Inside gold mining’s cyanide detox playbook: three chemistries, one

Inside gold mining’s cyanide detox playbook: three chemistries, one relentless target

As discharge limits tighten to 1.0 mg/L total cyanide and 0.5 mg/L WAD cyanide, gold plants are leaning on three oxidation workhorses — SO2/air, hydrogen peroxide, and Caro’s acid — to turn lethal CN⁻ into far less toxic cyanate and ammonia.

Industry: Gold_&_Copper_Mining | Process: Tailings_Detoxification_&_Disposal

Gold extraction relies on cyanide (free CN⁻ and hydrogen cyanide, HCN) to leach precious metals, leaving tailings with soluble cyanide and metal–cyanide complexes. Discharge is now governed by strict ceilings (World Bank: total cyanide ≤1.0 mg/L, WAD cyanide ≤0.5 mg/L, free cyanide ≤0.1 mg/L; WAD = weak-acid dissociable cyanide) and many countries mirror these thresholds (studylib.net).

The industry’s go-to detox pathways are the INCO SO₂/air process, the hydrogen peroxide (H₂O₂, often Cu‑catalyzed) route, and Caro’s acid (H₂SO₅) oxidation. Each converts CN⁻ to cyanate (OCN⁻), which then hydrolyzes to ammonia (NH₃) and carbonate (CO₃²⁻) over hours, but their catalysts, operating windows, collateral reactions, and costs diverge sharply.

SO₂/air oxidation (INCO process)

The INCO process oxidizes cyanide in slurries by dosing sulfur dioxide (SO₂) and air (O₂) with a copper catalyst at alkaline pH. SO₂ hydrates to bisulfite and sulfite (SO₂ + H₂O → HSO₃⁻ + H⁺ → SO₃²⁻ + 2H⁺), and sulfite is the reducing agent. In the presence of dissolved copper, cyanide (free CN⁻ and metal–CN complexes) is oxidized to cyanate while sulfite is oxidized to sulfate: Cu–CN⁻ + SO₃²⁻ + O₂ → OCN⁻ + SO₄²⁻ (paperzz.com). Typical conditions: pH ~8–9 (maintained with lime), 0.5–2 hours residence, and oxygen mass transfer often the rate limit (paperzz.com) (paperzz.com).

Inefficiencies arise because sulfite can self-oxidize (2 SO₃²⁻ + O₂ → 2 SO₄²⁻) without destroying CN⁻ (paperzz.com). By the end of the reactor train, much copper precipitates as hydroxide and metal–cyanide complexes are destroyed (e.g., Fe–CN is reduced and Cu/Ni/Zn–Fe–CN settles out) (paperzz.com). The primary product, cyanate, hydrolyzes to ammonia and carbonate (relatively benign). Very little thiocyanate (SCN⁻) is oxidized (<10%) in this process (paperzz.com).

Operationally, SO₂/air is robust in slurries and continuous, typically arranged as cascaded stirred tanks with aerators. Unlike some oxidants, it removes “total cyanide” (free plus metal‑complexed) to OCN⁻ (Canadian Mining Journal). Reagent costs are low (SO₂/bisulfite and lime), and the method has dominated “North American gold mines in the last 20 years” (Canadian Mining Journal). Plants rely on accurate reagent addition; dosing is typically handled with precise liquid metering such as dosing pumps.

Achievable performance is strong: well‑designed INCO plants reliably hit WAD cyanide below ~0.5–1.0 mg/L (onemine.org) (docslib.org). The caveat: dissolved O₂ must be sufficient; if oxygen is limiting, CN can persist. A telltale sign is zero residual dissolved O₂ at the reactor exit, indicating more aeration is required (paperzz.com) (paperzz.com). Upsets — loss of sulfite or oxygen — can halt CN oxidation even if sulfite remains (paperzz.com) (paperzz.com). The process does not attack thiocyanate or thiosulfate (S₂O₃²⁻) appreciably, leaving thiosulfate as an oxidant sink.

Literature notes that free CN must be entirely metal‑complexed for SO₂/air to proceed efficiently; any free CN can stall oxidation while sulfite continues burning to sulfate (paperzz.com). Where oxygen and catalyst are adequate, weakly‑complexed CN is essentially fully converted to OCN⁻.

Costs are competitive. Historic operating costs (1980s) ran ~$0.7–1.3 per tonne of ore treated (≈$2–3/tonne in today’s dollars), comparable to alkaline chlorination at the time (studylib.net).

Hydrogen peroxide (H₂O₂/Cu) oxidation

Hydrogen peroxide oxidizes cyanide to cyanate rapidly in the presence of copper catalyst (Cu²⁺). The core reaction is Cu–CN⁻ + H₂O₂ → OCN⁻ + H₂O, with rates rising at higher Cu and H₂O₂ concentrations (paperzz.com). Stoichiometry is roughly 1:1 for free CN, but once CN is gone, excess H₂O₂ decomposes (2 H₂O₂ → 2 H₂O + O₂), boosting dissolved oxygen without adding salts (paperzz.com).

Typical operation is pH 9–10 (faster near ~10) and 30–180 minutes contact. Free CN and copper–CN complexes oxidize quickly; after CN is consumed, Cu precipitates as Cu(OH)₂ (pale yellow) and residual peroxide breaks down, so dissolved O₂ spikes are common (paperzz.com). Iron cyanide complexes also precipitate with Cu or Fe as hydroxides. Thiocyanate is not oxidized by H₂O₂ in practice (paperzz.com).

Peroxide is sensitive to solids — it reacts with particulates — so it is mainly applied to clarified solutions (clarified liquor, mill circulation water, pond return water) rather than raw tailings slurry. Many operations polish recycled water streams with H₂O₂ instead of dosing the slurry, where mineral oxidation would squander peroxide. Clarification ahead of peroxide is commonly achieved with sedimentation; equipment such as a clarifier is used to remove suspended solids.

Performance can be strong. At the Ok Tedi gold mine (Papua New Guinea), handling 22,500 t/d of ore, a Degussa‑designed H₂O₂ plant cut WAD cyanide from ~100–300 mg/L to ≈1–10 mg/L total; free CN fell from ~50–100 to undetectable; easily‑liberatable CN dropped from ~90–200 to <0.5 mg/L (docslib.org). The product is cyanate, later hydrolyzing to ammonia and carbonate; excess H₂O₂ simply yields oxygen, adding no salts (paperzz.com) (docslib.org).

Upside: no gas aeration, no toxic sulfur species added, and effective oxidation of free CN and many metal–CN complexes (Cu, Zn, Cd, Ag), precipitating metals as hydroxides (paperzz.com). Constraints: it does not oxidize thiocyanate or ferri/ferrocyanide complexes such as Fe(CN)₆³⁻ (docslib.org), and without Cu catalyst it reacts very slowly with CN (paperzz.com). Dose control matters: stoichiometric addition often leaves residual CN; plants frequently run two stages — a main oxidation shot and a smaller polishing dose for residual OCN⁻ or CN (paperzz.com).

On costs, peroxide is pricier per unit oxidizing power than sulfite or sulfuric acid, and supply logistics can bite at very large flow. One case demonstrated that switching from H₂O₂ to Caro’s acid cut raw chemical cost by ~60% (onemine.org). Even so, for modest CN loads in water and where solids handling penalizes INCO, peroxide can be economical — as at Ok Tedi, where capital and O&M were justified by meeting free CN <0.05 mg/L at the compliance point (docslib.org).

Caro’s acid (H₂SO₅) oxidation

Caro’s acid, peroxymonosulfuric acid (H₂SO₅), is generated on‑site by reacting concentrated hydrogen peroxide with concentrated sulfuric acid (H₂O₂ + H₂SO₄ → H₂SO₅ + H₂O). Typical ratios are ~0.3–0.6 moles H₂O₂ per mole H₂SO₄; “hot” mixing yields 25–45% H₂SO₅ and “cold” mixing yields 70–80% by weight (paperzz.com). Because H₂SO₅ is unstable, plants generate it in‑line and consume it immediately.

Oxidation of CN is extremely rapid and needs no catalyst: H₂SO₅ + CN⁻ → OCN⁻ + H₂SO₄, with cyanate forming within minutes (paperzz.com). Copper–CN complexes react similarly quickly. Reactions are typically run at slurry pH ~9–10, but the acid addition causes local pH dips; if bulk pH falls too low (risking HCN formation), efficiency drops and OCN⁻ oxidation can accelerate, so tight pH control keeps the system near 9–10 (paperzz.com). Lime is added after mixing to neutralize the H₂SO₄ formed.

Caro’s acid also consumes thiosulfate and thiocyanate, unlike INCO and H₂O₂. Experiments show thiosulfate is usually consumed roughly 3:1 by H₂SO₅, while thiocyanate oxidation is slower but significant given enough oxidant (paperzz.com) (paperzz.com) (paperzz.com). The main product remains cyanate, with sulfate generated as H₂SO₄ is re‑formed; cyanate’s further oxidation is slow at high pH but accelerates if acidified (paperzz.com).

Systems are compact — reaction times are seconds to minutes, so many plants dose directly into the tailings pipeline, then add lime downstream to neutralize and precipitate metals (paperzz.com). Chemical dosing precision matters here as well, and liquid feeds are often metered via dosing pumps for control.

Field performance is compelling: in one gold plant demonstration, injecting Caro’s acid dropped WAD CN from ~40–108 mg/L to below 20 mg/L at a 1:1 H₂SO₅:CN molar ratio, and to ~4 mg/L at 1.8:1 — orders of magnitude faster than H₂O₂ alone (onemine.org). In practice, Caro’s often reaches <1–2 mg/L CN in a single pass for moderate criteria; for ultra‑low targets, additional polishing (e.g., a small H₂O₂ dose post‑Caro’s) can remove residual cyanate or CN. The trade‑off: significant sulfate generation and metal‑hydroxide sludge in final slurries.

Advantages include speed, solids tolerance, and no need for copper; unlike SO₂/air, Caro’s will destroy free CN even with limited Cu (paperzz.com). A transient pH drop also accelerates cyanate hydrolysis. Limitations include handling strong oxidants (concentrated H₂SO₄ and H₂O₂), tight control to avoid HCN formation, and high oxidant demand on sulfide‑rich ores due to thiosulfate/thiocyanate consumption. Though often viewed as high‑cost because of the H₂O₂ content, a plant trial reported ~60% lower raw chemical cost versus an H₂O₂‑only scheme thanks to rapid detox at moderate dose (onemine.org).

Environmental outcomes and by‑products

All three pathways convert CN⁻ to OCN⁻, which slowly hydrolyzes to NH₃ and CO₃²⁻; the resulting small ammonia residual is far less toxic than cyanide. None of these oxidants generate persistent organic toxins or high‑chloride waste. The main residual concerns are elevated sulfate and metal hydroxide sludge (especially after Caro’s acid).

Process comparison and selection factors

Mechanisms converge on CN⁻ → OCN⁻, but catalysts, selectivity, and by‑products differ:

SO₂/air: requires Cu and air; effective for slurries; ineffective on SCN⁻ and thiosulfate; low reagent cost; needs aeration hardware (paperzz.com) (paperzz.com).
H₂O₂/Cu: requires Cu; effective on free CN⁻ and CuCN⁻; ineffective on SCN⁻; generates O₂ with no added salts; good for low‑solids streams (paperzz.com) (docslib.org).
Caro’s acid: no catalyst needed; oxidizes CN⁻, CuCN⁻, SCN⁻, and S₂O₃²⁻; very fast; produces sulfate acid; higher chemical usage (paperzz.com) (onemine.org).

Co‑contaminants matter. Thiosulfate and thiocyanate drive oxidant demand up across the board (sulfite losses in INCO, extra H₂O₂, etc.). Caro’s acid will actively consume them at ~3:1 ratio for thiosulfate, while INCO and H₂O₂ largely leave SCN⁻ intact (paperzz.com).

Feed concentration influences selection. For high initial CN (e.g., >100–200 mg/L WAD), a single‑pass INCO or Caro’s reactor can cut levels to single‑digit mg/L; H₂O₂ would require large doses (with oxygen venting). Where very high throughput is needed, the continuous INCO reactor with high oxygen rates often wins on capacity. For moderate/low CN (e.g., <50 mg/L), any method can work; H₂O₂ or Caro’s often achieve targets with a small footprint.

Final targets can push plants to hybridize. Ultra‑low goals (free CN <0.1 mg/L, total CN <0.5 mg/L) are often met by bulk removal with INCO followed by polishing (Caro’s or H₂O₂). Some operations inject Caro’s acid directly into the tailings line to meet stringent limits. H₂O₂ polishing after INCO is also used to remove residual CN if O₂ was insufficient upstream (paperzz.com).

Cost and operations break down as follows. SO₂/air tends to have lower reagent costs but higher equipment costs (aeration tanks/blowers). H₂O₂ systems have moderate capital but higher reagent cost per kg CN destroyed. Caro’s acid has small capital (mixers/pipelines) but uses costly co‑reagents (H₂O₂ + H₂SO₄). Reported figures: INCO/chlorination in the 1980s cost ~$0.7–1.3 per tonne ore (≈$2–4 updated; studylib.net); Caro’s vs H₂O₂ showed ~60% lower raw chemical cost for Caro’s in one trial (onemine.org). H₂O₂ circuits require frequent replenishment and sometimes catalyst management. Continuous INCO circuits carry high downtime costs if upset, so design robustness is critical (onemine.org) (paperzz.com).

When plants pivot to H₂O₂ on clear streams, compact clarification upgrades such as a lamela settler can increase settling capacity without expanding footprint.

Industry uptake and multi‑step trends

Surveys and field records indicate SO₂/air remains the most widely installed method for large‑scale tailings detoxification in North America (Canadian Mining Journal) (onemine.org). Hydrogen peroxide technologies saw early adoption at Ok Tedi (operational since 1982; 20+ million m³ treated) and continue in water‑treatment roles (docslib.org) (docslib.org). Caro’s acid has fewer full‑scale references but is drawing interest; at least one trial recorded ~60% lower reagent cost compared to peroxide (onemine.org).

As standards tighten, multi‑step circuits are increasingly common. Reviews describe INCO plus oxidant polishing as “critical” to hit the lowest discharge targets (paperzz.com) (onemine.org).

Selection guide: feed, targets, and cost

No single process is universally best. SO₂/air (INCO) is generally first choice for large, particulate‑bearing tailings due to maturity and low operating cost (Canadian Mining Journal). Hydrogen peroxide is favored for polishing solutions or modest flows where oxygen generation is acceptable. Caro’s acid is selected where rapid or deep detox is needed (high cyanide loads, presence of thiocyanate) and handling of strong oxidants is feasible. Evaluators should model reagent stoichiometry and cost for the exact CN species present and pilot if needed to confirm WAD/free CN meet local floors (often 0.1–1.0 mg/L). As one best‑practice note puts it, “the correct reagent selection and supply…are important to an operable process with minimal operator workload” (onemine.org).