An ore sorting sensor that uses landmark radio frequency research could be a cost-effective innovation. CSIRO radio frequency researchers have used a magnetic resonance technique to develop a sorting sensor that works by ‘listening in’ to bulk ore as it passes along a conveyor belt.The sensor, believed to be the first of its type in the world, can rapidly measure and distinguish the quality or grade of minerals by detecting their crystalline structure. The resonance technique is reported to provide accurate and clear mineral discrimination and measures it directly. The online, real-time measurement capability enables low grade batches of ore to be discarded, making mining operations significantly more efficient and sustainable.

The sensor encircles an appropriate feed conveyor. Importantly, it can work at very high throughputs.

The sensor measures consecutive short sections or ‘batches’ of ore on a primary conveyor, each with a mass of about 0.5 t. Low grade batches can be diverted using simple traditional technologies such as conveyor flop gates and splitters.

By discarding the lower-grade ore at the front of the process, CSIRO estimates efficiency could be improved by at least 20%. Deposits previously considered uneconomic could become viable through this increased productivity. The sensing occurs deep inside the bulk ore, which avoids the laborious job of separating individual rocks before measurement.

The sensor can be retrofitted to existing primary conveyors, significantly reducing the capital cost of installing the sorting infrastructure.

CSIRO radiophysicist Dr David Miljak and his team at Lucas Heights (Sydney) have built a laboratory prototype of the sensor. Newcrest Mining funded the construction and deployment of the first field test prototype, which is scheduled for installation at an Australian mine by the end of 2012.

The prototype has a large aperture, suited to measure ore depth of up to 400 mm. It is configured to detect chalcopyrite, a dominant copper ore mineral in its own right that can also be used as a ‘tracer mineral’ for gold.

The sensor exploits the natural magnetic properties of chalcopyrite. The nuclei of copper atoms in the mineral act like tiny magnets, which tend to line up. The sensor blasts half-tonne batches of the ore with short pulses of radio waves, momentarily exceeding the power output of a typical radio station. These blasts push the nuclei out of alignment. When the nuclei relax to their original condition, they emit radio waves. The strength of the signal reveals how much chalcopyrite is in the ore.

The mineral response, which is 15 orders of magnitude weaker in power, is detected after a short time by an advanced radio receiver.

In trials last year, signals were obtained from small samples in the laboratory prototype, validating the fundamental measurement concept.

Miljak says the technology is related to medical MRI – the non-invasive ‘magnetic resonance imaging’ technology used to image the body. “We use radio waves at specific frequencies that penetrate deep into rock,” he says. “Advanced radio receivers listen for faint electronic ‘blips’ that allow us to obtain accurate information about the mineral mix.”

“A lot of effort involving sampling and detailed block models goes into understanding ore deposits, but no information has been rapidly available at the scale of a few tonnes,” he says.

The sensor could help develop a management system that makes decisions online – and quickly.

Miljak leads the CSIRO team that developed the technology and that took out the minerals and energy category of the inaugural Innovation Challenge awards last year. The awards are run by The Australian newspaper in association with Shell and supported by the Department of Innovation, Industry, Science, Research and Tertiary Education.

He says the breadth of experience and knowledge of his four-strong team – including physics, electronics, electrical engineering, minerals and mining practice – together with close industry ties, drove the technology development.

The challenges of developing the prototype to operate at mines are formidable. The method must be accurate enough to use in unforgiving mining environments and large engineering scale-up is needed. Most ‘zero field’ magnetic resonance experiments are performed in solid state physics research with sample sizes of a few cubic centimetres.

The prototype has a sensing volume 100,000 times that size. Miljak and his team have been working on the technology for about six years, but it has been only recently that they have begun to “think big.”

The Innovation Challenge win convinced them to start considering new ways of developing the technology and the sensor. Other potential uses have been identified in iron ore mining.