Simulation of rare earth element solvent extraction processes for plant design and optimization is complex, requires extensive bench scale and pilot scale testing, and is traditionally performed using proprietary algorithms. The complexity of simulation is mainly due to traditional simulation software solvent extraction modules not being designed to simulate rare earth separation processes because of the interaction between each element. This paper presents an alternative method for designing and optimizing rare earth element solvent extraction separation processes using METSIM to converge each stage of the battery individually until the full battery reached ultimate convergence. A feed-forward controller is used to control the organic to aqueous flow ratio feeding the solvent extraction modules. This method minimizes the complexity of simulating such processes by using inexpensive batch solvent extraction data, and allowing for rapid theoretical optimization, by varying the process configuration and parameters from a graphic interface instead of the traditional code-based interface of most proprietary models. Comparison of the model predictions to pilot plant data reveals that the method can be used to accurately model rare earth solvent extraction processes, as illustrated by its application in the development of Rare Element Resources innovative process.

The Rare Earth Elements (REEs) consist of 16 metals, specifically the 15 lanthanide elements and yttrium, which is typically found in the same mineral deposits as the lanthanides. Scandium can be associated with the rare earth elements, but is usually only present in trace amounts, and therefore omitted. The term “rare earths” originates from an old term used to designate minerals (earths) and their believed scarcity upon their initial discovery. The rare earths designation is now considered a misnomer because it was demonstrated that the crustal abundance of some rare earth elements, such as cerium, is known to be higher than many common metals, such as copper. However, enriched deposits are fairly uncommon compared to other similarly abundant metals. The REEs are typically divided into groups commonly consisting of light rare earths (LREEs), medium rare earths (MREE), and heavy rare earths (HREEs). Light rare earth elements include lanthanum to neodymium, medium rare earths include samarium to gadolinium and the heavy rare earth elements include terbium to lutetium, and yttrium. The REEs exhibit similar chemical characteristics and are usually extracted from the ore and processed into a mixed rare earth concentrate. The rare earth concentrate requires further processing in order to separate the REEs from each other and produce a series of single element products. Although many technologies are currently in development for the separation of REE, the only current commercial separation process is solvent extraction. Only a few such facilities are currently in operation, none of which are operating in North America (Gupta & Krishnamurthy , 2005). The two main barriers to entry in the rare earth separation industry are the very low publicly available literature on REE separation process design and the extremely high capital costs associated with such processes. In order to overcome these hurdles, potential new entrants need to first perform extensive and expensive bench scale and pilot scale test work in order to design and scale up their own solvent extraction process as well as raise substantial capital once the process is designed. Simulation of such processes have been addressed previously in the literature (Bazin, Boudrias-Chapleau, & Ourriban, 2013), but most of the proposed models have been programmed using proprietary algorithms, directly related to the solvent extraction modules. The present paper proposes a method for significantly reducing the schedule and cost associated with the development and design of REE separation processes by solvent extraction using a commercially available simulation software: METSIM. In this method, the solvent extraction modules are directly integrated in the overall process flowsheet. The proposed method was successfully used to assist in the development of Rare Element Resources’ patent pending zero-discharge solvent extraction separation process. Although a comprehensive model was built for the various phases of the process development program, only a single battery of solvent extraction cells, from a section of the process will be discussed herein. METSIM was selected because it is extensively used in the mining and metallurgical industry, and is easily customizable (Qiuyue, Zhang, Lv Guozhi, & Xiaofeng, 2012).


Solvent extraction (also referred to as liquid-liquid extraction) is a separation method based on the difference of solubility of ions or compounds in different solvents. In the solvent extraction process, ions or compounds contained in an aqueous phase migrate to an organic phase when the two phases are mixed. The organic phase is selected based on three criteria: its capacity to selectively solubilize the targeted ions or compounds, its immiscibility with the aqueous phase, and the absence of a stable emulsion phase between the two solvents. The latter two properties enable a simple gravity separation of the liquids (Aguilar & Cortina, 2008). Solvent extraction is typically carried out in a group of mixer-settlers, collectively referred to as a battery. Because high loading rates are achieved in the organic phase, the mixer-settlers are usually small in size.

Separation of various elements is achieved by the difference of mass transfer between the phases for each element. The ratio of concentration between the organic and aqueous phases is named “Distribution Ratio”, and is often referred to as the equilibrium constant Kd. The second most important parameter is the Separation Factor “SF”, defined as the ratio of distribution ratios, illustrating the preferential extraction of one element versus another element. Consecutive REEs tend to have very similar distribution ratios combined with low separation factors, and as such require many stages of solvent extraction for complete separation.