ENGINEERING EXPLAINED / EARTHS
The hard part is pulling them apart.
Explore how a small change in selectivity becomes dozens of ideal stages, inspect a mixer-settler, and compare cerium with copper in the crust.
RARE LABS
From lab chemistry to production.
We are building an autonomous scale-up system for chemistry using robotics and AI, starting with peptides and pharma.
01 / FIFTEEN CLOSE NEIGHBOURS
Small differences. Long trains.
These are the fifteen lanthanides. Similar solution chemistry makes neighbouring metals difficult to separate. The full rare-earth family also includes scandium and yttrium; promethium is radioactive and is not a normal commercial separation target.
A separation factor α compares how strongly two metals favour the extracting phase. As α approaches 1, the separation becomes harder. Use the slider to see an idealised binary stage estimate.
Each tile represents one whole ideal stage in this analogy, not a designed mixer-settler.
Why the reel says “about 64”
N = ln[(p / (1 − p))²] / ln(α)
At p = 0.999 and α = 1.24, N = 64.2 theoretical stages; a whole-stage count rounds up to 65. At α = 2.83, it is 13.3 theoretical stages, or 14 whole stages. “13 to 64” is a rounded comparison of theoretical values.
This is a Fenske-style, constant-selectivity limiting analogy, with symmetric 99.9% end-product purities. It is not an operating solvent-extraction design. Actual extraction and scrubbing stages require phase ratios, equilibrium isotherms, feed composition, recovery targets, loading, acidity, entrainment and stage efficiency. A purity number alone does not determine the train.
02 / MIX, THEN LET THE PHASES SPLIT
What one stage looks like.
The mixer brings aqueous solution and organic extractant into contact. The settler gives the two immiscible phases room to separate. Repeated extraction, scrubbing and stripping steps build the separation.
Swipe the drawing to inspect the labels →
*Layer order shown for an organic phase less dense than the aqueous phase. Phase continuity and density depend on the actual chemistry. Flow connections are schematic.
A stage count is not a startup time. Residence time, tank inventories, recycles and control strategy determine how a real train settles. This page does not multiply a nominal settler time by stage count and call it a dynamic prediction.
03 / ABUNDANCE IS NOT SEPARABILITY
The word “rare” misses the problem.
These reference estimates are crustal mass abundances, not ore grades or recoverable reserves. Abundance estimates vary with the crustal model. Finding atoms is only the beginning: concentration, mineral processing and separation determine whether a deposit can become a useful product.
Sources & boundaries
- USGS: cerium geochemistry sheet and copper geochemistry sheet — the 66.5 and 60 ppm reference values shown above.
- USGS: rare-earth statistics and information — rare-earth family, uses and abundance context.
- Kukkonen et al. (2022): extraction and precipitation chemistry review — aqueous/organic extraction, selectivity and the role of extractant chemistry.
- University of Manchester: separation-process notes — Fenske minimum-stage relationship. The extraction comparison here is explicitly a limiting analogy.
The α range is a teaching sweep, not a universal range for every neighbouring pair or extractant. No current national supply-share claim is needed for this calculation.
RARE LABS
From lab chemistry to production.
We are building an autonomous scale-up system for chemistry using robotics and AI, starting with peptides and pharma.
