Direct uranium isotope ratio analysis of single micrometer-sized glass particles

Abstract

1 Introduction The isotopic fingerprints of uranium (U) and plutonium (Pu) are of particular interest for international safeguards (Axelsson et al., 2009) and nuclear forensics (Mayer et al., 2007) as the knowledge of these signatures enables to link nuclear material to its respective nuclear processes and activities (Donohue, 2002). U and/or Pu isotopic signatures are stored in micrometer-sized particles (Donohue, 1998) that can be emitted during nuclear processes. The absolute amount of nuclear material in such particles usually ranges from picograms to nanograms (Axelsson et al., 2009). Sampling of such material is performed in nuclear facilities (e.g. enrichment facilities, hot cells, etc.) and the nearby environment by means of swipes (Donohue, 1998, 2002). Analyzing individual particles has the advantage that possible signatures of unknown isotopic compositions can be detected, even if they are masked by dust particles or other matrices that have declared or natural isotopic signatures. In comparison, bulk analysis, which comprises the analysis of the entire swipe, would only yield an average value of different isotopic signatures present on the swipe (Donohue, 1998). Fission track-thermal ionization mass spectrometry (FT-TIMS) (Esaka et al., 2004; Lee et al., 2007) and secondary ionization mass spectrometry (SIMS) (Betti et al., 1999; Ranebo et al., 2009; Tamborini, 2004) are usually applied for the isotopic characterization of single, micrometer-sized particles from safeguards samples (Donohue, 1998). The performance of TIMS in terms of accuracy and precision of isotope ratio measurements is unquestioned (Heumann et al., 1998); however, the need for a nuclear reactor for irradiation (Lee et al., 2007) of the FT detector with thermal neutrons is the main drawback of FT-TIMS. SIMS offers the advantage of combining the localization of particles collected by means of swipes and the determination of the isotopic information in one instrument (Betti et al., 1999). Moreover, scanning electron microscopy combined with energy-dispersive X-ray spectrometry (SEM-EDX) is applied for the localization of particles and the determination of the elemental composition (Ciurapinski et al., 2002; Donohue et al., 2008), as well as morphological characterization of the particles (Kips et al., 2007; Ranebo et al., 2007). Even though both TIMS and SIMS are state-of-the-art techniques, the International Atomic Energy Agency (IAEA) is pursuing improvements and new method developments in order to obtain a complete picture of a particle's history and to help verifying the absence of undeclared activities. Recently, laser ablation-inductively coupled plasma mass spectrometry (LA-ICP-MS) was applied for direct actinide isotope analysis of single particles (Boulyga and Prohaska, 2008; Lloyd et al., 2009; Varga, 2008). Varga analyzed U isotope ratios of U3O8 powder particles (i.e. depleted, natural, low-enriched and highly enriched U) – having lateral dimensions of 10–30 μm – by employing a high resolution double-focusing ICP sector-field mass spectrometer equipped with a single collector. The U amount in a 10 μm particle was estimated to approximately 0.46 ng, assuming a spherical particle (Varga, 2008). Depleted uranium oxide particles (i.e. larger than 20 μm), embedded in dust and surface soil, were directly analyzed by Lloyd et al. (Lloyd et al., 2009) by employing LA-multi collector (MC)-ICP-MS for the determination of 235U/238U and 236U/238U. The sampling volume corresponded to approximately 4 ng U (Lloyd et al., 2009). The applicability of LA-MC-ICP-MS for isotope ratio analyses of U and fission products of micro-samples (i.e. dimensions ranging from 100 μm to 1 mm) collected in the vicinity of Chernobyl was demonstrated by Boulyga and Prohaska (Boulyga and Prohaska, 2008). The present work describes the application of ns laser ablation coupled to MC-ICP-MS for the direct determination of 234U/238U, 235U/238U and 236U/238U isotope ratios in single, U-doped, 10–20 μm-sized glass particle reference material (Raptis et al., 2002). The purpose of the glass particles applied in this study is the simulation of environmental samples (e.g. soil, sediment or dust) that contain hot particles (Raptis et al., 2002), which is of interest considering a future application of the presented method for such samples. Hence, method development and the merits of this analytical technique are demonstrated by means of this reference material, in which absolute U amounts are in the picogram range. To the authors' knowledge it is the first time that a certified reference material is used for the validation of LA-MC-ICP-MS for U isotope ratio analyses in single particles. The full evaluation and validation of the presented method is described. In addition, a novel approach for operator-independent data processing, including non-laborious external correction of the isotope ratios of interest is presented in this work. 2 Materials and methods 2.1 Reagents, standards and certified reference materials Analytical reagent grade nitric acid (65% (m/m), Merck KGaA, Darmstadt, Germany) underwent sub-boiling distillation twice (Milestone-MLS GmbH, Leutkirch, Germany) prior to use. 1% (m/m) HNO3 was prepared by diluting purified 65% (m/m) HNO3 with reagent grade type I water (18.2 MΩ cm at 25 °C, Ultra Clear Basic Reinstwassersystem, SG Wasseraufbereitung und Regenerierstation GmbH, Barsbüttel, Germany) that was also purified by sub-boiling distillation (Milestone-MLS GmbH, Leutkirch, Germany) prior to use. Certified isotope reference materials – CRM U500 (New Brunswick Laboratory, U.S. Department of Energy, Washington, DC, U.S.) and IRMM-187 (European Commission-JRC, Institute for Reference Materials and Measurements, Geel, Belgium) – were applied for external correction and the determination of the secondary electron multiplier yields. An in-house prepared mixture, having a 236U/238U isotope ratio of 8.96 × 10−7, of IRMM-184 (European Commission-JRC, Institute for Reference Materials and Measurements, Geel, Belgium) and U500, was used for optimizing the deceleration filter. This mixture was obtained by diluting and gravimetrical mixing of both IRMM-184 and U500 in order to obtain a solution with a 236U/238U isotope ratio that is approximately twice as high as the 236U/238U abundance sensitivity that was determined without applying the deceleration filter. Determination of the abundance sensitivity for uranium isotopes at the mass m-2 u was employed with a 233U isotopic spike solution ((99.4911 atom percent of 233U; CRM 111-A (New Brunswick Laboratory, U.S. Department of Energy, Washington, DC, U.S.)). A natural U solution (IRMM-184) was used for the determination of the UH+/U+ hydride ratio when introducing liquid standards. Dilution of the standards was accomplished with 1% (m/m) HNO3 in order to get solutions exhibiting concentrations in the low ng g−1 range (i.e. smaller than 10 ng g−1). The isotope amount ratios of the certified reference materials used in this study are given in Table 1. S1 and S3 glass particles that are doped with U of certified isotopic composition (European Commission-JRC, Institute for Reference Materials and Measurements, Geel, Belgium) were used for method development, optimization, validation and proof of principle. The particles were produced by the IRMM for the IRMM support programme to the International Atomic Energy Agency (IAEA) and for the IRMM external NUSIMEP quality control programme (Nuclear Signatures Interlaboratory Measurement Evaluation Programme). Matrix glass, consisting of 70% SiO2, 15% B2O3, 10% Na2O, 4% CaO and 1% Al2O3 (i.e. borosilicate glass), was mixed with U3O8, in order to produce U-containing glass particles. The U-doped glass was blended with matrix glass in order to simulate environmental samples (e.g. soil, sediment or dust) that contain hot particles. A detailed description of the preparation of the used glass particles is given in Raptis et al. (2002). The analysis of the matrix glass enabled to exclude interferences occurring from matrix elements (e.g. Al, Pb, Si, etc.). The size of the glass particles was between 10 and 20 μm (Raptis et al., 2002). Certified isotope amount ratios of the utilized particles are given in Table 1. The estimated U amount per 10 μm-sized glass particle was less than 100 pg when assuming (1) a spherical shape, (2) a borosilicate glass density of 2.5 g cm−3 and (3) an U3O8 amount of 5% (m/m) (Raptis et al., 2002). About 1.1 mg of each glass particle standard were distributed on a cellulose acetate membrane filter (OE 67, Whatman GmbH, Dassel, Germany) and treated in a closed glass Petri dish with acetone vapor for about 30 min. The transparent membranes were then affixed, using customary glue, to a glass object plate. Both the membrane and the glass object plate were screened for their U blank. 2.2 Instrumentation All isotope ratio measurements were accomplished with a double-focusing high resolution sector-field MC-ICP-MS (Nu Plasma HR, Nu Instruments Limited, Wrexham, U.K.) A solid state nanosecond laser ablation system (UP 193, ESI-NWR Division, Electro Scientific Industries, Inc., Portland, CA, U.S.) was coupled to the MC-ICP-MS in order to perform direct analyses of the glass particles of interest. Both Ar and He were employed as carrier gas. A membrane desolvation system (DSN-100, Nu Instruments Limited, Wrexham, U.K.) was connected in parallel to the LA-MC-ICP-MS setup. This setup enables liquid sample introduction without the necessity of de-coupling the laser system for the measurement of the liquid U reference material. No liquid solution was aspirated during laser ablation of the glass particles. A schematic diagram of this setup is given elsewhere (Boulyga and Prohaska, 2008). The desolvation of the liquid samples by means of the membrane system allows introducing a dry aerosol into the ICP. Hence hydride generation is reduced compared to the introduction of a wet aerosol. Instrumental parameters are summarized in Table 2. 2.2.1 Collector block configuration The ‘Nu Plasma HR’ MC-ICP-MS is equipped with 12 Faraday cups and three discrete-dynode secondary electron multipliers. This configuration enables the simultaneous determination of the isotopes of interest. The major U isotopes, 235U and 238U, were measured with Faraday cups (i.e. L1 and L3), whereas the minor isotopes, 234U and 236U, were analyzed with secondary electron multipliers (i.e. IC0 and IC1). IC0, the secondary electron multiplier with which 236U was measured, has a deceleration filter installed to eliminate ions with low kinetic energies in order to improve abundance sensitivity. Yield variations of the secondary electron multipliers (i.e. IC0 and IC1) were determined by means of U500 and IRMM-187, respectively, according to standard-sample bracketing before and after the analyses of the glass particles. The UH+/U+ hydride ratio for both liquid measurements and laser ablation was assessed by analyzing 238U with L3 whereas its respective hydride was measured with IC0 at m/z = 239. The determination of the 236U/238U abundance sensitivity was accomplished with a 233U isotopic spike solution. Ion currents of ions having a mass-to-charge ratio of 231 and 233, respectively, were measured with the same secondary electron multiplier (i.e. IC0) and Faraday cup (i.e. L1), respectively, that were employed for the determination of 236U and 238U. 2.3 Data processing Ar or He was measured as blank for about 10 min prior to the analyses of the S1 and S3 glass particles. Blank determination for the liquid certified reference materials (i.e. U500 and IRMM-187) was performed by measuring 1% (m/m) HNO3 for 10 min. The blank signal of each isotope of interest was calculated as an average of six blocks. Prior to this each block had been assessed as an average of 100 single data points. The blanks of the isotopes of interest were then subtracted from the single recorded signal intensities of the glass particles and the liquid certified reference materials, respectively. Afterwards, the signal intensities, recorded in V, were converted into counts per second (cps) by using a factor of 6.24150965 × 107. Since 234U and 236U were measured with discrete-dynode secondary electron multipliers, dead time correction of these intensities was performed according to Eq. (1) (Nelms et al., 2001). (1) I t = I 0 ( 1 − I 0 τ ) where I t is the intensity (in cps) corrected for the dead time, I 0 is the raw intensity and τ is the dead time of the secondary electron multiplier. Dead times of 34 ns (U c = 9 ns) and 4 ns (U c = 1.5 ns) were applied for IC0 and IC1, respectively. The dead time and the linearity of the secondary electron multipliers were determined according to the dynamic procedure published by Richter et al. (Richter et al., 2009). Different concentrations of IRMM-073/7 (European Commission-JRC, Institute for Reference Materials and Measurements, Geel, Belgium) were used for determining the deviation of 233U/235U isotope ratios from the certified ratio at increasing 235U count rates. Mass bias correction was performed by measuring 235U/238U in these solutions. The 234U/238U, 235U/238U and 236U/238U isotope ratios of the liquid certified reference materials were calculated by dividing the average signal intensities, which were obtained with respect to the procedure already described for assessing the average blank signals, of the isotopes of interest. The calculation of the external correction factors is given in Eq. (2) (2) CF U X / U 238 = ( U X / U 238 ) certified ( U X / U 238 ) Bl and  τ corrected ,  lq X = 234 , 235 , 236 where CFUX / U238 CF U X / U 238 is the external correction factor, (UX / U238)certified ( U X / U 238 ) certified is the isotope ratio of the certified reference material and (UX / U238)Bl and τ corrected,  lq ( U X / U 238 ) Bl and  τ corrected ,  lq is the blank and dead time corrected isotope ratio of interest of the liquid (lq) sample. The external correction factor was determined twice a day according to standard-sample bracketing. The average of the daily determined external correction factors was applied for the correction of the glass particle isotope ratios. Signal intensities of the time resolved ablation profiles were corrected for blank and dead time. An ablated single particle and a typical ablation profile are given in Fig. 1. In order to enhance the precision of the background, the gas blank, used for blank correction, was determined by measuring it for 10 min as already described above. Additional gas blanks were recorded for about 15 s prior to the particle ablation and after it in order to monitor the complete wash-out of the previously analyzed samples. The blank and dead time corrected signal intensities were subject to a so-called logical test, which aimed at setting an operator-independent threshold for the selection of intensities used for further evaluation. The threshold was set to twenty times the standard deviation of the determined blank signal. The signal intensities of the isotope that yielded the smallest number of single data points above this threshold, defined the peak area used for further evaluation. Then the intensities of each isotope were integrated over the defined peak area, and the 234U/238U, 235U/238U and 236U/238U isotope ratios were calculated by dividing the integrated signal intensities of the isotopes. Integration was performed by summing up the single recorded signal intensities (i.e. recorded in time resolved analysis mode with an acquisition time of 1 s per data point) in Microsoft Excel 2007. Integrating over the defined peak area was performed because of counting statistics, thereby reducing the influence of small count rates on the isotope ratio. Finally, the blank and dead time corrected isotope ratios (UX / U238)Bl and τ corrected, GP ( U X / U 238 ) Bl and  τ corrected , GP of the glass particle (GP) were multiplied with the respective external correction factors CFUX / U238 CF U X / U 238 in order to obtain the final glass particle isotope ratio (UX / U238)corrected, GP ( U X / U 238 ) corrected , GP (see Eq. (3)). (3) (UX / U238)corrected, GP = (UX / U238)Bl and τ corrected,  GP x CFUX / U238 X = 234, 235, 236 ( U X / U 238 ) corrected , GP = ( U X / U 238 ) Bl and  τ  corrected ,  GP x CF U X / U 238 X = 234 , 235 , 236 2.3.1 Uncertainty evaluation Expanded total combined standard uncertainties (U c) were calculated according to ISO/GUM (ISO/IEC Guide 98-3, 2008) and EURACHEM (EURACHEM/CITAC Guide CG 4), using the GUM Workbench Pro software (version 2.4, Metrodata GmbH, Weil am Rhein, Germany). The single parameters that were propagated for computing the expanded total combined standard uncertainties of the 234U/238U, 235U/238U and 236U/238U isotope ratio measurements of the glass particles are given in Table 3. No significant contributions to the expanded total combined standard uncertainties could be identified with respect to Faraday cup gain calibration and different signal response/decay times of Faraday cups and secondary electron multipliers. Thus, these effects were not included in the uncertainty budget. 3 Results and discussion 3.1 Selection of isotope standard for calibration U500 and IRMM-187 were investigated for the determination of the external correction factor. The utilization of U500 was determined as this certified reference material exhibits an almost 1:1 235U/238U isotope ratio, which is regarded as preferable for the determination of mass bias effects. The applicability of IRMM-187, which is a low-enriched U isotopic reference material, was additionally studied as its U isotope amount ratios have lower expanded uncertainties compared to U500. The external correction factor corrects in one step for all effects present during the analysis, except for the liquid–solid matrix dissimilarity. The effects that are considered in the external correction factor are: mass bias (for 234U/238U, 235U/238U and 236U/238U), secondary electron multiplier yield and its variation (for 234U/238U and 236U/238U), peak tailing from 238U+ ions (for 236U/238U) and 235U1H+/235U+ hydride rate (for 236U/238U). The 235U1H+/235U+ hydride rate for both laser ablation and liquid measurements (i.e. introduction of dry aerosol) was about 1 × 10−6. The standard solution concentrations in the low ng g−1 range were applied in order to not overload the secondary electron multipliers. The maximum count rate determined for 234U was set to approximately 9.5 × 105 cps. Applying an external correction factor determined by means of IRMM-187 did not result in a significant difference, compared to U500 (see Fig. 2). Correction with IRMM-187 yielded a bias of 0.28% of the average (n = 25) 235U/238U glass particle isotope ratio, whereas the average of 235U/238U isotope ratios that were corrected with U500 showed a bias of 0.20% from the certified value. Both corrections resulted in accurate 235U/238U glass particle isotope ratios (see Fig. 2). The lower uncertainties of the isotope amount ratios of IRMM-187, compared to U500, were not reflected in the expanded total combined uncertainty as the main contribution to the uncertainty results from the repeatability of the glass particle measurements (see below). It was decided to apply U500 for external correction as determined mass bias effects are independent of detector non-linearities. We also investigated the applicability of using lower concentrations of U500 for calibration in order to account for low laser ablation signals. The final solution concentration was selected around 4 ng g−1 in order not to overload IC detectors and as lower concentrations led to higher uncertainties (i.e. 2.8% and 1.5% for 234U/238U and 235U/238U when applying a 0.4 ng g−1 standard compared to 2.6% and 1.4% at 4 ng g−1). In Fig. 3 both 234U/238U and 235U/238U glass particle isotope ratios that were corrected with U500 (4 ng g−1) are shown. These ratios were determined by applying spot sizes of 10 μm and 15 μm. The average 234U/238U and 235U/238U isotope ratios (see Fig. 3) measured in 28 S1 particles yielded biases of −0.51% and 0.15%, respectively. The maximum observed 238U signal intensities ranged from about 0.14 V to 2.1 V for the analyzed glass particles. Considering the combined standard uncertainties, 2.6% and 1.4% expanded uncertainties (k = 2) were computed for 234U/238U and 235U/238U isotope ratio measurements. Comparing the biases and the uncertainties of the isotope ratios, it can be concluded that the isotope ratios are in good agreement with the certified range. Moreover, an underestimation of the individual uncertainty contributions can be excluded. The main contribution to the uncertainty budget is produced by the repeatability of the measurements (i.e. 85.5% and 96.5% for 234U/238U and 235U/238U). In addition, it is evident that the different spot sizes do not influence the final results and can be chosen upon the size of the analyzed particles. The same applies for the other ratios (data not shown). 3.2 Laser ablation carrier gas It is well known (Eggins et al., 1998; Guillong and Günther, 2002; Günther and Heinrich, 1999; Horn and Günther, 2003) that the application of He as laser ablation carrier gas has certain advantages (e.g. smaller particle size distribution, enhanced signal intensities, increased transport efficiency, minimized elemental fractionation) for particular sample matrices over Ar, especially if a laser wavelength of 193 nm is applied. In this work both carrier gases were investigated with respect to their influence on the accuracy and precision of the isotope ratio analysis of glass particles. The analyses of the glass particles were, apart from the used carrier gas, performed under the same measurement conditions, applying spot sizes of 10 μm and an external correction factor determined by means of U500. Laser ablation under He atmosphere resulted in an average 235U/238U isotope ratio (n = 31) of 0.007 33(10), which deviated 0.41% from the certified ratio. In comparison, the average 235U/238U ratio (n = 25) of 0.007 32(10) that was obtained by using Ar as carrier gas showed a bias of 0.20% from the certified value. The given uncertainties of the average isotope ratios are total combined uncertainties. The applied carrier gas was not reflected in the expanded uncertainty, which was 1.4% for both applications. Therefore, no significant difference in the 235U/238U isotope ratios was observed (see also Fig. 4). This was additionally confirmed by the fact that no significant signal enhancement on the average laser signals of the particles was observed when He was used. Applying a spot size of 10 μm, the maximum observed 238U glass particle signal intensities ranged from 0.2 V to 0.7 V, both for Ar and He. Isotope ratios that were determined by means of ablation under He atmosphere exhibit a slightly better repeatability (i.e. 0.49%) compared to particles ablated under Ar atmosphere (i.e. 0.60%). However, the use of He was not regarded to be advantageous for the presented purpose. Therefore, Ar was used as carrier gas for further investigations. 3.3 Determination of 236U/238U isotope ratios The determination of low 236U/238U isotope ratios is usually difficult because of significant peak tailing from 238U+ ions on the mass of 236U, and interference by 235U1H+ ions. The peak tailing in a magnetic sector-field mass spectrometer mainly results from low-energy ions, which originate from ion collisions with residual gas molecules and ion energy variations after the ESA and magnetic sector (Boulyga et al., 2006). In order to reduce the amount of these low-energy ions and improve the 236U/238U abundance sensitivity, a deceleration filter was applied for the determination of 236U. The optimization of the deceleration filter is described in detail elsewhere (Boulyga et al., 2006). The ablation of the S1 glass particles was accomplished using Ar as the carrier gas. External correction was performed by means of U500 CFU236 / U238 CF U 236 / U 238 correction factors. Applying the deceleration filter improved the 236U/238U abundance sensitivity by almost one order of magnitude (i.e. 1.26 × 10−7 compared to 1.06 × 10−6 with the filter turned off) and led to a reduced bias from the certified range and significantly reduced uncertainties. Results that were obtained with and without deceleration filter are given in Fig. 5. The average of 61 particles of 236U/238U isotope ratios yielded a bias of −2.48% from the certified value. In contrast, the average of 28 particles, which were analyzed with the deceleration filter turned off, yielded a bias of 16.3%, which was mainly caused by the poor abundance sensitivity. The calculation of combined standard uncertainties for 236U/238U glass particle isotope ratio measurements yielded an expanded uncertainty of 20% (k = 2) for measurements with the deceleration filter turned off (see Table 4). The main contribution to the expanded total combined uncertainty budget was assigned to the peak tailing from 238U+ ions, which accounts for 78.9% of the expanded total combined standard uncertainty (see Table 4). However, the application of the deceleration filter reduced the expanded uncertainty to 5.8% (k = 2) and the contribution of the 238U+ peak tailing to the uncertainty budget to 19.4% (see Table 4). In the latter case, the repeatability of the measurements was the main contributor to the expanded total combined uncertainty budget as it accounted for 26.4%. 3.4 Reproducibility 235U/238U isotope ratio measurements of S1 glass particles, applying Ar as the carrier gas and spot sizes of 10 μm, were performed within two independent measurement series for investigating the reproducibility of the method. The second measurement series was repeated after nine months. The reproducibility, calculated as RSD of forty-two 235U/238U isotope ratio measurements, is 0.62%. An external repeatability (calculated as RSD) of 0.65% (n = 17) and 0.60% (n = 25), respectively, was observed for the 1st and the 2nd measurement series for 235U/238U. 234U/238U and 236U/238U isotope ratio measurements yielded an external repeatability of 0.71% (n = 17) and 1.84% (n = 22), respectively, for analyses using 10 μm spot ablation and Ar as the carrier gas. A summary of the U isotope ratio measurements of S1 glass particles is given in Table 5. In comparison, Varga reported precisions of 0.9–5.1% for the 235U/238U isotope ratio analysis of individual U3O8 particles, with lateral dimensions ranging from 10 to 30 μm. The reported precision for the single collector measurements was achieved by performing partial ablation of the particles together with applying low laser energy (Varga, 2008). A relative precision of 0.22% (2σ) for 235U/238U LA-MC-ICP-MS isotope ratio measurements, applying spots of approximately 25 × 14 μm, of 138 natural uraninite grains was reported by Lloyd et al. (Lloyd et al., 2009). 3.5 Demonstration of applicability of developed method by means of S3 glass particle measurements S3 glass particles were analyzed under optimized conditions (i.e. carrier gas: Ar; external correction: U500; application of deceleration filter) to demonstrate the applicability of the developed method for the investigation of low-enriched uranium particles. Spot sizes of 10 μm and 15 μm were applied for ablation of S3 glass particles. The maximum 238U signal intensities ranged from about 0.05 V to 1.5 V. The determined 234U/238U, 235U/238U and 236U/238U isotope ratios of the S3 glass particles, which are shown in Table 6, were in good accordance with the certified value. Relative expanded uncertainties (k = 2) of 3.8% (234U/238U), 0.8% (235U/238U) and 6.7% (236U/238U) were calculated for the isotope ratio measurements (see Table 4). In case of 234U/238U and 235U/238U, the main contribution to the expanded total combined uncertainty budget (i.e. 89.9% and 75.9% for 234U/238U and 235U/238U, respectively) resulted again, as already observed for S1 particles, from the repeatability of the glass particle measurements. Considering the relative distribution of parameters propagated for the expanded total combined standard uncertainty of 236U/238U isotope ratio measurements, the main contributors are the variation of the IC0 yield (i.e. 46.3%) and the repeatability (i.e. 15.7%). In contrast to S1 glass particles, no contribution was assigned to the peak tailing from 238U+ ions. 238U peak tailing does less affect the determination of 236U of S3 glass particles as they exhibit an about 42 times higher 236U/238U isotope ratio than S1 particles (see Table 1). 4 Conclusions We demonstrated the applicability, reliability and robustness of LA-MC-ICP-MS for the direct analysis of distinct individual particles with respect to their U isotopic composition (i.e. 234U/238U, 235U/238U and 236U/238U). Glass particles, which were produced by the IRMM in order to simulate environmental samples containing hot particles, were used for the full evaluation of the presented method. The used particles are certified with respect to their U isotopic composition, including 236U, and they are exhibiting a particle size down to 10 μm with U amounts in the low picogram range. As U-containing particles were blended with inactive matrix glass, matrix interferences resulting from, for example, Al, Pb or Si could be excluded by analyzing the inactive matrix glass. Thus, the glass particles were ideal for performing a full validation of LA-MC-ICP-MS for the presented aim by means of certified reference materials. The application of a deceleration filter was proven to be a prerequisite in particular for the accurate determination of very low 236U/238U isotope ratios (i.e. 10−5). Expanded total combined standard uncertainty calculations revealed the main contributors to the uncertainty budgets and helped to optimize the analytical procedure. It is evident that combined uncertainties have to be evaluated according to the abundance ratio as contributors may change significantly when ratios vary within orders of magnitude. The presented semi-automatic evaluation approach may result in a future use of this technique in safeguards and/or (nuclear) forensics with respect to the analysis of environmental samples. However, pre-selection of U-containing particles by means of, for example, SEM/EDX prior to LA-MC-ICP-MS analysis is regarded to be advantageous for the application of the presented method for routine analysis.