Five major projects, five very different routes from rock to gold. The difference starts with what the ore will—and will not—give up.
Gold plants are often compared by their size and equipment, but the more interesting story is in the ore. Where the gold sits, what surrounds it, and how readily it can be freed determine the route from mine to final product.
Consider two extreme cases. At one project, fractured rock can be stacked on a heap almost as it comes from the mine. At another, most of the gold sits within arsenopyrite, where cyanide cannot readily reach it until the sulphide has been broken down. A conventional gold plant cannot solve both problems in the same way.
That leaves a practical question at every stage: where does the gold go next? It may enter a gravity concentrate, a copper concentrate, a leach solution, or a tailings stream that needs more treatment. The five proposed plants below show how mineralogy and testwork determine those routes. Their daily capacities give a sense of scale, but none of the quoted recovery figures is an achieved operating result.
Black Pine: leave the rock whole, control the heap
Black Pine’s gold sits largely in oxidised carbonate rock cut by fractures and zones of decalcification. Those natural openings matter. Leach solution can reach much of the gold without the mine first building a crushing and grinding plant. Tests found that making most samples substantially finer brought only modest extra extraction, although one sample was an exception. Harder, less-fractured dolomitic material will need more attention to fragmentation during mining.
So the proposed route begins with run-of-mine ore truck-stacked directly on a heap, at an average of about 74,200 tonnes a day. Cyanide solution percolates through the rock; gold-bearing solution is collected and passed through carbon adsorption, stripping and electrowinning before being refined into doré. There is no mill between the pit and the pad.
That apparent simplicity shifts much of the engineering into heap management. Clay-rich ore can obstruct flow, while blocky ore can let solution channel past gold. Material will be blended and the stacked lifts ripped to improve contact. More importantly, ore is divided into metallurgical classes according to how much of its gold dissolves in cyanide. The more readily leached material occupies most of the pad; lower-response material is placed separately. Lower-grade circulating solution is kept away from zones where dissolved gold could be lost through preg-robbing. Here the geology influences not only recovery, but the shape of the heap and the routing of its pipes.
The estimated life-of-mine gold recovery is 70.2%. Large-scale tests of run-of-mine material were still in progress when the study was completed, so that estimate should not be confused with performance demonstrated on an operating heap.
Donlin: float the sulphides, then break them open
At Donlin, roughly 80–90% of the gold is associated with arsenopyrite, much of it within the mineral’s crystal structure. Free gold makes up less than 1%. Direct cyanidation of unoxidised material recovered only 5–30% in tests. The challenge is therefore not just to expose more surface by grinding: the gold-bearing sulphides themselves must be oxidised before leaching.
The proposed plant first crushes the ore and grinds it in a semi-autogenous (SAG) mill and a ball mill. After a first flotation stage removes accessible sulphides, its tailings are ground more finely and floated again. Tests of this staged arrangement improved gold recovery by about 1.8 percentage points over the comparison circuit at the same concentrate sulphur grade. This sequence avoids overgrinding sulphides recovered early while giving the harder-to-liberate portion another chance to float.
The two flotation stages produce a sulphide concentrate. Only that concentrate is sent through pressure-oxidation autoclaves, where the minerals are broken down; it is then washed, neutralised and treated by carbon-in-leach to recover gold. At an average mill rate of 53,500 tonnes a day, concentrating the sulphides before oxidation is a fundamental part of making the route workable. Some flotation tailings also provide carbonate for neutralising the acidic oxidation products.
The ore’s gold deportment explains the flotation–oxidation–leaching sequence. It does not, by itself, require a SAG mill: the grinding arrangement was also chosen for cost, clay handling, climate and operating flexibility. Nor is this an entirely new circuit; much of the underlying testwork and design work was done years earlier.
Troilus: catch free gold and float the copper-gold minerals
Troilus has a different split to manage. Some gold occurs as native gold or electrum and can be recovered by gravity; another portion travels with copper-bearing sulphides into a flotation concentrate. Tests support grinding the main ore stream to about 75 micrometres P80, then regrinding the much smaller concentrate stream to about 20 micrometres P80. P80 means 80% of the material passes that size.
The proposed 50,000-tonne-a-day plant crushes ore, passes it through high-pressure grinding rolls (HPGRs), and feeds two ball-mill lines. Gravity units recover free gold within the milling circuit before copper-gold flotation. The flotation concentrate is reground, given another gravity opportunity and cleaned into a saleable copper concentrate with gold credits. Gravity concentrates are smelted into doré. The low-grade flotation tailings are thickened for storage rather than sent to a cyanide-leach plant; further leaching was judged uneconomic.
Ore variability still affects how robust that design must be. The Southwest Zone supplies much of the early feed, and additional tests on it raised the adopted ball-mill hardness basis from 13.8 to 16.2 kWh/t. That is a grindability measure used to size and assess milling duty, not a whole-plant power bill. Examinations of cleaner tailings also found free gold, including flattened grains that may have circulated through regrinding. This makes recovery from the regrind stream a sensible focus, although further improvement has not been quantified by a completed pilot programme.
Whistler: recover copper first, then leach its tailings
Whistler’s copper is overwhelmingly in chalcopyrite. In a tested ore composite, about 95% of the copper occurred in that mineral, but at a grind of 159 micrometres P80 only 55% of its copper sulphides were liberated. The plant therefore cannot make a clean copper product simply by floating a coarse grind. It needs a finer step—but only for the material that becomes copper concentrate.
The planned 40,000-tonne-a-day circuit crushes and HPGR-treats the ore, then ball-mills it to about 120 micrometres P80. Copper rougher flotation splits the flow. On one branch, the concentrate is reground to roughly 15 micrometres, cleaned and filtered for sale. On the other, the copper rougher tailings are thickened to remove residual flotation reagents, then leached with cyanide. Carbon-in-pulp adsorption, elution and electrowinning recover that gold as doré.
Why build a second recovery route for tailings? Leach tests showed that meaningful gold remains after copper flotation, even though different mineral fractions respond very differently. The proposed leach gives the rougher tailings 48 hours of residence time, matching the duration used in the key tests. A separate pyrite-flotation stage was dropped in favour of leaching the copper rougher tailings directly. This is still an early-stage design; its modelled recoveries need the scrutiny that comes with further development.
Cerro del Gallo: make a finer heap feed, then recover the cyanide
Cerro del Gallo’s gold is very fine and occurs in quartz veins and sulphide inclusions. Weathering creates oxide, mixed and fresh-sulphide ore types that do not leach alike. Gravity performed poorly, while flotation tests on weathered and mixed-oxide material gave disappointing results. Column leaching provided the stronger route, but the way the ore is crushed matters, particularly for mixed and sulphide material.
Unlike Black Pine, Cerro del Gallo is designed to fine-crush its heap feed. Primary and secondary crushers are followed by HPGRs, producing material around 5 millimetres P80. Cement agglomeration then helps maintain permeability before the ore is conveyor-stacked in lifts. The planned 16,667-tonne-a-day crushing and stacking operation leaches each lift for 120 days: a higher solution application rate for 40 days, followed by a lower rate for 80. The projected field gold recovery differs by domain—about 74% for oxide material and 53% for fresh sulphide—rather than following one mine-wide number.
The ore also contains cyanide-soluble copper. As copper dissolves, it ties up cyanide and complicates gold recovery from solution. The proposed answer is a SART circuit: it precipitates copper and silver as a saleable product and releases cyanide for reuse. Only then does the solution go to carbon adsorption and the gold-recovery circuit. Thus the ore dictates two plant decisions at once: finer preparation of the solid heap feed and treatment of the liquid leaving the heap. The crushing tests show a benefit under some conditions, not a universal HPGR recovery uplift for every ore type.
The design lesson
These plants differ most in where they split the flow. Black Pine leaves the rock uncrushed but separates heap zones and circulating solutions. Donlin concentrates sulphides so a smaller stream receives pressure oxidation. Troilus removes free gold by gravity and sells a copper-gold concentrate. Whistler follows copper flotation with a separate gold-leach path for the tailings. Cerro del Gallo treats the whole ore on a prepared heap, then separates copper and recycles cyanide before recovering gold.
That is the practical link between geology and plant design: first establish where the valuable metal sits and how each ore type responds; then decide which stream needs the next treatment. Throughput and equipment names matter, but they make sense only after that question is answered. All five are proposed designs, and test extractions or projected recoveries should not be read as operating guarantees.
Technical reports consulted
These are five large projects with reports published in 2026, not an exhaustive ranking of the world’s largest gold plants. Daily rates reflect each study’s planning basis and are not directly comparable operating achievements.