Purification of cerium, neodymium and gadolinium for low background experiments

Cerium, neodymium and gadolinium contain double beta active isotopes. The most interesting are Nd and Gd (promising for 02 search), Ce (2 candidate with one of the highest Q2). The main problem of compounds containing lanthanide elements is their high radioactive contamination by uranium, radium, actinium and thorium. The new generation 2 experiments require development of methods for a deep purification of lanthanides from the radioactive elements. A combination of physical and chemical methods was applied to purify cerium, neodymium and gadolinium. Liquid-liquid extraction technique was used to remove traces of Th and U from neodymium, gadolinium and for purification of cerium from Th, U, Ra and K. Co-precipitation and recrystallization methods were utilized for further reduction of the impurities. The radioactive contamination of the samples before and after the purification was tested by using ultra-lowbackground HPGe gamma spectrometry. As a result of the purification procedure the radioactive contamination of gadolinium oxide (a similar purification efficiency was reached also with cerium and neodymium oxides) was decreased from 0.12 Bq/kg to 0.007 Bq/kg in Th, from 0.04 Bq/kg to <0.006 Bq/kg in Ra, and from 0.9 Bq/kg to 0.04 Bq/kg in K. The purification methods are much less efficient for chemically very similar radioactive elements like actinium, lanthanum and lutetium.


Introduction
The study of the double beta (2) decay gives important information about the properties of neutrino and of weak interactions [1,2]. There are many potentially 2 active isotopes among the lanthanide elements [3] (cerium, neodymium, samarium, gadolinium, dysprosium, erbium and ytterbium, see Table 1), the most interesting are 150 Nd and 160 Gd (promising to search for neutrinoless 2 decay), 136 Ce (one of the highest Q 2 among 2 + candidates), 156 Dy and 164 Er (possible resonant 02 process).
The isotope 150 Nd is one of the most favourable nuclides to search for 02 decay thanks to a high released energy (Q 2 = 3371 keV), a considerably high isotopic abundance (5.6%) and promising theoretical estimations (see [2] and references therein). The isotope 136 Ce is one of the most interesting 2β + active nuclei with the theoretical predictions for the two neutrino 2 decay half-lives on the level of 10 18 -10 22 yr [3,4,5,6]. The isotope 160 Gd is interesting object for double beta decay experiments thanks to a high isotopic abundance (21.9%) and promising theoretical predictions for the 02channels [7,8], while 152 Gd is considered as a perspective nuclide to search for resonant neutrinoless double electron capture with a predicted half-life on the level of 8×10 23 -8×10 26 yr for an effective neutrino mass of 1 eV [9,10].
Both scintillation [7] and HPGe [11,12,13] detectors can be used to search for 2 decays of lanthanides. The main problem of compounds containing lanthanide elements is their high radioactive contamination by uranium, radium, actinium and thorium, as well as by radioactive lanthanides containing radioactive isotopes ( 147 Sm, 138  Here we present results of R&D to purify cerium, gadolinium and neodymium oxides for 2 experiments.

Purification method
In this study we have used CeO 2 and Gd 2 O 3 powders provided by Stanford Materials Corporation (USA), while the high purity neodymium oxide (Nd 2 O 3 ) was produced by Soviet Union industry in the seventies.
The samples were purified with a combination of physical and chemical methods in several steps. At first the rare earths' oxides were dissolved in high purity mineral acids. Then obtained acidic aqueous solutions were used for the main purification procedures: liquidliquid extraction and re-extraction. The liquid-liquid extraction (solvent extraction) is a method by which a compound is pulled from a solvent A to a solvent B, while the solvents A and B are not miscible (usually water and an organic solvent). Two immiscible liquids are mixed together, then the solutes with higher polarity dissolve in the solvent with a higher polarity, while the solutes with lower polarity solve in a solvent with lower polarity. This technique is widely utilised in chemical laboratories by using separatory funnels (see Fig. 1).
The solutes of lanthanide oxides were dissolved in water (phase A), while phosphor-organic complexing compound trioctylphosphine oxide (TOPO) in toluene was used as solvent B. At these conditions elements with a higher oxidation move to the organic phase with a higher distribution level rather than elements with a lower oxidation. Such a procedure allows to separate elements with different oxidation states [18]. Hydroxides of the purified elements were precipitated from aqueous solutions by neutralization using gaseous ammonia.

Phase B
Phase A Finally cerium, neodymium and gadolinium oxides have been obtained by decomposition of the hydroxides at high temperature.

Dissolving of Gd 2 O 3 and Nd 2 O 3
Nd 2 O 3 and Gd 2 O 3 oxides were dissolved in hydrochloric acid solution. Amounts of concentrated hydrochloric acid (of the super pure quality grade) and demineralized water (18.2 MΩ•cm) were calculated to obtain a solution with concentration of LnCl 3 20% and pH  0.1.
Weighted powders of gadolinium and neodymium compounds were put into Teflon beaker. The hydrochloric acid solution was added very slowly to avoid boiling point due to exothermic reaction: were Ln is Nd or Gd. 04001-p.2

Extraction
In case of gadolinium and neodymium extractions, the solvent A was 20% acidic solution of gadolinium or neodymium chloride. We have used TOPO with concentration 0.1 mol/L as the solvent B. Equal volumes of two phases were shacked in Teflon separatory funnel for one minute. In 0.5 hour the solutions were completely stratified. Extraction of thorium and uranium have been occurred by the reaction: Then purified NdCl 3 and GdCl 3 solutions were separated and rinsed with toluene to reduce TOPO content in aqueous phase.

Precipitation
Further purification and separation were carried out by precipitation of the hydroxides by increasing of the pH level up to 7 [19]. We have used gaseous ammonia to neutralize the acidic media: Amorphous hydroxides of gadolinium and neodymium were obtained.

Dissolving
CeO 2 powder with mass of 500 g was taken for purification. A mixture of concentrated nitric and hydrofluoric acids was used to dissolve CeO 2 : The reaction occurred only after heating up to 80-90 o C over at least 24 hours for complete dissolution of cerium oxide.
Some part of the cerium sample has been lost due to formation of insoluble cerium fluoride. Initial amounts of CeO 2 and HNO 3 were calculated in order to obtain a solution with a 10% concentration of Ce(NO 3 ) 4 and 5 mol/L of nitric acid.

Extraction
In case of cerium oxide, the solvent A was a solution of Ce(NO 3 ) 4 with a high content of nitric acid while the solvent B was 33% solution of TOPO in toluene. Stirring of the solutions leads to extraction of cerium to organic liquid together with thorium and uranium traces: However the efficiency of cerium extraction was very low due to the lack of TOPO molar concentration in the solvent B (~ 1.03 mol/L).

Re-extraction
Re-extraction of cerium from the organic phase was performed into low acidic water solution with simultaneous decreasing of the Ce oxidation level. Hydrogen peroxide was utilized as a reducing agent: Since solubility for Ce 3+ in organic phase is less than for Ce 4+ , cerium concentration in aqueous phase was increased. However, a substantial part of cerium left in organic phase.

Precipitation
The same technique as described in section 3.3 was applied for cerium sedimentation. Gaseous ammonia was passed through the Ce(NO 3 ) 3 aqueous solution with an aim to neutralize the acidity of the solution. At the same time the oxidation level of cerium is increased due to the excess of hydrogen peroxide after the re-extraction procedure, which leads to amorphous hydroxide production:

Oxides recovering
Stoichiometric oxides CeO 2 , Nd 2 O 3 and Gd 2 O 3 were reduced from the cerium, neodymium and gadolinium hydroxides at high temperature. The amorphous hydroxides sediments were rinsed several times with ultrapure water and placed into quartz beakers for drying and annealing. The backers were put into muffle, heat up to 900 o C and kept for 4 hours. The outputs of the purified oxides were ~ 90% for Nd 2 O 3 , Gd 2 O 3 and ~ 20% for CeO 2 .

Measurements of radioactive contamination
The radioactive contamination of the samples before and after the purification was tested by using ultra-lowbackground HPGe gamma spectrometry deep underground (3600 m w.e.) at the STELLA facility in the Gran Sasso National Laboratories of the INFN (Italy). The detectors, GeBer (244 cm 3 ) in case of CeO 2 , Gd 2 O 3 and GePaolo (518 cm 3 ) for Nd 2 O 3 , were located inside a passive shield made of low radioactivity copper (10 cm), lead (20 cm) and borated polyethylene (10 cm). The setups were flushed by highly pure nitrogen to remove radon. The energy resolution of the spectrometers are about 2 keV for 1333 keV  quanta of 60 Co. The samples were placed directly on the end cap of the detectors. Detection efficiency has been calculated with the help of the GEANT4 code [20].
Radioactive contamination of the samples before and after purification is presented in Table 2. 04001-p.3 Table 2. Radioactive contamination of lanthanide samples before and after purification (mBq/kg). Upper limits are given with 95% C.L., while uncertainties are presented with 68% C.L. The masses of the samples and the time of measurements are also given.

Conclusions
The liquid-liquid extraction technique was used to purify CeO 2 , Gd 2 O 3 and Nd 2 O 3 . The radioactive contamination of the samples before and after the purification was tested by using ultra-low-background HPGe gamma spectrometry at the Gran Sasso National Laboratories of the INFN (Italy). Gadolinium oxide has been purified most effectively: radioactive contamination was decreased to <0.004 Bq/kg in 228 Th, to <0.008 Bq/kg in 226 Ra, and to <0.04 Bq/kg in 40 K. The purification methods are much less efficient for chemically very similar radioactive elements like actinium, lanthanum and lutetium. Further R&D of purification methods and preparation of experiments to search for 2 decay of several lanthanide isotopes is in progress.