Investigation of chromatographic procedures for the analysis of cationic impurities in uranium and plutonium matrices by ICP-OES and ICP-MS

The goal of this study is to evaluate and optimize chromatographic separation procedures for the analysis of cationic impurities in uranium and plutonium matrices using ICP-OES and ICP-MS. Four separation protocols were tested using UTEVA and TEVA resins with various eluents. The protocols successfully retained the U-Pu matrix on the resin while allowing quantitative recovery of 29 cationic elements.

Based on performance and practicality, two UTEVA-based protocols using HNO₃ with oxalic acid or hydrogen peroxide eluents were selected and validated on a uranium oxide reference material. These methods enabled accurate determination of 23 impurities with good recoveries. The results demonstrate the potential to analyze up to 41 impurities in U-Pu materials, supporting quality control in nuclear material purity with detection limits at the μg g⁻¹ level.

The original article

Investigation of chromatographic procedures for the analysis of cationic impurities in uranium and plutonium matrices by ICP-OES and ICP-MS

Marion Hernandez, Alexandre Quemet, Laure Montreuil, Christophe Maillard, Sarah Baghdadi 

Spectrochimica Acta Part B: Atomic Spectroscopy, Volume 225, 2025, 107136

https://doi.org/10.1016/j.sab.2025.107136

licensed under CC-BY 4.0

Selected sections from the article follow. Formats and hyperlinks were adapted from the original.

Throughout the nuclear fuel cycle, from mining to reprocessing, uranium undergoes several chemical, physical and isotopic transformations. The production of plutonium, a by-product of the nuclear reaction occurring in uranium fuel, is an integral aspect of the spent fuel reprocessing process. It is extracted from used fuels recycled as (U,Pu)O₂ mixed oxide (MOX) fuel. At each processing and reprocessing step, impurities can be introduced. They can affect the material properties such as neutron flux during the chain reaction and phase transition temperature [1]. Ensuring the concentration in metal impurities in the uranium and plutonium matrices is necessary for material conformity to the nuclear grade specifications. Thus, trace elemental analysis is important to characterize and certify materials, such as pellet before irradiation [2].

Cations analysis in uranium and plutonium matrices are commonly performed using spectrometric methods such as inductively coupled plasma – optical emission spectrometry (ICP-OES) and mass spectrometric methods such inductively coupled plasma – mass spectrometry (ICP-MS) [[3], [4], [5], [6], [7], [8]]. Each of these methods has its own advantages, limitations and type of interferences (i.e. spectral interferences for the ICP-OES or isobaric interferences for ICP-MS). The key point for the selection of ICP-MS (for concentration of ppb or lower level) and ICP-OES (for concentration of ppm or more high level) is the concentrations of the target elements. MS offers better detection limits, but there could be isobaric and polyatomic as well as double-charge ions interferences due to the elements present and plasma gas used. Using argon as a plasma gas at ambient pressure makes it difficult to analyze certain isotopes such as ⁴⁰Ca, ⁵⁶Fe and ²⁸Si, which are mainly interfered by ⁴⁰Ar, ⁴⁰Ar¹⁶O and ¹⁴N₂, respectively. Impurities measurement at trace or ultra-trace levels in the presence of uranium or plutonium at several g L⁻¹ levels is very challenging due to space charge effects: light ions are more unfocused than heavy ions [9]. Its effect result in a suppression of analyte signals with light masses by uranium and plutonium. Although the detection limits of ICP-OES are higher than those of ICP-MS, they remain satisfactory for the analysis of impurities. However, the presence of elements with a rich emission spectrum, such as uranium and plutonium, makes the determination of trace impurities by ICP-OES challenging due to spectral interferences [10,11].

Indeed, analyzing cations at trace and ultra-trace levels becomes difficult when a sample is concentrated over 200 mg L⁻¹ of uranium and/or plutonium (U-Pu). Matrix effects are generally reduced by sample dilution, resulting in an increase of the detection limit.

To determine elemental concentrations at μg g⁻¹ levels in U-Pu matrices, the detection limit has to be as low as possible. Such specifications can only be achieved by separating cations from the U-Pu matrix to analyze them separately. Several techniques have been developed to separate different elements from actinides prior to their determination, including precipitation, solvent extraction or chromatography [2,[12], [13], [14]]. Chemical precipitation and solvent extraction are time-consuming and involves using costly solvent. It generates a big volume of effluents, also.

Chromatographic separation is a method offering selectivity, ease, and speedy implementation, while also ensuring reliability. TEVA (TEtraValent Actinides) and UTEVA (Uranium and TEtraValent Actinides) resins are well known for their abilities to retain actinides [[15], [16], [17], [18]]. TEVA resin, contains an aliphatic quaternary amine, having an extracting role of tetravalent actinide ions and Tc over a wide range of nitric acid concentrations [19]. UTEVA resin, made of dipentyl pentylphosphonate, extracts tetravalent actinides and hexavalent uranium as nitrato complex. The nitrate concentrations impacts the formation of nitrato complexes and therefore the actinides retention [15]. An important point is these two resins only retain tetravalent actinides (i.e. Th, Np and Pu) and uranium in HNO₃ media. Elements with another valence than +IV (except for U) are not retained onto the resin. TEVA and UTEVA are routinely used in the laboratory to separate Am, U and Pu [20,21]. In this study, a single-column separation system was used with either TEVA or UTEVA resin to retain uranium and plutonium and let through cations. Using a nitric and hydrofluoric acids mixture as conditioning, feeding and/or scrubbing phases has been reported in several studies [7,14]. It has been proven efficient to separate cations from uranium. However, fluoride ions lead plutonium to its trivalent state, which can no longer be fixed onto the resins. To fix the whole plutonium onto the resin, hydrofluoric acid must be avoided. Concentrated nitric acid, 8 mol L⁻¹, was proven to be enough to separate cations impurities from uranium using UTEVA resin [22,23]. The ATSM standard test method describes a method for quantifying 25 elements in a Pu-based material [24]. In this standard test method, Pu-based material is dissolved and converted into a 8 mol L⁻¹ HNO₃ solution prior to its separation on an AG MP-1 resin. The AG MP-1 resin is made of a quaternary ammonium, like the TEVA resin. After loading the column with the 8 mol L⁻¹ HNO₃ sample solution, the column was scrubbed with a 8 mol L⁻¹ HNO₃/0.006 mol L⁻¹ HF acid mixture.

This paper presents separation protocols developed to collect cations and U-Pu separately for routine laboratory use using TEVA or UTEVA resins, combining ICP-OES and ICP-MS. The results obtained on a certified reference material (CRM) will be discussed in terms of detection limit and recovery.

2. Experimental

2.2. Instrumentation

Trace cations analyses by ICP-OES were performed using two Optima 8300 DV (Perkin Elmer, USA). In the absence of U-Pu, an Optima ICP-OES set up on a lab bench was used to analyze the solutions with greater speed and ease. For radionuclides analysis, an Optima 8300 DV (Perkin Elmer, USA) nuclearized in a glove box was used. The glove box holds the introduction system and an autosampler. Both instruments were equipped with a micro-nebulizer and a small volume gutter cyclonic nebulization chamber. Sc (5 mg L⁻¹) was used as internal standard for ICP-OES measurements. It helps monitoring matrices effects but no correction was applied.

For ICP-MS determination, an Elan DRCe (Perkin Elmer, USA) nuclearized in a glove box was used to operate in a radioactive environment. The glove box holds the introduction system and the autosampler. Samples were introduced into argon plasma through a Meinhard concentric nebulizer and a small volume baffled cyclonic spray chamber. An ASX-260 autosampler was used. The sample size is 3 mL minimum. ⁴⁵Sc, ¹³⁹La and ²⁰⁵Tl (1 μg L⁻¹) were used as ICP-MS internal standards. Corrections were applied (i.e. corrected intensity = measured intensity / internal standard intensity).

Single elements standards diluted in 0.5 mol L⁻¹ HNO₃ were used to prepare an external calibration curve of both spectrometric techniques. Calibration curves accuracy was checked using control solutions prepared independently. Indium is interfered at mass 113 and 115 by ¹¹³Cd and ¹¹⁵Sn, respectively. As the isotope composition is natural, isobaric interference corrections were used to quantify properly In [27].

3. Results and discussion

3.1. Protocols selection

Separation results for the 14 cations with protocols A, B, C and D are summarized in Table 3. Cations recoveries were between 80 and 120 % for the 14 cations (Al, B, Ca, Cd, Cr, Cu, Fe, Ga, Mn, Mo, Ni, Pb, Si and Zn) except for Ca in the protocol D (130 %). The RSD for the 14 cations were below 10 % for the protocols B and C. For protocol A and D, the RSD were also below 10 % except for B and Ca for when the RSD were between 13 and 32 %. It should be noticed the first series of separation were performed and analyzed before the second and third series. The results of the first series for Ca were incoherent due to a pollution and not taken into account. Thus, the results presented here for Ca are the average of the second and the third separations. These results showed that the 14 cations were not retained on the resin for protocols A, B, C and D. Separation protocols were found to be effective for these cations.

The four protocols have similar performances, although protocols B and C appear to have better repeatability. Indeed, the choice of protocols used for the rest of this study was based on practical considerations. Protocol A required two scrubbing steps, compared to only one for the other protocols, that are more time-consuming for routine analysis. Protocol C involves a valence adjustment of uranium and plutonium by hydrogen peroxide and have already proved efficient to fix it [21]. The protocol C is routinely applied in the laboratory to separate U and Pu. Thus it will help separating the cations, U and Pu on the same aliquot sample saving a considerable amount of time. Protocol D using nitric acid and a TEVA resin should not retain all the plutonium unless Pu valence adjustment is performed using the Fe(II)/NaNO₂ method [19]. Therefore, the Fe and Na analysis is not possible with protocol D. Protocols B and C seems to be more adapted to separate the cations from uranium and plutonium in routine work. Therefore, they are used for the rest of the study.

Then, 6 new cations (Ag, Mg, Na, Sn, Ti and V) were added to create the “multi 20c” solution. Using protocol B and C, the six cations recovery were between 85 and 110 % with a RSD between 1 and 3 % (Table 3). Finally, separation tests were performed on 9 new cations (multi9c solution). Using protocols B and C, the nine cations (Ce, La, Nd, Pd, Rh, Sb, Se, Sr and Y) were recovered between 95 and 110 % with a RSD between 2 and 10 %. The 29 cations were not retained onto the resin for protocol B and C.

3.4. Comparison of the detection limits

Detection limits were estimated for each separation method and compared to the ones obtained using the dilution method, which is the method currently used in the laboratory. Detection limits are summarized in Table 7.

Spectrochimica Acta Part B: Atomic Spectroscopy, Volume 225, 2025, 107136 - Table 7. Detection limits in mg kg−1 of U of the Morille CRM with “/” where elements cannot be analyzed with the protocol.

Detection limits obtained are between 0.05 and 3 mg kg⁻¹ of U using separation and between 0.05 and 35 mg kg⁻¹ of U for the sample dilution. Similar detection limits were obtained for 20 cations (Ag, Ba, Be, Bi, Cd, Cr, Dy, Eu, Gd, In, Li, Mo, Pb, Sm, Sn, Th, Ti, V, W and Zr). Separation helps obtaining better detection limits for Co (sample dilution: 0.5 mg kg⁻¹; separation: 0.05 mg kg⁻¹), Cu (sample dilution: 35 mg kg⁻¹; separation: 1 mg kg⁻¹) and Ni (sample dilution: 16 mg kg⁻¹; separation: 0.03 mg kg⁻¹).

In another study, better detection limits were obtained: LOD as low as 0.001 mg kg⁻¹ [14] and LOD as low as 0.05 mg kg⁻¹ in this study. The factor 50 on the LOD can be explain by (1) the ICP-MS used in this study is nuclearized and of an old generation, that limits is sensitivity. Typical sensitivity for ¹¹⁵In is about 40,000 cps (μg L⁻¹)⁻¹ in our instrument whereas sensitivity upper than 1,000,000 cps (μg L⁻¹)⁻¹ can be obtained with new ICP-MS. (2) The blanks of the instrument used are not as clean as desired due to its long use with very different matrices. It is possible to achieve better detection limit by using a cleaner instrument. However, rinsing it during a long time is prohibited, as the effluent volume will be too high and is considered radioactive and needs to be recycled before evacuation. Using a cleaner and a more sensitive instrument will help to obtain detection limit comparable of the previous study, which used an ICP-MS on a lab bench [14].

Another way to decrease the detection limit is to reduce the sample dilution required to its analysis. As HNO₃ concentration has to be adjusted at 0.5 mol L⁻¹, elution phase (loading and washing) are diluted 6 and 8 time for protocols B and C before the ICP-MS analysis, respectively. The elution phase can be evaporated to dryness and dissolved again in the appropriate volume of 0.5 mol L⁻¹ HNO₃, which can help to reduce the detection limit by a factor of 10. It will be necessary to check that the cations have not been lost during the evaporation step before using this approach.

4. Conclusions

This study aims at developing chromatographic separation procedure to separate cations from U-Pu. The protocols were tested on simulated solutions of 29 cations in absence of U-Pu and 14 cations in a U-Pu matrix. Uranium and plutonium were fixed onto the resin using protocol B (UTEVA resin with HNO₃/H₂C₂O₄ mixture as eluent) and C (UTEVA resin with HNO₃/H₂O₂ mixture as eluent). Separation protocols were found to be effective with recovery yields between 80 and 120 %.

Then, the separations performed on the Morille CRM showed satisfactory results for most of the elements. A Ba pollution from oxalic acid used in protocol B was observed. Using protocol C, Zr is fixed onto the resin. To eluate Zr, oxalic acid needs to be used. Moreover, the Th behavior, a tetravalent actinides, did not help us to recover it with the other cations in the feeding and scrubbing phases. To recover thorium independently of U and Pu, using hydrochloric acid is required. Since hydrochloric acid is corrosive to glove boxes, its use is avoided. Combining the studies on the CRM and on the simulated solutions, it is possible to analyze 41 cations (Al, Ag, B, Ba, Be, Bi, Ca, Cd, Ce, Co, Cr, Cu, Dy, Eu, Fe, Ga, Gd, In, La, Li, Mg, Mn, Mo, Na, Nd, Ni, Pd, Pb, Rh, Sb, Se, Si, Sm, Sn, Sr, Ti, V, W, Y, Zn and Zr) with the two selected protocols in a U-Pu matrix.

Without separating U from the cations it is impossible to analyze Be, Dy, Eu, Gd and Sm in the CRM at these concentrations.

These separation protocols were implemented to certify a Pu-based reference material and to samples for irradiation experiment. Detection limits in the order of μg g⁻¹ of powder were obtained. Furthermore, as Th and Np are fixed onto the resin, these protocols could be extended for cationic impurities analysis in Th or Np-based materials. It should be notice Np is only fixed on the resin using protocol C. The oxalic acid used in the protocol B is a complexing agent of Np.