Abstract
Crystallization is widely used as a final recovery step in the production of battery-grade nickel and cobalt sulphates. However, its potential to function as an effective and selective purification operation is often not fully understood and hence underutilized in process design.
To meet the stringent battery-grade specifications, especially for products derived from refining mixed hydroxide precipitate or from recycling end-of-life battery feedstocks, flowsheets often incorporate extensive upstream impurity removal stages. When the behaviour of crystallization with respect to impurity partitioning and rejection is not well understood or integrated into the process design, there is more reliance on upstream purification methods which would lead to capital and operating cost-intensive process designs. A comprehensive understanding of behaviour of individual impurities during crystallization can therefore enable more targeted and efficient design of upstream purification steps.
This paper presents a systematic evaluation of impurity incorporation behaviour during nickel and cobalt sulphate crystallization, based on multiple test work programs conducted on solutions derived from mixed hydroxide precipitate refining and battery recycling pilot campaigns. The study examines both the tendency of impurities to incorporate into the crystal and their removal efficiency through post-crystallization washing steps. Particular emphasis is placed on understanding which impurities are effectively rejected during crystallization from those that exhibit strong incorporation, with the aim of informing purification strategies and process design.
The results demonstrate that crystallization purification performance is highly impurity specific. Impurities like sodium are separated efficiently, while impurities like magnesium show limited separation effects. Recognizing and accounting for this behaviour enables flowsheets to be designed to explicitly leverage crystallization selectivity, thereby reducing redundant purification steps and supporting lower-cost and more efficient process flowsheets.