Characterization of Gaseous UF6 LIBS Plasma Near the U I 646.498 nm Emission Line for Nuclear Safeguards Applications

Nuclear safeguards and security depend critically on accurate measurement of uranium-235 enrichment at processing facilities. While current destructive analysis methods like COMPUCEA provide high accuracy, they require labor-intensive wet-chemistry sample preparation in the field. Laser-induced breakdown spectroscopy (LIBS) offers a promising alternative for gaseous UF₆ enrichment measurements due to its isotopic sensitivity, direct gas-phase compatibility, and minimal sample preparation requirements. Building on previous feasibility studies using the U II 424.437 nm line that revealed self-absorption bias, this study investigates the U I 646.498 nm line as an alternative analytical target.

A study conducted at Lawrence Berkeley National Laboratory (Berkeley, California), with collaboration from the University of Michigan (Ann Arbor, Michigan), presented a comprehensive characterization of the gaseous UF₆ LIBS plasma behavior, examining the effects of laser pulse width and wavelength on spectral characteristics and fundamental plasma properties through temporally resolved analysis, Boltzmann-plot temperature determination, and electron number density evaluation. Spectroscopy spoke to George Chan of the Lawrence Berkeley National Laboratory and corresponding author for the paper¹ resulting from this work.

What spectroscopic features enable LIBS to distinguish between uranium isotopes in gaseous UF₆, and how do isotope shifts compare to other line-broadening mechanisms in this system?

As with other uranium-containing samples, isotopic analysis of gaseous UF₆ by LIBS relies on isotope shifts in atomic uranium emission lines. The chemical form of uranium (gaseous UF₆ versus solid compounds) does not alter the underlying measurement principle. Certain uranium transitions exhibit comparatively large isotope shifts, often exceeding 10 pm. Under optimized plasma and detection conditions, these shifts are significantly larger than the contributions from other line-broadening mechanisms such as Stark, collisional, and Doppler broadening. As a result, isotope shifts dominate the observed line structure and enable isotopic ratio measurements when suitable spectral lines are selected.

Why is the U I 646.498 nm line considered a viable alternative to the U II 424.437 nm resonance line for UF₆ enrichment measurements, and what spectroscopic trade-offs are involved in selecting non-resonant lines?

The selection of the U I 646.498 nm line as an alternative to the commonly used U II 424.437 nm resonance line is based on our previous systematic evaluation of uranium emission lines between 280 and 745 nm in UF₆ LIBS plasmas.² This wavelength range was dictated by practical constraints: gaseous UF₆ exhibits strong absorption below 300 nm,³ while the quantum efficiency of most ICCD detectors decreases significantly beyond approximately 745 nm. In that study,² uranium lines were evaluated and shortlisted based on their isotope shifts and signal-to-background ratios (SBRs), and then further tested using UF₆ samples with different enrichment levels. Among all candidates, the U I 646.498 nm line showed the best overall analytical performance. The primary trade-off in using a non-resonant line is its higher excitation energy, which typically results in weaker emission and shorter persistence in the plasma. However, non-resonant lines are generally less affected by self-absorption, making them advantageous for isotopic measurements.

How does self-absorption manifest in LIBS emission spectra, and why is it particularly problematic for accurate isotopic ratio determination in uranium plasmas?

Self-absorption (SA) in LIBS emission spectra occurs when cooler atoms at ground or lower excited states within the plasma absorb photons emitted by hotter, excited atoms of the same species along the optical collection path. Although SA is a common phenomenon in plasma, it becomes analytically problematic when the distortions of emission signals become notable. Self-absorption affects isotopic measurements because it is a non-linear function of atomic (isotopic) number density. As such, the two isotopic emissions are affected to different extent. Because ²³⁸U is far more abundant than ²³⁵U in low-enriched UF₆, the ²³⁸U emission experiences stronger SA. This preferential suppression causes the ²³⁵U signal to appear enhanced on a comparative basis, resulting in a positive bias in the inferred enrichment. The SA non-linearity means that simple empirical correction factors are inadequate. We have formulated and validated a dedicated iterative correction approach to compensate for the SA effect.

The study reports a long-lived continuum attributed to overlapping molecular emission rather than ion–electron interactions. What spectroscopic evidence supports this interpretation, and how does it differ from classical LIBS continuum behavior?

Several experimental observations support the interpretation that the long-lived continuum observed in UF₆ LIBS plasmas arises primarily from overlapping molecular emission rather than from classical ion–electron interactions. It is well established that dense, strongly overlapping emission features—whether atomic or molecular—can produce a pseudo-continuum background. A well-known example is the elevated background observed in ICP spectra of samples containing high concentrations of rare-earth elements, which is attributed to the overlap of line wings from the line-rich rare-earth species.⁴

In conventional LIBS plasmas, continuum emission is typically dominated by Bremsstrahlung and recombination processes, both of which depend on the electron number density. However, our experimental results show that the continuum intensity in UF₆ plasmas does not follow the expected relationship with measured electron density, indicating that ion–electron interactions alone cannot account for the observed behavior.

Furthermore, unlike typical LIBS plasmas where continuum emission decays rapidly with the electron number density as the plasma cools, UF₆ plasmas exhibit a long-lived continuum. Given the strong tendency of uranium and fluorine atoms to recombine into molecules as the plasma cools, molecular emission provides a consistent explanation for this persistence. The recombination reaction is well supported by the fact that most of the UF₆ molecules in our sealed sample chamber appear to be non-destructive by the LIBS plasma, even after years of experiments. Although the molecular recombination saves the UF₆ samples for repeated use, the associated long-decaying plasma continuum in UF₆ LIBS degrades SBR and makes the selection of detection timing for optimal SBR more challenging.

How does time-resolved spectroscopy help disentangle atomic, ionic, and continuum emission in a UF₆ LIBS plasma, and why is this especially important for optimizing signal-to-background ratios?

Atomic, ionic, and continuum emissions in a LIBS plasma evolve on different time scales. Time-resolved spectroscopy exploits these differences by allowing the detector gate delay and integration window to be optimized for the analytical signal of interest. Continuum emission is generally strongest at early times, before the ionic and atomic emission develop. If data acquisition begins too early, when continuum emission is still intense, the signal-to-background ratio of atomic or ionic lines is significantly degraded, reducing both sensitivity and precision. Proper temporal gating suppresses the continuum contribution while preserving analytical line emission, thereby improving analytical performance. This approach is particularly important for UF₆ LIBS, where the continuum persists longer than in typical LIBS plasmas.

What spectroscopic differences are expected between femtosecond- and nanosecond-laser-induced plasmas, and how do the observed emission decay characteristics inform the choice of laser pulse duration for fieldable UF₆ measurements?

Prior to this study,¹ little information was available on differences between femtosecond- and nanosecond-laser-induced plasmas in gaseous UF₆, as most comparative LIBS studies focus on solid samples. For solids, reduced matrix effects are often reported for femtosecond LIBS, but this advantage is almost irrelevant for gaseous UF₆, of which the sample matrix is well defined. Our comparison of UF₆ LIBS with femtosecond- and nanosecond-lasers was therefore motivated primarily by scientific curiosity and a desire to understand any fundamental or analytical differences. Experimentally, we observed no clear analytical advantage of femtosecond-LIBS for UF₆ enrichment measurements. For fieldable measurements, laser selection must balance spectroscopic performance with system size, robustness, and operation reliability. At the current level of technological maturity, the greater complexity of femtosecond laser systems makes them significantly less practical for field deployment. The absence of measurable analytical benefits in our study reinforces the choice of the well-established nanosecond-laser technology for field-use UF₆ LIBS measurements.

How does increasing UF₆ pressure influence collisional line broadening and isotope shift resolvability, and what spectroscopic considerations govern the optimal pressure range for enrichment analysis?

Increasing UF₆ pressure significantly alters the plasma characteristics. At higher pressures, the plasma becomes visibly brighter due to an increase in electron number density, which in turn enhances uranium emission intensity. At the same time, increased collisional interactions lead to broader emission lines. These two effects compete with each other: stronger emission improves signal quality, while increased line broadening degrades isotope-shift resolvability. As a result, operation at elevated pressure requires a careful compromise. The optimal pressure range is determined by balancing emission intensity against maintaining a line width sufficiently narrow that the ²³⁵U and ²³⁸U peaks can still be reliably decomposed through spectral fitting. In practice, this balance is established empirically by optimizing pressure, laser energy, and detection timing to preserve isotopic discrimination.

How are electron number density and excitation temperature extracted from LIBS spectra, and what assumptions underlie the use of Stark broadening and Boltzmann plots in this gaseous UF₆ plasma?

Electron number density and electronic excitation temperature were extracted using well-established protocols for LIBS diagnostics. Emission lines were fitted with Voigt profiles. The Lorentzian component was used to determine Stark broadening, while the spectrally integrated line intensity was used for Boltzmann-plot analysis.

Electron number density was derived from Stark broadening of a selected fluorine emission line rather than from uranium line for two reasons. First, Stark broadening parameters for fluorine are more reliably characterized in the literature. Second, fluorine emission lines originate from higher-energy states closer to ionization, resulting in larger Stark widths that can be measured more accurately. The standard assumptions were adopted: Stark broadening is the dominant collisional broadening mechanism, and ion-induced Stark contributions are negligible compared with electron-induced broadening.

Electronic excitation temperature was obtained using a Boltzmann plot constructed from a group of selected U I emission lines. Strictly speaking, the derived temperature describes the population distribution among the selected U I levels only. Extending this temperature to the full U I population assumes a single Boltzmann distribution, while extension to other species (e.g., U II) would require the stronger assumption of local thermodynamic equilibrium. These assumptions are standard in LIBS diagnostics and are appropriate within the context of this study.

What does the deviation of the U I 646.498 nm line from the Boltzmann distribution suggest about plasma optical thickness or population kinetics, and how might this impact quantitative isotopic measurements?

The deviation of the U I 646.498 nm line from the expected Boltzmann distribution suggests either inaccuracies in the spectroscopic constants used for this transition or the presence of SA within the plasma. Through additional experimental investigation, we confirmed that this line does experience measurable SA, even though it is non-resonant. Self-absorption distorts line intensities and therefore introduces bias into quantitative isotopic measurements, particularly enrichment assays. Recognizing this effect, we have developed and tested a correction methodology over the past couple of years to compensate for the SA effect. Incorporation of this correction significantly improves the accuracy of isotopic analysis and mitigates the bias associated with optically thick plasma conditions.

From a spectroscopic standpoint, why does the ns-Nd:YAG laser at 1064 nm offer advantages for field-deployable UF₆ LIBS systems, despite the availability of shorter wavelengths and ultrafast lasers?

From a practical standpoint, robustness is the primary requirement for field-deployable UF₆ LIBS systems. Our goal is a technique that can be reliably operated by technicians in the field without specialized laser expertise. Modern nanosecond-pulsed Nd:YAG lasers are mature, turn-key systems that are compact, reliable, and effectively maintenance free. In several field-test experience of which we shipped our LIBS system to Oak Ridge National Laboratory, our laser system performed flawlessly without requiring realignment after shipment.

We deliberately retained operation at the fundamental wavelength of 1064 nm, even though shorter wavelengths can be generated through frequency doubling or mixing. Nonlinear crystals introduce additional complexity and are prone to drift, often requiring periodic retuning—an undesirable burden in field environments. Ultrafast laser systems, while powerful research tools, are inherently more complex and mechanically delicate, increasing the risk of misalignment during transport. Our design philosophy is to solve the analytical problem using the simplest and most robust tool available. For UF₆ enrichment measurements in the field, a nanosecond-pulse Nd:YAG laser at its 1064-nm fundamental wavelength best satisfies this requirement.

References

1. Garrett, L. J.; Jovanovic, I.; Chan, G. C.-Y. Spectral Emission Characteristics Near 646 nm and Plasma Properties of Laser-Induced Plasma of Gaseous Uranium Hexafluoride. Spectrochim. Acta Part B 2026, 236, 107357. DOI: 10.1016/j.sab.2025.107357
2. Chan, G. C.-Y.; Martin, L. R.; Russo, R. E. Characterization and Optimization of a Spectral Window for Direct Gaseous Uranium Hexafluoride Enrichment Assay Using Laser-Induced Breakdown Spectroscopy. Appl. Spectrosc. 2023, 77, 819−834. DOI: 10.1177/00037028221112953
3. DePoorter, G. L.; Rofer-DePoorter, C. K. The Absorption Spectrum of UF₆ from 2000 to 4200 Å. Spectrosc. Lett. 1975, 8, 521−524. DOI: 10.1080/00387017508067354
4. Boumans, P. W. J. M.; Vrakking, J. J. A. M. High Resolution Spectroscopy Using an Echelle Spectrometer with Predisperser-III. A Study of Line Wings as a Major Contribution to Background in Line Rich Spectra Emitted by an Inductively Coupled Plasma. Spectrochim. Acta Part B 1984, 39,1291−1305. DOI: 10.1016/0584-8547(84)80212-8