Tandem analytical approaches to characterize challenging separation systems

Abstract

Characterization of challenging separations systems is central to enabling their execution or development. Some separations systems can be challenging due to unknown or variable matrices. An example of this would be separations relevant to remediating the U.S. Department of Energy’s Hanford site, where tanks being readied for remediation present unknown and difficult to characterize matrices due to decades of variable waste management practices and high radiation fields, respectively. Other separations systems can be challenging due to the system analytes being intrinsically difficult to separation from each other. An example of this would be intra-lanthanide separations, where the lanthanides have such similar chemistry that a ligand’s ability to discriminate between complexes is minimal in many instances. In these systems, multiple analytical approaches may provide value by either directly aiding a separation in situ or providing characterization of the system to enable general separations development. This dissertation will focus both on enabling in-process separations and separations development using tandem analytical approaches, where two analytical tools are used to describe system chemistry. The first two chapters of the dissertation will focus on real-time, on-line monitoring to improve the efficiency of nuclear relevant separations that pertain to either remediation of the Hanford site or reprocessing of used nuclear fuel. Current analytical approaches for characterizing the constituents in separations media for these types of separations typically involve discontinuous, off-line analyses of manually acquired samples well-after their acquisition. This type of measurement and evaluation does not provide immediate feedback on the stream’s condition or prevent deviations from optimal conditions from being quickly discovered and corrected. To expedite characterization, immediate feedback can be acquired by direct, spectroscopic interrogation. Spectroscopy can be readily utilized in on-line processes environments, but analysis of the real-time spectroscopic data using univariate analysis can be complicated in multicomponent solution streams due to significant spectral band overlaps and interference effects often present in solutions with mixtures of analytes. Chemometric analysis, which uses multivariate analysis of multiple wavelengths/wavenumbers to quantify analytes in a spectrum, can be used in tandem with spectroscopic data to derive accurate assessment of analyte concentration. To date, limited comparisons are available in the published literature regarding the ability of and appropriate statistical treatments for chemometrics to deconvolute overlapping Raman and UV-vis spectra. This work considers chemometric models in the use of Raman and UV-vis spectra to describe the chemistry of separations relevant to the Hanford site and used nuclear fuel. Regarding Hanford site work, Raman spectra with interfering analyte bands were assessed with three different validation sets. The approach using a partial least squares (PLS) statistical treatment was the most effective at quantifying analytes for all validation set samples. The PLS model performance was then further evaluated with optical spectroscopy on a simulated used nuclear fuel solution that had extreme interference effects in its UV-vis spectra, such as large baseline perturbation and spectral overlaps, that were different than those in the vibrational (Raman) spectra. This work showed the PLS model is able to quantify analyte concentrations in off-normal solution conditions and was also able to identify those off-normal samples. The final section completed a tandem assessment of cerium (Ce) chemistry in potassium carbonate media using UV-vis and electrochemistry. High concentration potassium carbonate media has shown some potential to support tetravalent praseodymium and terbium, which could be used to develop a high-performance, advanced lanthanide separation scheme. Studies here focused on more readily oxidizable cerium to understand solution conditions and electrode potential required to oxidize cerium to the tetravalent state in carbonate electrolytes. The changes of oxidation states were monitored simultaneously using spectroelectrochemistry as the spectral features of the two different oxidation states are different. In addition, the speciation of the cerium carbonate complexes was evaluated with the aid of computational modeling. It was found that higher temperatures and carbonate concentrations can help to lower the electrode potential for the Ce(III)/Ce(IV) couple. The compiled work shows the benefits of tandem analytical approaches in enabling the execution and development of challenging separations systems. Spectroscopic on-line monitoring using chemometrics in tandem was shown to support real-time analysis of analyte characterization for Hanford waste and used nuclear fuel. The tandem use of spectroscopy and electrochemistry provided early understanding of factors impacting lanthanide oxidation in high concentration potassium carbonate media.