Mixer Settler Efficiency: Design for 92% Stage Recovery

Last updated: 2026-10-28

Mixer settler efficiency — expressed as Murphree stage efficiency — tells you how close each contact stage comes to equilibrium, and it typically runs 85–95% in well-designed hydrometallurgical service. Designing each stage for a target of 92%, instead of treating efficiency as an afterthought, is what separates circuits that hold nameplate recovery from those that quietly leak metal into raffinate for years.

This guide is written for process engineers who must commit a stage count, mixer size, or settler area to paper. It sits inside our solvent extraction resources library, alongside the wider liquid-liquid separation equipment family, and it assumes you already know what a mixer-settler is — if not, start with our mixer settler design guide and come back.

Key Takeaways

• Mixer settler efficiency is measurable: 85–95% Murphree efficiency per stage is typical, 92% is a sound design target.
• Four levers set it: mixer power intensity, residence time, settler flux, and phase continuity.
• Pilot stage testwork, not vendor claims, should fix the design efficiency number before procurement.
• Four stages at 92% beat five stages at 80% on capital, power, and organic inventory.
• Every percentage point of lost efficiency shows up as recoverable metal stranded in raffinate.

What Is Mixer Settler Efficiency and Why Does It Cap Recovery?

Every solvent extraction flowsheet is built on an equilibrium calculation, but no real tank ever reaches equilibrium. Mixer settler efficiency is a fractional measure that expresses how far an actual contact stage progresses toward equilibrium, typically 85–95% of the way in hydrometallurgical duty. That shortfall sounds small, yet it compounds stage after stage and ultimately caps the recovery the plant can achieve, no matter how generous the reagent dose or how many operators watch the interface.

Four definitions anchor the rest of this article:

Murphree stage efficiency is a ratio that compares the actual concentration change achieved across one stage with the change that perfect equilibrium would deliver, typically 0.85–0.95 for copper oxime systems in extraction duty. It is the workhorse number for converting theoretical stages into physical tanks.

Settler flux is a hydraulic loading rate that states how many cubic metres of combined phase flow each square metre of settler plan area must handle per hour, typically 3–8 m³/h·m² in copper SX. It controls whether the dispersion band behaves or floods.

Phase continuity is a dispersion condition that determines which liquid forms the continuous background and which travels as droplets, typically organic-continuous in copper extraction and aqueous-continuous in stripping. Continuity governs droplet size, mass transfer area, and how the stage responds to O/A ratio drift.

Mixer settler efficiency, then, is the practical combination of all three: a whole-stage performance number that the mixer creates, the settler preserves, and the phase continuity pattern enables. A stage can mix beautifully and still underperform if the settler re-entrains droplets that never finished transferring their solute.

Why does this number cap recovery? Because the McCabe-Thiele stage count that underpins every SX flowsheet assumes equilibrium stages. When real stages deliver only 92% of a theoretical stage each, four physical tanks behave like 3.7 theoretical ones. If your flowsheet needs 3.6 theoretical stages, the design closes with margin. If efficiency sags to 85%, those same four tanks deliver 3.4 — and the raffinate tenor rises. Efficiency is therefore not a trim on the design; it is the multiplier sitting between your isotherm work and the steel you order. Note the asymmetry, too: lost efficiency cannot be recovered by adding reagent, because the constraint is contact effectiveness, not capacity. How to measure it honestly, and how to design for it deliberately, is the subject of the rest of this guide.

Key Design Parameters and Typical Ranges

Mixer settler stage design couples chemical kinetics with hydraulics, and the parameters below move together rather than independently. All values are typical industry ranges for copper oxime and similar acid-extraction systems; each one must be confirmed by stage efficiency testwork on your actual liquor and reagent pair before it becomes a procurement specification.

Three of these deserve interpretation beyond the table.

Mixer power intensity and residence time trade against each other. A mixer running at 1.4 kW/m³ with 150 seconds of residence will usually beat one running at 0.8 kW/m³ for 180 seconds, because mass transfer area — total droplet surface — is what the oxime reaction actually sees. But pushing past roughly 2 kW/m³ buys nothing: kinetics plateau while droplet size keeps falling, and the settler inherits the problem. The efficient design point is the lowest power intensity that reaches equilibrium in the residence time you can afford to build. In retrofit work we routinely find mixers running 20–30% above their original power intensity because operators chased recovery with speed; the recovery never came, but the entrainment did.

Settler flux is the most site-specific number on the list. Two plants with identical reagent systems can show a factor-of-two difference in allowable flux because of colloidal silica, clay fines, or reagent degradation products. Design at the value your dispersion band test curve supports, with margin, and treat industry averages as a starting hypothesis only.

Dispersion band depth is your early-warning instrument. A stable band at 200–300 mm means the settler is keeping up. A band creeping toward 350 mm and beyond means droplets are arriving finer than they are leaving — usually a mixing problem upstream, not a settling problem in the box. Operators who log band depth daily catch efficiency loss weeks before raffinate assays confirm it.

How Do You Size Stages for a Target Mixer Settler Efficiency?

Designing to a stated efficiency is a different exercise from designing to an O/A ratio or a stage count. The procedure below is how we approach it on new trains; the arithmetic of the isotherm itself is covered in our McCabe-Thiele walkthrough, so we pick up from there.

1. Fix the design efficiency target. Choose 92% Murphree efficiency as the basis for a new copper extraction train — ambitious enough to avoid padding the stage count, conservative enough that gentle mixing can deliver it. Record the basis in the design file so later optimization has a reference.
2. Determine theoretical stages from equilibrium. Run shake-out tests, plot the isotherm, and step off theoretical stages at the chosen O/A ratio. Suppose the construction requires 3.6 theoretical extraction stages for the target raffinate tenor.
3. Convert theory to hardware. Divide theoretical stages by design efficiency: 3.6 ÷ 0.92 = 3.9, so specify 4 physical extraction stages. Repeating the same division at 80% efficiency gives 4.5, which rounds to 5 stages — the entire capital argument for designing to efficiency in one line of arithmetic.
4. Set mixer power intensity and residence time. Size each mix compartment for 2–3 minutes at 1.0–1.5 kW/m³, then check impeller tip speed stays low enough to protect droplet size. These two settings are what physically deliver the 92% target.
5. Size the settler from disengagement testwork. Run dispersion band tests on real PLS and organic, select a flux with margin — say 4.5 m³/m²·h against a tested 6 — and fix the plan area so the band sits mid-window at design flow.
6. Choose and lock phase continuity. Operate extraction organic-continuous for droplet stability, set the O/A control loop accordingly, and confirm the impeller type suits that continuity. Continuity flips are the most common self-inflicted efficiency loss in commissioning.
7. Verify with a pilot stage before committing. Run a single full-scale-geometry stage — or a geometrically similar mini-plant stage — on real feed, measure aqueous in and out, and back-calculate the Murphree efficiency. Adjust power intensity or residence time until the measured value clears 92% with margin, then freeze the design.

Three scenarios show how the same procedure lands in different places:

• For high-flow, low-grade copper PLS — say above 400 m³/h at under 1.5 g/L Cu — a four-stage train designed at 92% efficiency is preferred because each extra physical stage adds concrete, agitators, and organic inventory that the metal value of a lean liquor struggles to repay; see our copper solvent extraction guide for the circuit context.
• For high-O/A cobalt and nickel polishing duty, a five-stage train with higher power intensity mixers is preferred because separation factors with D2EHPA versus Cyanex 272 class reagents are low, each theoretical stage is precious, and the smaller aqueous flow keeps mixer volume affordable.
• For small batch duty — a few cubic metres per batch on a campaign basis — a compact two-stage skid with generous residence time is preferred because operator attention is intermittent, so the design must reach equilibrium even at conservative mixing rather than depend on tight around-the-clock control.

Why Does Mixer Settler Efficiency Drop in Operation?

Efficiency is not a constant; it is a state. These six faults account for most of the deterioration we are called to diagnose. Each is listed as symptom, then cause, then corrective action.

1. Raffinate tenor creeping upward. Stage-wise aqueous assays show one or more stages losing 3–5 efficiency points over weeks → impeller wear reducing pumped flow, or extractant concentration drifting below specification as inventory ages → inspect impellers during the next shutdown and restore diameter and pitch; analyze organic for active extractant content and top up to the tested concentration.
2. Dispersion band deepening stage by stage. Band grows past 350 mm while interface stays on level control → power intensity has crept up with VFD changes, or feed fines are stabilizing fine droplets → trim mixer speed back toward design and review upstream solution clarity; if solids are the driver, see our crud treatment in SX resource before the band floods the weir.
3. Phase reversal in one mixer. Violent agitation, swinging levels, and a sudden step change in that stage’s efficiency → O/A ratio drift from a failed flow meter or a surging upstream pump has flipped continuity → restore ratio control and dampen the surge; re-establish the intended continuous phase at startup flow before returning the stage to service.
4. Organic carryover rising in raffinate. Visible sheen and ppm trends climbing at the raffinate outlet → excessive tip speed shearing secondary droplets the settler cannot catch → reduce impeller speed, verify against the entrainment guidance in our SX entrainment resource, and confirm the skim launder recovers what is already circulating.
5. First-stage efficiency chronically low after startup. Stage 1 delivers 80–85% while downstream stages meet design → air entrainment from a vortexing feedwell, or inlet piping delivering unequalized flow to the impeller → fit or repair the vortex breaker, baffle the feedwell, and balance the inlet distribution; efficiency usually recovers within days.
6. Efficiency differing between parallel trains on the same feed. Train A holds 92% while Train B sits at 87% with identical reagents → maldistributed flow or a worn drive on one train, often invisible because total flow looks correct → meter each train’s phases individually, compare absorbed power on the agitator drives, and correct the distribution before blaming the chemistry.

A disciplined plant logs per-stage aqueous assays weekly and dispersion band depth daily. With those two records, every fault above announces itself before it costs a campaign. Note also that several of these faults travel together: worn impellers raise droplet size variance, which deepens the band, which raises entrainment, which degrades the organic, which lowers kinetics again. Fix the mechanical root cause first, because chemical fixes layered on top of a mechanical problem are money spent twice.

What Does Mixer Settler Efficiency Cost in Power and Reagent?

Efficiency has a price, but it is small against what inefficiency costs. The figures below use explicit assumptions so you can rescale them: a 500 m³/h PLS stream at 1.9 g/L Cu, 1:1 O/A ratio (1,000 m³/h combined flow), six stages total (four extraction, two strip), mixer power intensity 1.2 kW/m³ over ~12 m³ of mixed volume per stage, power tariff US$0.10/kWh, copper at an assumed US$9,000/t, and extractant at an indicative US$9/kg. All figures are indicative and to be confirmed for your flowsheet.

Read the table from the bottom up. Designing the mixers to deliver 92% rather than 85% costs perhaps 0.02 kWh per m³ of combined feed in extra agitation — a few thousand dollars a year. The extractant inventory those mixers hold up is a working-capital number, not a consumption number, and it is identical whether the stage runs at 85% or 92%. Meanwhile every single point of efficiency the circuit fails to deliver strands copper in raffinate that the leach has already paid to dissolve: roughly 12–20 t of copper per year per point at these flows, worth US$110,000–190,000 annually. On this arithmetic, spending engineering effort — and modest testwork budget — to hold design efficiency is among the highest-return activities in the plant. The same logic applies downstream, where efficiency losses in stripping quietly inflate the acid balance of the copper SX-EW circuit.

Field Data: Four Stages at 92% Versus Five at 80%

The contrast between designing to efficiency and accepting whatever the mixer gives you is clearest in paired project experience. The data below comes from a copper SX circuit in Central Africa, anonymized as a matter of policy, and should be read as indicative of the pattern rather than as a guarantee.

A four-stage extraction train designed to our 92% basis recorded a 92.4% average Murphree efficiency over a 30-day campaign, measured by stage-wise aqueous analysis — composite samples taken at each stage’s aqueous outlet every four hours and back-calculated against the isotherm. Settler flux averaged 4.6 m³/m²·h with the dispersion band holding at 240 mm, and final raffinate tenor averaged 0.09 g/L Cu against a 0.12 g/L design promise. The circuit met its recovery target with margin to spare.

The comparison case was an older five-stage train elsewhere in the region, originally built when its designers assumed 80% efficiency. On paper, five stages at 80% deliver 4.0 theoretical stages — more than the 3.7 the 92% train achieves. In practice the older plant never saw that paper performance: two stages sagged to 74% during a reagent supply change, overall extraction fell below specification, and the plant responded by raising extractant concentration — buying with reagent what the mixers failed to deliver. Its combined mixer volume, agitator power, and organic holdup ran roughly 25% higher than the four-stage design for the same duty.

The lesson is not that five stages are wrong; some chemistries genuinely need them. The lesson is that a measured, designed efficiency lets you buy the fewest stages that meet recovery — and hold that recovery when reality intrudes. An 80% assumption forces you to buy a fifth stage as insurance against inefficiency you could instead have engineered away for the cost of one properly tested mixer.

FAQ

What is a good mixer settler efficiency for a new copper SX plant?
Design to 92% Murphree efficiency per extraction stage, with a proven achievable band of 85–95% for well-mixed gravity mixer-settlers on copper oxime systems. A 92% basis usually lets a four-stage train close the flowsheet where an 80% basis needs five. Confirm the number with a pilot stage on your actual liquor before procurement, because kinetics, temperature, and solids all move the achievable value.

How is Murphree stage efficiency measured in an operating plant?
Sample the aqueous phase entering and leaving each stage simultaneously, analyze for the metal of interest, and compare the actual change with the equilibrium change predicted from the isotherm at that composition. Stage-wise aqueous analysis over a 30-day composite campaign gives a robust average. The method needs no shutdown, only good sampling discipline and current isotherm data for your reagent batch.

Can stage efficiency exceed 100%?
Occasionally, yes, and it is a data-quality flag rather than a triumph. Apparent values above 100% usually mean the equilibrium reference used in the back-calculation no longer matches reality — a changed reagent batch, a temperature shift, or maximum-loading behavior in stripping duty. Investigate the isotherm basis before celebrating. Sustained values well above design deserve the same scrutiny as sustained losses.

Does higher mixer power intensity always raise stage efficiency?
No. Efficiency improves with power intensity only until equilibrium is reached within the residence time — typically somewhere between 1.0 and 1.5 kW/m³ for copper oxime systems. Beyond roughly 2 kW/m³, kinetics plateau while droplets keep getting finer, so the settler floods, entrainment rises, and overall stage performance falls. More power helps only when testwork shows the stage is kinetic-limited, not hydraulic-limited.

How does phase continuity affect mixer settler efficiency?
Continuity decides which phase disperses into droplets, and droplets are where mass transfer happens. In extraction, organic-continuous operation gives stable droplets and cleaner raffinate; aqueous-continuous stripping gives different, equally valid behavior. A continuity flip — usually caused by O/A ratio drift — changes droplet size, interfacial area, and settling behavior simultaneously, and typically costs a stage 5–10 efficiency points until it is corrected.

How many stages do I need at 90% efficiency versus 92%?
Divide your theoretical stage count by the efficiency. A flowsheet needing 3.6 theoretical stages requires 4 physical stages at 90% (3.6 ÷ 0.90 = 4.0) and 4 at 92% (3.9) — the same tank count. But at 85% it needs 4.3, which rounds to 5. The practical rule: near 92%, small efficiency changes are absorbed by rounding; below about 87%, they start buying whole extra stages of concrete and agitators.

When should stage efficiency be re-verified by testwork?
Re-run stage efficiency testwork whenever the reagent system changes, when feed composition or temperature shifts materially, after major mixer mechanical work, and at least on an annual cadence for the design file. Efficiency is a property of the specific liquor-impeller-settler combination, so a number borrowed from another plant or from a five-year-old campaign is a hypothesis, not a design basis.

Conclusion

Mixer settler efficiency is the number that converts your equilibrium chemistry into a stage count, your mixer volume, and ultimately your recovery — and treating it as a designed 92% target rather than an assumed constant is what separates lean, robust circuits from padded ones. Measure it in a pilot stage, deliver it with disciplined power intensity and residence time, guard it with a healthy dispersion band, and re-verify it whenever the chemistry moves. If you are sizing a new train or chasing lost points on an existing one, our engineers can review your isotherm, testwork, and mixer duties and request a technical proposal for a circuit designed to the efficiency you actually need.

Related reading: browse the full Lexin-tech resource hub for guides across every hydrometallurgical unit operation.

References

1. B.A. Wills and J.A. Finch, “Wills’ Mineral Processing Technology”, 8th edition, Elsevier.
2. T.C. Lo, M.H.I. Baird, and C. Hanson (eds.), “Handbook of Solvent Extraction”, John Wiley & Sons.
3. G.M. Ritcey, “Solvent Extraction: Principles and Applications to Process Metallurgy”, revised 2nd edition, published by the author.
4. Society for Mining, Metallurgy & Exploration (SME), “Mineral Processing Plant Design, Practice, and Control”, SME.
5. “Perry’s Chemical Engineers’ Handbook”, 9th edition, McGraw-Hill.
6. U.S. Geological Survey, “Mineral Commodity Summaries”, USGS, latest edition. USGS Mineral Commodity Summaries
7. “Hydrometallurgy” (journal), Elsevier — stage efficiency and mixer-settler scale-up literature.

Changelog

2026-10-28 — First published.