Utilization of traceable standards to validate plutonium isotopic purification and separation of plutonium progeny using AG MP-1M resin for nuclear forensic investigations

2.1 Sample preparation

The first set of experiments to test the effectiveness of the resins in purifying the Pu isotopic fractions analyzed by TIMS utilized traceable plutonium isotopic CRMs from New Brunswick Laboratory (NBL). Model ages which correspond to the age of purification of these standards cover a range from 1963 through 1971. Thus, significant amounts of the plutonium progeny are present in these standards and these standards are ideal for evaluating the complete removal of the isobaric interferences (²³⁸U and ²⁴¹Am are both isobaric interferences in a Pu isotopic measurement) possible for Pu isotopic measurement. The CRM standards were prepared as follows: Starting with solutions with a concentration of ∼100 mg/mL Pu, aliquots of approximately 20 μg of Pu were taken out for separations as described below. We have demonstrated that effective separation is possible for Pu amounts up to 2 mg using these columns and volume of chemicals. ¹⁷ The Pu solutions were dried down and dissociated by fuming with a couple of drops of concentrated HNO₃.

The plutonium fraction is separated from the decay products and from Ga using ion exchange chromatography. Aliquots used for Pu isotopic measurements must be separated from the decay products of the Pu isotopes in order to eliminate isobaric interferences. For Pu materials, two isobaric interferences come into play as decay products: ²⁴¹Am (produced by decay of ²⁴¹Pu) interfering with ²⁴¹Pu and ²³⁸U (produced by decay of ²⁴²Pu or present as an impurity in the initial sample) interfering with ²³⁸Pu. As the first step in the resin test, ∼20 μg aliquots of two high burnup CRMs (Pu isotopic CRMs 136 and 137), one low burnup CRM (Pu isotopic and assay standard CRM 126-A), and two aliquots each from two different Plutonium Metal Standard Exchange Program (PMSEP) metals were separated using the 50–100 mesh & 100–200 mesh resins from Bio-Rad ¹⁶ . Similar to the current analytical process using Lewatit resin ^{18 , 19} a water slurry of the Bio-Rad resins is loaded into a Nalgene-type disposable dropper with the tip of the bulb end cut and the dropper end plugged with a small ball of glass wool (approximately 0.8 g of the resin is used for each separation). The resin is conditioned with 4 mL of 12 M HCl (the conditioning and elution steps for the AG MP-1M resin stayed the same as that being currently used for the Lewatit MP 800 resin). 20 μg aliquots of the Pu (in 12 M HCl) was transferred to the column. The americium is removed by adding ∼12 mL of 12 M HCl. The Pu fraction is reduced to Pu(III) and eluted with 4 mL of 12 M HCl + 0.2 M HI solution. The uranium and gallium fractions remain on the column (see refs. 15] & 16] for additional details on the separation technique). During the Pu isotopic purification process, both 50–100 mesh and 100–200 mesh resins from AG MP-1M were used. The 100–200 mesh resin took approximately five times longer for the elute to drain. So, for all subsequent testing (Am, Pu, and U element content analyses by the IDMS technique) the 50–100 mesh resin only was used. All three Pu isotopic CRMs mentioned earlier and two other PMSEP Pu metals were separated on two different days (using the 50–100 mesh AG MP-1M resin) to demonstrate repeatability of the Pu isotopic separation.

2.2 UV–vis determination of UO₂²⁺/Ga³⁺ content

Alizarin red S (ARS) was used as a colorimetric dye to determine the behavior of UO₂²⁺ and Ga³⁺ on AG MP-1M resin with various acidic solutions as eluents. A 0.2 M stock solution was prepared by dissolving ARS (6.8 mg, 0.02 mmol) in pyridine in a calibrated volumetric flask (100 mL). An aliquot (1 mL) of this solution was transferred to the vials containing 10 μL aliquots of the elution fractions. The samples were mixed gently by inverting the capped vials several times and then aged at ambient temperature for 2 h to ensure that complexation equilibrium was reached. Having an excess of ARS with respect to the metal analytes was desirable to ensure that the dye could bind both UO₂²⁺ and Ga³⁺ (if both were present in the same sample) and to ensure consistency in metal-to-ligand binding ratios (e.g., 3:1 ARS:Ga³⁺). The samples were then transferred to matched UV quartz cuvettes (FireflySci, 1.4 mL capacity, 1.00 cm path length). UV–vis spectra were collected on an Agilent Cary in a 750–350 nm window. Samples were prepared in unbuffered deionized water at concentrations that maintained absorbance between 0.0 and 1.0 (unitless). Spectral details of relevant species dissolved in water are as follows: uncomplexed ARS λ_max = 420 nm, ε₄₂₀ = 2,830(10) M⁻¹ cm⁻¹; Ga(ARS)₃ λ_max = 480 nm, ε₄₈₀ = 4,300(30) M⁻¹ cm⁻¹; UO₂(ARS)^±_(pv) λ_max = 565 nm, ε₅₆₅ = 2,220(30) M⁻¹ cm⁻¹. Unfortunately, poor resolution of these peaks complicated quantitative analysis by UV–vis spectroscopy, but these results were informative in planning automated separation experiments by HPLC.

2.3 TIMS instrumental

Plutonium isotopic analytical data presented here were collected using cup-configuration shown in Table 1a using TRITON TIMS instruments (Manufacturer: ThermoElectron, Bremen, Germany). For the analysis presented here, each Faraday cup detector was connected to 10¹¹ Ω amplifiers. The “0” idle time chosen for the total evaporation (TE) analysis precludes use of the 10¹² Ω amplifiers in the isotope ratio measurements. Two different TRITON instruments, both performing at similar levels for isotope ratio measurements, were used for this investigation.

For isotopic analysis using TIMS instrumentation, the total evaporation (TE) technique has been demonstrated to yield more accurate major isotope ratios for actinide elements U and Pu ^{20 , 21 , 22 , 23 , 24} due to the ability to perform mass fractionation correction. Standardization of the analytical technique has further helped in improving the accuracy, precision, and overall uncertainty of the ²³⁵U/²³⁸U and ²⁴⁰Pu/²³⁹Pu major isotope ratio measurements. Analytical method development efforts at NBL and IRMM, specifically the development of the modified total evaporation (MTE) technique ²⁵ accomplished a consistency in the tailing correction approaches and standardization of these correction methodologies, resulting in a vast improvement in the accuracy and precision of the minor isotope ratio data for uranium. The MTE technique has become the state-of-the-art analytical technique for U isotope amount ratio measurements.

TE analysis utilized the double filament configuration with 20 ng Pu, 5 ng Am, and 200 ng U loaded onto zone-refined rhenium filaments. Cup-configurations used for Am and U isotope ratio measurements are shown in Table 1b and c, respectively. Before starting the analyses of the turret as an auto sequence, ¹⁸⁵Re and ¹⁸⁷Re isotopes as well as ²³⁹Pu, and ²⁴⁰Pu (Pu analysis), or ²³⁵U and ²³⁸U (U analysis), ²⁴¹Am and ²⁴³Am (Am analysis) isotopes were focused and calibrated (magnet mass calibration) by manually heating the ionization and evaporation filaments to obtain 100 mV signals of ¹⁸⁷Re and ²³⁹Pu (or ²⁴¹Am, or ²³⁵U). During analysis fine changes in the mass calibration curve were captured by peak centering on ¹⁸⁷Re and ²⁴⁰Pu (isotope in the center cup) for each sample/standard before data collection.

3.1 Performance for ²⁴⁰Pu/²³⁹Pu major isotope ratio

The certified (or known, for the PMSEP metals) Pu isotopic composition of these standards (samples) was decay corrected to the date of separation and compared with measured isotope amount ratio data.

Table 2 show the performance characteristic observed for two Pu isotopic turrets analyzed on two different Triton instruments. The isotopic measurements were performed within a few days of separation of the Pu progeny. Turret 1 was analyzed on Triton SN747 and Turret 2 was analyzed on Triton SN748 utilizing the TE methodology described in Section 2. Turrets 1 and 2 included three aliquots each of CRMs 136, 137, and 126-A. Turret 1 also included aliquots from two different dissolutions of two PMSEP metals analyzed in duplicate. In order to demonstrate the repeatability of the isotopic separations, Turrets 3 and 4 following similar analysis schemes as Turrets 1 and 2 were made on two different days. Two other PMSEP metals were processed and analyzed along with the CRMs 136, 137, and 126-A for Turrets 3 and 4. Turret 3 was analyzed on Triton SN747 and Turret 4 was analyzed on Triton SN748. For all Turrets, precision (percent relative standard deviations, %RSDs) of <0.01 % was observed on the ²⁴⁰Pu/²³⁹Pu major isotope ratio for all Pu isotopic standards. The %RSDs observed in all cases were better than the acceptance criteria of 0.024 % for the ²⁴⁰Pu/²³⁹Pu major isotope ratio analysis by the TE method (2-σ limit on the control chart for ²⁴⁰Pu/²³⁹Pu isotope ratio using TE).

Note that for accuracy, the nature of the TE analytical methodology implies that the sum integrated ²⁴⁰Pu/²³⁹Pu major isotope ratios overlap the true values of these ratios within an uncertainty limit of ± 0.030 %. ^{24 , 26} The ²⁴⁰Pu/²³⁹Pu isotope ratio is the most diagnostic for safeguards applications to decipher the intended use of the Pu material ²⁷ and has smaller accuracy and precision target values from Nuclear Material Control & Accountability (NMCA) and safeguards perspectives. ²⁸ For the traceable Pu standards and for the PMSEP metals included in this study, the certified (or known, for PMSEP metals) values can be decay corrected to the date of purification of the Pu isotopic fraction and the percent relative deviations (%RDs) of these values from the decay corrected values can be calculated. In all cases, these %RDs are <0.030 %, consistent with expected trends for TE isotopic analysis of Pu. For PMSEP metals the %RDs are with respect to isotope ratio values (decay corrected to the date of separation) measured on Triton SN747 & VG-4 TIMS instruments that were independently obtained following the task area measurement practices using Lewatit resin.

Table 2 also show that the standards ran acceptable number of cycles (>100 1 s integrations are considered acceptable for a SUM – (²³⁹Pu + ²⁴⁰Pu) – target intensity of 6 V) considering the amounts of Pu loaded on the filaments and the analysis intensity used for the TE method. On an average, both turrets ran similar number of cycles. This demonstrates that ionization inhibiting elements, if any, present in the standards/sample solution were removed equally efficiently by both mesh sizes of the AG MP-1M resin.

For Turret 1, the Pu isotopic fraction was separated using 50–100 mesh resin on 8/30/2022 and analyzed using Triton SN00747T on 9/1/2022 and for Turret 2, the Pu isotopic fraction was separated using 100–200 mesh resin on 8/29/2022 and analyzed using Triton SN00748T on 9/1/2022. For Turret 3, the Pu isotopic fraction was separated using 50–100 mesh resin on 9/27/2022 and analyzed using Triton SN00748T on 9/29/2022 and for Turret 4, the Pu isotopic fraction was separated using 50–100 mesh resin on 10/12/2022 and analyzed using Triton SN00747T on 10/13/2022. A %RSD of <0.030 % based on the control charts for high burnup and low burnup CRMs is used as the acceptance criteria for TE analysis.

3.2 Performance for ²³⁸Pu/²³⁹Pu, ²⁴¹Pu/²³⁹Pu, and ²⁴²Pu/²³⁹Pu minor isotope ratios

Table 3 show the precision characteristic observed at the minor isotope ratios associated with the major ratio data shown in Table 2. For the minor isotope ratios, systematic biases arising from peak-tailing from the major isotope peak (²³⁹Pu, in this case) has been corrected using the data analysis technique described in literature. ²⁹ The relative deviations of the tail corrected minor ratio values from the certified values decay corrected to the date of separation (expressed as a percent – %RDs) as well as the %RSDs of the replicate measurements within the turret are also shown. For PMSEP metals the %RDs are with respect to isotope ratio values (also decay corrected to the date of separation) measured on Triton SN747 & VG-4 TIMS instruments that were independently obtained following the task area measurement practices using Lewatit resin.

The bias is statistically significant if the %RD values are larger than the %RSD values (i.e., bias is significant if within the precision observed, the %RDs are not overlapping “0” values).

4.1 Americium

²⁴¹Am produced from decay of ²⁴¹Pu isotope was separated as described in Section 2.1 and analyzed by TIMS instrument (VG-4). For the Am content measurements by IDMS, ²⁴³Am was used as the tracer. The ²⁴³Am tracer solution is calibrated against SRM 4322d to establish traceability. Table 4 shows the results of the Am element content measurements. Two separate aliquots from each of the CRMs 136, 137, and 138 were independently taken through the spiking and separation process using the Lewatit and AG MP-1M resins and measured in triplicates/duplicate on the VG-4 TIMS instrument using the TE methodology. The analytical technique followed the analysis scheme described in Section 2.2 and explained in more detail elsewhere. ³⁰

The three traceable CRMs included in this investigation contain ∼1.8 wt%, 1.95 wt%, and 0.32 wt% ²⁴¹Am based on the model purification dates ^{1 , 31} and the certified Pu isotopic abundances. Figure 1 shows that the separations using both Lewatit and AG MP-1M resins yielded Am contents consistent with these expected values. Within the observed precision of the analytical data, no difference between the Am content values obtained using the Lewatit and AG MP-1M resins are observed.

Figure 1:

Measured Am element contents in CRMs 136, 137, and 138. Both the Lewatit resin and 50 to 100 mesh AG MP-1M resin yielded identical results for each CRM. The two high-burnup CRMs (CRMs 136 and 137) have Am contents in the wt% range, indicating the old age & higher burnup of these Pu isotopic standards.

Within the uncertainties realized by the Am IDMS technique, no cut-to-cut differences are present in the Am content data for all three standards. An expanded uncertainty of ∼0.12 % is realized (uncertainties stated are the method uncertainties estimated from quality control charts covering long periods – years) by the IDMS technique for the Am element content data.

4.2 Uranium

Uranium isotopes ²³⁴U, ²³⁵U, ²³⁶U, and ²³⁸U are produced by α-decay of the Pu isotopes ²³⁸Pu, ²³⁹Pu, ²⁴⁰Pu, and ²⁴²Pu. U amount content was analyzed using ²³³U as the traceable spike standard. Two separate aliquots from the CRM standards 136, 137, and 138 were independently spiked with the ²³³U tracer and taken through the column chemistry as described in Section 2.1 using Lewatit and AG MP-1M resins. The separated U fraction was analyzed in duplicate using the Triton TIMS instrument as described in Section 2.2. Table 5 provides a summary of the results for all three CRMs.

The three traceable CRMs included in this investigation contain ∼1,300 μg/g, 1,430 μg/g, and 970 μg/g U based on the model purification dates ^{1 , 31} and the certified Pu isotopic abundances. Figure 2 shows that the separations using both Lewatit and AG MP-1M resins yielded U contents consistent with these expected values. Within the observed precision of the analytical data, no difference between the U content values obtained using the Lewatit and AG MP-1M resins are observed.

Figure 2:

Uranium element content data on CRMs 136, 137, and 138. Both the Lewatit resin and 50 to 100 mesh AG MP-1M resin yielded identical results for each CRM. The two high-burnup CRMs (CRMs 136 and 137) have higher U contents, indicating the old age & high burnup of these Pu isotopic standards.

For all three CRMs, within the precision of the data for the U element content results, there is no cut-to-cut difference between the U data. An expanded uncertainty of 0.20 % is obtained (uncertainties stated are the method uncertainties estimated from quality control charts covering long periods – years) for the trace U element content measurements using the IDMS technique.

4.3 Plutonium

For Pu element content measurements by the isotope dilution, ²⁴⁴Pu was used as a tracer. Two separate aliquots each from each of the three CRMs 136, 137, and 138 went through a double dilution before being mixed with ²⁴⁴Pu spike standard and taken through the separation process using Lewatit and AG MP-1M resins. Table 6 shows a summary of the Pu IDMS results in CRMs 136, 137, and 138.

Within the precision of the data, cut-to-cut differences are evident in the data. As expected, the most likely cause of this cut-to-cut difference is the weighing accuracy. Note that for the Pu element content measurements by the IDMS technique, the solution aliquot received from the dissolution task area undergoes a double dilution (by a factor of at least 1,000) before the aliquot for spiking are taken. Thus, any small accuracy issues with the weighing during the dissolution and initial aliquoting gets amplified by the dilution factor.

The precision of the aliquot averages for CRMs 136 and 137 suggest an expanded uncertainty of 0.30 % for the Pu element content measurements by the IDMS technique. However, due to cut-to-cut differences the uncertainty obtained for the grand average is ∼0.50 %. Only for CRM 138, the data did not indicate a cut-to-cut difference and an expanded uncertainty of ∼0.16 % is obtained for the Pu element content of this CRM (Figure 3).

Figure 3:

Plutonium element content data on CRMs 136, 137, and 138. Both the Lewatit resin and 50 to 100 mesh AG MP-1M resin yielded identical results for each CRM. Though these CRMs are not certified for Pu assay, values for the Pu assay have been established through an inter-comparison study [1] performed as part of a multi-laboratory radio-chronometric investigation.

Model purification dates for the Pu CRMs 136, 137, and 138 may be estimated by combining the Pu, U, and Am results presented in Sections 4.1–4.3. A model purification date of 4/16/1970 ± 72 days, consistent with published model purification date of 5/23/1970 ± 146 days and with manufacturer’s production date of 3/30/1970 ± 60 days (see [1], for additional information), is estimated for CRM 136. A model purification date of 8/27/1970 ± 52 days, consistent with published model purification date of 12/25/1970 ± 295 days and with manufacturer’s production date of 9/15/1970 ± 60 days (see [1], for additional information), is estimated for CRM 137. A model purification date of 9/22/1963 ± 476 days, consistent with published model purification date of 10/11/1963 ± 536 days and a manufacturer’s production date of 9/15/1963 ± 180 days (see [1], for additional information), is estimated for CRM 138. For all three CRMs, the agreement of the calculated model production dates with values available in literature obtained through an inter-laboratory comparison study [1] demonstrates the completeness of the element recoveries using the Lewatit and AG MP-1M resins.

4.4 Evaluation of U/Ga separation by UV–vis

A key challenge in this study was the separation of uranyl (UO₂²⁺) and gallium (Ga³⁺) on the AG MP-1M anion exchange resin. Under the separation conditions described here, UO₂²⁺ and Ga³⁺ co-elute from the column. Although it was determined that isotope ratio measurements by TIMS of uranium and gallium likely do not interfere with one another (this was demonstrated by independently analyzing the U and Ga isotope ratios in a U + Ga mixture of CRMs 112-A and SRM 994), we were interested in achieving this difficult (anionic) separation from a fundamental perspective. Uranyl and gallium are readily separated on cation exchange resins; ^{12 , 13 , 14 , 15 , 32 , 33} the +3 charge of aqueous gallium, Ga(OH₂)₆³⁺_(aq) allows this ion to interact with resin-bound anions more strongly than aqueous uranyl, which has a +2 charge. However, for anion exchange resins such as the one used in this study, the cationic charge is irrelevant, and instead the formation of differently charged anionic complexes determines the resolution power. This property makes the separation of UO₂²⁺/Ga³⁺ on AG MP-1M more challenging.

Upon reviewing dissociation equilibrium constant (K_d) measurements pioneered at LANL in the 1970s, ³³ it was realized that there was an opportunity to achieve the resolution of uranyl and gallium on AG MP-1M using nitric acid (HNO₃). Uranyl and gallium have maximally differing affinities for anion exchangers in 8 M HNO₃ (K_d,U:K_d,Ga ≈ 10:0). We found that eluting mixtures of UO₂²⁺ and Ga³⁺ with HNO₃ resulted in separation of the cations, but it was difficult to quantitatively determine the extent to which the ions were partitioned.

We developed a colorimetric technique for quantifying the amount of UO₂²⁺ or Ga³⁺ in fractions collected from the anion exchange resin columns (Figure 4) based on a literature method for detecting Ga³⁺. ³⁴ Due to high demand for measurements on programmatic samples, TIMS and ICP-MS/AES analyses were not available for use. Instead, we turned to UV–vis spectroscopy as an analytical tool. This technique has the advantages of fast sampling time (about 1 min per sample) and facile operation. While UO₂²⁺ does have a distinct UV–vis signature (vibronic progression from 470 to 590 nm), ²⁹ Ga³⁺ has no optical or UV absorption features, rendering it invisible to UV–vis measurements. Instead, we used a commercial, water-soluble dye (Alizarin Red S, ARS) that undergoes dramatic color changes in the presence of polyvalent metal ions (Figure 4). ³⁵ That is, solutions of the free ARS dye are orange with a maximum absorption occurring at 420 nm (ε₄₂₀ = 2,820(10) M⁻¹ cm⁻¹). When Ga³⁺ was added to these solutions, a distinct color change to pink-red was observed (λ_max = 480 nm, ε₄₈₀ = 4,300(30) M⁻¹ cm⁻¹). In the presence of UO₂²⁺, ARS takes on a deep purple color (λ_max = 565 nm, ε₅₆₅ = 2,220(30) M⁻¹. cm⁻¹). ³⁵

Figure 4:

Example of how ARS changes color in the presence of metal ions. ARS is an orange-yellow color in its free form, takes on a pinkish hue with Ga³⁺, and attains a purple color with UO₂²⁺. See Figure 5 for corresponding UV–vis spectra.

Figure 5:

Representative UV–vis spectra of fractions of a UO₂²⁺/Ga³⁺ mixture eluted from AG MP-1M with 8 M HNO₃. Spectra were selected to span a range of a purely Ga³⁺-containing fraction (green), a purely UO₂²⁺-containing fraction (navy), and a fraction containing both Ga³⁺ and UO₂²⁺ (teal). The corresponding spectrum of free ARS dye (gray) is included for comparison. Spectra were recorded in unbuffered deionized water. ^{13 , 36}

We sought to take advantage of the sensitivity of ARS’s color profile to Ga³⁺ and UO₂²⁺ ions. In principle, it is straightforward to measure molar extinction coefficients of the ARS–metal complexes and use these values to fingerprint the composition of solutions coming off of the anion exchange column. However, these studies were complicated by several factors. For instance, absorption at 480 nm became non-linear in the presence of excess gallium (Figure 6). That is, sub-stoichiometric complexes with putative formulas Ga(ARS)²⁺ and Ga(ARS)₃ did not adhere to Beer’s law, which obfuscated quantitative measurements. We tried to overcome this issue by ensuring that ARS was always in excess of Ga³⁺, but in these instances the resolution of absorption features corresponding to free ARS versus Ga(ARS)²⁺ versus UO₂(ARS)^±, which might all be in solution simultaneously was non-trivial. The broad peak shapes associated with each species made absolute deconvolution of the spectra infeasible. Consequently, we view the results presented in Figures 4 and 5 as qualitative assessments of the separation of gallium and uranyl cations on AG MP-1M.

Figure 6:

The titration of ARS (2.47 × 10 – 4 M) with varying equivalents of Ga³⁺ in water produces absorbances that deviate strongly from linearity at Ga:ARS ratios greater than 1. A logarithmic function fits the data reasonably well, but at high equivalences of Ga³⁺ the slope of this function becomes very small, making it difficult to accurately calculate gallium concentration from an absorbance measurement.

These results could be leveraged in testing separation conditions on an HPLC column of AG MP-1M. We anticipate that HPLC will afford even better separations of UO₂²⁺ and Ga³⁺ since Brownian motion (i.e., diffusion) will be decreased greatly compared to a gravity column. In sum, the preliminary results obtained indicate that the sequential separation of actinides (and gallium) on AG MP-1M is likely to succeed, and we expect to achieve this goal in planned follow-on studies.