Shining a light on uranium enrichment

8 April 2013

A new method of radioisotope sampling for nonproliferation safeguards assessments of gas centrifuge-based uranium enrichment plants aims to offer onsite analysis, which would be less expensive and faster than current techniques.

A new method of radioisotope sampling for nonproliferation safeguards assessments of gas centrifuge-based uranium enrichment plants aims to offer onsite analysis, which would be less expensive and faster than current techniques.

Safeguards technologies detect and deter illicit efforts to produce fissile material by countries that have committed to forgo nuclear weapons. The International Atomic Energy Agency monitors countries' compliance with international safeguards agreements. A number of types of nuclear facilities, including fuel-cycle plants, are monitored to verify that material from civilian facilities is not being diverted for weapons purposes.

One of the more difficult safeguards challenges is Gas Centrifuge Enrichment Plants (GCEPs), which can produce both LEU for civilian nuclear energy and other peaceful uses, and HEU for weapons. Conventional safeguards at GCEP facilities rely heavily on a combination of non-destructive and destructive analytical techniques. Non-destructive techniques can evaluate without damaging the material assessed, but the results are not always conclusive. NDA techniques are generally used to measure declared enrichment of UF6 cylinders. Enrichment is deduced by measuring the total intensity of the U-235 186 keV gamma ray at the cylinder wall. The cylinder wall thickness must be measured using an ultrasonic transducer to correct for variations in attenuation from the non-uniform cylinder wall thickness. Assay values are based on the assumption that the sample is homogeneous.

The existing destructive analysis technique for GCEPs, environmental sampling, is very accurate, but also less timely and cost-effective. The current practice of ES in gaseous centrifuge enrichment plants includes swipe sampling of surfaces in process areas and on occasion within the cascade halls during a low-frequency unannounced access (LFUA). Swipe samples are shipped by formal chain of custody to IAEA's Seibersdorf Analytical Laboratory (SAL) in Austria, where, depending on the type and number of samples, they are parsed and shipped to various laboratories in the Network of Analytical Laboratories (NWAL).

The IAEA evaluates the ES data (including SAL and NWAL reports) to determine whether particulate assays are fully consistent with facility declarations. Swipe samples are received at SAL or NWAL and processed in a Class 100 clean room. The ES particles are recovered from swipes by ashing, or physical removal, using ultrasonication, which can be done in solvent or by hand with a vacuum impactor. The fission-track method is used to irradiate particles mounted on a Lexan plate in a reactor using thermal neutrons. Fissile particles produce damage tracks that are evaluated by a trained analyst. Particles of interest are selected and mounted onto a filament to produce ions for element identification and enrichment analysis using thermal ionization mass spectrometry (TIMS). Secondary ion mass spectrometry (SIMS) is also routinely used for measuring isotopic composition of ES particles directly. While mass spectrometry analysis can achieve excellent relative abundance precision on uranium particles, misuse detection using the regiment of swipe sampling and offsite laboratory analysis is both expensive and time-consuming.

Norm Anheier, a staff scientist at Pacific Northwest National Laboratory (PNNL), is developing a unique destructive analytical capability to provide timely detection of undeclared HEU production. In addition to its scientific novelty, this technique -- called Laser Ablation Absorbance Ratio Spectroscopy Environmental Sampling (LAARS-ES) -- has many practical advantages. Most notably, sample collection and analysis can be performed much more quickly and inexpensively than the conventional techniques mentioned above.

How it works

Environmental aerosol sampling media is first introduced into a small, reduced-pressure chamber (Figure 1). The LAARS-ES laser enrichment measurement approach uses laser vaporization to prepare the sample (for example, collected aerosol particles), followed by wavelength-tuned laser diode spectroscopy to characterize U-235 abundance relative to U-238. The vaporized sample material is ejected from the sample surface to form a high-temperature (~50,000 K) plasma through interaction with the intense (~32 GW/cm2) pulsed laser radiation. Laser vaporization of the sample serves two important functions. First, the high-temperature plasma effectively atomizes and ionizes most of the uranium sample. No further sample manipulation or chemical processing is required to complete the enrichment measurement. The focused ablation laser spot size also defines the sampling spatial resolution (~20 µm) and the subsequent stepwise scanning (~30 µm) leads to high spatial resolution isotope analysis across the entire sample surface. This characteristic provides LAARS-ES with the ability to detect and analyze trace collections of uranium particles intermixed in a large excess of background particles, consistent with environmental samples likely collected within a GCEP.

Shortly after the sample vaporization, the plasma cools through supersonic expansion and conduction with the surrounding cover gas (argon, ~1300 Pa) to form a ~3mm-diameter hemispherical plume containing neutral uranium atoms. Two wavelength-tunable diode laser beams (~405 nm and ~415 nm, respectively) are combined, and then directed through the plume of neutral uranium atoms to selectively probe (via precise wavelength tuning) the U-235 and U-238 isotopes. The lasers excite the atoms from the same ground or first metastable state to different upper atomic transitions for each isotope. The laser beams pass through the plume and exit the sample chamber through a window port. The combined beams are separated using a holographic diffraction grating and each beam is then directed to a compact photodetector that measures the transmitted laser intensity. Comparison of the intensity immediately prior to the ablation pulse with the intensity at the time of maximum atomic column density, typically 15 µs after the ablation laser pulse, yields a precise absorbance signal that is directly proportional to the atom concentration of each isotope. The measured absorbance signals are processed to provide the U-235 relative abundance determinations each time the ablation laser fires (~200 times a second).

In the LAARS-ES experimental instrument, the sample substrate is mounted to a miniature XY translation system that scans the sample at the focal plane of the focused ablation laser. The laser pulse repetition is synchronized to the X-axis translation to sample a ~20 µm spot at ~30 µm step increments at 200 samples/sec. At this sampling rate, 120,000 discrete enrichment measurements can be completed in 10 minutes. The future LAARS-ES onsite instrument will use a rotary aerosol impactor drum, rather than XY translation, to scan the sample.

The envisaged implementation of LAARS-ES uses a stand-alone aerosol collector based on a rotating drum impactor (RDI). Aerosol collectors could be secured in tamper-indicating enclosures and staged at strategic locations within the GCEP. Under automated operation, the RDI could collect airborne particles for a pre-set time interval, at which point the drum could rotate to expose a new rectangular strip on the drum. A drum having a modest diameter (~100 mm) could provide time-resolved particle sampling and integration on a daily or weekly basis over a one-year period. During onsite inspections, the aerosol collectors could be retrieved and inserted into the LAARS-ES instrument to rapidly screen for undeclared enrichment levels. It is envisioned that the LAARS-ES instrument would be installed within an IAEA instrument rack, onsite in a secure area of the GCEP. A 40-mm length at the centre of the 63.5mm-long rectangular impaction strip could be raster scanned using LAARS-ES to characterize the collected particle distribution. The drum could be rotated and the next row of the strip is scanned. This process could be repeated until the entire impaction strip is analyzed. Within 10 to 20 minutes, each impaction strip could be completely scanned and the uranium particle count and U-235 abundance distribution could be determined. Each collected strip could be sequentially analyzed until all impaction strips are assayed. The unmeasured impaction strip regions could be archived or laboratory analyzed at a later date.

Experimental results

During the past four years, extensive laboratory investigations have been conducted to develop LAARS-ES laser-based isotope analysis method and design the aerosol sample collector. The research path concentrated first on establishing the feasibility of conducting isotope analysis on aerosol-sized particles. Initial studies used prepared gadolinium particles as a surrogate. While the chemistry and atomic spectroscopy of U and Gd are quite similar, Gd is non-radioactive and has only limited health concerns. These properties significantly simplified our initial feasibility investigations. Follow-on studies demonstrated the ability of LAARS-ES to detect and analyze mixtures of Gd particles with varying levels of enrichment, which were mixed in a large excess of background media (that is, nuisance dust). The final surrogate sample studies successfully detected just a few highly enriched particles in a large excess of low enriched particles, which are in turn intermixed in a large excess of 'dirt' particles. Current LAARS-ES investigations are now focused on uranium particle measurements and evaluation of the custom LAARS-ES aerosol sampler. Highlights of key experimental results are presented below.

To simulate environmental samples collected under an undeclared HEU production scenario, surrogate (Gd-based) particles were gravimetrically prepared from certified reference materials. Small amounts of these materials were mixed in varying proportions to yield desired isotope ratios to simulate enriched U-235. The surrogate particles were then mixed in a background matrix (dust) to better simulate process conditions. The highly-enriched particles of the surrogate can be clearly seen in both the spatial isotope map and distribution (Figure 3C and 3E). Laboratory testing indicates that HEU particles can be detected in environmental samples mostly composed of dust and low concentrations of uranium-bearing particles from normal GCEP processing. This sensitivity level is a critical factor that demonstrates the feasibility of isotope abundance determination on collected aerosol particles.

Another sample was prepared to assess the trace detection performance for the LAARS approach. This sample consisted of a well-characterized reference matrix containing a low concentration of surrogate (Gd) particles. Laboratory tests of the surrogate material demonstrate that the LAARS-ES system is capable of differentiating isotopic concentrates of the surrogate material that simulate trace quantities of high-enriched uranium in a matrix that is representative of expected GCEP environmental samples. The laboratory testing also identified a bias that is ~25% below the values expected from the surrogate blend. Figure 4 clearly demonstrates that the LAARS approach is capable of detecting just a few highly enriched particles (counts near 0.5 isotope ratio) in a large excess of surrogate materials that represent LEU and natural uranium, which are in turn intermixed in a large quantity of non-radioactive matrix.

Figure 5 shows a typical example of the LAARS-ES uranium sample analysis. This sample was prepared by pipetting 90 µl (2.7 µg total uranium) of dissolved autunite (hydrated calcium uranyl phosphate) onto the surface of a glassy carbon planchet. The spatial map is an XY intensity plot, where the colour scale is mapped to combined absorbance signal of both isotopes. The spatial sample concentration is immediately apparent. The ring-like structure is due to capillary flow by the differential evaporation as the solution dried on a hot plate. The raw and discriminated (red line on the raw data plot) U-235 relative abundance is shown plotted as a function of Ac. Outliers are easily observed on the raw data plot and the distribution shape on the discriminated plot. The discriminated data is processed into a histogram plot, which is fit with a Gaussian. The analysis results provide the number of valid measurements, U-235 relative abundance estimate, as well as the standard deviation. A subset of this full abundance analysis could be presented to the onsite inspector to facilitate quick assessment of potential facility misuse. This analysis simply reports the number of valid enrichment measurements above and below a preset threshold, typically the facility maximum enrichment declaration.

The project

A laser-based concept for uranium enrichment analysis was first developed at PNNL during the early 1990s. This concept was intended to address the need for a field-portable, uranium-isotope analysis capability to support field inspections. This early work formed the technical basis of the current LAARS technology. In 2008, PNNL invested in the development of a GCEP laser-based uranium enrichment analysis concept, and in 2011 the National Nuclear Security Administration's Office of Nonproliferation and International Security selected the project for further development of a LAARS-ES instrument prototype that features an integrated environmental aerosol sampling system. In 2012, PNNL evaluated designs and components for the prototype LAARS spectrometer and aerosol collector design. The laboratory has started to seek engagement opportunities to conduct technology demonstrations at domestic and international GCEPs.


Norm Anheier is a staff scientist at the Pacific Northwest National Laboratory with more than 22 years in applied and engineering physics.

FIgure 1: LAARS experimental setup
Figure 2: LAARS diagram
Rotating drum impactor and nozzle
Figure 3: Measurement of Gd-152 isotopic abundance
Figure 4: Trace enrichment detection performance
Figure 5: An example of LAARS-ES uranium sample analysis computer software

Privacy Policy
We have updated our privacy policy. In the latest update it explains what cookies are and how we use them on our site. To learn more about cookies and their benefits, please view our privacy policy. Please be aware that parts of this site will not function correctly if you disable cookies. By continuing to use this site, you consent to our use of cookies in accordance with our privacy policy unless you have disabled them.