Thickener Flocculant Dosage: Cut Cost, Keep Clarity

Last updated: 2026-10-27

Most thickener flocculant dosage problems come down to three failure modes — under-dose, over-dose and polymer destroyed by shear before it reaches the slurry — and plants that diagnose all three typically cut polymer spend by 20–40% while holding overflow turbidity below 200 NTU and underflow density within 2% w/w of target. This guide gives you the diagnostic logic, the dose units and the audit routine.

It is part of our wider solid-liquid separation coverage; sizing theory for the vessel itself is deliberately not repeated here — see our high rate thickener selection and thickener design calculation guides for area, torque and flux design.

Key Takeaways

• Under-dose gives turbid overflow; over-dose gives clear overflow but thin, slimy underflow.
• Dose in g/t of dry solids and control it on solids mass flow, never on slurry volume.
• Final polymer dilution near 0.05% w/w outperforms a 0.5% make-down in almost every duty.
• A centrifugal pump on the dosing line can destroy polymer; use gravity or a progressive cavity pump.
• Clay-rich feed and pH swings can shift the optimum dose by a factor of three — re-test, do not guess.

What Is Thickener Flocculant Dosage and How Is It Measured?

Thickener flocculant dosage is a mass ratio that states how many grams of polymer reagent are fed per tonne of dry solids entering the vessel, written as g/t. It is the operating variable with the highest leverage in the whole thickening circuit: a change from 15 to 25 g/t can transform settling flux, while an uncontrolled drift of 5 g/t can silently cost a mid-size plant five figures a year. The polymer used in hydrometallurgical duty is almost always a high-molecular-weight anionic polyacrylamide, supplied as dry powder, emulsion or bead, and activated (not merely dissolved) before use.

Make-down is the dilution and activation step that converts concentrated polymer product into a homogeneous, fully uncoiled solution the dosing pump can meter accurately. Ageing is a rest period that follows make-down — typically 30–60 minutes — during which the long polymer chains finish hydrating and reach their full extension length. A solution dosed before ageing is complete behaves like a partially effective reagent: the plant compensates by increasing dose, and the extra spend reads as an “ore problem” on nobody’s action list.

Overflow turbidity is an optical measurement that estimates suspended solids in the thickener liquor from light scattering, reported in nephelometric turbidity units (NTU) and used as the fastest real-time proxy for clarity. It is a proxy, not a measurement of mass concentration — the NTU-to-mg/L relationship changes with particle size and colour, so calibrate it against periodic suspended-solids (TSS) samples. Underflow density, usually reported in % w/w from a nuclear or ultrasonic density meter on the discharge line, is the second half of the performance pair: the thickener is paid to deliver clear overflow and dense underflow simultaneously, and dosage is one of the few levers that moves both.

Units and conversion deserve one clean example, because this is where most dosing records go wrong. Made-down solution at 0.05% w/w contains 0.5 g of polymer per litre. A plant treating 120 t/h of dry solids at a target dose of 25 g/t needs 120 × 25 = 3,000 g/h of neat polymer, which at 0.5 g/L equals 6,000 L/h of solution — 6 m³/h, which is why final-dilution dosing skids look like water pumps, not chemical pumps. Conversely, if the record shows a dosing pump delivering 4,200 L/h of 0.05% solution to a 105 t/h solids stream, the actual dose is 4,200 × 0.5 / 105 = 20 g/t. Convert from L/h to g/t on dry solids, every time, and write the conversion on the dosing skid itself.

One control principle overrides everything else in this section: dose must be ratio-controlled on dry solids mass flow, not on slurry volume. A volume-based loop holds polymer flow constant while feed density wanders between 22% and 31% w/w, so the true dose swings by roughly 40% without anyone touching anything. A mass-flow loop divides polymer flow by the product of volumetric feed flow and percent solids (from a density meter), and holds the g/t figure constant. Plants that make this one control change typically recover 10–20% of their polymer spend in the first month, before any other optimisation.

Why Does a Clear Overflow Sometimes Mean Trouble? The Three Failure Modes

Operators learn to watch the overflow launder, and that instinct hides half the story. The three canonical failure modes of flocculant dosing produce three different symptom signatures, and only one of them looks obviously wrong.

Mode 1 — under-dose. Polymer supply falls short of what the slurry needs, so flocs are small, irregular and slow. The visible signature is a turbid overflow, rising NTU, visible fines carryover into the process water circuit, and a thin, soupy bed that gives low underflow density. Downstream, the fines load returns as scale and grit in water pumps, and in a counter-current decantation (CCD) circuit as muddy intermediate liquors that cost wash efficiency — see our CCD wash ratio guide for that penalty quantified.

Mode 2 — over-dose. This is the expensive surprise. Excess polymer flocculates the fines so well that the overflow runs brilliantly clear, but the floc network becomes over-connected: a voluminous, gelatinous bed that traps water and refuses to consolidate. Underflow density falls even though the overflow looks better than ever, the rakes slog through a slimy blanket with torque creeping upward, filter cloths downstream blind with sticky polymer-coated fines, and the plant pays twice — once for the extra reagent and once in lost underflow density that the filters must then correct. Any plant that judges its dosing purely on overflow clarity is almost certainly over-dosing.

Mode 3 — destroyed polymer. High-molecular-weight anionic polymer chains are long, fragile structures; a pass through a centrifugal pump impeller or a throttled globe valve breaks them into short fragments with almost no bridging power. The symptom is maddening: the dose record says 30 g/t is being delivered, the jar test says 30 g/t should be ample, yet the thickener behaves as if it were under-dosed. The plant responds by pushing dose to 45 g/t and beyond, and the polymer bill climbs while clarity stays mediocre. The fix is hydraulic, not chemical — see the dosing-point architecture section below.

The table below is the first-check triage we use when walking up to an unfamiliar thickener.

Notice the discipline in the “action” column: every move is a step change with a settling period, not a continuous trim. Flocculant response is not linear and not immediate — a dose change needs one to two bed turnovers, often 4–12 hours, before its full effect shows in underflow density.

Make-Down, Ageing and Dilution Water: Why 0.05% Often Beats 0.5%

Why would a plant deliberately prepare a weaker solution and pump five times more liquid? Because polymer performance is chain length, and chain length is wasted when chains cannot uncoil and distribute individually.

At a 0.5% w/w make-down, an anionic polyacrylamide solution is close to a gel: viscosity runs into the thousands of centipoise, the chains are entangled with their neighbours, and when this syrup meets fast-moving slurry it cannot spread — it arrives as globs that wet unevenly and entrap a fraction of their own active polymer inside the glob. Mixed into a feedwell at any realistic energy input, only part of the purchased polymer ever does bridging work. At 0.025–0.05% w/w final dilution, the chains are individually extended, solution viscosity is near that of water, and contact with slurry distributes them in seconds. The same grams per tonne simply work better; in our pilot work the effective dose at 0.05% final dilution is commonly 30–50% lower than the dose needed to achieve the same settling rate from a 0.5% solution.

The practical architecture that reconciles storage economics with dosing physics: prepare and age a stock solution at 0.25–0.5% (kept warm, agitated gently, consumed within 24–48 h), then dilute it in-line with process water to 0.05–0.1% immediately upstream of the injection point, using an eductor or a dedicated dilution manifold. What must never happen is dosing neat 0.5% stock into the feedwell with no final dilution — that is the single most common make-down fault we see on site audits, and it alone can double apparent reagent demand. The quality of thickener flocculant dosage is largely set here, before the dosing pump ever starts.

Ageing discipline is cheap and non-negotiable. Two-stage make-down plants (wetting/dispersion stage, then gentle ageing with a slow paddle) should hold a minimum 30 minutes between the two; powder products generally need longer than emulsions. A plant that draws solution off the top of a tank still being filled with fresh polymer is dosing a blend of aged and unaged product, and its dose record will look like noise. Age the batch, then dose from it; stagger two tanks if the plant cannot tolerate batch gaps.

Dilution water quality matters more than most reagent schedules admit. High hardness (calcium and magnesium) compresses the electrical double layer around anionic chains and partially collapses them, cutting effective molecular weight; some iron-rich or high-chloride waters do worse. As a working rule, if dilution water exceeds roughly 500 mg/L total hardness, expect noticeably poorer polymer performance than the jar test performed with deionised or soft water predicted — the fix is softening the dilution supply or accepting a higher dose knowingly, with the cost on paper rather than hidden. Temperature matters too: below about 10 °C, hydration slows dramatically and ageing times should roughly double.

Dosing-Point Architecture: Feedwell Injection, Multiple Points and Static Mixers

Polymer is only as good as the fluid mechanics between the injection point and the floc’s arrival at the mud bed. Three architectures cover essentially every plant, and each has a failure signature worth knowing.

Feedwell injection — a single point (or two opposed points) dosing into the thickener feedwell, where incoming slurry provides the mixing energy. It is the standard architecture and the right answer for most duties. The engineering judgement that matters here: the feedwell must provide enough energy to distribute the polymer in about one residence time, but not so much that it shreds the flocs it just built. Modern high rate thickeners handle this with internal dilution and baffled, radial-inflow feedwell designs that decouple the mixing zone from the settling zone. If your flocs look fine and dusty at the feedwell exit and dense and grapelike only at the lab, your feedwell is over-shearing — the dose is fighting the hardware.

Multiple injection points — dosing split between the feedwell and one or more points further downstream (into the feed line, or into a dilution zone at the top of the bed). The purpose is staged flocculation: a small primary dose builds seed flocs in the feedwell, and a secondary dose grows them where shear energy is lower. Staged dosing of roughly 60–70% primary and 30–40% secondary frequently beats the same total dose at a single point on clay-bearing feeds, because early-formed flocs get to survive. It also adds two failure modes — now there are two dosing pumps, two calibration points and twice the chance that one leg silently starves — so it only belongs where someone owns the routine verification.

In-line static mixer — a motionless mixing element in the feed pipe, with polymer injected just upstream. This is the right architecture where the feedwell is old, oversized or short-circuiting, and it gives the most repeatable mixing energy of the three options. Its failure mode is overdose of energy for fragile flocs: a high-shear element appropriate for a 20% solids residue can demolish flocs on a 45% solids concentrate feed. Match element type to duty, and always install a sampling port just downstream of the mixer — it is the only place in the plant where you can watch flocculation happen in real time.

Whatever the architecture, two hydraulic rules are absolute. First, polymer must never pass through a centrifugal pump: the impeller’s high shear fields cleave the chains in a single pass, and the fix is either a gravity-fed header from the make-down plant (preferred — elevation costs nothing to maintain) or a progressive cavity (helical rotor) pump, whose low-shear progressing cavity action preserves molecular weight. Second, no throttling valves, sharp restrictions or half-open gate valves anywhere downstream of injection; if flow must be controlled, control it at the low-shear pump or with a variable-speed drive. Plants that inherited a centrifugal dosing pump and replaced it with gravity feed have, in documented cases, cut their polymer dose per tonne by 30–40% with no other change at all.

How Do You Diagnose Thickener Flocculant Dosage Faults in Seven Steps?

This is the routine we run, in order, when a thickener’s clarity or underflow density will not hold spec. It deliberately starts with arithmetic, not chemistry, because most “flocculation problems” are record-keeping problems.

1. Verify the dose arithmetic. Convert the actual dosing pump output (L/h) and solution concentration (g/L) against the current dry solids feed rate (t/h) into a true g/t figure, measured — not taken from the control system’s setpoint. Check that the dosing loop is ratio-controlled on solids mass flow. In roughly one audit in three, the real dose differs from the recorded dose by more than 20%, and the whole mystery ends here.
2. Verify the solution, not just the number. Sample the made-down solution at the injection point: measure viscosity against a freshly made standard, check concentration by draw-down or gravimetric method, and confirm the batch age is within its 24–48 h working life. A solution that is half-aged, half-fresh, contaminated with hardness, or older than its working life will underperform at any dose.
3. Run a confirmatory jar test on today’s feed. Use current slurry at current pH and temperature, current polymer, and a make-down and dilution protocol that mirrors the plant (final dilution 0.05% w/w). Step doses across a 3–5× range and record settling rate, supernatant clarity after 5 and 30 minutes, and (if possible) settled-solids volume. The jar optimum is your reference point, not your answer — see the scale-up trap below.
4. Diagnose shear damage in the field. If the plant’s true dose (step 1) is far above the jar optimum (step 3) while flocs at the feedwell are small and weak, polymer is being destroyed between the make-down plant and the slurry. Inspect the entire dosing line for centrifugal pumps and throttling valves. This diagnosis pays for itself faster than any other step: removing the shear element routinely allows a 25–40% dose reduction.
5. Fix make-down and dilution before touching dose. If the plant doses neat stock (0.25–0.5%) without final dilution, install the in-line dilution manifold first and hold the g/t constant while the solution quality changes. Expect several days of improving performance at constant dose; only then re-optimise the dose downward.
6. Optimise the dose in controlled steps. Move the dose in 10–15% increments, one change per day, holding feed conditions as steady as the concentrator allows. Track the full KPI set — overflow turbidity, underflow density, bed level, rake torque, and g/t cost — after each step. Stop when underflow density plateaus or overflow NTU starts to drift upward: the optimum is a range, and the bottom of the range is where the money is.
7. Re-check the feed story. If optimum dose has drifted far from the design or historical figure, characterise the feed: particle size distribution, clay and ultrafines content, and pH at the thickener. A shift from silica-dominated tailings to clay-rich feed, or a pH excursion of one to two units, moves the optimum dose by a factor of up to three. Document the new baseline so the next drift is caught by trend, not by complaint.

The Jar-Test-to-Plant Scale-Up Trap

The jar test is indispensable and misleading in equal measure. A jar test at 500 mL scale with gentle paddle mixing and idealised contact rarely reproduces the energy history a floc sees between injection and the mud bed — and the error is not random, it is systematically optimistic. Jar flocs are larger, more regular and settle faster than plant flocs receiving the same dose, so transferring the jar optimum directly to the plant usually produces a dose that looks right on paper and underperforms in the tank.

Three specific traps recur. First, contact energy: jars mix polymer and slurry at a controlled, moderate gradient, while a real feedwell mixes at whatever the feed pump and geometry deliver — which is why staged dosing and static mixers exist. Second, dose–underflow decoupling: jar tests rank doses by settling rate and clarity, but they say little about how the dose affects bed compaction and underflow density, which is often where the plant optimum differs most. Third, sample degradation: slurry stored for two days oxidises, pH drifts and fine particles age surface-chemically; a jar test on stale sample can easily return an optimum 30–50% off today’s feed.

The defensible procedure is a two-stage scale-up. Use the jar test to bracket the range and screen polymer types — it is excellent for relative ranking. Then confirm the top two or three dose levels on continuous pilot or on-site testwork, at plant energy history, before locking an operating figure. Our pilot testing services exist precisely for this step, and our slurry rheology guide covers the compaction side that jar tests cannot see. Confirm the number before locking an operating figure for thickener flocculant dosage; the pilot work is a rounding error against one year of dosing at the wrong figure.

What Does Thickener Flocculant Dosage Cost, and How Do You Audit It?

Cost first, with assumptions stated. Take a plant treating 2,000,000 t/y of dry solids through thickening. At a 25 g/t dose and an anionic polyacrylamide delivered price of US$3,000/t (typical range US$2,200–4,200/t for dry powder, freight and duties aside), annual flocculant spend is 2,000,000 × 25 g = 50 t × US$3,000/t = US$150,000. Doses outside 10–40 g/t (and clay duties far beyond) move this figure proportionally. A dosing upgrade — mass-flow ratio control, in-line final dilution, a low-shear (gravity or progressive cavity) delivery — typically costs US$60,000–150,000 installed. For an over-dosed or shear-damaged plant saving 20–30% of spend, payback runs roughly 1–3 years on reagent alone, and substantially faster once avoided underflow-density losses and downstream filter blinding are counted. Table 2 gives indicative doses and costs by material type; treat every row as a starting bracket, to be confirmed by test work.

The KPI set for any dosing program is short and should be trended, not spot-checked: overflow turbidity in NTU (with periodic TSS calibration in mg/L), underflow density in % w/w, flocculant cost in US$/t of dry solids, bed level, and rake torque. Bed level and torque together tell you whether the bed is compacting or building a voluminous gel — the two states that separate correct dosing from over-dosing when overflow clarity looks fine in both cases. For circuits feeding filtration, add filter cycle time and cake moisture to the trend; our filter cake moisture guide explains that linkage, and our thickener vs filter press comparison shows where the dose penalty lands economically.

The monthly dose audit is a one-page routine that belongs to one named person:

• Recalculate true dose in g/t from pump strokes/L and calibrated solution strength against weighed or metered dry solids — compare to setpoint and to last month.
• Cross-check solution inventory: litres of make-down consumed versus theoretical demand at target dose; a persistent gap means dose is not what the DCS believes.
• Trend overflow turbidity and underflow density monthly means and their variability; rising variability usually precedes rising mean.
• Re-jar-test against a fresh feed sample if the dose trend has moved more than 10% in three months.
• Walk the dosing line: check for new valves, pump swaps, or bypassed dilution — hardware drifts back toward convenience unless someone looks.
• Report one number upward: flocculant cost per tonne of solids, with the NTU and % w/w context that stops it being optimised blindly.

A structured audit is a small fraction of a shift per month, and it is the difference between a dosing system and a dosing opinion.

Field Data: An Indicative Dose-Optimisation Result

The figures below are indicative, drawn from anonymised commissioning and audit experience and rounded; they describe a gold tailings thickener treating approximately 180 t/h of dry solids at 28% w/w feed, and are not attributable to any named client.

At first survey the plant was dosing 46 g/t of anionic polyacrylamide, delivered through a centrifugal dosing pump with no final dilution — neat 0.4% stock straight to the feedwell. Overflow turbidity averaged 380 NTU, underflow density averaged 52% w/w against a 58% w/w target, and rake torque ran 25% above its commissioning baseline. Annual polymer spend at these conditions was approximately US$287,000.

The intervention sequence was: gravity-fed header replacing the centrifugal pump (zero reagent effect intended, one-day outage), in-line dilution to 0.05% w/w final (dose held at 46 g/t for one week), then stepped dose reduction over three weeks. Final steady state: 23 g/t dose, overflow turbidity 85 NTU, underflow density 57.5% w/w, rake torque within 5% of baseline. Polymer spend fell to approximately US$144,000 — an indicative saving of US$143,000 per year, or roughly half the original bill, with both overflow clarity and underflow density improved rather than traded. The reagent chemistry never changed; the delivery of thickener flocculant dosage did.

Frequently Asked Questions

Why is a 0.05% make-down often better than a 0.5% solution?

At 0.5% w/w the polymer solution is highly viscous and the long chains remain entangled, so the reagent arrives at the slurry as poorly distributed globs and only part of it does bridging work. At 0.025–0.05% the chains are individually extended and disperse into the slurry within seconds, so the same grams per tonne achieve a higher settling rate. Practically, prepare a 0.25–0.5% stock for storage economy and dilute in-line to 0.05–0.1% immediately before injection.

How do I convert L/h of made-down solution into g/t of dry solids?

Multiply the dosing flow in L/h by the polymer concentration in g/L (0.05% w/w = 0.5 g/L), then divide by the dry solids feed rate in t/h. For example, 3,000 L/h of 0.05% solution delivering to 120 t/h of solids equals 3,000 × 0.5 / 120 = 12.5 g/t. Always use measured dry solids, not design tonnage, and record the conversion where operators can see it.

Why does a centrifugal pump destroy flocculant, and what is the fix?

High-molecular-weight polymer chains are fragile, and the intense shear gradients inside a centrifugal pump impeller cleave them in a single pass, reducing bridging performance dramatically even though the mass flow of reagent is unchanged. Replace the pump with a gravity-fed header from the make-down plant, or with a progressive cavity pump whose low-shear operation preserves chain length. Also eliminate throttling valves and sharp restrictions downstream of the injection point.

Why does my overflow stay clear while the underflow gets thinner when I over-dose?

Excess polymer over-connects the flocs into a voluminous, gelatinous network that traps water and resists compaction, so the bed consolidates poorly and underflow density falls — while overflow clarity, which only needs small particles captured, looks excellent. The plant pays for extra reagent and then pays again downstream: filters and CCD circuits receive a thinner, more dilute feed and must remove the water the thickener failed to. Judge dosing on both clarity and underflow density, never clarity alone.

Why should dose be controlled on solids mass flow rather than slurry volume?

Flocculant demand scales with the mass of solids to be flocculated, not with the volume of water carrying them. A volume-based loop holds polymer flow constant while feed solids content wanders — a drift from 22% to 31% w/w swings the true dose by roughly 40% with no alarm. Ratio-controlling polymer flow against the product of feed flow and density from a mass flow meter holds the g/t figure constant and typically recovers 10–20% of polymer spend on its own.

How can clay content and pH change my optimum dose?

Clay and ultrafine particles present vastly more surface area per tonne than silica sand, so they adsorb polymer and demand proportionally more of it; a shift from coarse tailings to clay-rich feed can triple the optimum dose. pH matters because it controls particle surface charge and polymer configuration — anionic polyacrylamide performs best in mildly alkaline duty and loses efficiency in acidic feeds. Re-run jar tests whenever mineralogy or pH shifts materially, because the optimum thickener flocculant dosage moves with them — often by a factor of three, not a few percent.

How long does made-down polymer solution stay good for?

Plan on a working life of 24–48 hours after full ageing for anionic polyacrylamide solutions, with emulsion-derived products at the shorter end and always within the supplier’s stated window. Beyond that, microbial degradation and slow chain scission cut molecular weight, and the plant quietly compensates with higher dose. Two ageing tanks on staggered batches avoid dosing half-aged solution, and gentle agitation — never a high-shear mixer — keeps aged product homogeneous.

How often should we re-run jar tests on a running plant?

As a minimum, quarterly against a fresh feed sample, plus immediately after any ore-source change, grind-size shift, or pH regime change. Between tests, the monthly dose audit’s trend review acts as the trigger: if true dose, turbidity or underflow density have moved more than about 10% from baseline, re-test rather than re-trim. A jar test costs a technician half a shift; three months of dosing at a stale optimum typically costs far more.

Conclusion

Thickener flocculant dosage is where the cheapest wins in solid-liquid separation hide: a plant that doses on dry solids mass flow, dilutes polymer properly, protects it from shear and audits the g/t figure monthly will typically spend 20–40% less on polymer than its neighbour — with clearer overflow and a denser underflow, not instead of them. The three failure modes in this guide are all diagnosable in a week with a density meter, a jar kit and the discipline to convert L/h into g/t honestly. If you want the optimum dose confirmed at plant energy history before you commit, our pilot testing team can run the confirmation on your slurry, and the rest of our technical guides are in the resource library.

References

1. Wills, B. A. & Finch, J. A. (2016). Wills’ Mineral Processing Technology, 8th edition, Chapter 15 — Dewatering. Butterworth-Heinemann, Oxford.
2. Hogg, R. (2000). “Flocculation and dewatering.” International Journal of Mineral Processing, 58(1–4), 223–236. Elsevier.
3. Moody, G. (1992). “The use of polyacrylamides in mineral processing.” Minerals Engineering, 5(3–5), 479–492. Elsevier.
4. Owen, A. T., Fawell, P. D. & Swift, J. D. (2007). “The preparation and ageing of acrylamide/acrylate copolymer flocculant solutions.” International Journal of Mineral Processing, 84(1–4), 3–14. Elsevier.
5. Fawell, P. D., Nguyen, T. & et al. — CSIRO research on feedwell design and radial inflow flocculation in gravity thickeners, published across Minerals Engineering and International Journal of Mineral Processing (2000s series).
6. Society for Mining, Metallurgy & Exploration (SME). SME Mineral Processing Handbook — sections on sedimentation, thickening and flocculant practice. SME, Littleton, CO.
7. U.S. Geological Survey (2026). Mineral Commodity Summaries 2026. Reston, Virginia. usgs.gov

Changelog

2026-10-27 — first published.