Dry Stack Tailings Design for Hydrometallurgical Plants

Filtered, dry stack tailings have become the default risk answer for new hydrometallurgical plants: dewatering to 85–90% solids w/w yields a soil-like cake that is compacted into a stable, unsaturated stack — no water-retaining dam — while recycling up to about 80% of process water. This guide sets out the flowsheet, water balance, design rules, costs and permitting issues owners should expect.

Key Takeaways

• Dry stack tailings remove the wet-dam failure mode and simplify governance under the GISTM framework.
• Filtration is the heart of every solid liquid separation flowsheet; start dewatering selection at flowsheet stage, not detail design.
• Typical scheme: high-rate or paste thickener, filter press, conveying, then compacted lifts.
• Arid-region projects recover 75–85% of process water, a decisive advantage in Chile, Peru and Central Asia.
• CAPEX is higher than a conventional dam; realistic OPEX sits at roughly USD 2–7 per tonne of dry solids.

What Dry Stack Tailings Means

Dry stack tailings — often called filtered tailings or dry stacked filtered tailings (DSTF) — are tailings that have been dewatered to the point where they behave as a moist, granular soil rather than a slurry, and are then transported and compacted in engineered lifts. A dry stack facility is a tailings storage method that contains no permanent water-retaining dam; the stored mass is self-supporting, unsaturated and drained from beneath.

ASME VIII Div 1 vs Div 2 explained

Three concepts anchor the terminology:

1. Filtration. A filter press is a batch pressure-filtration machine that squeezes dilute tailings slurry between cloth-covered plates and delivers a handleable cake, typically at 80–90% solids w/w. Vacuum belt filters and plate-and-frame presses perform the same duty continuously or in cycles, and equipment choice is discussed in our filter press overview.
2. Pre-thickening. A paste thickener is a deep-bed thickening vessel that raises the solids content of the feed ahead of filtration, stabilising the filter cycle time and reducing pump and cloth loads; the trade-offs against a high-rate unit are compared in our paste thickener overview. Where water recovery is the dominant driver, a high-rate thickener upstream of the filters still captures most of the free water at far lower capital cost per square metre.
3. Stacking. Filtered cake is placed by conveyor or truck, spread in lifts — commonly 5–15 m — tracked or compacted, and shaped so that rainfall runs off rather than infiltrates. The finished landform resembles an engineered fill or a waste-rock dump far more than a dam.

The practical consequence is risk: because the stored tailings are unsaturated and drained, there is no phreatic surface and no large impoundment whose failure could flow like a liquid. That is the essential difference between a dry stack and every variant of wet impoundment, conventional, centre-line or downstream. It is also why regulators, insurers and lenders — and communities remembering Brumadinho — now ask “why not filter?” as the first question in any new tailings review. The remainder of this article explains how to answer it with numbers: flowsheets, water balances, cake property targets and cost ranges.

The Flowsheet: From Thickener to Stack

A dry stack flowsheet is a dewatering train followed by an earthworks operation. The equipment is conventional; the integration is where projects succeed or fail. The typical train for a copper or gold hydrometallurgical plant is:

1. Thickening. Tailings slurry at 25–40% solids w/w reports to a high-rate or high-density thickener, flocculated with 10–30 g/t of polymer. Underflow leaves at 55–70% solids w/w as a high-density slurry, releasing the majority of recovered water before filtration. Thickener sizing — see our high-rate thickener page — is driven by settling flux and rake torque limits, not by clarification, so testwork on the real tailings is mandatory.
2. Filtration. Underflow is pumped to filter presses (or, for coarser and higher-tonnage duties, vacuum belt or horizontal pan filters). The press fills in 2–6 minutes, squeezes at 8–16 bar, cores and discharges a cake in a 10–20 minute cycle. Cycle time sets the number of plates and presses, which is why feed consistency from the thickener matters more than any single machine specification.
3. Cake handling. Cake falls onto a belt feeder and main stacking conveyor. Brittleness and slope angle at discharge determine whether cake breaks into lumps (good) or smears onto belts (bad); both are functions of the clay and ultra-fines content, and both must be seen in pilot filtration runs, not assumed.
4. Stacking and construction. Cake is spread in lifts of 5–15 m, with 10–20 m typically chosen for large arid-region facilities. Each lift is compacted with dozers and rollers to a target dry density; freeboard to the final landform, haul ramps and dump faces are designed like a waste-rock facility with geotechnical acceptance criteria per lift.
5. Underdrainage and water management. A network of graded underdrains, or a free-draining starter layer of coarse waste, keeps the base unsaturated. Stormwater is separated from any seepage, channelled to a small pond, and returned to plant; because the stack is unsaturated, this pond is a fraction of the size a wet impoundment would need.
6. Progressive closure. Each completed lift area is shaped, capped where required and revegetated in step with operations. Closure spending is spread over life-of-mine instead of deferred to a decommissioning liability, which materially improves the facility’s position in any GISTM-aligned governance review.

Indicative stream values across the train are summarised below; each project must be confirmed with bench and pilot testwork on representative samples.

The two design levers that dominate everything downstream are underflow density and cake moisture. Raise underflow solids by five points and the filter area shrinks measurably; miss the cake moisture target and stacking equipment bogs down, lift productivity collapses and the schedule slips. This is why we treat thickener and filter selection as one optimisation, not two procurement packages.

Water Recovery: Why Arid-Region Projects Choose Dry Stack Tailings

In the Atacama desert of Chile, the Peruvian Andes, Sonora in Mexico, Nevada and the Gobi — wherever copper and gold occur and water does not — the water balance is the project driver. A conventional tailings pond loses water to evaporation across a large wetted area and traps it permanently in the stored slurry voids. Dry stack tailings invert that picture: water is released at two engineered points, the thickener overflow and the filter filtrate, both of which are captured, piped and returned to the plant, while the wetted surface shrinks to a small storm pond.

A simplified annual water balance for a 3 Mtpa plant in an arid climate illustrates the scale:

The remaining 15–25% leaves the circuit in three ways: moisture locked in the stacked cake, evaporation from the small pond and active faces, and humidity losses from cooling and dust suppression circuits. Even so, recovery of 75–85% compares with roughly 50–65% for a conventional impoundment in the same climate, and the difference compounds every year of operation. For a plant consuming, say, 2.5 million m³ of raw water annually, the delta is close to 500,000 m³ per year — in Chile or Mongolia, enough to determine whether a project clears its water-rights permitting at all.

Two second-order benefits matter to owners. First, pumped water volume falls: recovered water returns by gravity or short lifts instead of long reclaim pipelines from a distant dam, cutting both energy and maintenance. Second, the facility’s water inventory — the volume that could escape in a worst-case event — is small, which simplifies emergency planning and community disclosure. Broader context on facility selection and monitoring is in our tailings management application note, and our filter cake moisture resource explains how residual moisture shifts the balance further.

Design Considerations: Cake Properties, Stack Geometry and Climate

Design starts with the cake. A target gravimetric cake moisture below about 15% w/w — equivalently 85% solids w/w — is the usual acceptance threshold, because at that consistency filtered tailings can be hauled, spread and compacted immediately in almost any weather. Below 80% solids, cake tends to rut under dozers; above 92%, dust generation and conveyor spillage grow. The target is set by the fines: clay-rich residues from oxidised gold ores or clays in porphyry copper weathering zones hold water tenaciously and may need longer squeeze times, finer cloth or steam-assisted cycles.

Fines handling extends beyond the press. Ultra-fines that escape thickening can blind cloth and destabilise lifts, so many plants split the stream — hydrocyclone sands to the filters, slimes to a separate small impoundment or to paste backfill. For high-rainfall sites with fine, low-permeability tailings, a hybrid thickened or paste approach is preferred because a filtered stack built from saturated, low-permeability material can consolidate slowly and shed shear strength in prolonged wet seasons; Indonesia, Ecuador and parts of West Africa routinely push designers this way.

Stack geometry follows from cake quality. Coarse, free-draining cake accepts 15–20 m lifts and steep overall slopes, maximising capacity per hectare — useful where land is scarce around a leach pad. Finer or wetter cake calls for 5–10 m lifts, wider benches and lower stack angles. For seismically active regions such as the Andes, Turkey or Central Asia, lower lifts with systematic compaction and generous drainage are preferred because densification and low saturation both raise the cyclic resistance of the stacked mass; a dry stack is not automatically immune to seismic design, it is simply far easier to engineer safely than a saturated dam. In every case the foundation and underdrainage do the quiet work: a free-draining starter layer, graded collector drains and monitoring piezometers keep the base unsaturated for the life of the facility.

Climate closes the loop. In arid Western Australia or northern Chile, cake dries further in the stack and gains strength; in wet climates the same cake may never drain below its placement moisture, so the design moisture margin, lift height and cover strategy must all be chosen conservatively.

Is Dry Stacking Suitable for Every Mine?

No — and honest screening saves budgets. Very high tonnages (above roughly 8–10 Mtpa of fine tailings) drive filter counts, floor space and OPEX upward quickly, and may favour large thickened facilities instead. Extremely fine or clay-rich tailings can produce cake that is hard to stack economically. High-rainfall sites lose part of the water-recovery advantage and face tougher placement windows. Conversely, water-scarce, seismic, land-constrained or socially sensitive settings — the majority of new copper and gold projects today — are precisely where filtered storage outperforms on risk, water and closure cost.

Cost: CAPEX and OPEX of a Dry Stack Facility

Owners should expect a dry stack facility to cost more upfront and less over life-of-mine than a wet impoundment of the same capacity. The capital premium sits in the dewatering plant: thickener, filter presses, conveyors, stacker and the civil works beneath them. The offsets appear in smaller land disturbance, no embankment raises, lower closure bonding, reduced water purchases and — increasingly — a smoother permitting path.

Indicative cost ranges, to be confirmed by vendor proposal and site conditions, are:

Assumptions behind these ranges matter. Power at USD 0.05–0.12 per kWh, flocculant at USD 2.0–3.5 per kg, cloth life of 3,000–8,000 cycles and a 10–20 year life of facility all shift the totals. Throughput is the strongest lever: a 100 t/h gold plant and a 600 t/h copper concentrator sit at opposite ends of nearly every range, and duty-cycle strategy (how many presses run, how many standby) can swing CAPEX by a third.

Permitting interacts directly with cost. Because the facility stores no water, many jurisdictions assess it as a landfill- or fill-like structure rather than a dam, shortening review — but arid-region water extraction, dust and progressive closure plans still attract scrutiny, and lenders applying GISTM-aligned expectations will demand documented governance either way.

How Long Does Permitting a Dry Stack Facility Take?

Realistically, 12–36 months depending on jurisdiction, with the dry stack advantage showing in review quality rather than only speed. Regulators in Chile, Peru, Kazakhstan and Western Australia are familiar with filtered facilities, so the technical basis — cake testwork, stability runs, water balance — is usually accepted without the multi-year impoundment debate. What extends timelines is rarely the stack itself: water rights, dust management and community consultation dominate the critical path.

What Should You Test Before Committing CAPEX?

Three test campaigns de-risk the investment. Bench settling and rheology define thickener flux and underflow density. Pilot filtration — running your slurry through a full-size press plate at representative cycle times — fixes cake moisture, cloth selection and press count. Stacking trials on the actual cake confirm lift compaction, trafficking and drainage. Skipping the middle one is the most expensive shortcut in this flowsheet.

Regulation and Risk After Brumadinho

The January 2019 Brumadinho dam collapse in Brazil, which killed 270 people, transformed tailings governance worldwide. Within months, institutional investors launched the Church of England Pensions Board–led global disclosure initiative, and in August 2020 the International Council on Mining and Metals (ICMM), the United Nations Environment Programme (UNEP) and the Principles for Responsible Investment (PRI) jointly published the Global Industry Standard on Tailings Management (GISTM).

The GISTM is a governance and assurance framework built on seven principles spanning affected communities, integrated knowledge base, design, construction and ongoing management, ongoing change management, monitoring and control, and closure and post-closure. It does not mandate a specific technology and contains no rule that says “filter your tailings.” In practice, however, it requires operators to justify their facility choice against alternatives, to disclose facility details publicly, and to demonstrate that consequences of failure were honestly assessed. A facility with no retained water pool and a small consequence-of-failure classification answers most of those questions cheaply — which is why banks and insurers now read a dry stack decision as evidence of strong governance, not merely an engineering preference.

National regulation has moved the same direction. Brazil legislated upstream-dam decommissioning after Brumadinho; several South American and Central Asian permitting authorities now request filtered alternatives be evaluated in tailings alternatives analyses. Owners should therefore budget for a documented alternatives study regardless of which option they finally choose — and for the reality that, in water-stressed regions, the filtered option increasingly wins on water security and social licence before cost is even discussed.

Project Data Point (Anonymized)

The following anonymised project illustrates the numbers in practice. A 2.4 Mtpa (dry solids) gold hydrometallurgical plant in Central Asia — a steppe climate with under 300 mm annual rainfall and severe winters — selected a filtered facility after a water-balance-driven alternatives study.

Testwork on the flotation-and-leach residue showed a P80 near 75 µm with moderate clay. A 24 m diameter high-density thickener, flocculated at about 18 g/t, produced underflow at 68% w/w solids (basis: plant commissioning data, averaged over the first operating year), releasing roughly 0.62 m³ of water per tonne of dry solids before filtration. Eight chamber filter presses, each fed at 60% w/w solids, delivered cake at 11.5% gravimetric moisture (basis: daily 1 kg grab samples, 12-month average, n≈3,600), comfortably below the 15% w/w stacking target and stable enough for immediate conveyor transfer even at −25 °C.

Overall process water recovery measured 82% (basis: plant-wide water metering reconciled monthly against slurry tonnage and raw-water inflow), marginally above the design case of 80% because winter evaporation losses were lower than assumed. Cake was stacked in 12 m lifts with dozer compaction to a dry density above 1.9 t/m³ (basis: annual density testing programme); the stack had reached lift 4 at the time of writing with measured pore pressures near zero, consistent with the free-draining underdrainage design.

The owner’s reported outcomes: zero make-up water restrictions through two dry summers, a dam-classification consequence rating far below the wet alternative, and dewatering OPEX of about USD 3.1 per tonne of dry solids — within the ranges quoted earlier. The plant’s dewatering train was piloted for six weeks before purchase; both the vendor and the owner cite that pilot as the reason the facility met specification from month one.

Frequently Asked Questions

What is the difference between dry stack tailings and thickened or paste tailings?
Thickened and paste tailings are still stored as a slurry or paste, usually behind some form of containment or within a purpose-shaped impoundment, and retain a saturated phreatic surface. Dry stack tailings are filtered to a handleable cake and compacted in lifts as an unsaturated, free-draining landform with no retained water. The filtering step is the distinction; thickening alone does not produce a stackable product.

How much water can a dry stack tailings facility recover?
Arid-region facilities typically recover 75–85% of the water entering with the tailings slurry, measured annually. Losses are residual cake moisture (10–15% w/w), evaporation from a small pond and dust conditioning. Conventional impoundments in the same climate often recover only 50–65%, so the difference can exceed 500,000 m³ per year for a mid-size plant.

What cake moisture is needed for dry stacking?
The industry working target is gravimetric moisture below about 15% w/w, equivalent to 85% solids w/w. At that consistency the cake can be conveyed, hauled and compacted without rutting. Coarser tailings may stack acceptably at 18–20% w/w moisture; clay-rich tailings may need below 12%. The correct target comes from pilot filtration and stacking trials on your actual residue.

Are filtered facilities safe in seismic or high-rainfall regions?
They are engineered successfully in both, but differently. In seismic regions, lower lifts, systematic compaction and robust underdrainage provide the necessary cyclic resistance; saturation is the enemy, and filtered stacks are far easier to keep unsaturated than dams. In high-rainfall regions, the challenge is placement windows and slow consolidation of fine cake, which is why hybrid or paste alternatives are often screened there.

How much does a dry stack tailings facility cost?
Expect roughly USD 0.2–0.6 million of CAPEX per tonne-per-hour of dry solids filtration capacity, and USD 2–7 per tonne of dry solids in total OPEX, dominated by power, flocculant and cloth. These are typical ranges to be confirmed by vendor proposal and site conditions; closure provisions and water purchase savings usually favour the filtered option over life-of-mine.

Which equipment produces filtered tailings?
The standard train is a high-rate or high-density thickener followed by filter presses for fine, moderate-tonnage duties, or vacuum belt and pan filters for coarser or higher-tonnage feeds. Conveyors, stackers and compactors complete the system. Equipment selection depends on particle size, throughput and clay content, and should be confirmed by bench and pilot testwork before procurement.

Can an existing wet dam be converted to dry stacking?
Retrofitting is possible but is really a decommissioning-plus-replacement project: new filtered tailings are diverted to a new dry stack while the legacy impoundment is dewatered, stabilised and progressively covered. Owners typically justify it on closure liability reduction, water recovery and consequence-of-failure rating. It is a multi-year, capital-intensive transition best planned with pilot filtration data and a staged tailings alternatives study.

Conclusion

Dry stack tailings have moved from niche to default for new hydrometallurgical plants in water-scarce and risk-sensitive settings. The engineering is not exotic — thicken, filter, convey, stack — but the margin between a facility that meets specification from month one and one that frustrates for years is decided in testwork: cake moisture response, cycle times at your solids, stacking behaviour of your clays. A filtered facility at 85–90% solids w/w eliminates the wet-dam failure mode, recovers 75–85% of process water and positions the project strongly under GISTM-aligned governance; the price is a dewatering plant that must be sized on evidence, not vendor brochures.

Solid-Liquid Separation explained

Lexin-tech supplies and integrates the full solid–liquid separation train for copper and gold plants — thickeners, filter presses, testwork and stack planning — and every design we release is backed by pilot data on your tailings. If you are screening filtered storage for a new or retrofit project, request a technical proposal through our pilot-testing service: send us a representative sample and your plant basis, and we will return cake quality, water balance and sizing numbers you can take to your board. A machine-readable summary of this article is available at https://www.lexin-hydromet.com/llms.txt.

Related reading: explore more guides in our resources hub, from cake moisture optimisation to thickener selection.

References

1. Wills, B. A., & Finch, J. A. (2016). Wills’ Mineral Processing Technology, 8th edition. Elsevier. — Chapters on dewatering, thickening and filtration of processing tailings.
2. ICMM, UNEP & PRI (2020). Global Industry Standard on Tailings Management. Global Tailings Review. — Governance framework of seven principles for tailings facility design, management and disclosure.
3. Davies, M. P., Lighthall, P. C., Rice, S., & Martin, T. E. (2010). “Dry stack tailings — design considerations.” Tailings and Mine Waste ’10. CRC Press.
4. ICOLD & UNEP (2001). Tailings Dams: Risk of Dangerous Occurrences — Lessons Learnt from Practical Experiences, Bulletin 121. — Compilation of tailings dam incidents underpinning post-failure risk frameworks.
5. Dunne, R. C., Kawatra, S. K., & Young, C. A. (Eds.) (2019). SME Mineral Processing & Extractive Metallurgy Handbook. Society for Mining, Metallurgy & Exploration. — Sections on plant dewatering circuits and water management.
6. U.S. Geological Survey (2025). Mineral Commodity Summaries 2025. — Production and processing context for copper and gold hydrometallurgical operations.

Changelog

• 2026-10-20: first version.

Related reading

• Practical guide: Recovery