Abstract
This study sets out to demonstrate the capability of using electrochemistry and laser-induced breakdown spectroscopy (LIBS) for concentration prediction of multiple species in a molten salt system at 773 K. Samples contained UCl3 ranging from 0 to 10 wt%, GdCl3 ranging from 0 to 5 wt%, and MgCl2 ranging from 0 to 1.5 wt%, with LiCl-KCl eutectic salt as the remainder. Multivariate models were produced using semi-differential cyclic voltammograms (SDCVs) and normalized spectra acquired from LIBS. The SDCV model best predicted UCl3 levels, while the LIBS model best predicted GdCl3 and MgCl2 concentrations. A third model was developed by fusing the SDCV and LIBS signals. This model predicted UCl3 well and predicted GdCl3 and MgCl2 better than previous models. This model was then evaluated by using blind samples. The model predictions correlated well with inductively coupled plasma mass spectroscopy measurements, passing a t-test at a 95% confidence level.
Original language | English |
---|---|
Pages (from-to) | 617-626 |
Number of pages | 10 |
Journal | Nuclear Technology |
Volume | 207 |
Issue number | 4 |
DOIs | |
State | Published - 2021 |
Externally published | Yes |
Keywords
- Cyclic voltammetry
- laser-induced breakdown spectroscopy
- online monitoring, pyroprocessing
- sensor fusion