Vanadium Solvent Extraction: Leach Liquor to V2O5

Last updated: 2026-11-04

Vanadium solvent extraction is what turns a leach liquor holding 2-6 g/L vanadium into a strip solution clean enough to crystallize ammonium metavanadate and calcine to V2O5 of commercial purity. The flowsheet pivots on one variable most metals never present: oxidation state. Control V(IV) against V(V) correctly and a four-stage circuit credibly returns better than 94% of the vanadium to product (indicative).

Key Takeaways

• Oxidation state is the design variable: V(IV) extracts as a cation, V(V) as an anion.
• D2EHPA systems take V(IV) at pH 1.5-2.5; amines take V(V) from alkaline or weak-acid liquors.
• Reduce Fe(III) to Fe(II) before D2EHPA extraction — iron is the main selectivity thief.
• Three to five extraction stages at O/A 1.0-1.5:1 are typical; pH control between stages matters more than stage count.
• A 30 m3/h, 4 g/L V(IV) liquor yields about 4.8 t/d of V2O5 (synthetic worked example below).

Why Is Vanadium Solvent Extraction Different From Other Metal SX?

Ask a copper SX engineer to design a vanadium circuit and their first question — “what form is the metal in solution?” — has three answers instead of one. Vanadium solvent extraction is a liquid-liquid separation process that exploits a changeable oxidation state rather than a fixed ionic form, which makes redox control as much a design variable as extractant choice. In copper SX the ion is always Cu2+; here the metal can arrive as V(III), V(IV), or V(V), and the industrial extractant families split along exactly that line.

The feed sources decide the chemistry. Stone coal and black shale operations that roast and acid leach their ore deliver vanadium as a V(IV)/V(V) mixture riding on high iron and aluminum. Vanadium recovery from steel slag and vanadium slag usually yields V(IV)-dominant sulfate liquor after reducing roast-leach treatment. Spent catalyst recycling — dehydrogenation and desulfurization catalysts above all — produces the strongest liquors, often 5-20 g/L V (deposit- and stream-specific), with molybdenum as the awkward co-solute. All three feed families funnel into the industry segment covered by our vanadium and manganese industry applications page, and all three depend on the general solvent extraction principles summarized on that pillar page.

Now the chemistry that makes vanadium unusual. The vanadyl ion, VO2+, is a stable doubly charged cation that dominates V(IV) speciation in acidic sulfate liquor below roughly pH 3 — precisely the window where cation exchangers of the D2EHPA family work best. Oxidize vanadium to the pentavalent state and the picture inverts: V(V) exists in solution as anionic species — H2VO4- and related monomers near neutral pH, and at pH below about 2 the polymerized decavanadate ion, V10O28^6-. Anions are invisible to cation exchangers but are exactly what primary, tertiary, and quaternary amine extractants are built for, by ion-pairing.

So oxidation state control is THE design variable in vanadium SX. Choose the cation route and you must hold the liquor reducing enough that vanadium stays as VO2+ — adjusted with reductants chosen by the project (the usual families being iron-based reductants, sulfur dioxide, or bisulfite). Choose the anion route and you must oxidize fully to V(V), with generic oxidants dosed to a controlled residual. Half-oxidized liquor satisfies neither extractant: extraction efficiency sags, scrubbing loads up, and the purity of the eventual V2O5 suffers. No other common hydrometallurgical SX duty makes the flowsheet this sensitive to a valve left open on a redox reagent line.

Which Parameters Govern Vanadium Solvent Extraction?

Once the route is chosen, a vanadium SX flowsheet is governed by a short list of parameters, summarized in the table below. Every range is typical or indicative — test work on your own liquor sets the final numbers, because stone coal leachates, slag liquors, and catalyst liquors differ by a factor of two or more in impurity load at the same vanadium grade.

Two parameters deserve special commentary. First, pH. D2EHPA is an organophosphorus acid extractant that exchanges its hydrogen ion for the vanadyl cation, so extraction consumes acid as the organic loads — pH rises as liquor moves down a countercurrent cascade unless each stage is trimmed. That one sentence explains most operating grief in vanadium circuits, and it is why stage-by-stage pH instrumentation appears again in the pitfalls section below. Second, the O/A ratio: because loaded organic capacity is modest (2-5 g/L V, indicative) relative to a 3-6 g/L feed, the organic flow must be larger than the aqueous flow, which is unusual compared with copper duty and directly shapes settler sizing. A disciplined scrub stage follows extraction: a scrub is a short organic-aqueous contact that washes co-extracted iron, aluminum, and phosphate off the loaded organic with a small volume of dilute acid or other generic scrub liquor, returning them to the circuit instead of the product. Stage hydraulics, mixer power, and settler geometry for all of this follow the mixer settler design guide on our solvent extraction resource hub, which also carries the strip and precipitation stages in more depth.

How Do You Design the Flowsheet for Different Vanadium Feeds?

Seven steps take any vanadium solvent extraction project from liquor sample to flowsheet verdict:

1. Characterize the liquor, not the ore. Measure V grade, V(IV)/V(V) split, free acidity or carbonate alkalinity, Fe(II)/Fe(III), Al, P, Si, Ca, and Mo. Valence speciation on a fresh sample is non-negotiable — valence shifts in the sample bottle otherwise lie to you.
2. Fix the valence target and extractant family. Reducing, acidic liquor with workable iron levels points to V(IV) plus D2EHPA-family cation extraction; oxidizing, alkaline-roast carbonate liquor points to V(V) plus amine-family anion extraction. This single decision fixes most of the rest.
3. Design impurity preconditioning. For the cation route this usually means reducing Fe(III) to Fe(II) to a controlled residual; for the amine route it means polishing out silica and controlling molybdenum.
4. Set pH windows, O/A ratio, and stage count. Bench isotherms and a McCabe-Thiele construction give the theoretical stages; add one practical stage, and verify with mini-cascade test work.
5. Choose scrub chemistry and stage position. Match the scrub to the specific co-extracted impurity — dilute acid scrubs for Fe and Al, targeted washes where P or Mo pressure exists.
6. Choose strip and downstream crystallization. Carbonate or ammonia-based strip (generic), then ammonium metavanadate precipitation and calcination to V2O5.
7. Close the water and salt balance. Strip reagents add sodium or ammonium salts; bleed streams and crystallization duty must be designed, not discovered.

The three feed scenarios then resolve into distinct verdicts, summarized with their indicative liquors here:

For stone coal acid roast liquor, the verdict is clear: run the D2EHPA cation route at pH 1.5-2.5 with deliberate Fe(III) reduction ahead of extraction. The liquor is already acidic and reducing-leaning; fighting it with an amine route would demand full oxidation and large neutralization reagent flows. For alkaline roast liquor, the verdict is equally clear: extract V(V) directly with an amine family from the carbonate liquor — the metal is already anionic, iron is naturally low, and acidifying just to switch to the cation route would burn money on reagents and gypsum disposal. For spent catalyst liquor, the verdict is conditional on molybdenum: where Mo matters commercially, the amine route after full oxidation is preferred because Mo and V both report to the amine and separate in strip or scrub; where Mo is negligible, D2EHPA after valence conditioning is the simpler plant. In all three cases the verdict, not the option list, is the deliverable — a flowsheet report that ends without one is unfinished.

What Goes Wrong in Vanadium Solvent Extraction Before the Assay Tells You?

1. pH creep between stages kills extraction before the assay proves it. The phenomenon: raffinate vanadium climbs over several days while extractant concentration assays look normal. The cause: D2EHPA extraction consumes acid, so pH drifts upward stage by stage unless each one is trimmed; past roughly pH 2.5-3 (indicative), aluminum hydrolysis flocs and crud formation compound the loss of interfacial area. The lab assay lags by a shift or more, so the first reliable alarm is the in-line pH trend, not the metal balance. The fix: pH measurement and acid trimming at every stage, with alarms set on drift rate rather than absolute limit.

2. Overnight oxidation-state drift in the feed tank ruins the scrub circuit. The phenomenon: after a weekend, the scrub stage suddenly carries iron into the strip liquor and AMV purity dips. The cause: the feed sat 12-48 hours, residual reductant was consumed, dissolved oxygen and residual oxidant pushed Fe(II) back to Fe(III) — and Fe(III), not Fe(II), is what D2EHPA loads strongly and scrubs poorly. The fix: dose reductant to a measured residual, not a recipe; sample the feed tank redox every shift; re-condition feed immediately before extraction, never the evening before.

3. Silica-driven interfacial crud. The phenomenon: a stable rag layer builds at every interface within weeks of startup. The cause: stone coal and slag liquors carry colloidal silica and aluminosilicate fines that collect at the organic-aqueous boundary; pH excursions accelerate it. The fix: polish the liquor upstream (clarification or filtration on the feed side), keep pH out of the hydrolysis window, and schedule crud removal as routine maintenance rather than an emergency.

4. Phosphate rides all the way to AMV. The phenomenon: strip liquor chemistry looks fine but ammonium metavanadate filtration slows and product analysis shows phosphate. The cause: P co-extracted in trace amounts, or carried physically in entrained liquor, concentrates through strip and co-precipitates. The fix: audit P at every stage boundary, strengthen the scrub where the audit points, and polish strip liquor before precipitation.

5. Gypsum scale in mixers and settlers. The phenomenon: impeller mass and shortened settler residence within months. The cause: pH adjustment with lime-based reagents in sulfate liquor supersaturates gypsum exactly where energy is highest. The fix: prefer acid-forward pH control within the extraction train, place neutralization where scaling does the least harm, and design vessels for cleanout access from day one.

6. Emulsification and organic inventory loss. The phenomenon: organic entrained in raffinate and strip liquor forces continual make-up. The cause: fine solids, degraded organic, and over-agitated mixers. The fix: control solids entering the circuit, monitor phase-disengagement time as a leading indicator, and keep mixer tip speeds within the design envelope rather than chasing throughput with rpm.

Worked Example: From 30 m3/h of Leach Liquor to V2O5 Tonnes

Consider a synthetic but fully redoable duty: 30 m3/h of V(IV) leach liquor at 4 g/L V, pH 2.0, from a stone coal roast-leach operation, treated through a D2EHPA cation circuit. The arithmetic below is internally consistent and every efficiency is marked indicative.

The extraction arithmetic rewards a second look. With ~93% transfer per stage, perfect countercurrent theory puts raffinate at 4 x 0.07^4 g/L — about 0.1 mg/L, a number no real plant reports. The honest credit is 96.5% overall, equivalent to a raffinate of roughly 0.14 g/L V: the gap between theory and credit is where entrainment, stage inefficiency, and imperfect valence control live. Downstream, 114.6 kg/h of vanadium at 99% strip recovery gives 9.6 g/L in a compact 12 m3/h strip liquor — concentrated enough for economic ammonium metavanadate precipitation. AMV carries a vanadium mass fraction of 50.94/116.98 = 0.4355, and calcination (2 NH4VO3 yields V2O5 plus ammonia and water) converts at a 181.88/233.96 = 0.7774 mass ratio, landing at 201.5 kg/h of V2O5 — call it 4.8 t/d from a plant treating one 30 m3/h stream.

What Does a V2O5 Recovery Circuit Cost to Build and Run?

Capital and operating costs for a vanadium SX circuit scale with liquor flow and impurity pressure more than with vanadium grade, so the ranges below are stated for the synthetic 30 m3/h case with explicit assumptions. Indicative installed capital for the extraction, scrub, strip, and precipitation sections — excluding leach, roast, and calciner — commonly falls in the range of roughly USD 4-9 million for a mixer-settler based plant (synthetic planning range, single train, standard materials of construction), with the spread driven by stage count, settler volume, and the sophistication of feed preconditioning. Operating cost is dominated by three lines: strip and precipitation reagents (carbonate- and ammonia-family consumption often lands near 2.0-2.5 t of sodium carbonate per tonne of stripped vanadium, indicative), acid for pH control, and organic make-up, which plant trials — never brochures — must quantify because entrainment behavior is site-specific.

Equipment selection follows directly from the process character. Vanadium duty means modest flows, tight pH windows, fast extraction kinetics, and crud sensitivity: compact mixer-settler cascades with short residence and per-stage pH trimming fit that character well. Where plot space, inventory, or organic entrainment dominate the trade, a centrifugal extractor is the alternative — short contact time, small holdup, fast phase disengagement — and our 50mm rotary drum centrifugal extractor is the typical pilot-to-commercial stepping stone for exactly this duty. The verdict for most greenfield vanadium plants: start with a compact mixer-settler cascade, because per-stage pH and redox interventions are easier to retrofit there, and reserve centrifugal extractors for space-constrained revamps or high-crud liquors where rapid disengagement pays. Finally, do not ignore the salt end: strip and precipitation generate sodium- or ammonium-rich bleed streams, and an MVR evaporation crystallization unit closes the water balance while recovering saleable salts instead of sending reagent value to the pond. Budget its duty in step 7 of the design procedure, not as a late addition.

FAQ

Why is V(IV) extracted by D2EHPA while V(V) needs amines?

Because the two oxidation states carry opposite charge. V(IV) exists in acidic liquor as the vanadyl cation VO2+, which cation exchangers like D2EHPA capture by exchanging a hydrogen ion. V(V) exists as anionic species such as H2VO4- or decavanadate, which amines capture by ion pairing. One metal, two extractant families — and no extractant handles both well, which is why oxidation state is fixed before extraction, not during.

What pH should a vanadium solvent extraction circuit run at?

For the D2EHPA route treating V(IV), the workable window is roughly pH 1.5-2.5 (typical, confirmed by test work). Below 1.5, the hydrogen ion competes with VO2+ for the extractant and efficiency drops; above about 3, aluminum hydrolysis, gypsum, and crud formation degrade settling and purity. Amines on alkaline liquors are far less pH-sensitive, which is one quiet advantage of the anion route.

How many extraction stages does a vanadium circuit need?

Three to five countercurrent stages are typical, with four as the common first-pass design (indicative). Bench isotherms and a McCabe-Thiele plot give the theoretical count; add one practical stage for axial mixing and pH drift. Going below three stages leaves the circuit hostage to feed variation; beyond five or six, each added stage buys vanishingly little and adds crud-prone interface area.

Why is reducing Fe(III) to Fe(II) so important before D2EHPA extraction?

Fe(III) is extracted by D2EHPA at least as readily as vanadyl at pH 2, while Fe(II) is essentially left behind. An unreduced feed therefore loads iron onto the organic, contaminates the strip, and exhausts the scrub. Reducing Fe(III) to Fe(II) ahead of extraction — with a reductant dosed to a measured residual — is the single cheapest selectivity measure in the whole flowsheet, and it protects vanadium recovery and product purity simultaneously.

How is ammonium metavanadate precipitated from strip liquor?

The carbonate- or ammonia-based strip liquor is heated to typically 60-95 degrees C (indicative), ammonium ion is added as an ammonia- or ammonium-salt-family reagent, and pH is adjusted into the window where NH4VO3 solubility is lowest. Seeding and slow agitation coarsen the crystals for filtration. The AMV is then washed, dried, and calcined at roughly 450-550 degrees C to V2O5, releasing ammonia that well-run plants recover and recycle.

Can one plant treat both stone coal leach liquor and spent catalyst liquor?

Technically yes, practically rarely. The two liquors want different valence targets, extractant families, and scrub chemistries, so a shared plant spends much of its life reconditioning. It is workable when campaigns are long (weeks per feed) and valence control is automated; it is a false economy when feeds alternate daily. Campaign operation with deliberate changeover procedures is the only version we would recommend building.

When are centrifugal extractors better than mixer-settlers for vanadium?

Three situations: severely limited plot space, high crud or solids load where rapid phase disengagement limits entrainment, and high-value liquors where small organic inventory matters. Centrifugal contactors hold seconds of residence instead of minutes, so they also suit the fast VO2+ kinetics. The offset is mechanical complexity and per-unit cost, which is why most greenfield vanadium plants still choose compact mixer-settler cascades and reserve centrifuges for special cases.

Conclusion

Vanadium solvent extraction is a flowsheet designed around one insight: the metal’s oxidation state is a process variable you control, not a fact you receive. Hold V(IV) as the vanadyl cation and a D2EHPA circuit at pH 1.5-2.5 with iron reduced ahead of it will deliver; push vanadium to V(V) and the amine route takes over; get the valence wrong and no extractant will rescue you. Around that core sit a handful of decisions — O/A ratio, three to five stages, a disciplined scrub, a carbonate or ammonia strip, ammonium metavanadate precipitation, calcination — each with a concrete verdict available from test work, as the synthetic 4.8 t/d example above shows. If you are scoping a stone coal, slag, or spent-catalyst vanadium project, send us the liquor analysis and target purity, and request a quote for a bench extraction program and equipment sizing — the flowsheet verdict you need is a few weeks of disciplined test work away.

References

• U.S. Geological Survey, Mineral Commodity Summaries — Vanadium chapter.
• Moskalyk, R.R. and Alfantazi, A.M., “Processing of vanadium: a review,” Minerals Engineering.
• Gupta, C.K. and Krishnamurthy, N., Extractive Metallurgy of Vanadium, CRC Press.
• Ritcey, G.M., Solvent Extraction: Principles and Applications to Process Metallurgy, revised ed.
• Lo, T.C., Baird, M.H.I. and Hanson, C., Handbook of Solvent Extraction, Krieger Publishing.
• Wills, B.A. and Finch, J.A., Wills’ Mineral Processing Technology, 8th ed., Butterworth-Heinemann.
• SME Mineral Processing and Extractive Metallurgy Handbook, Society for Mining, Metallurgy and Exploration, Englewood, CO.
• Habashi, F., A Textbook of Hydrometallurgy, Metallurgie Extractive Quebec.

Changelog

• 2026-11-04: Initial publication.

Related equipment

• Pilot-Scale Continuous Solvent Extraction Mini-Plant — Liquid-Liquid Separation (Solvent Extraction)