Turning rocks into riches with smart sorting that cuts waste, boosts recovery and powers the next generation of leaner, cleaner mining

As ore grades decline and sustainability pressures rise, mining operations are being forced to find new ways to optimise resource extraction. One approach that’s gaining traction across the globe is sensor-based ore sorting - a technique that uses real-time imaging and physics-based detection to identify and remove waste before it reaches the processing plant.

At the forefront of this shift is Jordan Rutledge, an area sales manager based in Perth. With a background in extractive metallurgy and over six years working across TOMRA Mining’s global offices, Jordan has seen the evolution of ore sorting from a niche tool to a mainstream solution.

“Sorting isn’t about fixing recovery losses,” Jordan explains. “It’s about changing the whole equation - recovering more metal, using fewer resources, and improving how we design and operate plants from the outset.”

Getting the most from the rock

Sensor-based ore sorting, often installed between primary crushing and grinding, works by scanning particles on a conveyor belt and separating them based on measurable differences like atomic density, physical appearance or reflectivity. Using technologies like X-ray transmission (XRT), laser or colour sensors, systems can distinguish between ore and waste with a high degree of accuracy.

Jordan says the value of sorting becomes apparent when it’s integrated early in a project:

“If you reject waste before grinding, you’re not just saving energy - you’re reducing chemical use, water demand, wear on equipment, and even tailings volumes.

“You’re also potentially recovering more metal, especially if you’re dealing with declining grades where sorting allows you to process previously uneconomical material.”

This approach has clear environmental and economic implications. By reducing the amount of barren material entering the processing plant, operations can improve feed grades, reduce comminution requirements, and often downsize equipment needs entirely.

A Transformational greenfield example in Saudi Arabia

One example Jordan refers to frequently is the Ma’aden Phosphate project in Saudi Arabia. Located in the remote northern desert, the Wa’ad Al Shamal facility uses XRT sorters to remove high-silica waste rock - mostly chert - from phosphate ore at an early stage.

“They’re sorting out about 700,000 tonnes of silica every year before the material hits the flotation plant,” Jordan says. “That allowed the plant to be built about 40 percent smaller than originally planned.”

The savings extend well beyond capital cost. According to project data, energy consumption in the crushing circuit was significantly reduced, while wear and maintenance costs on downstream mills dropped by over USD 35 million per year. Flotation reagent usage was also lowered, cutting chemical costs by an additional USD 4.2 million annually.

But Jordan says the most interesting benefit isn’t always in the numbers. “One thing I find really important is how sorting allows for less selective mining. At Ma’aden, they can now excavate closer to chert-rich zones without worrying about sending low-grade material to the plant, because they know the sorters will handle it. That’s extending the life of mine in a very practical way.”

Tin recovery at altitude: The MINSUR case

In a completely different geological and operational context, the San Rafael tin mine in Peru offers another compelling example. Operated by MINSUR and located at altitudes above 4,500 metres in the Andes, San Rafael is the world’s largest underground tin operation. When faced with declining head grades and a bottleneck in their wet processing plant, the company looked to ore sorting as a way to relieve pressure and improve recovery.

“They fast-tracked that project in just over a year,” says Jordan, “and they paid back the capital investment in four months. That kind of turnaround is rare.”

At San Rafael, the sorters are rejecting sub-economic coarse particles before they reach the concentrator. By removing waste early, MINSUR was able to increase throughput from 2,950 to 3,200 tonnes per day and improve overall tin recovery from 90.5 to 92.5 percent. Perhaps more importantly, around a quarter of the feed now comes from material that previously would have been excluded for falling below the plant’s cut-off grade.

“The takeaway here,” Jordan notes, “is that ore sorting changes how you think about ore and waste. Material that was once uneconomic becomes viable. That shifts the resource base, and in some cases, it can extend mine life or justify further investment.”

Building the world’s largest lithium sorting plant

Back in Western Australia, Pilbara Minerals’ P680 Expansion Project is pushing ore sorting into new territory. Commissioned in mid-2024, it’s currently the largest lithium ore sorting plant in the world. The facility treats spodumene ore at over 1,000 tonnes per hour using a combination of XRT and colour-based sorting systems.

“Pilbara had a very clear goal - to remove contamination early, so their wet plant wouldn’t waste energy or reagents processing barren rock,” says Jordan. “That’s exactly what the sorting plant is achieving.”

The plant uses ten sorters in total, configured across fines, mid-size, and coarse particle streams. The early rejection of waste has helped reduce energy consumption by an estimated 8 to 15 gigawatt-hours annually, while also stabilising product quality and improving recovery downstream.

Jordan credits the success of the project to long-term planning and close collaboration. “This wasn’t an afterthought - they tested the ore back in 2017, designed the system based on real data, and partnered closely across teams to get it running. It shows what’s possible when sorting is built into the process, not bolted on at the end.”

The role of AI and data in ore sorting

Beyond the hardware, Jordan sees exciting developments in how data and artificial intelligence are improving sorter performance.

“We’ve been using AI in our systems for over a decade,” she explains, “but now we’re moving into deep learning with tools like CONTAIN^TM. That’s helping us identify inclusion-based mineralisation - things like tin or gold trapped in dense zones within rock - that older systems might miss.”

Another software tool, OBTAIN^TM, helps the system digitally separate touching particles on the belt to ensure each rock is assessed independently. This is particularly useful at high throughput where rocks crowd together.

“These tools are allowing us to look at each particle more accurately,” says Jordan, “and they also help generate real-time operational data. That data can flag feeding issues, monitor ore grade trends, and inform upstream or downstream decisions. It’s a much more integrated approach to mineral processing.”

Looking ahead

As environmental constraints tighten and resource complexity increases, Jordan believes ore sorting will become a core component of mineral processing strategy - not just a niche technology.

“There’s a growing recognition that sorting supports both economic and environmental goals,” she says. “You’re using less energy, generating less tailings, and getting more out of your deposit.”

And scale is no longer a limiting factor. “We’re working on larger, faster systems that can handle broader size ranges,” she says. “There’s no reason this can’t be part of Tier 1 operations.”

From phosphate in the desert to lithium in the Pilbara and tin at altitude, sensor-based ore sorting is proving itself across diverse geologies. As Jordan puts it: “It’s not about replacing traditional processing - it’s about rethinking how we approach it, starting with the rock itself.”