Trace element analysis in lithium matrices using micro-discharge optical emission spectroscopy

The increasing demand for high purity battery elements and the necessity to reliably determine trace concentrations of impurity metals have triggered recent development of new analytical methods. Both in battery metal production and recycling there is a growing need for new fast, precise and easy-to-use analytical methods, especially for the on-line and on-site analysis of lithium salt solutions, whose derivative products are used in the battery industry. Typically used established techniques, such as ICP-OES or ICP-MS, are usually limited to laboratory use due to high plasma gas flow rates and power consumption, making them unsuitable for real-time analysis and monitoring of industrial processes on-site. Therefore, a fast and precise on-site method which allows trace element analysis would be preferable. Here we have investigated the potential of micro-discharge optical emission spectroscopy (mu DOES) for the given challenge of on-line quantification of impurity metals in lithium matrices. The technology is based on a micro-plasma, which is directly created inside an aqueous sample without any carrier gas by using electrodes and high voltage pulses. In this study, the impurity elements Na, K, Al, Fe and Zn were simultaneously measured both in lithium carbonate and lithium hydroxide solutions. For this purpose, the lithium concentrations were varied between 0.3 and 2 mg L-1 and those of the contaminants between 0 and 50 mu g L-1. Calibration series and long-term stability measurements were carried out, whereby various parameters such as the plasma discharge energy, signal integration setting and sample electrical conductivity were optimised. Micro-discharge optical emission spectroscopy proved to be useful for the fast and precise main component and trace analysis of saline solutions. A relative standard deviation of 3% was achieved for the lithium concentration in long-term measurements over 9 h. For the trace impurity metals (Na, K, Al, Fe, and Zn) high coefficients of determination (R-2 > 0.99) and limits of detection in the low mu g L-1-region, comparable to ICP-OES, were obtained. Multi-linear regression models were used to correct for cross-element correlations that may occur at increasing lithium concentrations due to ionisation effects. Industrial process samples were measured on-site and the results were validated using laboratory ICP-OES.