EDITOR: | June 23rd, 2014 | 3 Comments

Separation of Rare Earths – Art vs. Science (VIII)

| June 23, 2014 | 3 Comments

Steve MackowskiThis 8th 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:

Separation of Rare Earths – Art vs Science (I), (II), (III), (IV), (V), (VI) and (VII)

Last week I left you with a question. If you have the rare earths spectrum quoted in article (VII) as typical for a LREO Plant and have an annual rate of 10,000 tonnes of REO, how big is your HREO separation section? Trust you all did your homework as this leads onto this weeks section.

A 10,000 tpa REO separation plant (using my generic model in VII) handles all of the REO in the first solvent extraction circuit. This circuit removes the SEG (MREO) + HREO. It processes all of the REO as an aqueous stream. Let’s assume a feed concentration of 200 gm/litre. The extraction stage has a mixer and a settler. The mixer is where the aqueous and organic are mixed together and mass transfer occurs. Typically the mixer has a residence time of 5 minutes, that is the average time the aqueous and organic are being mixed, and from the organic : aqueous ratio (as determined experimentally) a mixer volume can be calculated. Let’s say this section has a mixer volume of 100 units. The mixer is followed by a settler where the fine dispersion is allowed to disengage, that is the aqueous and organic form layers. Just as oil on water. Again, the size of the settler area needed to effectively allow clean organic and clean aqueous to be obtained can be calculated.

Now again, let’s assume that the SEG+HREO is precipitated and re-dissolved to get 200 gm/litre. Again using 5 minutes mixing time, the 700 tpa (500 SEG + 200 HREO) will require mixers of volume 7% of the volume of the previous units handling 10,000 tpa. Similarly, when we get to the separation circuit for HREO the volume of the mixers will be 2% of the units handling 10,000 tpa. So we have three types (banks) of circuits. 10,000 tpa circuit using 100 sized units, SEG + HREO using 7 sized units, and the HREO using 2 sized units. This configuration is pretty fixed in the design. It is also possible that different chemicals are used, so having these circuits in different warehouses/locations is all possible.

As an aside, the issue of the 5 minutes mixer residence time is a relic of the past. In discussions with Gordon Ritcey, he told me that he had been plagued with the number since he believed for the most part he was responsible. It came about during early laboratory experiments (50s and 60s) where the volume of aqueous and the volume of organic are put into a bottle and shaken to get the dispersion. How long do you shake for? Gordon always used 5 minutes to ensure enough time. He felt that this number has stuck to this day! This is interesting since mixing time can be a very important variable during the experimentation stage. Back to the main story.

So you will have three sections, one comprising volume units of 100, another at 7 volume units, and the last at 2 volume units. Now a few weeks ago, Jack Lifton mentioned that the Solvay separation facility at La Rochelle was not at full capacity and that it was therefore a potential toll separation opportunity. You should be able to see that a key issue is which part of which circuit is under utilized? Using my generic circuit, a 50% utilization using the TREO spectrum in the example would only have 100 tpa HREO spare separating capacity. Obviously circuits can be upgraded or modified but this requires CAPEX. The design spectrum and the new feed spectrum need full analysis before simple arithmetic principles will give the correct answer.

So is tolling the answer? Yes, tolling is certainly possible. Due to the above, it needs careful design to allow the variation in feed stocks to be effectively managed. This design pretty much needs to cover the full operating life of the facility. It requires careful day to day quality control to ensure the TREO design spectrum range is not extended which would allow REOs to appear in their non-required products, or result in loss of recovery (saleable value).

So should you toll or should you process yourself? It’s a matter of risk management. An old adage was that you only use contractors (eg. tolling) if you can’t do a better job yourself! So on the yes side of tolling you may have:

  • Less CAPEX for your operation
  • Improved technology, skills and a dedicated workforce from day one (an assumption of course)
  • Less circuit development time
  • Centralized waste management with benefits of scale which may lead to reduced OPEX
  • Centralized chemical logistics with benefits of scale which may lead to reduced OPEX
  • Possible strategic tolling plants with beneficial financing arrangements

But on the no side of tolling, you may wish total control yourself. You may now have

  • The opposite to all of the above plus
  • Higher revenue since you are not paying the tolling margin

There is no clear-cut right answer. It will depend on your access to technology, access to CAPEX funds, geographical and environmental factors, and probably many others that are specific to your project. Obtaining the answer is a part of the project development process that requires considerable thought, planning and effective execution.

Look forward to next article where I will overview the issues about CAPEX and OPEX.

Steve Mackowski


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>

Copyright © 2016 InvestorIntel Corp. All rights reserved. More & Disclaimer »


  • Fred

    Keeping capex low by using tolling for concentrates is only a theoretical reference point. Most of the junior miners are incapable of financing even a low capex with stock flotation. The dilution would be so severe, the stock would resemble an IPO. It’s tough to raise more than a few million dollars from new stock based upon little more than core samples.

    Ultimately the big money comes from investment bankers. While they like low capex, they won’t lend you anything unless they’re looking at a business plan that they’d be comfortable explaining to their boss. So you need end users. This means big corporations. And for defense companies, government influence. Does the US want its fighter jets manufactured in China? Probably not. The big bucks necessary to raise the capex will be from linking up with the end users. And they could help raise enough money to build facilities to fully process the ore. Tolling may occur and even dominate, but it would mostly be the egg that gets hatched by the chicken. The REE miners should talk with companies like GE to get the process started.

    June 24, 2014 - 10:58 PM

  • Fred

    The following article gives insight into linking up with a big company like GE to sell it REE’s. The Motley Fool details how GE, on a massive implementation of 3D printing for manufacturing, is acquiring its metal powders.


    July 1, 2014 - 11:46 AM

  • Separation of Rare Earths – Art vs. Science (IX) | InvestorIntel

    […] Separation of Rare Earths – Art vs. Science (I), (II), (III), (IV), (V), (VI), (VII) and (VIII) […]

    July 2, 2014 - 8:28 AM

Leave a Reply

Your email address will not be published. Required fields are marked *