Technological breakthroughs

A research team at the University of Kelaniya, Sri Lanka, has published an integrated process that simultaneously produces nitrophosphate fertilizer and recovers rare earth elements (REEs) from Eppawala apatite — a 60-million-ton mineral deposit that has yet to be fully utilized in terms of value.

Background

Rare earth elements (REEs) — a group of 17 elements comprising the lanthanides along with yttrium and scandium — are present in many modern technological devices: electric vehicle batteries, wind turbines, and smartphones. China currently dominates the global supply chain, while many countries are seeking additional sources.

Sri Lanka holds the Eppawala apatite deposit in the North Central Province, containing measurable REE concentrations: cerium (Ce) at 1,124 ppm, lanthanum (La) at 482 ppm, neodymium (Nd) at 477 ppm, with total REEs around 2,465 mg/kg of ore. However, direct REE extraction from this deposit has been considered economically unviable due to the relatively low concentrations involved.

The research team led by Associate Professor Pradeep Wishwanath Samarasekere approached the problem differently: rather than treating REE extraction as a standalone process, they integrated REE recovery into an existing fertilizer production line, so that costs are shared and all raw material inputs are fully utilized.

A six-step process

Step 1 — Acid leaching with nitric acid: The apatite ore is ground to below 62 µm, then dissolved in 10M nitric acid at 70°C for 120 minutes at a pulp density of 2.4 mL/g. Under these conditions, REE leaching efficiency reached 99.7% and P₂O₅ decomposition efficiency reached 99.6%. The research team noted that acid concentrations either lower or higher than 10M both reduced efficiency, likely due to increased solution viscosity hindering the diffusion process.

Step 2 — Cooling crystallization for calcium removal: The leachate contains a large amount of Ca²⁺ ions, which interfere with subsequent REE recovery steps. The team cooled the solution to between -5°C and -20°C for 90 minutes, causing calcium to precipitate as Ca(NO₃)₂·4H₂O. This removed 42% of the calcium content, while REE co-precipitation remained below 2%. The recovered Ca(NO₃)₂·4H₂O can be repurposed as a nitrogen-rich fertilizer.

Step 3 — Partial neutralization to precipitate REE phosphates: The mother liquor was adjusted to pH 1.4 using 25% NH₄OH solution at 70°C. At this pH level, REE³⁺ ions selectively precipitate as REPO₄, achieving extraction efficiencies above 90% for all monitored rare earth elements.

Step 4 — Selective dissolution of REEs: The resulting precipitate still contains calcium and iron impurities. The team treated it with a phosphoric-sulfuric acid mixture to dissolve the REEs, while calcium converted to insoluble CaSO₄ and was removed. REE dissolution efficiency at this step exceeded 99%.

Step 5 — Double sulfate precipitation: REEs in solution were precipitated as NaREE(SO₄)₂·xH₂O by adding NaCl or Na₂SO₄ at 80°C. The resulting product contained 22.2% total REEs by weight.
Step 6 — Fertilizer production: The filtrate remaining after REE separation in Step 3 — still rich in nitrate and phosphate — was fully neutralized with ammonium hydroxide to produce the final nitrophosphate fertilizer product.

Recorded results

Regarding rare earth recovery, total REE recovery exceeded 90%, with individual element efficiencies in the order Pr > Nd > Ce > Gd > Sm > Y > Dy. Praseodymium and neodymium — two elements important in manufacturing magnets for electric motors and wind turbines — ranked among the highest in recovery efficiency. Dysprosium achieved lower efficiency during the neutralization stage but showed the highest recovery rate (59.5%) during the double sulfate precipitation stage, attributed to its smaller ionic radius and higher charge density compared to lighter REEs.

Regarding fertilizer quality, the nitrophosphate product met the following specifications: total nitrogen 18.2%, total phosphorus 13.9% (as P₂O₅), moisture content 0.6%, free phosphoric acid 0.1%. SEM-EDX analysis detected no traces of REEs or heavy metals in the fertilizer product.

Differences from existing methods

The two most common fertilizer production methods currently in use — SSP (single superphosphate) and TSP (triple superphosphate) — both generate phosphogypsum as a by-product, which can retain up to 85% of REEs and is difficult to process further. The nitrophosphate route used by the Sri Lankan team does not produce this by-product, allowing REEs to enter the solution and be recovered through the process described above.

The authors also noted that the CaO/P₂O₅ ratio of Eppawala ore is 1.53 — below the threshold of 1.6 commonly regarded as the economic boundary for wet-process applications — suggesting the ore is suitable for the proposed approach.

Remaining limitations

The authors acknowledge several areas requiring further improvement. Recovery efficiency at the double sulfate precipitation step was uneven across elements, with gadolinium reaching only 19.6%. The NaREE(SO₄)₂·xH₂O product still contains significant impurities including sodium (68.4%), iron, and aluminum, requiring additional purification steps if higher purity is needed.

Regarding deposit composition, the light/heavy REE ratio (LREE/HREE) at Eppawala is 17.3 — heavily skewed toward light rare earths. For comparison, a Swedish apatite deposit has a ratio of approximately 2.74 and contains a significantly higher proportion of heavy rare earths — the group currently commanding higher strategic value on the market.

The team recommends incorporating solvent extraction techniques in future studies to improve purity and enable separation of individual REE elements.

This research was published in the journal Sustainability (MDPI) in July 2025, funded by the Science and Technology Human Resource Development Project of Sri Lanka through the Asian Development Bank.