Application of Nanosecond-UV Laser Ablation ... - ACS Publications

Aug 29, 2011 - involved in safeguard analyses: (i) thermal ionization mass spec- trometry ... laser.32 However, only relatively “large” particles,...
0 downloads 0 Views 2MB Size
ARTICLE pubs.acs.org/ac

Application of Nanosecond-UV Laser AblationInductively Coupled Plasma Mass Spectrometry for the Isotopic Analysis of Single Submicrometer-Size Uranium Particles Fabien Pointurier,* Anne-Claire Pottin, and Amelie Hubert CEA, DAM, DIF, F-91297 Arpajon, France ABSTRACT: For the first time, laser ablationinductively coupled plasma mass spectrometry (LA-ICPMS) was used to carry out isotopic measurement on single submicrometer-size uranium particles. The analytical procedure was applied on two particle-containing samples already analyzed in the same laboratory by established techniques for particle analysis: combination of the fission track technique with thermo-ionization mass spectrometry (FT-TIMS) and secondary ion mass spectrometry (SIMS). Particles were extracted from their initial matrix with ethanol and deposited on a polycarbonate disk where they were fixed in a layer of an organic compound (collodion). Prior to the isotopic analysis, particles were precisely located on the disk’s surface by scanning electron microscopy (SEM) for one sample and using the fission track technique for the other sample. Most of the particles were smaller than 1 μm, and their 235U content was in the femtogram range. 235U/238U ratios were successfully analyzed for all located particles using a nanosecond-UV laser (Cetac LSX 213 nm) coupled to a quadrupole-based ICPMS (Thermo “X-Series II”). LA-ICPMS results, although less precise and accurate (typically 10%) than the ones obtained by FT-TIMS and SIMS due to short (2040 s), transient, and noisy signals, are in good agreement with the certified values or with the results obtained with other techniques. Thanks to good measurement efficiency (∼6  104) and high signal/noise ratio during the analysis, LA-ICPMS can be considered a very promising technique for fast particle analysis, provided that uranium-bearing particles are fixed on the sample holder and located prior to isotope measurement.

F

or safeguard and environmental monitoring purposes, it has become increasingly important to develop and apply analytical techniques to particulate matter collected at both established nuclear facilities and locations suspected of clandestine nuclear material handling.1 Indeed, this airborne fine particulate material is released in the handling of nuclear material. These highly mobile particles are found in many locations in the facility and their isotopic, elemental, and structural compositions provide specific information about the activities carried out in the installation.2 Therefore, the goal of the analyses is to detect the presence of any particles of nuclear materials (mainly uranium) and then determine their isotopic compositions. However, analytical techniques must be extremely sensitive as particle sizes are usually very low, most of the time in the submicrometer scale and at most only a few micrometers. Therefore, the particles analyzed in real-life samples contain a very low amount of uranium (in the femtogram to the nanogram range). Various analytical techniques can be used for the isotope ratio analysis of single particles. Before the isotopic measurement, uranium-bearing particles must be identified and located. This can be done using several techniques. The most straightforward is scanning electron microscopy equipped with an energy-dispersive or wavelength-dispersive X-ray analyzing system (SEM/EDX or SEM/WDX)3 and software for automated detection of particles (so-called “Gun Shot Residue” software, GSR).4 Secondary ion r 2011 American Chemical Society

mass spectrometers (SIMS) equipped with appropriate software can also be used for the initial detection and localization of uranium-bearing particles before isotopic analysis.57 Lastly, the fission track (FT) technique allows the specific detection and localization of particles which contain the highest proportions of fissile isotopes (235U and/or 239Pu).811 It should be noted that autoradiography methods can also be used but are not sensitive enough for long-lived α-emitting radio-nuclides like the main uranium isotopes. Therefore, this technique can only be applied for very large particles of uranium that are highly enriched in 235U, which are rarely present in real-life samples.6,12 To determine the isotopic composition of the uranium particles, two methods are routinely used by the few laboratories involved in safeguard analyses: (i) thermal ionization mass spectrometry (TIMS) following initial localization using the fission track technique and sampling and loading of the individual particles on a filament10,11 or possibly the location of uranium particles by SEM and isolation of individual particles using in situ SEM microsampling;1315 (ii) SIMS, directly at the surface of a carbon disk on which particles were deposited, generally after initial localization using SEM or SIMS itself.47,16,17 Received: June 22, 2011 Accepted: August 29, 2011 Published: August 29, 2011 7841

dx.doi.org/10.1021/ac201596t | Anal. Chem. 2011, 83, 7841–7848

Analytical Chemistry However, these methods present a number of limitations. For example, the implementation of the FT method requires a nuclear reactor with appropriate thermal neutron spectrum and flux. This hinders its widespread and routine use. Moreover, particle microsampling, loading on filaments, and TIMS isotopic measurements are relatively time-consuming. SIMS is a very sensitive surface analysis technique and has a significantly higher sample throughput compared to TIMS. However, molecular ions emitted from the analyzed area having the same mass as the analyte of interest (for instance, 207Pb27Al+ and 206Pb12C16O+ at m/z = 234, 207Pb12C16O+ at m/z = 235, 235UH+ and 208Pb12C16O+ at m/z = 236) may give rise to isobaric interferences.7,1618 Therefore, inaccurate results can be obtained, especially for the lowabundance 234U and 236U isotopes. Inductively coupled plasma mass spectrometry (ICPMS) is a very sensitive technique, with detection limits in the femtogram range for isotopes like 236U or 239Pu1926 and is increasingly used for the measurement of long-lived isotopes. Moreover, this technique can be easily coupled with laser ablation (LA) for direct measurement of solid samples. Thanks to a high-energy density laser beam, a very small amount of material, in the fgpg range, is ablated, evaporated, and carried by a gas flux (argon or helium) directly into the plasma. Detection limits for actinides are in the ng g1μg g1 range depending on sample type.2730 This technique offers a very good spatial resolution of up to a few micrometers, depending on the diameter of the laser beam. The potential advantages of LA-ICPMS are sensitivity and fast sample throughput, as only a few minutes may be enough to perform an isotopic analysis. The feasibility of LAICPMS for the analysis of single uranium particles has already been demonstrated by a few teams12,31 and by the authors through a collaboration with the Laboratoire de Chimie Analytique and Bio-Inorganique (LCABIE, Pau, France) equipped with a femtosecond-IR laser.32 However, only relatively “large” particles, typically a few tens of μm in diameter, were successfully analyzed.12,31 Such particles are rarely encountered in real-life samples. It is obvious that analysis of relatively large particles is easier than analysis of μm-size particles. In fact, a 20 μm-diameter particle of natural uranium such as U3O8 (density 8.22 g cm3) contains about 30 ng of uranium whereas a 1 μm-diameter particle of the same material contains only 3.6 pg of uranium, i.e., ∼25 fg of 235 U and ∼0.2 fg of 234U. The use of LA-ICPMS for isotope measurement of micrometersize uranium dioxide (UO2) particles was only briefly mentioned by Becker et al.33 To our knowledge, application of laser ablation ICPMS to the isotope analysis of submicrometer-size uranium particles has never been described. In this paper, we describe a fast and sensitive analytical method for measurement of uranium particles using a commercially available nanosecond-UV laser coupled with a quadrupole-based ICPMS and its application to the measurement of the 235U/238U uranium isotope ratio in single submicrometer-size uranium particles. These particles are extracted from two samples, one round-robin sample originally consisting of a graphite planchet on which uranium oxyfluoride (UO2F2) particles have been deposited and one cotton cloth sample (called “swipe sample”) used by IAEA inspectors for collecting dust in a nuclear facility. In this real-life sample, micrometer-size uranium particles are trapped, among a far larger number of other kinds of particles (iron oxides, lead, cellulose, etc.), in the cotton cloth fibers. The results for both samples were compared with those obtained in the same laboratory with the two established

ARTICLE

Table 1. Relevant LA and ICPMS Parameters for the Particle Analyses Described in This Paper LA Parameters Ablation rate

20 Hz

Spot size Pulse duration

100 μm ∼5 ns

Energy

∼2 mJ

Energy fluence

∼25 J cm2 (100 μm spot)

Ablation scheme

100 shots, single point

Total duration of the ablation

10 s

He flux

500 mL min1 ICPMS Parameters

ICPMS sensitivity (uranium) ICPMS dwell time

5  108 counts s1 per μg mL1 100 ms, peak jumping mode 235

U, 238U

Measured isotopes Total acquisition time

90 s

Ar flux

700 mL min1

techniques for particle analysis, FT-TIMS and SIMS. The overall measurement efficiency of the LA-ICPMS method is estimated, and relative benefits and drawbacks of the method with respect to the established techniques are discussed.

’ EXPERIMENTAL SECTION Instruments and Reagents. Isotopic analysis was carried out using a quadrupole-based ICPMS (“X-Series II”, ThermoFischer Scientific, Bremen, Germany). The operation and optimization of this instrument for the measurement of low amounts of actinides has already been described in detail in a previous publication.34 Briefly, this instrument is equipped with nickel cones, a platinum guard electrode and, with the S-option, an additional primary pump that decreases vacuum at the interface and provides an increase in sensitivity by a factor of 2 compared to an instrument not equipped with the S-option. A Teflon microconcentric nebulizer (“PFA 100”, Elemental Scientific, Inc., Omaha, NE, USA) in self-aspiration mode (∼100 μL min1) is used for sample introduction. Performance is optimized on a daily basis with respect to sensitivity, short-term stability (10 min), and background. Sensitivity is approximately 5  108 counts s1 for 1 μg mL1 of 238U. The background obtained with deionized water acidified to 2% with ultrapure grade nitric acid is about 0.5 count s1 for mass-to-charge ratios of 240, 241, and 242. For specific coupling with the laser ablation device, a two-inlet torch mixes the laser-generated aerosol with a liquid aerosol produced by means of the pneumatic micronebulizer described above. This dual-flow introduction system allows a complete and easy optimization of the LA-ICPMS coupling and ensures more stable signals when analyzing solid samples. The plasma was constantly maintained under these wet conditions by introducing an aqueous solution of deionized water acidified to 2% with ultrapure grade nitric acid. Laser ablations were carried out using a LSX-213 system (Cetac Technologies, Omaha, NE, USA) equipped with a Nd: YAG laser of a wavelength of 213 nm. The ablated material was transported into the plasma using helium as a carrier gas. Optimized ICPMS and LA parameters are summarized in Table 1. 7842

dx.doi.org/10.1021/ac201596t |Anal. Chem. 2011, 83, 7841–7848

Analytical Chemistry

ARTICLE

Ultrapure grade nitric acid (Merck, Darmstadt, Germany) and ultrapure deionized water (Milli-Q system, Millipore, USA) were used to prepare solutions for ICPMS performance optimization and calibration. Mass calibration of the ICPMS was checked using both a natural uranium standard reference material in aqueous solution (IRMM-184, Institute for Reference Materials and Measurements, Geel, Belgium) diluted to ∼1 ng mL1 and the NIST 612 glass reference material (NIST, Gaithersburg, MD, USA). Sample Preparation. Uranium particles were originally deposited on a graphite planchet (first sample) or embedded in a 10 cm  10 cm cotton cloth (Texwipes, Kernersville, North Carolina, USA) wiped on a smooth surface inside a nuclear facility (second sample). Particles are extracted from the graphite planchet simply by leaching a fraction of the planchet with ethanol (Normapur grade, VWR Prolabo, Fontenay-Sous-Bois, France). Particles are extracted from the cotton cloth in an ultrasonic bath with ethanol. Then, in both cases, the specific preparation procedure designed for the FT-TIMS particle analysis method is applied. In this way, particles are dispersed in a mixture of collodion (4% nitrocellulose in etheralcohol media, Merck) and ethanol and deposited on 30, 1 in. diameter polycarbonate disks (“Lexan”), called “deposition disks”. After evaporation of the ethanol, particles are fixed in a collodion layer with thickness of a few μm. Other polycarbonate disks which act as the fission track detectors are welded on each one of the deposition disks. After intense irradiation by a thermal neutron integrated flux of ∼1015 cm2 in the Orphee reactor (Saclay, France), the fission track detectors are separated from the deposition disks and etched in a concentrated sodium hydroxide solution (35 min at 60 °C). Latent damage in the detector material resulting from the interaction with the high-energy fission products along their flight path are enlarged by this process to become tiny holes which are visible under an optical microscope. The positions at which fissile radionuclide (239Pu, 235U, 233U) containing particles are detected as the origin of a group of fission tracks are projected back to the deposition disk. Correspondence between the positions of each fission track cluster and the positions of the corresponding particles are established thanks to landmarks and a relatively straightforward two-point algorithm. Precision of relocation is typically (2030 μm, which is enough for our analytical procedures. It should be noted that a better precision ((10 μm) can reportedly be obtained with a 3-point algorithm.4 Landmarks are very small holes formed by pushing a very thin tip through both the deposition disk and the corresponding fission track detector disk. When isotopic analysis by TIMS needs to be performed on particles, small squares of the collodion film containing the detected particles are cut out using a microhandling device (Leica, Wetzlar, Germany) under an optical microscope and attached to the TIMS filaments. All these treatments are carried out in a class10 (ISO-4) clean room facility. For LAICPMS analysis, the theoretical positions of the particles were calculated inside the ablation cell referential thanks to the aforementioned landmarks and algorithm. An interesting feature of this method is that the theoretical number nFT,2π of fission tracks recorded on the revelation disk (2π steradian) for a given particle is linked to the mass of 235U in the particle according to the following formula: nFT, 2π ¼ n235 U 3 Fth 3 σ th, 235 U

ð1Þ

where n235U is the mass of the 235U isotope, Fth is the integrated flux of thermal neutrons (8.2  1014 cm2), and σth,235U is the 235 U cross section for thermal neutrons (582  1024 cm2). We assume here that two fission fragments are produced through disintegration of a 235U nucleus. Thus, the number of observed fission tracks can be related to the apparent diameter of the particle (assuming the particle is spherical) by: π dUO2 NA 3 Φ3 nFT, 2π ¼ 1012 3 3 A235 U 3 Fth 3 σth, 235 U 3 6 MUO2 3

ð2Þ

where dUO2 and MUO2 are, respectively, the density and molar mass of the uranium compound (here UO2), A235U is the 235U abundance in the particle, NA is Avogadro's number (6.02  1023 atoms mole1), and Φ is the diameter of the particle expressed in μm. It is then relatively straightforward to calculate that one fission track recorded on the revelation disk is theoretically equivalent to 0.8 fg of 235U and to a spherical natural uranium UO2 particle of 0.28 μm. Uranium-bearing particles can also be detected using SEM/ EDX equipped with GSR software (Philips-FEI XL30 ESEM, Eindhoven, The Netherlands). An automated measurement session allows the identification of particles displaying a high backscattered electron signal related to the mean atomic number of the analyzed area. Therefore, the particles which contain high atomic number elements like uranium display the more intense signals and are selectively detected. With such a device, only particles larger than 1 μm can be identified if the whole surface of the deposition disk has to be scanned. However, smaller particles can be detected within a reasonable 12 h analyzing time if only a small portion of the disk (a few mm2) is scanned. In the same way, the theoretical positions of particles are calculated inside the ablation cell’s referential using the same algorithm and landmarks. It should be mentioned that use of collodion to fix the particles at the surface of the holder (polycarbonate disk) is a key point. Without this coating, micrometer-size particles may be blown out or moved by the laser ablation shots. Data Evaluation. Rapid ablation of the analyzed single particles results in short, very noisy, and transient signals.12,31 Under such conditions, it is very difficult to obtain true and precise isotopic ratios, especially if the isotopes of interest are measured sequentially using a single detector, which is the case for our quadrupole-based instrument. Only the 235U and 238U isotopes were measured. Minor 234U and 236U isotopes were not measured as corresponding count rates are expected to be very low, probably below the detection limits, for most of the analyzed particles. It is not easy to estimate uncertainty with very transient signals measured sequentially with a single detector. The method proposed by Varga,31 which consists of taking several replicates (35 spectra) by ablating only a fraction of each investigated particle, is not applicable to μmsize particles from real-life environmental samples, because the totality of the particulate material has already been consumed at the end of the first ablation. Raw results are treated as follows. First, signals at mass-tocharge ratios 235 and 238 are averaged over 1 s. A period of about 15 s before the ablation debris enters the plasma is used for the background correction. A raw signal is considered to be significant only if it is higher than the mean background plus three times the standard deviation of the background. Therefore, 7843

dx.doi.org/10.1021/ac201596t |Anal. Chem. 2011, 83, 7841–7848

Analytical Chemistry

ARTICLE

calculations are performed on n significant repetitions (n generally between 20 and 40 depending on the mass of ablated material) for which both signals at m/z 235 and 238 are significantly above background. 235U/238U atomic ratios are calculated simply by integrating the signals of the n significant repetitions obtained for each mass: n

R̅ ¼

∑ ðc0235, i þ c0238, i Þ 3 Ri i¼1 n

∑ ðc0235, i þ c0238, i Þ i¼1

ð3Þ

with c0 235,i = c235,i  c235,0, c0 238,i = c238,i  c238,0, and Ri = (c0 235,i)/ (c0 238,i); where c235,i and c238,i are, respectively, the count rates (counts s1) measured at mass-to-charge ratios which correspond to 235U and 238U during the repetition number i, and c235,0 and c238,0 are, respectively, the average count rates (counts s1) measured at m/z 235 and 238 during the baseline measurement before the particle’s signal. Therefore, results are not filtered or corrected, except for baseline correction. In particular, no mass bias correction is applied because it is regarded as negligible in comparison to other sources of errors. By the same token, the standard deviation is calculated over individual ratios obtained for each 1 s repetition, weighed by the sum of the count rates measured for the corresponding repetition: vffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffi u n 2 u ðc0 þ c0238, i Þ 3 ðRi  RÞ ̅ u 1 ui ¼ 1 235, i ð4Þ uðRÞ ̅ ¼ pffiffiffi 3 u n n t ðc0235, i þ c0238, i Þ  1

Figure 1. Typical example of LA-ICPMS signals at m/z = 235 and 238 obtained for a submicrometer-size uranium particle from the Nusimep-6 sample (particle #18).





i¼1

’ RESULTS AND DISCUSSION Analysis of the Nusimep-6 Round-Robin Sample. The LAICPMS technique was tested on the round robin sample “Nusimep-6” prepared by the IRMM (Institute for Reference Materials and Measurements, Geel, Belgium). This sample originally consisted of numerous and very small (most of them were submicrometer-size) certified uranium oxyfluoride (UO2F2) particles deposited on a graphite planchet.35 The certified 235U/238U atomic isotopic ratio (0.0070439 ( 0.0000070, uncertainty with a coverage factor of 2) is identical for all particles. It is very close to the natural uranium ratio (0.0072623 ( 0.0000021). This sample was already analyzed at CEA both by FT-TIMS and by SIMS. For the LAICPMS analysis, we reused one of the 50 Lexan disks prepared for FTTIMS analysis. In this case, U-bearing particles were located by SEM/EDX on a small part of the disk’s surface (a few mm2) with “extreme” conditions, so as to detect and locate submicrometer particles. Nineteen particles were located on the polycarbonate disk and were all successfully analyzed by LA-ICPMS. The LA-ICPMS isotopic analysis (ICPMS optimization, particles relocation inside the ablation cell, isotopic analysis) lasted one day. About one day was also necessary for identification and location of the particles by SEM. Only signals at m/z 235 and 238 were recorded. Sum of integrated count rates for both mass to charge ratios were between 10 000 and 250 000 counts and were below 20 000 counts for most of the particles. Typical ICPMS signals

Figure 2. Measured 235U/238U atomic ratios for the Nusimep-6 sample particles analyzed by LA-ICPMS. Expanded uncertainties are given with a coverage factor of 2. Particle #12 is extremely small, and its measurement results in a poor accuracy and a very large uncertainty.

recorded at mass/charge ratios corresponding to 235U and 238U are presented in Figure 1 for one of these particles (particle #18). Results of all 235U/238U ratio measurements for Nusimep-6 particles are plotted in Figure 2. Accuracy is acceptable as all of the errors bars (coverage factor of 2) overlap the target value. Average atomic ratio over the 18 particles is 0.00743 ( 0.00077 (coverage factor of 1), whereas the average ratio weighed by integrated count rates for all analyzed particles is 0.00712 ( 0.00041, i.e., closer to the certified value. Both values are in good agreement with the certified value. Relative uncertainties (k = 1) on the 235U/238U ratios are between 2.0% and 37% with an average value of 8.9%. LA-ICPMS results obtained for the 235U/238U ratios can be compared with those obtained in our laboratory by fission tracksTIMS and by SIMS (see Table 2). The standard deviations on the individual 235U/238U atomic ratios provide an indication of the reproducibility of the method. The average absolute errors are the averages of all the absolute deviations of each isotope ratio with respect to the certified 235U/238U atomic ratio. They give a good idea of the accuracy of the method. It appears that LA-ICPMS results, although they match the certified value, are less precise and accurate than the FTTIMS and SIMS results, as expected. LA-ICPMS accuracy and 7844

dx.doi.org/10.1021/ac201596t |Anal. Chem. 2011, 83, 7841–7848

Analytical Chemistry

ARTICLE

Table 2. Results obtained by LA-ICPMS (with Location of U Particles by SEM/EDX), FT-TIMS, and SIMS for the Measurement of 235U/238U Atomic Ratios on Particles from the “Nusimep-6” Round-Robin Samplea Techniques Number of particles Average isotope ratio (IR)

LA-ICPMS 19

FT-TIMS 18

SIMS 30

0.00743

0.00709

0.00703

Mean uncertainty

0.00077

0.00054

0.00012

Mean relative uncertainty

8.9%

7.6%

1.7%

0.00180 24.2%

0.00034 4.8%

0.00005 0.7%

0.00086

0.00019

0.00004

2.7%

0.5%

Uncertainty (k = 1)

Reproducibility (k = 1) Standard deviation on IR Relative SD on IR Accuracy Average absolute error Relative absolute error

11.5%

Figure 3. Typical example of LA-ICPMS signals at m/z = 235 (black line) and 238 (gray line) obtained for the micrometer-size uranium particle #12 from the IAEA sample #21954-11-01. An image of the corresponding fission track cluster is given on top right of the figure.

a

The SIMS is a single collector instrument. The TIMS is a multiple collector instrument but is used in this case in the single-collector peakjumping mode. It should be noted that mean uncertainty is higher than reproducibility for both TIMS and SIMS techniques. The reason for this unusual observation is that uncertainties for SIMS and TIMS isotope measurements are rather conservative so as to take into account possible biases mainly due to polyatomic interferences for SIMS and mass fractionation effect for TIMS.

reproducibility are typically in the 1020% range, whereas they are of only a few % for FT-TIMS and below 1% for SIMS. This is obviously due to the extremely noisy and transient nature of the LA-ICPMS signals. Analysis of an IAEA Swipe Field-Sample. We also analyzed a real-life swipe field-sample provided by the IAEA (#21954-1101). This sample was previously analyzed by FT-TIMS in our laboratory. Two isotopic compositions were evidenced by FTTIMS: one natural uranium composition and one low-enrichment uranium (LEU) (average 235U/238U atomic ratio of 3.2  102). Most of the particles were close to or below 1 μm. Again, we reused one of the 30 polycarbonate disks initially prepared for FT-TIMS analysis. For this sample, however, uranium-bearing particles were located thanks to the corresponding fission track positions on the revelation disk. Twenty-seven fission track clusters containing at least 4 fission tracks were observed, and the corresponding particles were located for further analyses by LA-ICPMS. Most of the clusters contain less than 10 fission tracks, although a few of them account for a few tens of tracks. In contrast, only the largest particles located on the 50 deposition disks are microsampled for TIMS analysis. Furthermore, it should be noted that counting the fission tracks is not easy, especially for the largest particles as tracks are very close to each other and can overlap. If it is assumed that the particles are made of UO2 (density of 10.96 g cm3) and according to formula 1, the corresponding 235U masses range from 3.5 fg to ∼200 fg. By reversing the formula 2, calculated equivalent diameters (assuming that particles are spherical) for natural uranium such as UO2 range from 0.45 to 1.8 μm, with an average equivalent diameter of 0.7 μm. Only mass-to-charge ratios of 235 and 238 were recorded, except in the case of particle #27, for which m/z values of 234, 235, 236, and 238 were recorded, because the highest number of fission tracks were observed (250 ( 50) for this particle. The LA-ICPMS isotopic analysis (ICPMS optimization,

Figure 4. Measured 235U/238U atomic ratios for the IAEA sample #21954-11-01 uranium particles analyzed by LA-ICPMS. Expanded uncertainties are given with a coverage factor of 2.

particle relocation inside the ablation cell, isotopic analysis, calculations) lasted one day. About 2 h were necessary for identification and location of the fission track clusters on the revelation disk. Signals recorded for 235 U and 238 U are also very short (between 20 and 40 s), noisy, and transient, but they are of higher intensities than those obtained with “Nusimep-6” particles. Particles from the IAEA sample are obviously larger. Integrated count rates ranged from 4800 to 4.3  106 counts. A typical example of the results obtained for one particle (here particle #12) is given in Figure 3. For this particle, the corresponding fission track group accounts for 5 fission tracks. All 235U/238U atomic ratios are also plotted in Figure 4. The two isotopic compositions are clearly identified: 3 particles show natural 235U/238U isotope ratios and 24 particles show LEU ratios ranging from 1.85  102 to 5.35  102. A good agreement is obtained between LA-ICPMS and TIMS results (see Table 3) although LA-ICPMS LEU isotope ratios showed a higher dispersion around the average value than in the case of TIMS (standard deviation on isotope ratio of 22.9% and 5.3%, respectively). This phenomenon can be explained by a lower accuracy and precision due to uneven signals and by a “mixing effect” due to the simultaneous analysis of several 7845

dx.doi.org/10.1021/ac201596t |Anal. Chem. 2011, 83, 7841–7848

Analytical Chemistry

ARTICLE

Table 3. Results Obtained by LA-ICPMS (with Location of U Particles by Fission Tracks) and FT-TIMS on the Measurement of 235U/238U Atomic Ratios on Particles from the IAEA Sample #21954-11-01 Technique

LA-ICPMS

TIMS

NU particles Number of particles

3

3

Mean isotope ratio (IR)

0.0085

0.0085

0.0012

0.0006

Figure 5. SEM image of the particle #18 from the Nusimep-6 sample.

Uncertainty (k = 1) Mean uncertainty (k = 1) Mean relative uncertainty Reproducibility (k = 1) Standard deviation (SD) on IR Relative SD on IR

12.7% 0.0017 20.1%

6.8% 0.0019 22.7%

Lowest ratio

0.0070

0.0073

Highest ratio

0.0103

0.0108

LEU particles Number of particles Mean isotope ratio (IR) Uncertainty (k = 1) Mean uncertainty Mean relative uncertainty

24 0.0338 0.0023 6.9%

19 0.0327 0.0037 11.1%

Reproducibility (k = 1) Standard deviation (SD) on IR Relative SD on IR

0.0077 22.9%

0.0017 5.3%

Lowest ratio

0.0185

0.0302

Highest ratio

0.0535

0.0386

micrometer-size particles in the same ablated area. Actually, a few ICPMS signals show several peaks, which may indicate the presence of as many particles in the analyzed area. Lastly, the measurement of minor uranium isotopes 234U and 236 U was attempted for particle #27, for which 250 ( 50 fission tracks were counted. 236U/238U could be measured neither by LA-ICPMS nor by FT-TIMS. However, a 234U/238U atomic ratio of (3.27 ( 0.47)  104 (uncertainty k = 1), in good agreement with the one measured by FT-TIMS ((3.17 ( 1.02)  104 on average for 11 particles), was measured. Estimation of the Overall Ionization Efficiency of the LA-ICPMS Technique. The overall ionization efficiency (i.e., ratio of the number of detected ions to the number of uranium atoms in the particle) can be estimated on the basis of the two series of analyses presented in the previous section of this paper. In the case of the “Nusimep-6” round-robin sample, the mass of uranium contained in a particle can be estimated on the basis of the apparent diameter measured on the SEM image of one particle, assuming that the particle is homogeneous, devoid of internal porosity, and is made of UO2F2 (density of 3.37 g cm3). This estimation makes sense only if the particle is spherical (i.e., if its volume can be calculated) and if it is sufficiently separated from the other uranium particles (so that this particle is the only one analyzed). These conditions are encountered only in the case of particle #18 (see Figure 5), which is apparently spherical with an observed diameter between 0.2 and 0.3 μm. Therefore, this particle theoretically contains between 14 and 48 fg of uranium, of which only between 0.10 and 0.33 fg are 235U. From these masses and the recorded 10 700 counts at m/z = 235 and 238, the overall calculated

Figure 6. Plot of the integrated numbers of counts registered at m/z = 235 by LA-ICPMS analysis of particles from the IAEA sample #2195411-01, versus the numbers of fission tracks counted for each analyzed particle.

measurement efficiency of the LA-ICPMS analysis is between 1.2  104 and 3.9  104. This value is at least equal to, but most probably greater than, the measurement yield for the usual introduction system (micronebulizer 100 μL min1, cooled spray chamber) which is on average ∼1.2  104, assuming a count rate of 5  108 counts s1 per μg mL1 for 238U. This is only a rough estimation as the diameter, chemical composition, and internal structure of the particle are not known with certainty. In the case of the IAEA swipe sample, an acceptable correlation is obtained between the numbers of integrated counts at m/z = 235 and the corresponding observed number of fission tracks (see Figure 6). One fission track is on average equivalent to 1290 ( 230 integrated counts (uncertainty k = 2) at m/z = 235. Therefore, as one fission track is also theoretically created by 0.8 fg of 235 U, the overall calculated measurement yield of the LA-ICPMS analysis is (6.3 ( 1.2)  104 . This value is higher than that derived from the particle analysis of the Nusimep-6 sample and also significantly higher than the measurement efficiency calculated for uranium with the same instrument using the classical liquid introduction system (∼1.2  10 4 ). More confidence is given to the measurement yield obtained with this method, as all particles were taken into account, whereas only one particle was considered in the case of the “Nusimep-6” sample.

’ CONCLUSION On the basis of the results described in this paper, the relative advantages and drawbacks of the three techniques (SIMS, FTTIMS, and LA-ICPMS) can be discussed. The SIMS technique by itself allows both fast location and isotopic analysis of particles 7846

dx.doi.org/10.1021/ac201596t |Anal. Chem. 2011, 83, 7841–7848

Analytical Chemistry

ARTICLE

Table 4. Estimated Analytical Delays (Days) for the Analysis of Two Samples (30 Uranium Particles Per Sample) with the Various Particle Analysis Techniques Used in This Worka SEM-LATechniques Deposition on disks

SIMS 1

ICPMS 1

Irradiation Particle deposition on filament Particle localization Isotope measurements Total

FT-LAFT-TIMS

ICPMS

1

1

10

10

3 3

1 1.5

1 6

4

3.5

21

0.5 1.5 13

These figures are given provided all instruments, equipments, and technical staff are available. Irradiation time includes transport and cooling of the samples. Isotope measurement time includes instrument optimization and calculations. a

(typically 1 week for the analysis of a 4-sample batch). However, it is not selective toward particles for which fissile isotope content is high (i.e., HEU particles). Therefore, if HEU particles are scarce among a large majority of NU particles, for instance, there is little chance that one or more HEU particles will be analyzed. In contrast, the FT-TIMS method allows analysis of the most 235 U-enriched particles, which are in many cases the most relevant for safeguards. However, this method requires long analysis times (4 weeks are necessary for the analysis of 2 samples). In fact, the steps responsible for this long analysis time are sample irradiation with an appropriate thermal neutron flux (in addition to the irradiation time, a cooling time of several days is mandatory to allow for the decay of short-lived activation products), the picking-up of particles to deposit them on TIMS filaments, and the isotopic measurements themselves. In this regard, one of the main advantages of LA-ICPMS is the short duration of the isotopic measurement: analysis of a single particle takes no more than 34 min, whereas tens of minutes are necessary for SIMS and 12 h for TIMS. However, a prerequisite for LA-ICPMS analysis is that the uranium particles must be located in advance, using either SEM or fission tracks. The estimated analytical delays for the analysis of 2 samples (30 uranium particles per sample) with the various particle analysis techniques used in this work are gathered in Table 4. In the case of the first sample (“Nusimep-6”), LA-ICPMS with localization of uranium particles with SEM can be compared with SIMS. Analysis time is slightly shorter, but precision and accuracy are significantly poorer. In the case of the IAEA field-sample, LAICPMS with localization of the uranium particles using the FT technique can be compared with FT-TIMS. Analysis time is significantly reduced as no microhandling of individual particles is required, the localization of the particles thanks to the observation of the fission tracks clusters is faster (only one or two disks are observed instead of 30), and the isotopic analysis is significantly shorter. Nevertheless, accuracy and precision are also poorer, although the difference is less marked than with SIMS. Moreover, sensitivity is probably significantly higher as submicrometer-size particles can be analyzed, which is currently rather difficult using TIMS. The main advantage of LA-ICPMS is its excellent overall measurement efficiency, which was estimated to be about 6  104, whereas the measurement efficiency with the same instrument and usual solution introduction conditions (micronebulizer, cooled spray chamber) is ∼1.2  104. Furthermore, LA-ICPMS acquisition times are very short,

typically between 20 and 40 s, which favors a high signal-tonoise ratio. Therefore, LA-ICPMS is a very promising technique for particle analysis, provided the uranium-bearing particles are fixed on the sample holder and are located prior to LA-ICPMS analysis. Therefore, the potential advantage of LA-ICPMS particle analysis combined with the FT technique can be expected for the situation in which only very small, submicrometer-size, highly enriched uranium particles are present, among a large majority of less enriched, natural, or even depleted uranium particles. HEU may not be detected and analyzed when SIMS is used, whereas TIMS may lack the necessary sensitivity for such tiny particles. However, more work is required before LA-ICPMS can be considered as a real alternative technique for particle analysis on micrometer-size particles. First, more analyses must be carried out on real-life samples that have also been analyzed with the established techniques so as to gain experience and confidence in the technique. Moreover, the particle relocation methodology should be improved so as to reduce the ablated area (for instance, from 100 to 50 μm) and, consequently, the “mixing effect” which appears when several particles of various isotope compositions are analyzed at the same time. Additionally, local cross-contamination effects (by deposition of ablation debris in the immediate vicinity of the ablated area) as well as the potential occurrence of polyatomic interference must be investigated. It would also be valuable to further optimize laser ablation parameters and the injection arrangement (ablation cell, tubing, etc.) so as to reduce the unevenness of the signals and improve precision. Lastly, as the simultaneous measurement of all isotopes of interest, including the minor uranium isotopes 234U and 236U, is crucial for achieving precise and accurate isotope ratios from transient signals, the application of multiple-ion counting in ICPMS is required.

’ AUTHOR INFORMATION Corresponding Author

*Phone: +33 1 69 26 49 17. Fax: +33 1 69 26 70 65. E-mail: [email protected].

’ ACKNOWLEDGMENT The authors are grateful to Olivier Marie for uranium particle localization by SEM, Anne-Laure Faure for fission track observation and localization by optical microscopy in clean-room facility, Christophe Moulin for support and advice, and Christophe Pecheyran and Fanny Claverie for fruitful collaboration about laser ablationICPMS. ’ REFERENCES (1) Donohue, D. L. J. Alloys Compd. 1998, 271273, 11–18. (2) Donohue, D. L. Anal. Chem. 2002, January 1, 25A–35A. (3) Ciurapinski, A.; Parus, J.; Donohue, D. J. Radioanal. Nucl. Chem. 2002, 251, 345–352. (4) Admon, U; Donohue, D.; Aigner, H.; Tamborini, G.; Bildstein, O.; Betti, M. Microsc. Microanal. 2005, 11, 354–362. (5) Tamborini, G.; Betti, M.; Forcina, V.; Hiernaut, T.; Giovannone, B.; Koch, L. Spectrochim. Acta B 1998, 53, 1289–1302. (6) Tamborini, G.; Betti, M. Mikromchim. Acta 2001, 132, 411–417. (7) Ranebo, Y.; Hedberg, P. M. L.; Whitehouse, M. J.; Ingeneri, K.; Littman, S. J. Anal. At. Spectrom. 2009, 24, 277–287. 7847

dx.doi.org/10.1021/ac201596t |Anal. Chem. 2011, 83, 7841–7848

Analytical Chemistry

ARTICLE

(8) Stetzer, O.; Bettin, M.; Van Geel, J.; Erdmann, N.; Kratz, J. V.; Schenkel, R.; Trautmann, N. Nucl. Instrum. Methods Phys. Res., Sect. A 2004, 525, 582–592. (9) Park, Y. J.; Song, K.; Pyo, H. Y.; Lee, M. H.; Lee, K. Y.; Kim, W. H. Nucl. Instrum. Methods Phys. Res., Sect. A 2006, 557, 657–663. (10) Lee, C. G.; Iguchi, K.; Inagawa, J.; Suzuki, D.; Esaka, F.; Magara, M.; Sakurai, S.; Watanabe, K.; Usuda, S. J. Radioanal. Nucl. Chem. 2007, 272, 299–302. (11) Shen, Y.; Zhao, Y.; Guo, S. L.; Cui, J.; Liu, Y.; Li, J.; Xu, J.; Zhang, H. Radiat. Meas. 2008, 43, S299–S302. (12) Boulyga, S. F.; Prohaska, T. Anal. Bioanal. Chem. 2008, 390, 531–539. (13) Shinonaga, T.; Esaka, F.; Magara, M.; Klose, D.; Donohue, D. Spectrochim. Acta B 2008, 63, 1324–1328. (14) Kraiem, M.; Richter, S.; K€uhn, H.; Aregbe, Y. Anal. Chim. Acta 2011, 688, 1–7. (15) Kraiem, M.; Richter, S.; K€uhn, H.; Stefaniak, E. A.; Kerckhove, G.; Truyens, J.; Aregbe, Y. Anal. Chem. 2011, 83, 3011–3016. (16) Esaka, F.; Esaka, K. T.; Lee, C. G.; Magara, M.; Sakurai, S.; Usuda, S.; Watanabe, K. Talanta 2007, 71, 1011–1015. (17) Esaka, F.; Magara, M.; Lee, C. G.; Sakurai, S.; Usuda, S.; Shinohara, N. Talanta 2009, 78, 290–294. (18) Esaka, F.; Watanabe, K.; Fukuyama, H.; Onodera, T.; Esaka, K. T.; Magara, M.; Sakurai, S.; Usuda, S. J. Nucl. Sci. Technol. 2004, 41, 1027–1032. (19) Pointurier, F.; Baglan, N.; Hemet, P. Appl. Radiat. Isot. 2003, 60, 561–566. (20) Baglan, N.; Hemet, P.; Pointurier, F.; Chiappini, R. J. Radioanal. Nucl. Chem. 2004, 261, 609–617. (21) Becker, J. S.; Zoriy, M.; Halicz, L. J. Anal. At. Spectrom. 2004, 19, 1257–1261. (22) Zheng, J.; Yamada, M. Talanta 2006, 69, 1246–1253. (23) Liao, H.; Zheng, J.; Wu, F.; Yamada, M.; Tan, M.; Chen, J. Appl. Radiat. Isot. 2008, 66, 1138–1145. (24) Lariviere, D.; Benkhedda, K.; Kiser, S.; Johnson, S.; Cornett, R. J. Anal. Methods 2010, 2, 259–267. (25) Pointurier, F.; Hubert, A.; Hemet, P. J. Anal. At. Spectrom. 2008, 23, 104–112. (26) Pointurier, F.; Hubert, A.; Faure, A. L.; Hemet, P.; Pottin, A. C. J. Anal. At. Spectrom. 2011, 26, 1474–1480. (27) Stetzer, M. D. Appl. Spectrosc. 2003, 57, 1173–1177. (28) Boulyga, S. F.; Desideri, D.; Meli, M. A.; Testa, C.; Becker, J. S. Int. J. Mass Spectrom. 2003, 226, 329–339. (29) Boulyga, S. F.; Tibi, M.; Heuman, K. G. Anal. Bioanal. Chem. 2004, 378, 342–347. (30) Mertz-Kraus, R.; Jochum, K. P.; Sharp, W. D.; Stoll, B.; Weis, U.; Andreae, M. O. J. Anal. At. Spectrom. 2010, 25, 1895–1904. (31) Varga, Z. Anal. Chim. Acta 2008, 625, 1–7. (32) Claverie, F. Developpements et applications d’un systeme laser femtoseconde infrarouge basse energie et haute cadence de tir pour l’analyse d’elements trace dans les solides par couplage ablation laserICP-MS. PhD thesis, Universite de Pau et des Pays de l’Adour, manuscrit 09PAUU3003, 30 January 2009. (33) Becker, J. S.; Sela, H.; Dobrowolska, J.; Zoriy, M.; Becker, J.Su. Int. J. Mass Spectrom. 2008, 270, 1–7. (34) Pointurier, F.; Hubert, A.; Baglan, N.; Hemet, P. J. Radioanal. Nucl. Chem. 2008, 276, 505–511. (35) Aregbe, Y.; Truyens, J.; Kips, R.; Richter, S.; Stefaniak, E. A.; K€uhn, H.; Kraiem, M. Nusimep-6: uranium isotope amount ratios in uranium particles; Interlaboratory Comparison Report EUR 23702EN; Official Publications of the European Communities  JRC Scientific and Technical Reports: Luxembourg, 2008.

7848

dx.doi.org/10.1021/ac201596t |Anal. Chem. 2011, 83, 7841–7848