Mining’s new math: use “just enough” leach reagent — and recycle the rest

Leaching is where profit meets chemistry for gold and copper. Globally, mines use roughly 60–70 kt/year of sodium cyanide in gold recovery — about 18% of world cyanide output — and many millions of tonnes of sulfuric acid for copper heaps (www.smenet.org). With NaCN at ~$1–2/kg and H2SO4 at ~$50–100/t, the winning strategy is “just enough”: dose to the minimum effective concentration, hold it there with online control, and recycle reagent from “barren” solution rather than waste it.

Reagent economics and usage baselines

Gold carbon‑in‑pulp and carbon‑in‑leach (CIP/CIL; adsorption processes) typically run dissolved cyanide at 100–500 ppm (0.01–0.05% NaCN) to deliver near‑quantitative dissolution without large excess (www.smenet.org). Drop much below ~100 ppm and gold can be left behind (www.process-analytical.co.za). Copper heaps are irrigated with dilute sulfuric acid, usually 0.5–5% w/w; consumption varies widely with ore mineralogy, from low kilograms per tonne for sulfides to very high where carbonates and iron hydroxides neutralize acid.

Gold leaching: online cyanide control

Modern plants use online cyanide analyzers to measure free (and sometimes “weak” complexed) cyanide in real time, with laboratory‑equivalent accuracy within minutes — OI Analytical’s CNSolution® is a commonly referenced system (www.azomining.com). Those measurements feed the distributed control system (DCS) or supervisory control and data acquisition (SCADA; plant control platforms), which then trims NaCN via a dosing pump. The result is a feedback loop that replaces operator “sawtooth” add‑and‑stop routines.

In one South African plant, manual control produced large swings, sometimes dipping below 100 mg/L — suboptimal for leaching — whereas automated control held cyanide within ±2–5% of setpoint and, during testing, never dropped below 100 ppm (www.process-analytical.co.za; www.process-analytical.co.za). Stability even allowed a lower setpoint margin — less reagent without sacrificing dissolution (www.process-analytical.co.za).

A Burkina Faso study echoed the gains: recalculating dose every 15 minutes from a calibrated online titrator and implementing a ±10 mg/L deadband cut overshoot/undershoot, tightening the cyanide curve to ±10 mg/L versus the prior ±50–100 mg/L swings; deviation correction time fell from 4–8 hours to about 1 hour (www.scirp.org; www.scirp.org). The authors concluded this “improves the variability of the ppm in the CIL” and will “certainly” yield cyanide cost savings (www.scirp.org).

Figures from that study show the setpoint held; trend plots indicate the controlled curve stayed almost flat while the earlier profile had sharp peaks and troughs. In a separate simulation, stable operation averaged 2.04 kg NaCN per tonne of ore (kg/t), while high‑consumption ores drove 30% higher usage — a penalty avoided with careful ore blending and automated feed control; the wrong ore mix caused abrupt 30% jumps in cyanide use with no added recovery (www.mdpi.com; www.mdpi.com).

Consumption metrics and setpoint discipline

Typical CIP/CIL plants target 1–3 kg NaCN per tonne of ore. At Florida Mine (Chile), ore types required ~1.75–2.24 kg/t, with a blended steady‑state ~2.04 kg/t (www.mdpi.com). Holding cyanide at the minimum effective concentration (often 200–300 mg/L) reduces loss; time spent above target or when leachable gold is low is pure waste (www.process-analytical.co.za). One illustration: processing 1,000 t/day at 2.0 kg/t and using only 0.1% less (2.0→1.998 kg/t) “saves ~0.1 t/day (∼35 t/year).”

Regulation, tailings limits, and cyanide recycling

Strict standards, including the International Cyanide Code, limit discharged cyanide; tailings solutions are typically held at ≤10–50 ppm (www.smenet.org). This drives internal recycling. Technologies such as weak acid dissociation/thiocyanate precipitation and SART (Sulfidization‑Acidification‑Recycle‑THickening) can convert copper–cyanide complexes to CuS and NaCN, recovering cyanide قبل final disposal. After carbon elution, hot caustic (NaOH) and steam return Au(CN)2⁻ to metallic gold and leave sodium cyanide in solution, which is electrowon and the barren liquor reused; Merrill‑Crowe filtrate still contains NaCN and is typically recycled (www.smenet.org). Acidification of WAD solution (pH<2) converts CN⁻ to volatile HCN, which can be captured and neutralized back to NaCN — a classical recovery step (www.unitedchemicalcn.com). Well‑run CIP/CIL trains often recycle >90% of cyanide between leach and adsorption; tailings often contain <10 ppm CN (www.smenet.org).

Copper heaps: acid monitoring and estimation

Heap or vat leaching of oxide/sulfide ores uses 0.5–5% w/w sulfuric acid delivered continuously (peristaltic or diaphragm pumps are typical). Consumption depends on ore: copper carbonates and limonites can demand tens of kg/t, whereas chalcocite/sulfide ores consume much less. An extreme laboratory case: dissolving pure malachite at pH≈0.6 required ≈986 kg H2SO4 per tonne; the same ore in a mild ammoniacal system recovered ~84% of copper with essentially zero acid (www.researchgate.net).

There’s no single “acid meter” in‑line. Plants triangulate using pH, oxidation‑reduction potential (ORP; redox proxy), copper concentration, and sulfate‑linked conductivity. Maintaining heap effluent pH ~1.5–2.0 via sensors in the raffinate tank keeps dissolution efficient without over‑acidifying. Some sites deploy in‑line titration or ion‑selective pH probes to track acid normality; the solvent extraction (SX; copper transfer to organic) stage provides feedback, with raffinate (barren leach solution) signaling how much acid was consumed. Cascade PI controllers adjust acid up if raffinate pH drifts above target and pull back if copper concentration falls.

Because in‑heap measurement is impractical, modeling augments control. A recent study applied Extended Kalman Filters (EKF; state estimator) to a one‑dimensional heap model, using outlet [Cu2+] and [H+] to estimate in‑heap copper grade, moisture, and acid consumption; the “virtual sensors” predicted acid use accurately and were shown to improve internal understanding, metal recovery, and reduce acid consumption in simulations (www.mdpi.com). In practice, a controller could trim feed so the heap just reaches breakthrough.

Acid recycling in SX/EW circuits

Acid is inherently cyclic in solvent extraction–electrowinning (SX/EW; copper recovery to cathode). After electrowinning, “barren electrolyte” recycles to strip; SX raffinate (copper‑barren leach liquor) returns to the heap — a closed loop where >90% of circulating sulfuric acid is reused, with fresh acid making up heap neutralization losses and bleed/evaporation (blog.emew.com). Impurities (iron, aluminum) are periodically precipitated (jarosite/hematite) to avoid buildup, allowing continued recovery of acid for further use.

Reagent recovery from barren streams

Cyanide (gold): Elution with hot caustic and steam converts Au(CN)2⁻ back to Au⁰, leaving sodium cyanide in solution; after electrowinning, the barren liquor is reused as make‑up cyanide. Merrill‑Crowe filtrate still contains NaCN and is recycled to leach. Tailings often retain “weak” complexes (thiocyanate, metal‑cyanide); oxidants or hydrogen peroxide (INCO process) convert these to cyanate or nitrate, and in some flowsheets the resulting solution is treated to free NaCN for reuse. Industry briefings note that “solutions containing leached material, water and residual cyanide are treated… different technologies are used to recover and reuse residual cyanide in… circuits” (www.smenet.org).

Sulfuric acid (copper): SX raffinate is routinely reused by irrigation; spent electrolyte cycles back to strip. Bleed streams rich in iron/alunogen are treated to recover some acid, for example via jarosite precipitation. In closed‑loop designs, spent acid from stripping is mostly regenerated internally (blog.emew.com).

Acid regeneration (emerging): Some flowsheets trial extracting sulfuric acid from wastes using anion resins to selectively recover sulfate as concentrated H2SO4 (e.g., Dow XUS‑43605 and similar). While not ubiquitous, this concept aligns with ion exchange systems; where adopted, it would rely on ion‑exchange resins rather than new chemical reagents.

Performance metrics and business impact

Operators track specific reagent consumption — kg CN or kg H2SO4 per tonne of ore, or per ounce of metal. In gold circuits, baseline consumptions are ~2–4 kg NaCN/t; blending and mode control delivered ~2.04 kg/t in one modeling case, and preventing 30% spikes by avoiding high‑demand ore can save ~0.6 kg/t per event (www.mdpi.com; www.mdpi.com). For a 1,000 t/day plant, cutting from 2.24 to 2.04 kg/t conserves 200 kg/day (≈73 t/yr), saving ~$70,000–$150,000 per year; the Burkina Faso team stated “it is certain that savings will be made on the cost of cyanide” after improving control (www.scirp.org).

In copper, acid savings are smaller percentage‑wise but scale fast: trimming from 30 to 27 kg H2SO4/t saves 3 kg/t; at 10 Mt/yr, that’s 30,000 t/yr acid — about $1–2 million in reagent. Process models (including Kalman filters) can trim 1–10% without eroding recovery (www.mdpi.com). In the malachite lab example, avoiding extremely low pH saved ~98% of acid (986→20 kg/t) with recovery falling only from ~99% to 84% — a trade‑off potentially tolerable on low‑grade ores (www.researchgate.net).

Operational stability is the hidden dividend. In [57], by preventing cyanide dips the plant ensured “all leachable gold is leached” continuously, while spikes above setpoint only add cost; automated control held within ±2–5% of target and is “no longer just the domain of large stable plants” thanks to modern analyzers (www.process-analytical.co.za; www.process-analytical.co.za). For acid circuits, stable pH control avoids acid breakthrough and unnecessary neutralization.

Implementation parameters and controls

Measure what matters: install online free‑thiosulfate cyanide analyzers (amperometric or colorimetric probes) and pH/ORP sensors on the key solution streams; ensure each leach tank or carbon column is tied into the DCS. Automate dosing with PID or model‑based control to trim cyanide or H2SO4 via a dosing pump, with deadbands, flow compensation, and integrator tuning to prevent oscillation; “Dynamic” cyanide PIDs have been reported to hold within ±5% of target (www.process-analytical.co.za).

Align setpoints to ore: hold cyanide just above the leaching threshold (as low as 200–300 mg/L when recovery is stable) and raffinate pH around ~1.8–2.0 for copper; re‑evaluate whenever feed changes. Monitor specific consumption — kg CN/oz gold and kg H2SO4/kg Cu — as KPIs to validate savings. Recycle aggressively: return barren solutions to leach (CIP tailings solution to process water pond; SX raffinate to heap), implement SART or acidification to reclaim cyanide where practical, and treat/reuse bleed without diluting raffinate (blog.emew.com; www.unitedchemicalcn.com; www.smenet.org).

Bottom line: data‑driven control of leach chemistry — holding reagent concentrations within tight bounds — consistently cuts consumption without sacrificing recovery in gold (www.process-analytical.co.za; www.scirp.org) and, with estimation and feedback, trims acid use in copper heaps (www.mdpi.com). Greener practice — meeting effluent limits with <0.05% cyanide — aligns with cost efficiency: use “just enough” reagent and recycle what remains (www.smenet.org).