Laser-Induced Breakdown Spectroscopy for Trace Technetium-99 Detection in Molten Salt Reactor–Relevant Systems

A recent study investigated the use of laser-induced breakdown spectroscopy (LIBS) as a viable technique for detecting and quantifying trace technetium-99 (^99Tc) for future molten salt reactor applications, where long-lived fission products must be carefully monitored. By immobilizing aqueous Tc in polyvinyl alcohol films, comprehensive emission surveys and calibration models were developed, achieving limits of detection as low as 0.710 µg mL⁻¹ even in the presence of interfering molybdenum. The results provide foundational Tc LIBS emission data and show strong potential for rapid, low-preparation monitoring of Tc in nuclear, environmental, and reactor-related applications. Spectroscopy spoke to Hunter Andrews of Oak Ridge National Laboratory (Oak Ridge, Tennessee), corresponding author of the paper ⁽¹⁾ resulting from their research, about their findings.

What plasma formation and excitation mechanisms in LIBS govern the emission intensity of trace radionuclides like technetium, and how do these differ from more abundant elements?

The measurement of elements via LIBS, radionuclide or abundant, is really the same with regards to the plasma formation and excitement. The goal here is to maximize the signal-to-noise (S/N) for the detection of the target species by tuning the excitation (for example, laser wavelength, energy, and spot size), as well as the emission (for example, spectrometer delay time, integration time, and intensifier gain). Here, because the primary target was trace Tc (sub 100 µg mL^-1), we used the Tc emission peaks to guide our experimental configuration.

How does the choice of laser parameters (wavelength, pulse energy, repetition rate) influence spectral resolution, signal-to-noise ratio, and limits of detection for Tc in LIBS?

Wavelength and pulse energy play a large role here. The instrument we used in this study (J200, Applied Spectra) utilizes a nanosecond pulsed 266 nm Nd:YAG. This laser wavelength couples very well with solids, providing precise spot size control, and passes through ultraviolet (UV)-transparent windows very well. This allowed us to analyze samples containing Tc within an inert atmosphere cell, and it allowed us to precisely raster the surface to measure many shots rather than dwell at a given location and only collect a few shots before the thin layer sample was ablated through.

What challenges arise in identifying Tc emission lines in LIBS spectra given the scarcity of reference data, and how can line assignment be validated?

This was a large goal of the study. Spark and Arc Tc emission line surveys exist, but within the LIBS literature, there was only one study that listed three UV emission lines. Here, we designed samples where Tc concentrations were varied, and matrix (Mo+PVA) was held constant. This allowed us to see which emission peaks changed with concentration and then assign them to Tc. We then cross-referenced these with arc/spark compilations and the National Institute of Standards and Technology (NIST) to verify their ionization state and energy level information. Now, future researchers can view our study to quickly gauge which strong lines to look for when measuring Tc.

How do matrix effects—such as the presence of molybdenum in molten salt reactor (MSR)-relevant samples—impact spectral interferences and quantitative accuracy in LIBS measurements?

There are two main challenges with complex samples and LIBS: spectral interferences and matrix effects. These are not to be confused as being the same thing. Spectral interferences are relatively straightforward. If there are two species emitting photons at or around the same wavelength, the user should either pick a different fully resolved peak, use peak fitting to distinguish the contributions from the overlapping peaks to their individual elements, or use a high-resolution spectrometer that can better resolve these emissions. Sometimes, the best solution could be a combination of these approaches. As for matrix effects, this is a more complex issue. Matrix effects arise from a sample’s change in composition effecting the plasma properties (for example, electron density or temperature). This is usually a factor when the bulk composition begins to change. Here, more complicated calibrations would be needed such as multivariate modeling. Regardless, accurate quantification is possible with a well-developed model.

Why is immobilization of aqueous Tc samples in polyvinyl alcohol (PVA) films beneficial for LIBS analysis, and what spectroscopic artifacts might this introduce?

The conversion of liquids into a solid form for LIBS has been shown to be a viable option for LIBS analysis of liquids many times in literature—sometimes being referred to as surface enhanced-LIBS. Many of these studies simply dry an aliquot of liquid onto a solid substrate and then analyze them, but this approach, while avoiding issues of liquids splashing, can still have negative effects, such as the coffee-ring effect, making the dried layer vary in thickness. We selected the PVA immobilization because we could do the liquid-to-solid conversion in a mold and then remove the produced thin layer and mount it on a microscope slide. The PVA used only introduced carbon peaks into our spectra, and we used copper tape beneath it to have an easily identifiable signature if we ablated through the thin layer as we rastered across the sample. The 266-nm laser coupled very well to the PVA thin layer and provided consistent spectra.

How do broadband and narrowband spectrometers differ in their ability to resolve Tc emission lines, and what trade-offs exist between spectral coverage and sensitivity?

This calls back to my answer about spectral interferences. Broadband spectrometers give you full coverage from ~180–1000 nm in a single shot, which is useful for most applications. However, this increased wavelength coverage comes at the cost of resolving power, even if they are designed as multichannel spectrometers. Narrowband spectrometers allow for a smaller region to be investigated with better resolution. An added benefit is that narrowband spectrometers can be equipped with intensified detectors where a weaker signal can be enhanced, leading to better detection capabilities. These detectors can also provide shorter delay and exposure times which can be important for increased signal-to-noise ratio. In this study, the broadband spectrometer detected many Tc emission peaks, but the narrowband spectrometer was needed for Tc lines with molybdenum (Mo) interferences and provided better detection limits. Fortunately, the system we used could operate both spectrometers simultaneously.

What strategies can be employed in LIBS spectral fitting and calibration modeling to discriminate overlapping emission peaks from Tc and chemically related species like Mo?

Here, we fit a Voigt peak to each Tc component and Mo component of the overlapping emission peaks. The peak fits were first configured on a sample with equal concentrations of Tc and Mo. Then, the samples with other concentrations of Tc were used to fine-tune the parameters such that the Mo fitted peak area remained relatively constant (since its concentration didn’t change) while the Tc peak fit scaled to capture the variance in the real data. Then, a linear calibration model was built for the Tc fitted peak area. This is one approach, but there are many others including multivariate methods.

How are figures of merit such as limit of detection (LOD), linear dynamic range, and precision spectroscopically determined in LIBS-based quantification of trace radionuclides?

Once a calibration model is built, these figures of merit can be determined or estimated. In LIBS calibrations, it’s not uncommon for different linear regimes to exist where the slope of the calibration linear model (that is, sensitivity) is smaller once a trace analyte reaches a given concentration for reasons such as the onset of self-absorption. In this study, we decided to calculate LODs from the prediction intervals of our calibration models. This approach considers the calibration uncertainty and incorporates signal dependent variance to provide a realistic LOD. Precision is simply determined from replicate measurements.

I add the caveat of “estimated,” because these figures of merit are dependent to the calibration model built and the experimental measurements. LIBS researchers are constantly improving upon each other’s work, so our figures of merit are not absolute for LIBS measurements of Tc.

Compared with inductively coupled plasma mass spectroscopy (ICP-MS) or liquid scintillation counting, what are the fundamental spectroscopic advantages and limitations of LIBS for in-situ or online monitoring of Tc in nuclear systems?

LIBS is well-suited for in-field or online measurements of nuclear systems versus these traditional measurement systems. This is because of the ability to build custom LIBS platforms, measure samples without preparation, and incorporate them into various systems. This measurement point at the system offers a big time-savings compared to pulling a sample and sending it to an analytical lab. For example, we are developing LIBS for MSR monitoring where we make an aerosol from the salt in the MSR system and sweep it through a LIBS cell for measurement.⁽²⁾ This allows for estimates of composition in real-time (as opposed to days). That being said, offline measurements allow for better sample control and preparation which can provide more accurate results.

What additional spectroscopic challenges would be expected when extending Tc LIBS analysis from aqueous or polymer matrices to complex molten salt or geological matrices, and how might they be mitigated?

Extending Tc detection via LIBS into these fields will result in more complex samples, so matrix effects will become a major factor. Only experiments will inform us how to properly calibrate for all these effects. Additionally, there will be much more competition in terms of emission wavelengths. It’s likely many of the peaks we detected in this study may see new interferences from additional elements, but we at least have a solid knowledge base to reference for line identification.

References

1. Andrews, H. B.;Murphy, Z.; Martinez, M. et al. Detection and Quantification of Trace Technetium in the Presence of Molybdenum Using Laser-Induced Breakdown Spectroscopy. J. Anal. At. Spectrom.2026. DOI: 10.1039/D5JA00319A
2. Andrews, H. B.; Kitzhaber, Z. B.; Orea, D. et al. Real-Time Elemental and Isotopic Measurements of Molten Salt Systems through Laser-Induced Breakdown Spectroscopy. J. Am. Chem. Soc. 2025, 147 (1), 910-917. DOI: 10.1021/jacs.4c13684