Separation of Rare Earths – Art vs. Science (PT VII)
This 7th article is part of a series of articles that flow from one to another and are almost prerequisite reading to understand the progression of my discussion. The previous articles can be accessed:
This week, I want to overview the separation flow sheet that we have progressively developed. In summary during previous articles, we have pieced together the feed preparation section where the mixed light rare earth carbonate is dissolved and any impurity management issues may be addressed; we have designed a number of solvent extraction sections (the gates) to get the different rare earths away from each; and we have finished the individual products to meet the requirements of the customer, be they for magnet market or for those customers with speciality surface property related requirements.
What we see in the separation flow chart above is a sequence of operational steps. There are horizontal progressions of steps where the individual rare earths are separated from groups of rare earths. Eg MREO is separated into it’s individual oxides of samarium, europium and gadolinium. Eg Didydium is separated into it’s individual oxides of neodymium and praseodymium. And there is a vertical progression of steps where the groups of rare earths are separated in bulk. Eg MREO+HREO is removed, then Nd/Pr, then Ce and leaving La. You have probably noticed that there are steps bolded and shaded in yellow. These are the steps where the sale of a product is possible. This could be a final separated rare earth oxide, eg neodymium, or it could be a group of rare earths being sold to a customer for further separation, eg HREO. In the following discussion, I will use actual company names to illustrate logic. I do not know the full details of their respective flow charts; I only have the limited amount of information available through their web sites on their REO spectrum, possible circuit configuration and potential products. I stress my statements are mine alone and are based on logic.
So a possible LREO separation plant starts with a mixed rare earth carbonate of ~50% rare earths. It is 50% because other than some minor impurities, the other 50% is the carbonate. An approximate spectrum of rare earths could be 45% cerium, 25% lanthanum, 23% neodymium + praseodymium, 5% MREO and 2% HREO. This spectrum is important as it affects not just the chemistry but the size of the various steps. I’ll come back to this later when I talk about CAPEX and OPEX.
Get our daily investorintel update
If this is a generic separation plant, who uses this series of circuits?
Obviously LREO projects, but not all use all of the processing steps, particularly at their primary manufacturing facility. I’ll explain with some examples.
Molycorp, Mountain Pass deposit contains majority LREO. It would be logical to expect the separation plant at Mountain Pass to have:
- Feed preparation where the RE Carbonate (or other form) is dissolved.
- MREO + HREO removal by solvent extraction (SX)
- Didydium removal by SX
- Final individual products of cerium and lanthanum produced for direct sale to the end user.
The MREO + HREO and the didydium would be ideal feed stocks for the Molycorp Chinese separation plants in eastern China, so these parts of the circuit may not be at Mountain Pass. Again, I stress I do not know the details but am presenting logic to serve as examples.
Lynas has publicly stated their product mix and their LAMP facility would have the MREO + HREO being precipitated and sent off site for further upgrading by others. The didydium, however, is further processed on site to produce separated neodymium and praseodymium.
So I would expect all LREO processors to be similar in structure (but the chemistry may be different) with differences depending on the degree of separation performed on site with the MREO, HREO and didydium. The other possible difference is a recent addition where to avoid the costs associated with producing relatively low value cerium and lanthanum, the vertical progression of processing steps is curtailed after the didydium gate. Such a circuit can be seen as a magnet feed only plant and would be valued as such.
So what happens to the MREO + HREO for example, precipitated as a solid and sold. It goes to another factory where it is dissolved and the processing steps continue horizontally as per the above model but only using those steps for MREO and HREO. The same for didydium. This could have been the case for the recent spilt material being exported from LAMP. A didydium product as a liquid chloride was going off site to be separated by others.
What is different for the different LREO separation plants is their size. That is the overall size of the plant and the size of the processing gates. To keep this simple, when the rare earth carbonate is dissolved, it is dissolved with enough acid to get a concentration of between 150-200 grams per litre. This gives the concentration driver to get good separation efficiency. So for a 5,000 tonnes per annum plant, a certain volumetric flow rate is required. Logically, then, a 10,000 tonnes per annum plant will process twice the flow rate. In equipment terms, the design here is important. For 10,000 do you have two units operating at 5,000, or one unit handling all of the flow? As you progress through the circuit, you can now imagine why the feed spectrum of rare earths influences the configuration and size of the processing equipment. This is not such an issue for an on-site separation plant, eg Arafura Resource’s Nolans project, as you only have your own feed material, but it is an issue where a number of different rare earth feed stocks are coming into one processing plant with a relatively fixed configuration. Question? If you have the rare earths spectrum quoted above as typical for a LREO Plant and have an annual rate of 10,000 tonnes of REO, how big is your HREO separation section?
This raises the question of multiple parties using the same separation facility, and the topical issue of tolling. This will have to wait until next week.
Mr Mackowski is a qualified engineer in mineral processing with over 30 years technical and operational experience in rare earths, uranium, industrial minerals, nickel, kaolin ... <Read more about Steve Mackowski>