 17 atoms of 165 Ho and 7 kBq (3.8x10 14 atoms) of 166m Ho from irradiating 30 mg of 20.4% enriched 162 Er for 54 d in the ILL high flux reactor.
In parallel to results published in [11], this work describes our efforts in the frame of the HOLMES project to define and tune the production and separation process of 163 Ho using two test batches of 162 Er. We also report on the analysis of a final 470 mg batch of 162 Er purchased in 2016 and give results on the separation of 163 Ho from irradiated test batches of 2014 and 2015. In contrast to the approach presented in [11], the chemical separation procedure was based on a two-step process involving cation exchange and extraction chromatography, which has already been successfully applied for the separation of neighbouring lanthanides (see [22] for more details). It is experimentally shown that recycling of large quantities of enriched 162 Er is feasible and might be continuously performed to satisfy needs of 163 Ho for HOLMES. Finally, an analysis of the purified material from both test batches representing in total 1.5 mg (or 5x10 18 atoms) of 163 Ho will be given. No analysis of the purified 163 Ho from the irradiated final 470 mg batch can be given since this material still awaits chemical separation.

Materials
Three different batches of 162 Table 1) was purchased from TraceSciences, Canada. This material, subsequently denoted as batch III, was not pre-purified and irradiated as delivered for 44 d. Each oxide from the described batches was provided to ILL in Suprasil 300 high purity quartz ampoules (Heraeus, Germany). Approximately 4 mg of batch III was kept for quantitative analysis of the material using ICP-OES, ICP-MS and neutron activation analysis (NAA). Another 400 μL of the stock solution was evaporated in a PE vessel and irradiated for 10 4 s in the neutron activation facility of the PSI spallation neutron source (SINQ). 3.9 mg of IRMM-527 (Al-0.1%Co) alloy, Sigma Aldrich, USA, was used as flux monitor. Gamma spectra of the 162 Er sample were recorded 5h, 24h and 4d after end of irradiation. Then the sample was irradiated with neutrons for another 10 4 s and immediately underwent chemical separation by ion exchange chromatography as described below. The separation was monitored by γ-spectrometry using a coaxial p-type HPGe detector, Mirion Technologies, USA. The Genie2000 software, Canberra, was used to evaluate the recorded spectra. Each sample was measured for at least 10 min to provide a statistical uncertainty below 1% for the peak area of interest. The separated fractions of Dy, Er and Yb were then measured for their isotopic composition using a sector-field based mass spectrometer Element 2, Thermo Fischer Scientific, Germany. The ICP mass spectrometer was operated in low resolution mode and wet plasma conditions. All measurement solutions were prepared from high purity nitric acid and water in polymer vessels.

Material separation
For the separation of 163 Ho from irradiated 162 Er of each test batch I and II, the ampoule containing the oxide was cracked, the material dissolved in 5 mL of 7 M HNO 3 and the pH adjusted to 4 with approximately 10 mL NH 4 OH. Approx. 0.5 mg of 162 Er activated at the PSI SINQ facility to yield 171 Er (t 1/2 = 7.52 h) was added prior to each separation to be able to monitor the elution of Er using γ-spectrometry. The solution was then loaded onto a column (L = 23 cm, d = 1 cm) containing 19 g of the cation exchange resin Aminex HPX87H from BioRad Laboratories, USA. Batch I containing 20.6 mg of 162 Er 2 O 3 was entirely loaded on the column, while the solution of batch II containing 119.5 mg of 162 Er 2 O 3 was divided in half to allow two independent separations. The elution of the lanthanides was performed at room temperature using increasing concentrations of α-hydroxy-isobutyric acid (HIBA) in a similar way as described in [22] via a peristaltic pump ISM834C (Ismatec, Switzerland). The gradual elution of the lanthanides was performed in 10 mL steps using a flow rate of 0.66 mL/min and was monitored by γ-spectrometry. All fractions containing chemically pure Ho were unified, the resulting 30 mL of solution acidified with 6 mL 1 M HNO 3 and loaded on a column (L = 16 m, d = 1 cm) containing 4.7 g of LN1 resin (Triskem, France) for final purification from HIBA and residual contaminants using increasing concentrations of HNO 3 . Fractions containing Er were unified to yield 70 mL of solution, acidified with 15 mL 1 M HNO 3 and similarly loaded on LN1 resin for HIBA stripping. After the elution of Er with 4 M HNO 3 , the acid is evaporated and the residue burned to the oxide in a quartz ampule (Heraeus, Germany) using a flame torch. The total recovery of recycled 162 Er is determined gravimetrically using an AE 100 balance, Mettler-Toledo, Switzerland.
The purified Ho fractions were unified, evaporated to dryness and redissolved in 10 mL of 1 M HNO 3 . This solution was subsequently investigated with ICP-MS and γ-spectrometry to determine the content of 163 Ho, 165 Ho and 166m Ho. The concentration of 163 Ho and 165 Ho was measured by dilution series using a Ho standard solution, Merck, Germany, as reference. All uncertainties are reported according to the "guide to the expression of uncertainty in measurement" (GUM) with a coverage factor of k = 1.

Material composition
The results of the ICP-MS, ICP-OES and neutron activation analysis of batch III is given in Table 1 together with the data from the original Certificate of Analysis provided by the supplier. The results obtained from all three analytical methods agree well within the statistical uncertainties of the measurement. The only significant deviation was found between results obtained by NAA and ICP-MS on Dy and Yb, since both of these impurity elements were found to have a non-natural composition.
The determined isotopic composition of Er tends to agree with the numbers provided by the supplier. It should be noted that the impurity content given by the supplier partially deviates from what was found by the analysis at PSI. This fact is due to isobaric interferences arising from Dy and Yb impurities, which are present in the material in the ‰ range. During the isotope enrichment process, not only the main product A Z N is enriched, but also its isobaric impurities such as A ZÀ 1 N or A Zþ1 N. Thus, in isotopically enriched samples of 162 Er, the Dy impurity gets enriched in 162 Dy. This fact has been confirmed by ICP-MS measurements of the Dy isotopic composition given in Table 2. Apart from much lower Gd content, the analysed sample also contained a significant amount (1530 ppm) of Yb enriched in 176 Yb, which has not been reported in the original analysis provided by the supplier. The Yb impurity is believed to originate from previous enrichment runs of 176 Yb operated at the same enrichment facility.
From the relative abundance of 164 Er and the concentration of the Dy and Ho impurities in the initial batch III material it is possible to deduce their expected contribution towards the 166m Ho production according to Fig 2. As calculations show, roughly 80% of parasitic 166m Ho will be formed from 164 Er, while 8% is expected to be produced from 164 Dy. The initial content of 235 ppm Ho as well as 163 Dy and 162 Dy almost equally contribute with 4% to the total 166m Ho inventory. Assuming a 163 Ho burn-up as given in [11], the expected 163 Ho: 166m Ho atomic ratio will be in the order of 3x10 3 according to calculations.

Ho separation
The separation profile of the batch II material irradiated for 53 d in 2015 is given in Fig 3. As it can be seen from Fig 3, a large number of radioisotopes is produced during the irradiation for 53 d at the ILL high flux reactor. Most elements may be efficiently separated on a cation exchange resin from the Ho product, although some overlap with Er and Dy was observed. This is due to the fact of high column loading that leads to inferior separation of the lanthanides [23]. 170/171 Tm, 154 Eu, 65 Zn and 60 Co, representing isotopes with the highest contribution to activity and dose rate, are efficiently separated. The absence of their γ-lines in the final Ho product yields decontamination factors exceeding 10 6 . The additional purification of the Ho product on the LN1 resin even further decreased the impurity contribution from Er and Dy. The corresponding separation profile is given in Fig 4. The final mass of isolated 163 Ho, the measured activity of 166m Ho and the impurity content originating from the batches I and II are given in Table 3. The cumulative amount of 163 Ho produced during these runs is approximately 1.5 mg, which represents roughly 27 MBq or 5x10 18 atoms of this isotope. The contribution of signals at m/q = {161-170} relative to m/ q = 163 is assigned to the sum of the respective isobars. Conservatively assuming that all signals at m/q = 166 are originating from 166 Er, the overall separation factor Ho/Er is calculated to be at least 10 5 and 3x10 4 for batches I and II, respectively. The cumulative separation factor for Ho/Dy after both chromatographic separations is difficult to assess, since signals at m/q = {162-164} have isobaric interferences from 162 Er, 163 Ho and 164 Er. The only Dy isotope free from isobaric interferences is 161 Dy, of which the abundance, however, was not measured. A more sophisticated analysis of the final product using resonant ionization mass spectrometry in connection with NAA as described in [11] would resolve the issue. In any case it can be stated with certainty that both 163 Ho samples processed so far contain less than 1% of (Dy+Er). This indicates that the separation procedure presented in this work yields 163 Ho of satisfactory quality similar to results obtained in [11].
Regarding the overall separation yield a clear dependence on the total mass of 162 Er is noteworthy. While for a separation of 163 Ho from 20.6 mg Er 2 O 3 (batch I) a total yield of 98.4% was achieved, the yield dropped down to 79.2% in case of the batch II material where 119.5 mg of Er 2 O 3 were processed. This is due to some unwanted overlapping of Er/Ho/Dy elution peaks as seen in Fig 3 as a result of higher column loadings with respect to the previous batch. In order to mitigate this problem in future, it is advisable to partition the column separation step into several runs. The recovery of 162 Er was successful in every separation and more than 90% of the material could be recycled and used for new irradiations.
As can be seen from numbers given in Table 3, the production of 163 Ho scales linearly with the amount of 162 Er irradiated for %50 days at the ILL reactor. Due to higher 162 Er enrichment of batch I, the total 163/165 Ho atomic ratio is consequently higher. Enrichments above 90% of 162 Er or much shorter irradiation times would be needed in order to make this ratio exceed 6. It is interesting to note that according to the measured data given in Table 3, the atomic ratio of 165 Ho to 166m Ho is constant through batch I and II and equals roughly 1.6x10 3 . This is due to a dynamic equilibrium reached after 50 d of irradiation in the production and destruction pathways leading to 165 Ho and 166m Ho, as has been mentioned earlier.
The 544.2 mg of enriched 162 Er 2 O 3 of batch III as well as 110.1 mg of recovered 162 Er 2 O 3 from batch II were irradiated at the high flux reactor at ILL in 2017 and will undergo chemical separation soon. The separated material is intended for neutron cross section measurements at the CERN n_TOF facility prior to final delivery to the HOLMES isotope separator. With a total amount of 6 mg 163 Ho, this new irradiated of batch III will fully cover the needs of this isotope within the frame of the HOLMES project [5].

Conclusions
We have successfully separated 1.5 mg of 163 Ho, corresponding to 27 MBq, from macro amounts of enriched 162 Er. This amount of 163 Ho is at least one order of magnitude higher than can be reasonably achieved by any other production route involving e.g. proton irradiations of Dy targets [10]. It was also shown that a recycling of 162 Er is easily achievable to meet with 163 Ho requirements for nuclear physics experiments devoted to measure the mass of the neutrino and to deduce the 163 Ho neutron capture cross section at stellar neutron energies. However, this production route comes with the disadvantage of parasitic formation of 166m Ho, which cannot be separated from the wanted 163 Ho by chemical means. A sophisticated, highly efficient and selective mass separation has to be performed for reliable removal of the isotopic contaminant in order to allow clean calorimetric measurements of the 163 Ho electron capture decay. A mass separation involving laser resonant ionization has already been successfully proven to reduce the 166m Ho amount by three orders of magnitude [21].