Recovery Rates in Magnetic Separation: Realistic Targets and How to Improve Them

Recovery rate is the fraction of valuable mineral that reports to the magnetic concentrate. It looks like a single number on a monthly report, but underneath sit four moving parts: the mineralogy of the feed, the liberation degree after grinding, the separator field and gradient, and the operating discipline of the crew. This guide walks through realistic recovery targets by mineral system, the levers that move them, and what numbers BAS engineers consider achievable in 2026 plants.

Realistic recovery ranges (by mineral system)

Numbers below assume well-liberated feed and properly tuned equipment. Real plants land within these bands; outliers usually mean a flowsheet problem upstream.

Iron ores

• Magnetite, dry drum LIMS on coarse feed: 95–99% recovery, concentrate 60–67% Fe.
• Magnetite, wet drum LIMS on ground feed: 96–99% recovery, concentrate 65–70% Fe with cleaner stage.
• Hematite, WHIMS at 1.0–1.5 T: 70–88% recovery, concentrate 60–65% Fe; dependent on liberation.
• Mixed magnetite-hematite (BIF): combined plant 88–95% Fe recovery with LIMS + WHIMS train.

Industrial minerals (deironing)

• Silica sand, rare-earth roll at 1.2–1.6 T: iron-bearing reject 80–95%; product Fe2O3 brought below 0.05–0.10%.
• Feldspar, WHIMS or rare-earth roll: iron rejection 70–90%; product Fe2O3 below 0.10%.
• Kaolin, WHIMS at 1.5+ T: iron-tinted impurity rejection 70–85% for paper-grade specs.

Other ores and slags

• Chromite (after gravity), WHIMS at 1.0–1.6 T: 75–90% recovery of paramagnetic chromite into final concentrate.
• Manganese ore, WHIMS at 0.8–1.2 T: 70–85% recovery.
• Iron-and-steel slag (coarse magnetic recovery): 85–95% Fe recovery at the magnetic stage; fine WHIMS lifts plant total to 90–97%.
• Aluminum dross beneficiation: metallic aluminum recovery 70–90% depending on shred size and grade.

Recycling and tramp duty

• Ferrous removal from shredded scrap, overband: 85–95% Fe extraction, depending on burden depth and magnet sizing.
• Ferrous polishing on aluminum lines: residual Fe brought below 0.5% of feed mass with basic-type metal separators.
• ECS aluminum recovery (downstream of magnets): 90–98% non-ferrous recovery when magnetic stage is sized correctly.

Why liberation dominates everything else

No magnetic separator can recover a grain that is locked inside gangue. Liberation degree — the fraction of valuable mineral that exists as free, single-mineral particles — sets the ceiling for recovery × grade trade-off:

• At 60% liberation, even perfect equipment gives you ~60% recovery at marketable grade; the rest is locked.
• At 90% liberation, equipment-driven losses dominate; tuning the separator and adding cleaner stages matter.
• At 95%+ liberation, multi-stage circuits push the last percent — diminishing returns vs grinding cost.

Plants that under-perform recovery targets usually have a grinding / classification gap, not a magnet problem. Verify with a sieve-by-sieve magnetic susceptibility analysis before adjusting equipment.

Multi-stage layouts that actually lift recovery

Single-pass magnetic separation rarely hits its ceiling. Standard practice:

1. Rougher stage — wide gap or low-intensity drum recovers most of the mass at relaxed grade target. Tailings go to scavenger.
1. Scavenger stage — same field family, recovers what the rougher missed. Tailings leave the circuit.
1. Cleaner stage(s) — higher selectivity, narrower gap or stronger gradient. Lifts grade of rougher concentrate.
1. Re-cleaner / cleaner-scavenger — final polish for industries with tight specs (BIF iron ore, glass-grade silica).

Adding a cleaner often costs less per unit recovery than upsizing a rougher. Always compare the two options on test data before specifying capex.

Instrumentation and sampling — measure before you tune

Plant teams chase recovery problems for months when the real issue is sampling error. Before blaming magnets, check:

• Cross-stream cutter samplers at every interface; spot-grab samples on slurries are unreliable.
• Mass flow meters on slurries and weightometers on belts; recovery is calculated, not measured directly.
• Magnetic content assays — XRF for total Fe, Davis Tube for magnetic Fe, satmagan for in-line monitoring.
• Particle-size monitor on the concentrate feed — drift in P80 silently changes recovery.
• Daily metallurgical balances — close the mass balance to within 2% before trusting any recovery number.

Practical levers when recovery dips

1. Re-check liberation with a fresh sieve+susceptibility analysis. Re-grind if needed.
1. Tune burden depth down within vendor envelopes; too thick a layer leaks magnetic mineral.
1. Adjust belt speed — slower for selectivity, faster for throughput; rebalance with stages.
1. For wet routes: tune slurry density and rinse water before increasing field current.
1. For overbands and drums: check air gap to belt; magnets dropped 10 mm too far halve the effective field.
1. For permanent magnets: check temperature — neodymium loses flux above ~80 °C operating temperature.
1. Check maintenance items — pegging on matrix, worn belt edges, and bearing slop all eat recovery.

When to call a lab

If recovery sits below catalog expectations after the basic levers are checked, run a representative-sample test program at the BAS solution center. Davis Tube, Frantz, lab WHIMS, and pilot drum tests typically resolve whether the gap is mineralogical (regrind required), equipment-related (upgrade required), or operational (procedure tuning enough). BAS supports plants on iron ore, industrial minerals, slag and recycling streams worldwide.

Frequently Asked Questions

What is a typical magnetite recovery rate?

95–99% with well-liberated feed and properly tuned LIMS. Below 95%, liberation or burden depth is usually the issue. Above 99% is unrealistic on industrial feeds.

Why is hematite recovery lower than magnetite?

Hematite is paramagnetic, so it requires high-intensity (1.0+ T) and is more sensitive to liberation, slime coatings, and slurry density. Realistic recovery sits at 70–88% in WHIMS plants, vs 95–99% for magnetite in LIMS.

How do I improve recovery without losing concentrate grade?

Add a cleaner stage rather than overdriving the rougher. A second high-intensity stage on the rougher tail captures missed mineral without sacrificing rougher grade. Verify on test data first.

Does adding more field strength always increase recovery?

Only up to the susceptibility ceiling of the target mineral. Beyond that, additional field captures gangue (chromite gangue in wolframite, biotite in feldspar). Higher field is not free — it raises power, capex, and selectivity loss.

How long does it take to optimise recovery on a new plant?

Commissioning recovery typically lifts from 80–85% to 92–96% over the first 6–9 months as liberation, burden, and maintenance procedures stabilize. Rushed startup numbers should not be used as the long-run benchmark.