Refractory Gold Ore Processing: Why It Resists Leaching

What makes gold ore refractory?

Refractory gold ore processing starts with a hard truth: some gold ores refuse to give up their gold to cyanide. You'll see three main culprits in the mineralogy. Sulphide encapsulation locks gold inside pyrite or arsenopyrite crystals, so cyanide ions never reach the metal. Preg-robbing carbon in the ore adsorbs dissolved gold from the pregnant solution, pulling it back out of solution. Tellurides are gold compounds that cyanide dissolves only very slowly. According to the USGS Mineral Commodity Summaries, gold in sulphide-bearing ores often needs more aggressive processing than free-milling oxide ore.

If you've worked with a cyanide leaching circuit, you know that a standard oxide ore can often yield high recovery in a bottle roll. Refractory ore doesn't. It might leach only a fraction of the gold, because the gold is physically shielded or chemically re-adsorbed. That's the precipitating event for every refractory project.

The financial consequence is direct: you're spending reagent and tank time on gold that never reports to the electrowinning cell. That's why owners often ask whether a refractory processing route can replace a simple CIP plant. The answer depends on testwork, not guesswork.

Diagnostic leach tests identify the resistance mechanism

Before you commit to any pretreatment, you need to know exactly why your ore is refractory. Diagnostic leach testwork is the standard way to do that. The sequence usually runs like this:

1. Head grade and mineralogy: assay the sample and run quantitative mineralogy to see how gold is locked.
2. Direct cyanidation bottle roll: the baseline. A low extraction signals refractory behaviour.
3. Gravity concentration test: check whether free gold can be recovered without cyanide.
4. Acid wash or carbon-in-leach test: if extraction jumps after removing carbonaceous matter, you have preg-robbing.
5. Oxidation test: if extraction improves after roasting or pressure oxidation, sulphide encapsulation is probable.

That ordered sequence gives you a decision tree instead of a guess. A plant running 1,200 t/d without a diagnostic programme often spends more on reagents and still loses gold. According to Xinhai's self-reported data, its CNAS-accredited laboratory performs about 5,000 elemental analyses per month and about 200 beneficiation testwork programmes per year, so testwork capacity is rarely the bottleneck. If you're designing a gold recovery flowsheet, start with this sequence, not with a vendor's preferred pretreatment.

Ultrafine grinding: expose the locked gold mechanically

Ultrafine grinding doesn't change the chemistry; it changes the physics. You grind the ore down to very fine particle sizes so that gold locked inside sulphide grains becomes exposed to cyanide. There's no oxidation, so the capital cost is lower than roasting or pressure oxidation, but power consumption rises sharply. If your ore is only mildly refractory, a regrind circuit might be enough. Many engineers compare the economics against a gold milling plant that already has spare grinding capacity.

One practical limit is that ultrafine grinding generates heat and consumes energy out of proportion to the extra recovery. You'll need to test the exact grind size against the recovery curve. Sometimes a moderate regrind plus longer leach time beats an extreme regrind. Don't skip the grind-recovery testwork; it's the cheapest optimisation you'll do.

Ultrafine grinding also helps with preg-robbing ores by exposing more gold for activated carbon competition, but it won't destroy the carbon. For that, you need oxidation or thermal treatment.

Roasting: burn off the sulphide and carbon

Roasting oxidises sulphide minerals to porous iron oxide, releasing the gold for cyanidation. It also destroys preg-robbing carbon in the same step. The trade-off is gas handling: roasting produces sulphur dioxide, which must be captured and converted to sulphuric acid or neutralised. You'll also need to control mercury and arsenic emissions. Roasting works well on ores with high sulphide sulphur, but it's thermally intensive and often requires a dedicated acid plant. For an existing operation, a gold recovery flowsheet may be retrofitted, but the gas cleaning system adds significant capital.

Roasting isn't a single recipe. Whole-ore roasting, partial roasting, and two-stage roasting each suit different sulphur and carbon contents. Testwork tells you which gas composition and temperature profile gives good gold liberation without over-roasting and forming refractory ferrites.

If your ore contains arsenic, roasting also volatilises arsenic trioxide, which is a serious occupational and environmental hazard. That alone often pushes a project toward pressure oxidation or bio-oxidation instead.

Pressure oxidation: aqueous oxidation at high temperature and pressure

Pressure oxidation (POX) runs in an autoclave at elevated temperature and pressure with oxygen. It oxidises sulphide minerals quickly, and the gold remains in the residue for cyanidation. POX handles highly refractory ores and achieves high gold recovery, but it's the most capital-intensive pretreatment route. You'll need exotic materials, oxygen supply, and skilled operators. The payback is often justified for ores with high gold grades or high sulphide content. According to the International Cyanide Management Code, cyanide handling downstream of POX requires the same environmental controls as any cyanide facility, which adds operating discipline.

POX is not a single autoclave size. Plant throughput, oxygen partial pressure, and residence time all scale with the sulphide sulphur grade. You'll run pilot tests before committing to a full autoclave, because the capital risk is too high for guesswork.

Some POX circuits use a two-stage process: a low-temperature oxidation for arsenopyrite and a high-temperature oxidation for pyrite. That flexibility is valuable, but it also increases instrumentation and control requirements.

Bio-oxidation: let bacteria do the work

Bio-oxidation uses acidophilic bacteria to oxidise sulphide minerals at ambient temperature and pressure. It's slower than roasting or POX, taking days rather than hours, but the energy cost is much lower. Bio-oxidation works well on arsenopyrite and pyrite ores with moderate sulphide content. You'll need a large tank farm, pH control, and a steady feed of nutrients. The bacteria also generate acid, so neutralisation is part of the circuit. If your ore responds to bio-oxidation in testwork, you can avoid the high-pressure equipment and gas handling of the other routes.

One subtlety is that bio-oxidation is a biological process, not a chemical reagent. You need to maintain a healthy bacterial population, which means stable temperature, pH, and dissolved oxygen. If the feed changes, the bugs can slow down. That's why bio-oxidation testwork runs for months, not weeks.

For ores with very high sulphide sulphur, bio-oxidation becomes impractical because the acid and heat generated can kill the bacteria. In those cases, roasting or POX is the safer engineering choice.

Cost and complexity: choosing among the four options

Each pretreatment route adds both capital and operating cost. Ultrafine grinding is often the cheapest to install but the most power-hungry per tonne. Roasting sits in the middle on capital, but gas handling adds operating complexity and environmental risk. Pressure oxidation has the highest capital, but it's fast and handles the widest range of refractory ores. Bio-oxidation has low operating cost but a larger footprint and longer retention time. According to Xinhai's published figures, the company has completed 600+ Mine EPC+M+O projects across 100+ countries, and its CIL/CIP circuits can reach 99% gold recovery on suitable ores. That experience matters when you're weighing testwork results against a full-scale plant.

If your ore responds to conventional cyanidation after oxidation, a gold CIL plant handles the downstream leaching efficiently. The key is to size the pretreatment for the actual ore, not for a generic flowsheet.

Many owners ask whether a modular plant can include pretreatment. The answer is yes for ultrafine grinding, and sometimes for bio-oxidation, but roasting and pressure oxidation are almost always site-built because of their gas handling and autoclave foundations. That's a capital-planning consideration you'll want on the table early.

Frequently asked questions