Introduction

Nuclear forensics is the examination of nuclear and other radioactive materials to identify the risk, origin and history of materials in order to support law enforcement investigations or to assist nuclear security inquiries [1,2,3]. Several sample characteristics (also known as signatures) can be used for such a purpose. These parameters (among others) include the major matrix composition (U and Pu for nuclear samples), minor constituents (organic or inorganic impurities, anions), isotope ratios of U, Pu, Pb, Sr, Nd, O or S, and morphology or texture analysis [2, 4,5,6,7,8]. In this set of sample characteristics, the age (production date) of the nuclear sample can also be determined: the radioactive parent nuclide (usually U or Pu) during the production of the nuclear samples is chemically separated (purified) from its daughter products. Thus, measuring the parent-to-daughter ratio, the elapsed time, i.e., the date of separation (“production”) can be calculated using the radioactive decay laws (Bateman equations). This unique information, which does not rely on any library or database for comparison and can be obtained directly from the calculation, is very important for nuclear forensics and was found to be very useful for nuclear forensic conclusions [3, 9,10,11,12].

Age dating calculations are based on the radioactive decay equations, assuming that the nuclear sample behaves as a closed system (i.e., no loss or increase for either the parent nuclide or a decay product). For U materials, the 230Th/234U or 231Pa/235U decay pair is used. If the initial concentration of the daughter nuclide is zero after the last chemical separation (i.e., the separation was complete and the initial 230Th daughter is zero), and the atom (amount) ratio of 230Th and 234U is measured, the elapsed time, thus the age of the sample (t) can be calculated as follows [13, 14]:

$$t = \frac{1}{{\lambda _{{U - 234}} - \lambda _{{Th - 230}} }}\ln \;\left( {1 - \frac{{N_{{Th - 230}} }}{{N_{{U - 234}} }} \cdot \frac{{\lambda _{{Th - 230}} - \lambda _{{U - 234}} }}{{\lambda _{{U - 234}} }}} \right)$$
(1)

where NTh-230 and NU-234 are the amount of 230Th and 234U, respectively, while λ is the respective decay constant. The measured 230Th/234U amount ratio (commonly called chronometer) is strongly sensitive to the purity of the material. Matrix interferences and impurities present in the sample can result in a biased (higher) age [15, 16] as the decay nuclides are at very low concentration (typically 10–4–10–7 less compared to the parent nuclides). Therefore, time consuming and extensive solution-based techniques are typically used to separate and purify the daughter product. In contrast to the generally used solution-based techniques, laser ablation (LA) sample introduction can be performed rapidly; it minimizes the measurement time, radioactive waste and sample consumption. Thus, the application of laser ablation is highly appropriate for nuclear samples, especially for nuclear forensic applications, where the investigated sample is an evidence, thus it has to be preserved to the highest extent. Using this LA technique, the 234U (parent) measurement usually due to the high concentration (μg/g level or higher) is not an issue (they are U materials) [11], but any interference on the measured lower-level 230Th daughter (usually present at pg level) can result in incorrect high age dating results. However, during laser ablation, when the original, intact sample is ablated (evaporated), no prior separations of the interferences are possible. Thus, a severe background caused by the polyatomic ions can be formed and can cause erroneous result. The 230Th daughter is at very low level: any additional intensities, e.g. molecular ions on m/z = 230, where the low concentration 230Th daughter appears, can cause an interference, leading to an mistakenly high signal.

The aim of the present study was to develop a reliable and validated method for the direct age dating measurement by laser ablation coupled to multi-collector inductively coupled plasma mass spectrometry (LA-MC-ICP-MS) for uranium samples. The obtained LA-MC-ICP-MS age results can be compared with the chemically separated, solution-based results, thus the accuracy of the method and the major factors affecting the accuracy can be assessed. The employed samples (ten U metals and uranyl nitrate with various 235U enrichments) in the present study were analysed before (also by other laboratories), thus their production date is well-known. The current systematic work was used to develop a robust age dating technique by LA-MC-ICP-MS. With this extensive set of samples the possibilities and disadvantages of LA-MC-ICP-MS for age dating can be assessed and completely exploited.

Experimental

Investigated samples

The studied age dating samples have different 235U enrichments, including depleted U (DU), natural, low-enriched U (LEU) and highly enriched U (HEU). The uranium samples originate from collaborative exercises (CMXs) organized by the Nuclear Forensics International Technical Working Group (ITWG). The four U metal samples derive from the 3rd (two HEU metals from the Round Robin-3: RR-A and RR-B samples) and the 7th ITWG exercises (CMX-7, two DU metals: ES2 and ES4 samples) [17]. Six uranyl nitrate samples were also analysed: three reference materials prepared at JRC-Karlsruhe with well-known ages (natural, LEU and HEU nitrate powders), validated also by other expert labs [18], one certified reference material (IRMM-1000 containing LEU nitrate) [19] and two DU metal samples from the CMX-7 exercise (ES-1 and ES-3 samples). All samples were analysed and dated before with the commonly applied age dating method (i.e., with chemical separation), thus the accurate model ages are known.

Reagents and materials

The U metal samples were measured directly by laser ablation without any sample preparation. From the uranyl nitrate powders a few grains were fixed on a transparent matrix by embedding them on a drop of transparent glue (UHU Plus 300) in order to avoid dislocation and possible contamination of the chamber. This step, however, can be avoided or may be modified if the sample must be kept for further study. U mass bias correction and ion counter gains were calculated using a NBS U-020 CRM (pressed as a pellet), currently distributed by New Brunswick Laboratory (NBL, Argonne, IL, USA) [20].

Instrumentation

A Nu Plasma™ (NU Instruments, Wrexham, United Kingdom) double-focusing multi-collector inductively coupled plasma mass spectrometer, equipped with 11 Faraday detectors and 3 discrete dynode electrode multipliers was used for the measurements. The MC-ICP-MS was connected to a NWR-213 laser ablation system (ESI, Huntingdon, UK). The MC-ICP-MS instrument was operated in low mass resolution mode (R = 300) and was optimised daily (torch position, gas flows, voltages) using a 30 ng g−1 multi-elemental solution (Inorganic Ventures, Christiansburg, USA). The optimisation aimed at achieving the highest sensitivity and stability of the measured 238U signal. The sensitivity was about 5.5 V for 238U+. The LA system was equipped with a two-volume cell (TV2 cell in this design) to achieve quick wash-out and to eliminate cross-contamination within the LA chamber. Further details of the instrumentation and optimization can be found elsewhere [20]. Before the LA analysis, the U metal samples were pre-ablated with the laser beam (i.e., the surface was cleaned and smoothened). The 234U isotope was measured on the Faraday detector (H1) for the high 234U signal, while the IC0 ion counter (equipped with a retardation filter to minimise the tailing of U signals) was used for the simultaneous measurement of the low abundant 230Th isotope. Laser beam diameter and fluence had to be adjusted for the respective sample to achieve maximum signal of 234U (~ 8–10 V). Approximately 5 ng of sample was consumed for each measurement. Detailed instrumental settings and data acquisition parameters are given in Table 1.

Table 1 MC-ICP-MS and LA operating parameters
Full size table

Results and discussion

Data treatment and evaluation

First, the raw intensity data obtained by LA-MC-ICP-MS analysis were corrected for the gas blank without ablating the material (Fig. 1). Then, the simultaneously measured 230Th/234U isotope ratios were calculated for the laser-ablated sample as the ratios of the background-corrected signals (Fig. 2). Ratioing was performed only during the time interval when the 230Th ion beam intensity (lowest signal) exceeded 10 cps. U mass bias correction with exponential correction and ion counter gains were calculated by measuring the NBS U-020 before the analysis of each sample.

Fig. 1
figure 1

Typical 234U signal (HEU metal, Round Robin-3 A sample) by LA-MC-ICP-MS

Full size image
Fig. 2
figure 2

Typical 230Th/234U amount ratio (HEU metal, Round Robin-3 A sample) by LA-MC-ICP-MS

Full size image

Age dating results were calculated using the relative sensitivity method as widely applied for LA studies: first, the 230Th/234U signal ratio was measured by laser ablation in a calibrant sample, and then compared to its known (e.g. certified or solution-based determined) 230Th/234U ratio to determine the relative sensitivity factor (RSF):

$$RSF = \frac{{\left( {\frac{{{}^{234}U}}{{{}^{230}Th}}} \right)_{KnownRatio} }}{{\left( {\frac{{{}^{234}U}}{{{}^{230}Th}}} \right)_{Laser Ablation} }}$$
(2)

This relative sensitivity factor (calibration factor) is applied then for an unknown sample to determine its corrected 230Th/234U ratio (i.e., multiplying the measured laser ablation 230Th/234U signal ratio for the unknown sample with RSF to obtain the corrected 230Th/234U ratio). The age of the U sample can be determined by Eq. 1 using this obtained RSF. It has to be emphasized that the calibrant sample has to be measured in one run prior to the unknown sample in order to keep the RSF constant. The calibrant should be as similar in composition to the unknown sample as possible as shown in the literature for the LA examinations [21, 22]. The 230Th/234U ratio in the calibrant has to be well-known: either a standard (e.g., CRM-125A or IRMM-1000) or a sample, where a chemical separation was applied before to determine the 230Th/234U ratio, can be an option.

This relative sensitivity factor (RSF) includes Th and U sensitivity differences (e.g., ionization potentials), mass bias and detector efficiencies, which is not taken into account by the NBS U-020 measurement. The method also corrects for the possible small variation of the measurement conditions during the sequence. Further details on the relative sensitivity method can be found elsewhere [22, 23]. For an interference-free measurement, i.e., where Th and U are measured without spectral interferences, the calibration by RSF is close to ~ 1 (0.7–1.3). Thus, high values of RSF (RSF≫1) indicate the presence of an additional signal on m/z = 230 and isobaric interferences.

The age dating measurement of the samples was done in one sequence to make sure that Th and U sensitivities, detector efficiencies and mass bias remained approximately constant during the measurement. The LA-MC-ICP-MS measurement was performed in the following order: U mass bias by U-020 for U—Calibrant (HEU)—Sample. Dilute HNO3/HF between samples was applied to remove the Th and U traces if needed.

All results reported hereafter are based on three replicate measurements of the samples. All uncertainties are combined standard uncertainties indicated in parentheses and include a coverage factor of 2 (k = 2, approx. 95% confidence interval).

Age dating results

The age dating results of the U metal samples (two HEU metals from the Round Robin-3 exercise and two DU metals from the CMX-7 exercise) are summarized in Table 2. The age results after chemical separation (“true values”, also called as model date in the literature) are shown in comparison to those measured by LA-MC-ICP-MS.

Table 2 Age dating results of the U metal samples
Full size table

Similarly, the age results obtained for the uranyl nitrate samples are shown in Table 3. For the uranyl nitrate samples, the HEU uranyl nitrate was used as calibrant due to the high 230Th signal and purity. The reference date for the measured age results is 11 November 2022.

Table 3 Age dating results of the uranyl nitrate samples
Full size table

Observations of age dating by LA-MC-ICP-MS

The higher the U enrichment the closer the LA-MC-ICP-MS obtained results are to the chemically separated (“true”) values. The measured LA-MC-ICP-MS age dating results are in good agreement if highly enriched U is measured, however, their uncertainty is higher (> four times). The simple reason for this is the elevated, easily measurable 230Th decay product. At lower 230Th concentrations the isobaric interferences at m/z = 230 (like 206Pb24Mg+, 206Pb12C2+, 202Hg14N2+) are increasingly significant. This fact is also reflected in the relative sensitivity factors: using Eq. 2 one can calculate the correct RSF factors knowing also the measured 230Th/234U obtained by chemical separation. The RSF values 1.36 ± 0.03, 1.39 ± 0.03, 106.9 ± 9.8 and 80.0 ± 7.5 for the RR-A (HEU), RR-B (HEU), CMX-7 ES-2 (DU) and CMX-7 ES4 (DU) U metal samples, respectively. In case of uranyl nitrate samples the RSF results are 10.1 ± 0.5, 2.85 ± 0.06, 1.02 ± 0.05, 4.78 ± 0.06, 4.28 ± 0.07 and 5.24 ± 0.08 for natural U, LEU, HEU, IRMM-1000 (LEU), CMX-7 ES-1 (DU) and CMX-7 ES-3 (DU), respectively. Therefore, for other than HEU samples, the additional signal (including interferences) on m/z = 230 caused by the interferences hinder the accurate measurement by LA-MC-ICP-MS in low mass resolution. This was also demonstrated before with a limited number of samples [24]. In order to overcome this issue, for instance, the use of higher mass resolution or reaction cell is needed in laser ablation to remove such polyatomic interferences [24, 25].

Note that different RSF values are obtained for HEU metal and uranyl nitrate samples (e.g. 1.36 ± 0.03 for RR-A U metal vs. 1.02 ± 0.05 for HEU uranyl nitrate): this means that the laser ablation efficiencies are different for the different chemical compounds. In consequence, the RSF factors obtained for the U metal should not be applied for uranyl nitrates, because that will result in erroneous age dating results. This observation was also demonstrated comprehensively in laser ablation [24, 25]. Also note that the age dating result for the natural and DU samples are obviously wrong (the ages are > 100 years), but for the low-enriched uranyl nitrate samples (LEU and IRMM-1000) a reasonable, although incorrect value is obtained (February 1991 ± 1.5 years and 1978 ± 2.9 years, respectively). This poses a high risk in the correct data interpretation.

Conclusions

LA-MC-ICP-MS can be used for the age dating of uranium materials and the developed method is capable to perform the direct analysis reliably for HEU samples. However, for lower U enrichment, incorrect and misleading age dating results were obtained due to the isobaric (polyatomic) interferences on m/z = 230 in the current configuration (i.e. in low mass resolution mode). As the investigated U samples were pure (the impurity content was < 0.03%, measured by U assay), a higher discrepancy is expected for impure U materials, as the bias is caused by the spectral interferences. The RSF calibration factors are demonstrated to be characteristic to the U compound, thus the RSF for a certain material is not transferable to other matrices. In order to remove the molecular interferences, a higher mass resolution or reaction cell is required to extend the method to lower U enrichments.