Why Spodumene Wastewater Is a Lithium Loss Problem, Not Just a Disposal Problem
Spodumene conversion plants lose roughly 5% of their total lithium load in the spent liquor — a figure that exceeds 4% on both the sulfuric acid bake and the soda ash/acid roast pathways, per Saltworks' 2024 hard-rock processing brief. For a 30,000 tpy LCE operation, that translates to 50-200 tpy of lithium carbonate equivalent (LCE) walking out of the plant in brine at 200-500 mg/L Li⁺ rather than ending up in product. The dominant conversion routes shape that loss in different ways. Sulfuric acid bake digestion yields a Na₂SO₄-rich, Li-bearing brine at pH 1-3 with sulfate above 20,000 mg/L. The soda ash/acid roast route produces a CaSO₄ + Li₂SO₄ brine at pH 6-9 with calcium and sodium as the balancing cations. In both cases, total dissolved solids (TDS) exceed 50,000 mg/L, leaving the spent liquor unsuitable for activated sludge: microbial kinetics effectively halt above 30,000 mg/L TDS, and there is no biodegradable BOD or COD load to drive a biological stage. Sulfate above 1,500 mg/L further rules out anaerobic biological sulfate reduction at reasonable reactor volumes. The economic pivot is straightforward — treating the brine as a lithium brine treatment problem rather than a waste disposal problem reframes the whole capex conversation. A 2026-compliant process train combines pH/oxidation adjustment, lime-soda softening, nanofiltration, reverse osmosis, and MVR-driven zero liquid discharge lithium crystallization that recovers the lost lithium as a saleable product.
Effluent Characterization by Conversion Route
Effluent profile is set by the conversion chemistry, and the wrong softening or membrane design follows from misreading it. The table below summarizes typical ranges for the three main pathways: sulfuric acid bake, soda ash/acid roast, and limestone roast.
| Parameter | Sulfuric Acid Bake | Soda Ash / Acid Roast | Limestone Roast |
|---|---|---|---|
| pH | 1-3 | 6-9 | 6-8 |
| TDS (mg/L) | 80,000-150,000 | 50,000-90,000 | 40,000-70,000 |
| SO₄²⁻ (mg/L) | 20,000-40,000 | 10,000-20,000 | 5,000-12,000 |
| Na⁺ (mg/L) | 20,000-30,000 | 15,000-25,000 | 5,000-12,000 |
| Ca²⁺ (mg/L) | 200-800 | 1,000-2,000 | 800-1,500 |
| Mg²⁺ (mg/L) | 100-500 | 50-200 | 200-500 |
| Li⁺ (mg/L) | 200-500 | 200-400 | 150-300 |
| SiO₂ (mg/L) | 50-150 | 80-200 | 100-200 |
| Fe/Mn/Al (mg/L each) | 5-50 | 5-30 | 5-20 |
Biological treatment is non-viable across all three brines. Above 30,000 mg/L TDS, monovalent salts disrupt the osmotic balance of microbial cells, suppressing substrate uptake rates to near zero; halophilic biomass would be required, and the absence of a BOD/COD load removes the driving force. The ore-derived metals — Fe, Mn, and Al in the 5-50 mg/L range — must be oxidized and precipitated before any membrane stage, since dissolved iron fouls polyamide RO elements irreversibly at concentrations above 0.1 mg/L in the feed. Silica at 50-200 mg/L SiO₂ becomes a membrane foulant above pH 8 when it polymerizes; antiscalant selection must therefore hold the operating pH at or below 7.5 in the RO loop, or include a magnesium oxide silica suppression step upstream. Heavy metals are not the only trace problem — strontium and boron at low mg/L levels also matter for battery-grade Li₂CO₃ specifications, and the process train must either reject them in the membrane stage or remove them by ion exchange downstream.
Step-by-Step Process Train: From Spent Liquor to Lithium Recovery
A defensible 2026 train sequences seven unit operations, each with measurable design targets. The order matters: skipping softening or sending unoxidized iron to NF/RO will shut a membrane rack down within days.
| Step | Unit Operation | Key Parameters | Design Target |
|---|---|---|---|
| 1 | Equalization + oxidation | pH 7-8 with lime/NaOH; ORP +200 to +400 mV via NaClO or H₂O₂ | Fe/Mn/Al precipitated; TSS 80-95% reduction |
| 2 | Lime-soda softening | 1.5-2.5 g Ca(OH)₂ per g Mg; 2-3 g Na₂CO₃ per g Ca | Ca <50 mg/L; Mg <10 mg/L |
| 3 | Multi-media filtration + cartridge/UF guard | 5 µm cartridges; UF at 0.1-0.2 µm | SDI <3; turbidity <1 NTU |
| 4 | Nanofiltration preconcentration | NF270/DK at 20-30 bar; 70-85% recovery | Divalent rejection 95-99%; monovalent passage 30-60% |
| 5 | Reverse osmosis / closed-circuit RO | Feed TDS 30,000-80,000 mg/L; 60-80% recovery | Monovalent rejection 95-99%; Li concentrated 2-5× |
| 6 | MVR / forced-circulation crystallization | 50-80°C; Na₂SO₄ removed as Glauber's salt | Mother liquor to Li₂CO₃ / Li₃PO₄ precipitation at pH 11-12 |
| 7 | Polishing and reuse | Mixed-bed ion exchange on MVR condensate | Conductivity <0.5 µS/cm; >95% water loop closure |
Step 1 — Equalization and oxidation: a 6-12 hour equalization basin evens out pH and flow surges from the conversion circuit. Lime or caustic raises pH to 7-8, and ORP is held at +200 to +400 mV with sodium hypochlorite (typical dose 5-15 mg/L as Cl₂) or 30% H₂O₂ to oxidize Fe²⁺ to Fe³⁺ and Mn²⁺ to Mn⁴⁺, which precipitate as hydroxides. A DAF clarifier for TSS and metal hydroxide removal takes out 80-95% of the suspended solids in the same stage, generating a sludge that downstream processing dewaters.
Step 2 — Lime-soda softening: this is the heart of the brine train. Lime dose 1.5-2.5 g Ca(OH)₂ per gram of magnesium precipitates Mg(OH)₂, and soda ash dose 2-3 g Na₂CO₃ per gram of calcium precipitates CaCO₃. Target residuals are Ca below 50 mg/L and Mg below 10 mg/L — anything higher will foul the downstream RO with CaSO₄ scale (Ksp ~4.9 × 10⁻⁵) and Mg(OH)₂. The softening sludge reports to a filter press for softening and metal hydroxide sludge at 25-35% dry solids.
Step 3 — Media filtration and cartridge/UF guard: a multi-media filter for SDI reduction brings the Silt Density Index below 3 and turbidity below 1 NTU, which is the warranty threshold for spiral-wound RO. A 5 µm cartridge and a UF polishing stage at 0.1-0.2 µm remove the fines that slip past the media filter. SDI above 5 will crush RO membrane life from 5+ years to under 18 months.
Step 4 — Nanofiltration preconcentration: NF270 or DK-class membranes run at 20-30 bar with 70-85% recovery. The divalent rejection of 95-99% strips Ca, Mg, and SO₄, while monovalent passage of 30-60% lets Na⁺ and Li⁺ bleed into the permeate. This is the nanofiltration lithium recovery step that lets a downstream RO see a softened, partially desalinated feed. For a closer look at the principles, see this nanofiltration working principle and industrial applications reference.
Step 5 — Reverse osmosis: the NF permeate enters an RO system for lithium preconcentration running at feed TDS of 30,000-80,000 mg/L and 60-80% recovery. Monovalent rejection is 95-99% for Na⁺ and Li⁺, so the retentate ends up 2-5× more concentrated in lithium than the feed. Closed-circuit RO (CCRO) is increasingly preferred over conventional RO above 50,000 mg/L feed TDS because it cuts energy 20-35% and reduces scaling frequency.
Step 6 — MVR crystallization: the RO retentate feeds a MVR crystallizer at 50-80°C under vacuum. Na₂SO₄ drops out as Glauber's salt (Na₂SO₄·10H₂O) below 32°C, or as anhydrous Na₂SO₄ above that temperature. The mother liquor, now enriched in lithium with sulfate stripped, reports to Li₂CO₃ or Li₃PO₄ precipitation at pH 11-12 using soda ash or trisodium phosphate, dosed via an automated chemical dosing for lime, soda, and Na₂CO₃ reagent addition system for stoichiometric accuracy.
Step 7 — Polishing and reuse: MVR condensate carries trace volatile organics and dissolved CO₂; a mixed-bed ion exchange polisher drops conductivity below 0.5 µS/cm, qualifying the stream as boiler feed or CIP water. Water loop closure exceeds 95%, which is the only way to meet inland Australian and Chinese discharge regulations on a zero-discharge basis. For facilities that run a small low-salinity side stream (lab wastewater, camp greywater), a MBR polishing step for any low-salinity side streams handles the biological load that the main brine train cannot.
Lithium Recovery: Choosing Between Li₂CO₃, Li₃PO₄, and LiOH Routes
The recovery step is where wastewater becomes revenue. Three precipitation routes dominate, and the choice is set by market, downstream use, and reagent logistics.
| Route | Reagent / pH | 2025-2026 Price | Li Recovery | Reagent Cost | Best Fit |
|---|---|---|---|---|---|
| Li₂CO₃ (battery grade) | Na₂CO₃ at pH 11-12, 80-90°C | $13,000-15,000/t LCE | 85-95% | $2,000-3,000/t LCE | NMC cathode, general battery market |
| Li₃PO₄ (high purity) | Na₃PO₄ at pH 9-10 | $25,000-40,000/t | up to 99.9% | Lower than Li₂CO₃ route | LFP cathode precursor |
| LiOH·H₂O | Causticization of Li₂CO₃ with Ca(OH)₂ | $18,000-25,000/t | 80-90% | Adds $1,500-2,500/t | High-nickel cathode, premium batteries |
Lithium carbonate precipitation at pH 11-12 and 80-90°C is the workhorse route. The reagent cost is well-characterized at $2,000-3,000 per tonne of LCE, and the saturated Na₂CO₃ mother liquor can be recycled upstream to the softening stage, dropping net reagent demand. The Li₃PO₄ route at pH 9-10 yields 99.9% Li recovery in a single stage and produces a precursor that feeds directly into LFP cathode production — attractive when the same operator runs both spodumene conversion and cathode active material plants. The LiOH·H₂O route is the highest value but adds a causticization reactor (Li₂CO₃ + Ca(OH)₂ → 2 LiOH + CaCO₃) and a calcium removal stage; it pays off only when battery-grade LiOH commands a sustained premium. Each kilogram of lithium recovered avoids 1.5-2.5 m³ of brine sent to ZLD, directly cutting crystallizer capex and steam demand — the lithium recovery from waste brine credit is the single largest lever on net treatment cost.
Zero Liquid Discharge Design and Operating Cost in 2026
The capital and operating cost of a high TDS wastewater train in 2026 hinges on the lithium recovery add-on and the feed concentration. The table below shows a 50 m³/h spodumene wastewater case (softening + NF + RO + MVR) at mid-2026 pricing.
| Cost Item | Benchmark 2026 | Notes |
|---|---|---|
| CAPEX (full train, 50 m³/h) | $8-15M USD | Includes softening, NF, RO, MVR; lithium precipitation adds $1.5-3M |
| OPEX (per m³ treated) | $1.50-3.50 | Steam 45%, electricity 20%, reagents 15%, membranes/maintenance 20% |
| Net OPEX after Li credit | $1.00-2.50/m³ | Lithium revenue offsets 20-40% of OPEX at 2025-2026 LCE prices |
| MVR steam economy | 8-12 kg water/kg steam | 60-80% thermal energy reduction vs multi-effect evaporation |
| MVR vs TVR preference | MVR above 20 m³/h feed | TVR viable 5-20 m³/h; below 5 m³/h, mechanical vapor recompression still wins on footprint |
CAPEX spans $8-15M USD for the full train at 50 m³/h, with the lithium precipitation package adding another $1.5-3M depending on whether the plant goes to Li₂CO₃, Li₃PO₄, or LiOH. OPEX runs $1.50-3.50 per cubic meter of treated brine, dominated by steam (45%) and electricity (20%); reagents and membrane replacement split the remaining 35%. At 2025-2026 LCE prices, the recovered lithium credits offset 20-40% of OPEX, so net treatment cost lands at $1.00-2.50/m³ — competitive with conventional ZLD-only designs that produce no saleable byproduct. MVR is preferred above 20 m³/h feed because steam economy of 8-12 kg water per kg steam translates to a 60-80% reduction in thermal energy versus multi-effect evaporation, which is the baseline against which any ZLD lithium plant is benchmarked. For a deeper dive on brine concentration economics, see the forward osmosis operating cost and ROI for brine concentration reference, and the broader high salinity wastewater treatment process methods and costs guide.
Frequently Asked Questions
What is the lithium concentration in spodumene processing wastewater?
Spent liquor from sulfuric acid bake digestion typically carries 200-500 mg/L Li⁺, representing roughly 5% of the plant's total lithium load across both the acid bake and soda ash/acid roast pathways. At 50 m³/h feed, that concentration yields 50-200 tpy of recoverable LCE for a mid-sized operation.
Why can't biological treatment handle spodumene wastewater?
TDS above 50,000 mg/L inhibits microbial kinetics beyond recoverable rates, and the brine has no biodegradable BOD or COD load to drive a biological stage. Sulfate above 1,500 mg/L further rules out anaerobic sulfate reduction at practical reactor volumes, forcing the train into a physico-chemical and membrane configuration.
What NF/RO recovery rates apply to lithium preconcentration?
Nanofiltration with NF270 or DK-class membranes runs at 20-30 bar with 70-85% recovery, rejecting 95-99% of divalent ions while passing 30-60% of monovalents. Downstream reverse osmosis at 60-80% recovery rejects 95-99% of the remaining monovalents, concentrating lithium 2-5× in the retentate before MVR crystallization.
How much does a 50 m³/h spodumene wastewater train cost in 2026?
A complete softening + NF + RO + MVR train for 50 m³/h runs $8-15M USD CAPEX in 2026, with an additional $1.5-3M for the lithium precipitation package. OPEX lands at $1.50-3.50 per cubic meter before lithium credit, dropping to $1.00-2.50/m³ net once the recovered LCE is sold at 2025-2026 market prices.
Which lithium salt should a spodumene plant precipitate from its brine?
Li₂CO₃ at $13,000-15,000/t LCE is the safest default with 85-95% recovery and well-understood reagent cost. Li₃PO₄ at $25,000-40,000/t and 99.9% recovery pays off when the same operator runs an LFP cathode line. LiOH·H₂O at $18,000-25,000/t requires an extra causticization step and pays off only where high-nickel cathode demand is firm.
