Analytical
characterization of chemical constituents is an essential
component of both industrial processes and fundamental studies. Applicable
techniques abound, but optical spectroscopy has the unique combination
of being easily incorporated into online (in situ) monitoring probes
and can also provide abundant chemical information (concentration,
oxidation state, speciation). However, this abundance of chemical
information can make applying optical spectroscopy to complex chemical
processes challenging. This is particularly evident in the bulk electrolysis
of U(VI) to U(IV), an essential preparatory step in some nuclear fuel
recycling schemes. Starting and product materials have interfering
spectral signatures in both electronic and vibrational spectroscopies,
making accurate and real-time analysis difficult to achieve. This
spectral complexity was overcome through the use of chemometric models,
multivariate analysis techniques that were built from spectral training
sets and applied to process data in real-time for immediate process
analysis. Here we discuss the application of optical spectroscopy
to online monitoring of a bulk electrolysis process and the methods
utilized to analyze data as well as characterize the efficiency of
U(IV) generation.