When crushed ore is placed in a liquid denser than water, the lighter particles float and the denser particles sink, and if the liquid density lies between the densities of two minerals, they separate cleanly. Spodumene, the principal lithium mineral, is denser than the quartz and feldspar that accompany it, so the method works well on coarse spodumene ore. In the laboratory, heavy liquid separation (HLS) uses a series of liquids of increasing density. In the plant, dense media separation performs the same function, using cyclones fed with ore suspended in a slurry of fine ferrosilicon that behaves as a heavy liquid.
HLS is the first test for any hard-rock lithium project considering DMS. It establishes the crush size, the cut density, the expected concentrate grade and recovery, and the proportion of lithium that will require flotation instead. Our article on spodumene processing covers the full flowsheet.
Test procedure
- Crush to the candidate DMS top size, for example 6.3 mm, and screen into size fractions. Fines below about 0.5 mm are removed, because DMS cannot treat them.
- Prepare the liquids at a series of densities, for example 2.70, 2.80, 2.90 and 3.00. Tungstate solutions such as sodium polytungstate and lithium heteropolytungstate are common because they are less toxic than the older organic heavy liquids, and the densest options are needed for the higher cuts. Check the density of each liquid before use.
- Perform the sink-float separation, starting with the lightest liquid. Stir the sample in, allow it to separate, and remove the floats and the sinks. The floats form one density fraction, and the sinks pass to the next, denser liquid. Repeat to the densest liquid.
- Wash, dry, weigh and assay every fraction for Li₂O, and for iron and the other impurities that affect concentrate value.
Heavy liquids should be handled with gloves and eye protection under good ventilation, and the liquid recovered from the washings for reuse.
Worked example: the washability table
| Density fraction | Mass (%) | Li₂O (%) | Cumulative sinks mass (%) | Cumulative sinks grade (% Li₂O) | Li₂O recovery to sinks (%) |
|---|---|---|---|---|---|
| Sinks at 3.00 | 15 | 6.0 | 15 | 6.00 | 59.8 |
| 2.90 to 3.00 | 7 | 4.5 | 22 | 5.52 | 80.7 |
| 2.80 to 2.90 | 8 | 2.0 | 30 | 4.58 | 91.4 |
| 2.70 to 2.80 | 12 | 0.6 | 42 | 3.45 | 96.1 |
| Floats at 2.70 | 58 | 0.1 | 100 | 1.51 | 100 |
where
- ,
- mass % and Li₂O grade of each density fraction, summed over all fractions heavier than the cut
- calculated head grade: the mass-weighted grade of all fractions (% Li₂O)
The calculated head is 1.505% Li₂O. At a cut density of 2.90, the sinks contain 22% of the mass at 5.52% Li₂O, calculated as (15 × 6.0 + 7 × 4.5) ÷ 22 = 5.52, with a recovery of (0.90 + 0.315) ÷ 1.505 = 80.7%.
The washability curve shows that a single DMS stage cut at about 3.0 produces a 6% Li₂O concentrate at about 60% recovery. A lower cut increases recovery but reduces the grade below specification. The 2.80 to 3.00 middlings, 15% of the mass and about 32% of the Li₂O, are where the main design decision lies: whether to re-crush and re-treat them, send them to flotation, or reject them.
Lithium conversions
where
- Li₂O content of pure spodumene, LiAlSi₂O₆ (%)
- lithium carbonate equivalent: mass of Li₂CO₃ the lithium could make; 2.473 is the ratio of molar masses
A 6.0% Li₂O concentrate is about 6.0 ÷ 8.03 = 75% spodumene, if all of the lithium is present in spodumene. One tonne of 6% Li₂O concentrate contains 60 kg of Li₂O, equivalent to about 148 kg of lithium carbonate. To convert lithium metal to Li₂O, multiply by 2.153.
Translating HLS results to the DMS plant
HLS represents an ideal separation, whereas DMS cyclones misplace some particles near the cut density. The sharpness of separation is described by the Ecart probable, Ep = (d75 − d25) ÷ 2, where d75 and d25 are the densities at which 75% and 25% of particles report to sinks; a smaller Ep indicates a sharper separation. A DMS cyclone also separates at a density somewhat higher than the medium density in its feed, so operators set the medium density to achieve the target cut.
Fine particles separate less sharply, and running HLS by size fraction shows which sizes DMS can treat effectively. Micas, tourmaline and iron-bearing minerals with densities of around 2.8 to 3.2 can report with spodumene, so the sinks should be assayed for Fe₂O₃ and other impurities as well as Li₂O. Coarse crushing reduces cost but leaves spodumene locked with gangue, and testing at two or three crush sizes quantifies the trade-off.
Glossary
- Heavy liquid separation
- A laboratory float-sink test that splits ore into density fractions using liquids of known density.
- Dense media separation
- The plant process using a heavy suspension, usually ferrosilicon in water, in cyclones or drums.
- Washability curve
- Cumulative grade and recovery of the sinks plotted against separation density.
- Cut density
- The density at which particles are split into floats and sinks.
- Sinks and floats
- Particles heavier and lighter than the liquid.
- Ecart probable, Ep
- A measure of DMS sharpness: half the density range between 25% and 75% reporting to sinks.
- SC6
- A spodumene concentrate grading about 6% Li₂O, a common commercial specification.
- LCE
- Lithium carbonate equivalent: the mass of Li₂CO₃ the contained lithium could make.
Frequently asked questions
What density separates spodumene from quartz?
Spodumene has an SG of about 3.15 and quartz 2.65, so cuts between about 2.7 and 3.0 are tested. The best cut depends on the amount of intermediate-density mica and iron minerals in the ore.
How much Li₂O does pure spodumene contain?
8.03%. A 6% Li₂O concentrate is therefore about 75% spodumene, if all of the lithium is in spodumene.
How does HLS differ from DMS?
HLS is a near-perfect laboratory separation in true liquids. DMS in the plant uses a ferrosilicon suspension in cyclones and misplaces some particles near the cut density, so plant results are somewhat lower.
Why are fines removed before HLS?
Dense media separation cannot treat fine particles efficiently, so the test is run on the coarse fraction that the plant would treat, and the fines are directed to other processes such as flotation.
Discuss your plant with a metallurgist
Tiegel is a Perth metallurgy company. We help with testwork, plant optimisation, metal accounting and investigations, and we build the Tiegel Suite software.
Talk to a metallurgistReferences and further reading
- Wills, B.A. and Finch, J.A. (2016). Wills' Mineral Processing Technology, 8th edition. Butterworth-Heinemann. Chapter 11, Dense medium separation.
- Tadesse, B., Makuei, F., Albijanic, B. and Dyer, L. (2019). The beneficiation of lithium minerals from hard rock ores: a review. Minerals Engineering, 131.
Worked examples and charts marked illustrative use invented figures, not data from any operation.