How Lithium Flotation Chemistry Sets the Ceiling on Project Returns

Lithium is abundant in African spodumene, the hard-rock mineral that most of the world’s lithium supply is dug out of. Yet at the Goulamina mine in Mali, a substantial share of that lithium was leaving the processing plant in the tailings dump rather than the concentrate. Before a reagent redesign, Goulamina’s original flotation circuit recovered just 75% of the lithium in its ore.

For an Australian investor weighing exposure to African lithium, that number matters more than the size of the resource underneath it. A large deposit that a plant cannot efficiently process is a development story that struggles to close; the stage where lithium is either captured or lost is flotation, and the chemistry that governs it has been one of the least examined variables in the whole equation.

This is an explainer about what flotation reagent design actually determines in practice. After reading, you will understand why conventional chemistry leaves lithium behind, what the pilot results at Goulamina and Manono show, and why those numbers matter when you are assessing project-level economics rather than headline resource grades.

Why conventional flotation chemistry keeps leaving lithium in the tailings

The intuitive assumption is that lithium flotation should be a solved problem. Spodumene is a well-understood mineral, oleic acid has been the standard collector for decades, and cheaper is usually better when a chemical does its job. That assumption holds right up until the ore stops being clean and uniform.

Flotation is the stage where lithium is separated from waste rock. Air bubbles are pushed through a slurry of crushed ore, and chemical collectors attach selectively to the lithium-bearing spodumene particles, carrying them to the surface as a froth while the gangue, the commercially worthless rock, sinks and is discarded.

Oleic acid became the default collector for a simple reason: it is inexpensive, renewable, and works adequately on clean, homogeneous ore, according to Shandong Zhuocheng statements published in Australian Mining in October 2026. The trouble is that real African deposits are rarely clean or homogeneous.

The same source identifies where the chemistry breaks down:

• It underperforms on ore of variable grade, where a collector tuned for one feed stops working when the feed changes
• It loses selectivity when gangue minerals are present, pulling waste into the concentrate and diluting the product
• The desliming stage required to prepare ore for it strips fine lithium-bearing particles out before flotation can act on them

Tightening regulations on reagent volumes are adding further pressure to the case against single-reagent systems. For you as an investor, this means reagent choice is not a procurement line item. It is a process design decision that sets the ceiling on how much lithium a plant can actually sell, and projects that have not addressed it carry a recovery risk that headline resource estimates rarely show.

What desliming costs a lithium circuit

Desliming is a preparation stage that removes very fine particles from the ore before flotation. The purpose is to cut reagent consumption, because fines soak up collector chemicals disproportionately. The cost is that those fine particles often contain recoverable lithium, and once they are stripped out, that lithium is gone before the circuit has a chance to capture it.

At Goulamina, this loss mechanism was a direct contributor to the 75% baseline recovery rate. One tonne of lithium in every four was effectively discarded before flotation began.

How site-specific reagent systems work and why the design logic is different

If one cheap reagent that works everywhere sounds efficient, the appeal is obvious. The problem is that it assumes every deposit behaves the same way, and ore variability is precisely the thing that breaks a universal formula. The alternative is not a better molecule. It is a system.

Tailored reagent systems use combinations of collectors, frothers, and depressants calibrated to the specific mineralogy of a single deposit, according to Shandong Zhuocheng statements in Australian Mining, October 2026. A collector gathers the target mineral, a frother stabilises the bubbles that carry it, and a depressant suppresses the gangue so it stays out of the concentrate. The mixture is tuned, not copied.

The design is built around three objectives working at once:

• Balancing recovery rate against selectivity, so the circuit captures more lithium without dragging in more waste
• Accommodating feed grade variability without destabilising the circuit
• Reducing the total volume of reagent consumed

The most consequential change is architectural. When a reagent system is selective enough to handle fine particles without removing them first, the desliming stage can be eliminated entirely. Every particle size, including the fines that previously went to waste, stays in the circuit, and the single largest lithium loss mechanism is removed from the flowsheet.

The core design principle is that the reagent mixture must be built around the deposit’s specific mineralogy. It cannot be lifted from a standard formula and applied to a different ore body.

This is why laboratory and pilot-scale characterisation of the actual ore has to come first. The Manono deposit in the Democratic Republic of Congo (DRC) illustrates the point: with an average feed grade of 1.3% Li₂O, an elevated lithium oxide content relative to typical spodumene, it needs a purpose-designed circuit rather than an off-the-shelf scheme.

For you, the implication is concrete. A reagent system cannot be transplanted between projects, so a company’s ability to do that characterisation work in-house is what determines whether it can close the gap between resource grade and saleable concentrate. When you read a project’s metallurgical test work, whether the flotation approach is generic or genuinely optimised tells you how much weight those numbers deserve.

What the pilot results at Goulamina and Manono actually show

Start with the numbers. At Goulamina, concentrate grade rose from 4.8% to 5.26% Li₂O and recovery climbed from 75% to 92%, with desliming removed from the circuit. At Manono, a 1.3% Li₂O feed was processed to a 5.2% Li₂O concentrate at 88% recovery, with whole ore entering flotation directly.

What each figure represents matters more than the figure itself. Consider what a move from 75% to 92% recovery does in process terms.

Goulamina (Mali): eliminating desliming to lift recovery

Goulamina’s original flowsheet ran on traditional oleic acid collectors and a desliming stage that stripped fine lithium particles before flotation. Shandong Zhuocheng’s intervention did two things: it tailored the reagent scheme to the ore, and it eliminated desliming altogether, letting all particle sizes enter the circuit.

The result at pilot scale was recovery rising from 75% to 92% and concentrate grade improving from 4.8% to 5.26% Li₂O. In plain economic terms, for every tonne of lithium oxide in the feed, the operator now captures far more saleable product from the same ore, with no additional mining cost. As of October 2026, the company reported the technology was being scaled up, with full-scale validation still pending.

Manono (DRC): whole-ore flotation on a high-grade deposit

Manono’s deposit carries an elevated 1.3% Li₂O average feed grade. Shandong Zhuocheng processed raw Manono ore in its own laboratory and pilot plant in China and developed a circuit that allowed whole ore to enter flotation directly, without any prior treatment stage.

At pilot scale, that produced a 5.2% Li₂O concentrate at 88% recovery. As of October 2026, the company’s technical team had deployed to the Manono site to begin preparations for large-scale trials, with additional process refinements identified as requiring evaluation under field conditions. The Manono processing plant itself entered production in May 2026, and its first certified export batch was approved in July 2026 by the DRC’s Centre d’Expertise, d’Évaluation et de Certification (CEEC).

These are company-reported pilot figures, not independently audited outcomes. All results originate from Shandong Zhuocheng’s own statements and have not been independently verified. Full-plant performance will determine their actual economic impact.

The distance between pilot and full-plant performance is precisely where project economics either hold or fall apart. The two case studies give you concrete reference points for what reagent optimisation can move at the process level, but the pilot caveat is the check that stops those numbers being read as guaranteed operating performance.

What processing efficiency means for African lithium project economics

The technical story becomes a strategic one at a single point: the gap between the lithium in the ground and the lithium a plant can sell. That gap is what separates a viable feedstock source from a resource that never quite converts into profit.

Three levers connect recovery and grade to realised economics:

• Higher recovery from the same feed means more saleable tonnes without a dollar of extra mining cost
• Higher concentrate grade means a more valuable product per tonne shipped
• A richer concentrate lowers the transport cost per unit of contained lithium

In the DRC, that efficiency is not merely an optimisation target. Manono is commissioning amid serious power, rail, and processing bottlenecks, according to IntelliNews (“DRC mining growth exposes power, rail and processing bottlenecks,” 18 September 2026). When physical logistics are a binding constraint, strong metallurgy at the plant becomes a competitive necessity for getting product to market at all.

Projects that achieve strong metallurgical performance gain a first-mover advantage in a supply chain where physical logistics remain a binding constraint.

The scale context sharpens the point. Manono’s Phase 1 is designed for roughly 5 million tonnes of ore throughput a year, targeting around 1 million tonnes of spodumene concentrate annually. A 2026 production target of approximately 30,000 tonnes of lithium carbonate equivalent has been attributed to Zijin Mining and multiple secondary reports, though this figure is unverified in open sources and should be treated as indicative. Zijin Mining holds a majority stake in Manono, reported variously between 54.9% and 61% across sources, with state miner Cominiere at approximately 35.1% and the DRC government retaining 10%.

For Australian investors, there is a supply-chain dimension worth holding onto. African projects increasingly feed Chinese refining operations, so reagent optimisation that improves feedstock reliability and consistency has value beyond any single mine. It reduces operational variability for downstream processors across the chain.

The practical takeaway is this: processing efficiency is the variable that separates a large resource from a profitable one, and it is consistently under-represented in the headline numbers most project coverage leads with. Whether you hold exposure to the African projects themselves or to the companies supplying reagent systems and process technology into them, plant-level metallurgical performance is where actual economic outcomes are decided.

From pilot to plant: the questions that will determine whether these results hold

Pilot data is a credible signal. It is not a guarantee, and the difference is where an informed assessment lives.

The core translation risk is one of scale. Laboratory and pilot circuits run on smaller, more controlled sample volumes; a full plant introduces feed variability, higher throughput rates, and equipment scaling effects that a pilot cannot capture. Processing innovations frequently prove narrower in impact at commercial scale than their early figures suggest.

The programme status, as reported by Shandong Zhuocheng in October 2026, reflects exactly that unfinished state. Goulamina is scaling up with full-scale validation pending. Manono is in a large-scale trial phase with further field assessment still required. Neither has published independently audited full-plant results.

There is also a source concentration risk worth naming plainly. All of the collector chemistry characterisation and every pilot figure in this story derive from Shandong Zhuocheng’s own statements. No independent expert commentary or third-party verification of the reagent mechanism or the pilot results was found in open sources at the time of writing.

Given that, here are the three things to watch:

1. Full-plant metallurgical test results that show whether pilot recovery and grade survive commercial throughput
2. Independently verified concentrate grade and recovery data from a third party
3. Updated feasibility or study documents that formally incorporate the new reagent schemes

What independent verification would change

Third-party auditing matters because it converts a company-reported pilot figure into a number that can be built into feasibility economics with confidence. A verified recovery rate can anchor a financial model; an unverified one can only inform expectations.

The absence of independent verification does not invalidate the Goulamina or Manono results. It sets the appropriate confidence level for how you should use them: as directional evidence of what reagent optimisation can achieve, not as bankable operating performance.

What the pilot data tells you, and what it does not

The central argument is straightforward. Flotation reagent design is a material determinant of lithium project economics, not a secondary technical detail, and the Goulamina and Manono pilot results show what site-specific optimisation can actually move: recovery from 75% to 92% in one case, an 88% recovery on high-grade feed in the other.

What the data confirms is narrow but real: reagent redesign can lift recovery and concentrate grade at pilot scale in these specific ore bodies. What it does not yet confirm is full-plant commercial performance, which remains pending at both sites and unverified by any independent party.

When you evaluate African lithium project disclosures, reagent strategy and the source of the metallurgical test work belong among the questions you ask, not among the things you assume are resolved. The informed position sits between uncritical enthusiasm and outright dismissal, pending full-scale confirmation.

This article is for informational purposes only and should not be considered financial advice. Investors should conduct their own research and consult with financial professionals before making investment decisions. Past performance does not guarantee future results. Financial projections are subject to market conditions and various risk factors, and the pilot-scale results discussed here are company-reported and have not been independently verified.