Gold ore gravity plant introduction

This article will describe units, circuits and strategies used to recover gold that is either liberated or present in very high-grade gold particles, referred to as gravity-recoverable gold, (GRG), as well as gold present in much lower grades in sulfidic particles, typically pyrite and arsenopyrite, that can in their own right be recovered by gravity. These particles will be referred to as gold carriers.

Recovery strategies for GRG and gold carriers vary, as only GRG can be substantially recovered at very low-weight recovery into the concentrate, or yield (less than 0.1%), typical of the type of semi-continuous units used today, whereas gold carriers such as pyrite and arsenopyrite are recovered by continuous units capable of much higher yields, which typically match or slightly exceed the sulfide content of the stream treated. Gravity separation has been utilized in gold plants as the primary recovery mechanism or alternatively ahead of other downstream processes such as flotation and cyanidation since the inception of mineral processing.

The recovery of free and sulfide (pyrite, arsenopyrite and telluride)-associated gold from the primary grinding circuit featured in all these installations. Virtually, almost every gold mine incorporated gravity recovery in the primary grinding circuit.

The treatment of the concentrates ranged from full gravity via tables through amalgamation and in many cases cyanidation was used to recover gold from the concentrates. The use of amalgamation was featured in many mines but has since been phased out due to health and environmental issues. Until recently, the only common option was the shaking table, despite its lower efficiency. Rotating devices are also used in a very limited number of plants. Intensive cyanidation, notwithstanding its higher recoveries, never achieved a high degree of acceptance, possibly because of the lack of a commercial unit.

Issues such as the poor operability, security and maintenance of these circuits combined with rapid advances and the elegance of the carbon-in-pulp (CIP) and carbon-in-leach (CIL) process, capable of achieving very high recoveries, saw a reduction in reliance on gravity as a primary means of concentration. This was amplified by the move towards simplified, low capital plants with low manning levels and automated processes. This drove down operating costs, which in turn made possible the treatment of lower grade ores.

However, some orebodies have been found to have attributes that do not lend themselves to high recovery through the direct cyanidation route. Coarse free gold and gold associated with complex sulfide minerals tend to complicate the cyanidation process. Coarse gold increases the residence time required to achieve high recoveries by cyanidation. Complex metallurgy can cause coatings on gold that render it impervious to cyanidation, while other forms of gold such as gold locked in a sulfide lattice as solid-solution gold or attached to a sulfide particle can report to the tails stream of a conventional cyanidation plant. These problems are generally amplified for the coarse grinds normally associated with low-grade ores. A better understanding of these problems and the development of larger, more reliable gravity units, as well as intensive cyanidation, have heralded a return to gravity recovery.

Gold recovery in comminution circuits

Hydrocyclones demonstrate a trait that is useful for gold mills – minerals with higher specific gravities having a progressively finer cut-size (whether this is desired or not). In plants with free-milling gold, this leads to gold concentrating in the milling circuits (see attached). On a size-by-size basis, concentrations of up to 100 times can occur in cyclone underflow. In this respect, hydrocyclones are outstanding rougher concentrators. Many operations take advantage of this effect by installing gravity-concentration units in the grinding circuit. Treating a split of the cyclone underflow is perhaps the most common, but ball-mill discharge and cyclone feed streams have also been treated.

There are a number of considerations for gravity circuits fed by bleed streams. First is the effect on the overall circuit water balance. As the gravity tailing (which represents nearly 100% of the feed mass) is typically returned to the ball-mill discharge sump, any water addition in the gravity circuit will affect the grinding circuit’s overall water balance. In most cases, the water balance can be adjusted with other streams in the ball-mill circuit, but the overall water balance should be carefully considered.

Most circuits treat cyclone underflow, but ball-mill discharge is also sometimes used as gravity-circuit feed. In either case, feed to the gravity circuit can be pre-concentrated to a certain extent. For cyclone underflow streams, weirs can be constructed in the underflow tubs. In this case, cyclone underflow is fed to a weired area in the underflow tub, and gravity-circuit feed taken from this area. Slurry excess to the capacity of the gravity circuit overflows back to ball-mill feed. For those circuits treating ball-mill discharge, a sluice can
often be arranged to provide some pre-concentration for the gravity circuit.

Flash flotation is also used in comminution circuits for gold recovery. The application on cyclone underflow is generally similar to integration of a gravity circuit. Milling-in-cyanide is also conducted by many operations. Cyanide and lime are added to the grinding circuit, often with trimming of both pH and cyanide levels during the subsequent leach process. There is relatively little information published concerning leach kinetics, but the addition of cyanide to the milling circuit clearly initiates the leach process earlier. This practice is not recommended for those ores that contain preg-robbers (which necessitates the use of carbon-in-leach (CIL)), or when pre-aeration is required to passivate cyanicides. In the case of cyanicides requiring pre-aeration, adding cyanide to the milling circuit can dramatically increase the total amount of cyanide required.

Gold plant ball mill circuit operation

Very often in SAG circuits, the ball-mill circuit is neglected. In many operations where economics dictate that throughput is worth more economically than the required sacrifice in grind, the focus shifts to throughput so much that the available ball mill power is not used to the fullest extent. Even in those operations where a firm grind target is not adhered to and attainable throughput governs the balance of the circuit, it is foolish not to take full advantage of the installed grinding power. Ores for which recovery is grind insensitive in the range of the typical operation are unusual.

Given that mills are charged to the target ball charge with reasonably sized media, and the feed to the ball mill circuit is not so coarse as to cause constant scatting, the key to efficient ball-mill circuit operation is efficient classifier operation. The standard classifier for ball mill circuits is the hydrocyclone. Ensuring that the finest and most efficient cyclone cut involves selecting the appropriate cyclone configuration for the ranges of grinds that will be encountered. The apex (spigot) size can be manipulated to deliver the maximum underflow density at the target operating conditions, with the vortex finder tailored for the desired product size. With a given configuration, adding the maximum amount of water (subject to cyclone feed-pump limitations, the minimum overflow density, and cyclone pressure) will generally result in attaining the finest possible grind. Employing a control system to maximize water addition to the cyclone feed pump (subject to pump capacities and downstream densities constraints) is often employed successfully to maximize ball mill circuit grind.

There is strong evidence supporting the concept of using a mixed-size make-up ball to attain incremental grinding efficiencies in ball mills. There is little reason to believe that the steady-state media size distribution resulting from the wear rate of the make-up ball size corresponds to the optimum ball size based on the mills’ feed and target grind. In general, a mixed make-up ball-charging regime improves grinding efficiency, with greatest benefits seen for single-stage milling applications with large size reductions.

Nonetheless, most operations tend to use a single-sized make-up ball for reasons of convenience. There is less conclusive evidence for the removal of fine steel from ball charges. While some operations claim an anecdotal benefit from removal of fine steel, unpublished studies by the author indicate a substantial benefit from the use of fine media (less than 12mm) as a supplement to a conventionally sized make-up ball when grinding a gold ore to an 80% size of 75 mm. It is possible that removal of ball chips, which may tend to float due to a shape factor and likely contribute very little to grinding, could result in an improvement in grinding efficiency.

Gold ore plant pebble crushing

The first consideration when discussing pebble crushing is why there is a need for the unit operation. Secondly, the configuration of the overall pebble circuit merits discussion. Pebble crushing, almost a standard for SAG circuits today, was controversial early in the development of autogenous grinding (AG) SAG milling. This was largely due to the fear of failing to efficiently separate grinding steel from the recycle load, with subsequent crusher damage. Magnet and metal detector manufacturers have minimized this difficulty, and today, more SAG circuits are constructed with pebble-crushing circuits than without. Making an efficient steel–magnetite separation, however, remains problematic for some producers.

The need for pebble crushing stems from two factors: a depression in SAG grinding rates at certain particle sizes, and the accumulation of a harder fraction in the mill load. These factors typically result in a mill throughput increase with the installation of pebble crushing that is larger than would be expected purely on the basis of the additional power. Overall, pebble crushing can increase throughput as well as decreasing the total power required to grind to a given size. Typically, pebble crushing also coarsens ball-mill circuit feed, a consideration if a ball-mill circuit is already taxed.

The definition of critically sized material is often misunderstood. Critical size particles are those where the product of the mill feed-size distribution and the mill breakage rates result in a build-up of a size range of material in the mill load; this critical size can be of any dimension. That said, the concept of critical size has become almost synonymous with pebble-crusher feed, and therefore it is typically referenced as the size range of 13–75 mm. Such a definition ignores larger critical-size material that cannot pass the mill grates– such a size often results with very hard ore types that have received insufficient breakage in blasting and primary crushing. Throughput with these ores can benefit from improved blast fragmentation, primary crushing, or SAG pre-crushing. Without such feed-size reduction, however, additional pebble crushing power may be of little benefit, because pebble generations can be quite low.

There are several critical design elements of a successful pebble-crushing circuit, including: material handling/diversion capabilities, metal removal, belt loading, pebble-crusher feeding, and return of crushed material to the circuit.

After classification, the SAG discharge oversize is conveyed to the pebble crushing circuit. The conveyors should have provision for returning the oversize to SAG feed (during pebble-crusher maintenance, for example, or during periods when metal is detected). Additionally, the ability to reject oversize material can also be useful. This is useful for diverting material after metal detects (discussed below), for sampling, or in cases where metallurgical work confirms grade depletion, and allows rejection of the stream to waste. A travelling chute (as opposed to flop gates) to separate the stream between pebble-crusher feed and return to SAG feed offers the greatest flexibility.

The design of an efficient metal-removal system is critical. The risk of inefficient metal removal from the pebble-crusher feed is obvious, and allowing excessive mill balls to the pebble crusher will rapidly damage both crusher manganese and other crusher components. The present industry standard for metal removal is the cross-belt magnet. In design of cross-belt systems, sufficient belt capacity should be designed so that belts can be run with lower volumetric loading. In other words, at a fixed belt size and loading, metal separation is better with relatively faster belt speeds versus relatively higher belt loadings. Designing to a Conveyor Equipment Manufacturers Association (CEMA) belt loadings of 65% or less has worked well at PTFI, with peak loadings of up to 85%. PTFI has used belt speeds up to 3.8 m/s (750 fpm) successfully. As an alternative to cross-belt magnets, manufacturers have recently fielded magnets fitted to remove tramp metal from directly screening oversize as the oversize material is loaded onto belts. After the magnets for steel removal, metal detectors should be installed to detect any metal that bypassed the magnet(s). Such detectors should be upstream of a diverter gate, so that a metal detect results in diversion of the material back to the SAG feed.

For the most efficient operation of a pebble-crusher, provision for a surge bin should be included. The use of a surge bin to allow full-choke feeding improves crusher performance and helps ensure that crusher components wear more evenly. Far steadier operation (in terms of maintaining high power draw without power spiking) can be maintained with a surge bin than without. Perhaps the ultimate ‘surge bin’ is a pebble stockpile with reclaim feeders. In addition to the advantages of surge bins, the use of a stockpile of sufficient capacity can allow the benefit of not having to recycle pebbles back to SAG feed during periods of crusher maintenance, and can allow mill throughput to be maintained at high levels even if the pebble-crushing circuit capacity cannot keep up. The pebble accumulation can then be worked through during periods of increased SAG capacity (due to softer, finer ore, or other reasons). Obviously, the pebble stockpile must stay in balance.

The last major decision for a pebble-crushing circuit is where to put the pebble-crusher product. Conventionally, crusher product was returned to SAG feed. Some designs, however, now allow pebble-crusher product to be returned to the SAG screens/discharge (allowing operation of the pebble crusher in closed circuit) or even to the ball-mill circuit. There is no ‘right’ answer for where the crusher product should be put 100% of the time. Sending the pebble-crusher product to the SAG discharge allows the material to be classified prior to going to the ball-mill circuit, and relieves the SAG mill of loading. Sending the pebble-crusher product back to the SAG circuit allows for attaining a finer SAG circuit product, and can relieve the ball-mill circuit. Sending the pebble-crusher product directly to the ball-mill circuit can reduce SAG discharge screening requirements, but if pebble-crusher product size is not well controlled, ball-mill scatting problems could result. Given sufficient screen capacity, perhaps the best combination is to allow for directing the crusher product either back to the SAG feed, or to the SAG discharge screens.