Ore sorting technology has moved from pilot curiosity to mainstream consideration. Multiple operations across Australia now run production-scale sorting systems. Equipment manufacturers have expanded offerings.

The promise is compelling: reject waste before it enters processing, reduce energy and water consumption, improve head grades. The reality, as always, is more nuanced.

What Ore Sorting Actually Does

Sensor-based ore sorting examines individual rocks or small fractions on a conveyor belt. Sensors detect properties correlated with ore value. Air jets or mechanical diverters separate ore from waste.

Common sensing technologies include:

• X-ray transmission (XRT) - density differences
• X-ray fluorescence (XRF) - elemental composition
• Near-infrared (NIR) - mineral identification
• Laser-induced breakdown spectroscopy (LIBS) - elemental analysis
• Colour and texture imaging

Different ore types suit different sensors. Sulphide copper ores respond well to XRT. Gold ores can be challenging without specific indicator minerals. Rare earths often require XRF or LIBS.

Where Ore Sorting Fits in the Flowsheet

Most sorting installations sit after primary crushing and before milling. The goal is to remove waste rock early, before expending energy grinding it.

Feed size matters. Current technology handles material from about 20mm to 150mm effectively. Below 20mm, particle spacing on the belt becomes challenging. Above 150mm, the physics of diversion gets difficult.

This means sorted material still needs secondary crushing and milling. Sorting doesn’t replace conventional processing—it optimises feed to it.

The Economic Case

Sorting creates value through several mechanisms:

Reduced processing costs. Less material through mills means less energy, water, and wear. If you reject 30% of feed as waste, you’ve cut those costs proportionally.

Improved grades. Concentrating ore before milling increases head grade. Higher grades typically mean better metallurgical recoveries.

Extended mine life. Material previously considered sub-economic might become viable with sorting. Lower cutoff grades expand the resource.

Deferred capital. If sorting allows an existing mill to handle a larger orebody, you’ve potentially avoided a mill expansion.

Against these benefits, count the capital cost (typically $10-30 million for a production-scale system), operating costs (mainly maintenance and labour), and metal losses (not all ore makes it to the accept stream).

What Makes Ore Sortable

Not every ore is suitable. Good sorting candidates have:

Distinct property differences between ore and waste. If the sensor can’t reliably distinguish them, sorting won’t work. This requires testwork, not assumptions.

Liberation at sortable sizes. If valuable minerals are finely disseminated through host rock, there’s no waste to reject. Sorting works best when ore and waste occur in distinct particles.

Consistent feed. Sorting systems need stable feed rates and size distributions. Highly variable material challenges the system.

Sufficient grade differential. The economic value of sorting depends on rejecting material significantly lower grade than the accept stream. If everything is similar grade, there’s nothing to sort.

Implementation Lessons

Operations that have deployed sorting successfully share common approaches:

Extensive testwork first. Laboratory sorting tests using actual samples from across the orebody. Not just one zone—the full range of material the operation will handle.

Pilot before production. Many operations run pilot-scale systems (processing a sidestream) before committing to production scale. This reveals problems at lower cost.

Operator buy-in. Sorting systems require operational attention. If plant operators don’t understand and accept the technology, it often runs poorly.

Realistic expectations. Sorting improves economics at the margin. It doesn’t transform uneconomic projects into bonanzas.

Where It’s Struggling

Certain applications have proven difficult:

Very fine-grained ores. The physics of sorting below about 15mm remain challenging. Fines-heavy deposits often can’t sort their full feed.

Complex polymetallic ores. When multiple valuable elements distribute differently, defining “ore” versus “waste” becomes complicated.

High-variability deposits. Orebodies where properties change rapidly across short distances challenge sorting system calibration.

Low-margin operations. The capital investment doesn’t pay back quickly enough if per-tonne margins are already thin.

The Australian Context

Australia has several production sorting operations, predominantly in iron ore, base metals, and gold.

Iron ore sorting in the Pilbara removes silica from hematite, improving product grades for low-grade deposits. Roy Hill and others have published results.

Base metal operations use sorting for pre-concentration, particularly where mineralisation is coarse and variable.

Gold sorting has had mixed results. Where indicator minerals like pyrite correlate well with gold, sorting works. Where gold is free and randomly distributed, the correlation breaks down.

Getting Started

If you’re considering ore sorting:

1. Commission testwork. Representative samples across your orebody, tested with relevant sensor technologies.

2. Build the economic model. What mass rejection rate is realistic? What value does that create given your specific costs?

3. Assess site constraints. Do you have space for a sorting facility? Can your crushing circuit deliver appropriate feed?

4. Consider a pilot. For significant capital decisions, real-world pilot data beats testwork predictions.

5. Plan for integration. Sorting isn’t standalone—it affects everything downstream. Milling rates, reagent consumption, tailings volumes all change.

Ore sorting isn’t universally applicable, but where it fits, it can materially improve project economics. The key is rigorous assessment before commitment.