How Particle Size Decisions Drive Mineral Processing Recovery

Grinding mills consume between 50% and 70% of the total energy used in a typical mineral processing plant, and almost every recovery failure, reagent overrun, and energy loss further down the circuit traces back to a single decision made at that mill: how fine to grind.

Particle size is not one metric. It is a variable that moves through comminution, flotation, leaching, and classification, changing how each stage performs as it goes. Get it wrong at the mill and you pay for it everywhere else.

This guide follows that chain from start to finish: why size is set during grinding, how it governs separation and reaction rates, and how modern circuits monitor and control it in real time.

By the time you finish, you will know how particle size decisions made at the mill determine what you recover at the end of the circuit, which specific size thresholds matter, and which control strategies separate well-run operations from those quietly bleeding energy and recovery.

Why grinding consumes so much energy, and what particle size has to do with it

Start with the scale of the problem, because it reframes everything that follows. Grinding, technically called comminution, is the process of crushing and milling ore to break mineral grains free from surrounding waste rock. It is also the single largest electricity consumer in most processing plants.

The anchor figure: Comminution accounts for roughly 50-70% of total plant energy, making the grinding circuit the dominant cost centre in most operations.

To put that in a wider frame, the numbers are striking:

• Grinding and comminution: approximately 50-70% of total plant energy (multiple 2025-2026 sources)
• Mining’s share of global energy use: approximately 6% (Sandvik Rock Processing, 2025)
• Comminution’s share of global energy use: approximately 2.5% (Sandvik Rock Processing, 2025)

That last figure is worth sitting with. A single class of equipment, grinding mills, consumes around 2.5% of all the energy used on Earth. For your operation, that means the grind target you set is, before anything else, the largest line on your energy bill.

Innovations in energy-efficient mining technology are increasingly relevant here, because the same circuit optimisation logic that governs grind target selection also determines which equipment upgrades deliver the highest return on a plant’s electricity bill.

How the size target drives the energy bill

Here is the relationship that governs that bill: the finer you grind, the more energy you spend per tonne, and the increase is not linear. Pushing the target size down by a modest amount can demand a disproportionately larger jump in specific energy, measured in kilowatt-hours per tonne (kWh/t).

That creates a dual penalty when you get the target wrong. Grind too coarse and valuable minerals stay locked inside host rock, reducing how much you can ever recover. Grind too fine and you waste energy, accelerate equipment wear, and create ultrafine particles that cause problems downstream.

The useful news is that the same target size can often be reached at lower energy cost. Mill selection, grinding media size, classification efficiency, and mill load can all be tuned so the circuit hits its grind target while consuming less specific energy. Treat grinding optimisation as a cost exercise, not just a metallurgical one, and it becomes one of the highest-leverage decisions on the plant.

The flotation window: what happens to recovery when particle size moves outside the optimal range

Once ore leaves the mill, the size you have produced meets its first hard test in the flotation cell. Froth flotation separates valuable sulfide minerals from waste by attaching them to rising air bubbles, and it only works well within a surprisingly narrow band of particle sizes.

For sulfide minerals in conventional mechanical flotation cells, that band sits at roughly 20-150 µm. A 2026 Springer chapter on fine and coarse particle flotation documents this range as the zone where recovery is maximised, and Metso’s 2024 technical guidance places the optimum between 25 and 150 µm. The agreement across sources is close.

The window exists because of physics, not convention. Particles have to collide with a bubble, attach to it, and stay attached through the turbulence of the cell. Too small, and they lack the mass to collide reliably. Too large, and hydrodynamic forces tear them off the bubble before they reach the froth.

Industry anchor: Copper sulfide operations commonly target a P80 of approximately 75-100 µm, placing the bulk of the ore squarely inside the flotation window to maximise copper recovery.

For any plant treating sulfides, these thresholds are not a marginal efficiency guide. They define the grind target that governs recovery. Operate outside them and valuable minerals report to tailings instead of concentrate as a structural outcome, not bad luck.

Coarse particle flotation technologies developed specifically for base metals are redefining where the upper boundary of that flotation window sits, with some circuits now recovering sulfide minerals at sizes well above the conventional 150 µm ceiling, which changes the grind target calculus for concentrator designers.

What over-grinding actually costs at the reagent and selectivity level

The penalty for over-grinding is worse than lost recovery alone. Particles below approximately 10 µm carry an enormous specific surface area relative to their volume, and that surface has to be coated with reagent.

The result is a double hit to your economics. Collector and frother consumption climbs because there is simply more surface to cover, while those same ultrafines promote mechanical entrainment of gangue, dragging waste minerals into the froth. Reagent cost rises and concentrate grade falls at the same time. Over-grinding does not just waste energy at the mill; it compounds the loss in the flotation cell.

How particle size controls leaching kinetics and what it means for hydrometallurgical circuits

Flotation is not the only downstream stage that reads particle size. In hydrometallurgy, where minerals are dissolved out of ore using chemical solutions, one relationship sits underneath everything: surface area to volume.

Smaller particles expose proportionally more mineral surface to the leaching reagent. Because dissolution happens at that surface, finer particles dissolve faster and lift extraction yields. This principle holds across the main leach circuit types:

1. Heap leaching: ore is stacked and irrigated with solution that percolates through the pile, dissolving minerals as it travels.
2. Tank leaching: finely ground ore is agitated in tanks with reagent, giving the fastest and most complete contact.
3. Pressure oxidation: ore is treated under elevated temperature and pressure to accelerate dissolution of refractory minerals.

Each of these responds to finer feed with faster kinetics. But finer is not automatically better, and heap leaching is where that trap appears.

The core trade-off: In hydrometallurgy, particle size is an economic decision that weighs grinding cost against improved leach performance, not a simple case of finer being better.

Grind too fine for a heap, and the particles pack tightly enough to cause solution channelling and permeability problems. Solution finds preferential paths through the pile and bypasses much of the ore, so the reagent never reaches the mineral surfaces you paid to expose. The surface area benefit is partly cancelled by poor contact.

Heap leach permeability is also being addressed at the blasting stage rather than the grinding stage, with micro-fracturing techniques designed to create internal crack networks that preserve solution flow paths through coarser ore without requiring a finer grind target.

For your circuit design, the read is this: the right size target balances faster dissolution against the practical reality of moving solution through the chosen leach environment. Misjudge it and you raise grinding cost and reagent consumption at once, with no yield gain to show for either.

Understanding particle size distribution: the concept behind the control target

So far the discussion has leaned on single numbers like P80. That is a useful shorthand, but it hides something important, and understanding what it hides is what separates partial control from full control.

A real grinding product is never one size. It is a spread of particles across a wide range, and that spread is called the particle size distribution (PSD). The PSD matters more than any single average, because the tails of the distribution, the very finest and very coarsest fractions, are usually where recovery is lost.

P80 is the standard control metric most plants use. It is simply the size at which 80% of the particles pass, a practical single number to aim the circuit at. Its limitation is that it says nothing about the tails: two grinding products with the same P80 can behave very differently if one contains far more ultrafines or coarse particles than the other.

That is why controlling only P80 leaves money on the table. The specific fractions that cause the worst problems are:

• P80: a useful guide metric for centring the distribution, but not the whole picture
• Ultrafines below approximately 10 µm: the reagent-hungry, entrainment-causing fraction that P80 does not capture
• Coarse tail above approximately 300 µm: poorly liberated particles that resist attachment and report to tailings

A plant that hits its P80 target but ignores those tails is managing only part of the problem. The recovery and reagent losses driven by the fines and coarse fractions will persist no matter how precisely the P80 is held.

Classification equipment and closed-circuit grinding as PSD management tools

This is where classification equipment earns its place. Hydrocyclones use centrifugal force to split a slurry stream by size, sending oversized particles back to the mill for another pass while letting correctly sized material move forward.

That feedback loop is the essence of closed-circuit grinding: oversized material is continuously returned for reprocessing, keeping the product size far more uniform than open-circuit operation, where ore passes through the mill once and moves on. The tightness of the resulting PSD depends directly on how well the classification stage performs. Poor classification efficiency lets the distribution accumulate the very tails that undermine flotation, which is why cyclone performance is not a side issue but a core lever on whether your grind actually meets the flotation window.

Continuous particle size monitoring: how real-time measurement changes what circuit control is possible

Knowing the right PSD target is only half the battle. Holding it through changing ore is the other half, and this is where traditional practice falls down.

The problem with intermittent manual sampling is speed. Samples are taken periodically, sent for analysis, and by the time size drift is detected, the circuit may have spent a long stretch swinging between under- and over-grinding. Recovery and concentrate grade fluctuate the whole time. The correction always arrives after the excursion has already cost you.

Continuous online measurement closes that gap. The clearest documented commercial example is CiDRA Minerals Processing’s CYCLONEtrac Particle Size Tracking System (PST), a sensor installed directly in the cyclone overflow stream. Its specifications, from CiDRA’s 2024 white paper, are as follows:

• Update rate: every 4 seconds
• Online availability: greater than 98%
• Minimum detectable particle size: typically 75 µm or greater
• Up to 5 particle sizes measurable simultaneously
• Wetted sensor with no moving parts, installed in the cyclone overflow stream

What real-time actually means: A 4-second update rate lets the circuit be controlled on the same timescale as its disturbances, rather than reacting long after the grind has already drifted.

The contrast with offline sampling is stark once you lay it side by side.

That greater than 98% availability at 4-second resolution is the difference between a circuit that drifts and one that holds its target through ore hardness changes and feed variability. For anyone weighing a grinding circuit upgrade, continuous size measurement is a direct route to cutting recovery variability and energy waste at the same time, rather than chasing them as two separate problems.

Continuous process intelligence platforms that integrate mill load, power draw, and classification data into a single real-time view extend what online size measurement alone can do, giving operators a connected picture of why the grind is drifting rather than just signalling that it has.

Turning particle size knowledge into better processing outcomes

Step back and the whole circuit resolves into a single picture. Particle size is the connective variable running from the mill through separation and leaching to the control room, and every trade-off you make on energy, liberation, reagent use, and recovery converges on it.

Three decisions determine whether your operation captures or wastes that value:

1. Set the grind target relative to the flotation window. For sulfides, aim the bulk of the distribution into the approximately 20-150 µm range rather than simply chasing maximum liberation.
2. Manage the full PSD, not P80 alone. Actively limit the ultrafine fraction below 10 µm and the coarse tail above 300 µm, because those tails drive the losses P80 cannot see.
3. Hold the target with continuous monitoring. Real-time measurement at 4-second resolution feeding closed-loop control keeps the grind stable through feed variability.

Get those wrong and the research is clear on what follows: recovery lost to size extremes, reagent over-consumption, higher energy use, process instability, and lost revenue with elevated operating costs. Those five risks are all symptoms of one poorly controlled variable.

That is the case for treating particle size as a circuit-wide control priority rather than a grinding department metric. With comminution consuming 50-70% of plant energy, it is usually the single largest lever for cutting cost and lifting recovery at the same time, which makes it the right starting point for any circuit optimisation review.

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.