Spray Dryer Fines: Causes, Recovery and Control

What Fines Are, and Why They Matter

In a milk powder spray dryer, atomised feed droplets dry into individual particles. The distribution of particle sizes is wide, and the smallest particles — commonly those below around 90 micrometres, though the exact cut-off depends on the product and the analytical method — are referred to as fines.^[1] Fines are not waste: they are fully dried product. But because they are so light, they are easily carried out of the chamber in the exhaust air rather than falling out as finished powder, and that is where the problem begins.

How much of the powder becomes fines is substantial and product-dependent. Pilot-scale work reported by Masters found that up to 35% of the powder could become fines depending on particle size and the outlet-air duct arrangement.^[2] At industrial scale, it is widely reported that 10–20% of the powder leaves the drying tower as fines in the exhaust air, which must then be recovered downstream.^[3] The fines loading reaching the bag filter, expressed as a fraction of the powder produced, can range from roughly 40% to 90% depending on the product and plant configuration.^[4] In other words, fines are not a minor side-stream — on many plants they represent a large proportion of total production passing through the recovery system.

Fines matter for four reasons, which this page addresses in turn:

• Yield and value — fines not recovered are finished product lost to the exhaust
• Product quality — the fines fraction affects bulk density, flowability, wettability and the success of agglomeration
• Emissions — fines escaping to atmosphere are a regulated particulate emission
• Safety — fine, dry powder is the most dust-explosion-sensitive material on the plant

How Fines Are Formed

Fines arise from two broad routes. Primary fines are simply the small end of the droplet-size distribution produced at atomisation: small droplets dry into small particles. Secondary fines are generated within the chamber and downstream — by attrition, by the fracture of fragile hollow particles, and by particles that fail to agglomerate. Both are controlled by the same family of variables.

Atomisation

Atomisation sets the droplet-size distribution, and therefore the baseline fines fraction. Rotary (centrifugal wheel) atomisers, pressure nozzles and two-fluid nozzles each give different distributions; rotary atomisers tend toward more uniform particle sizes, while pressure nozzles allow droplet size to be tuned through pressure and orifice selection.^[5] Higher atomisation energy — higher wheel speed or higher nozzle pressure — produces finer droplets and therefore more fines. Worn, eroded or incorrectly sized nozzles and wheels shift the distribution and are a common, overlooked cause of rising fines on an established plant.

Feed total solids (concentration)

This is one of the most important and most controllable levers, and it operates in the direction plant operators expect: lower feed solids tends to produce more fines. Studies of nozzle-zone behaviour show that a low total-solids feed forms a thin, flexible particle shell that can collapse, whereas a higher total-solids feed forms an earlier and more pronounced stable crust.^[6] A more concentrated feed also means each droplet carries more solid, so for a given droplet size the dried particle is larger and heavier. Concentrating the feed further in the evaporator before the dryer — within the limits set by viscosity and stickiness — therefore shifts the particle-size distribution coarser and reduces the fines fraction. It also saves energy, because removing water in the evaporator is far cheaper than removing it in the dryer.

Drying conditions and the glass-transition effect

The drying rate and the temperature history of each droplet govern whether it forms a robust particle or a fragile one prone to fracturing into fines. Crucially, the relationship between the outlet temperature and the powder’s glass-transition temperature (Tg) sets the stickiness of the particle surface. Operate too far below Tg — over-drying, running the outlet too hot or too dry — and particles are dry, hard and non-sticky, so they do not agglomerate and the fines fraction rises. Operate too close to Tg and particles become sticky, causing wall deposits and instability. The workable window sits between these, and finding it is central to controlling both fines and deposits. Patent literature confirms the mechanism from the other direction: adding sugar, which alters the glass-transition behaviour and makes the powder stickier in the dryer, has been found to reduce the volume of sub-90-micron fines by promoting agglomeration.^[1]

Agglomeration and fines return

The standard industrial method for controlling fines is deliberately to recycle them. Collected fines are returned to the top of the chamber, into the atomisation or “collision” zone, where the dry fines collide with newly formed, still-wet droplets and coalesce into larger agglomerates.^[7] Agglomeration occurs through collisions of wet particles (primary agglomeration) or wet and dry particles (secondary agglomeration), and by controlling chamber geometry and process parameters these can be forced to occur to varying extents.^[8] Well-managed fines return turns a fines problem into a product-property advantage: it produces the larger, more porous, more readily wettable agglomerated powders demanded for instant products. Poorly managed fines return simply recirculates fines without agglomerating them, wasting energy and stressing the recovery system.

Chamber and outlet-duct design

The chamber geometry, the position of the outlet air duct, and the presence of an internal fluidised bed at the base of the chamber all influence how much powder is entrained as fines and how the air flows. Modern dryers increasingly place outlet ducts near the top rather than the bottom of the chamber and incorporate internal fluid beds, and these features change the fines behaviour relative to older designs.^[2] On an existing plant these are largely fixed, but they explain why two dryers running the same product can have very different fines behaviour, and why a design review matters when fines are a persistent problem.

The Causes of Fines at a Glance

Recovering Fines: Cyclones and Bag Filters

Because 10–20% of production leaves the tower as fines, every milk powder plant needs a recovery system to capture that value and to keep particulate out of the exhaust.^[3] The dairy industry uses two technologies, usually in combination.

Cyclones

A cyclone separates powder from air by spinning the particle-laden air so that the heavier particles are thrown outward and downward against the wall, falling to the discharge, while the cleaned air leaves through a central upward vortex. Cyclones are simple, robust and have no moving parts. Plants do not use a single standard arrangement: depending on duty and age, a plant may run one cyclone, a bank of cyclones in parallel, cyclones in series (multi-stage), or high-efficiency multicyclones. Recovery therefore varies with the number and quality of the cyclone stages — a single cyclone is commonly quoted at around 90%, but a well-configured multi-stage or high-efficiency arrangement recovers considerably more.^[9] In normal operation, good cyclones bring fines losses down to a low level.

Cyclones have two inherent limitations. The first is physical: the very finest particles carry too little mass to be thrown out of the air stream, so no cyclone — however well designed — separates them efficiently; they follow the air and pass through. The second is condition-dependent and is the one that catches plants out in practice: cyclone performance collapses if air leaks into the cyclone or its discharge. The separation depends on a clean, stable vortex and on powder discharging freely against a sealed boundary. Any in-leakage of air — through a worn discharge gaiter or seal, a failed or worn rotary valve or airlock, a cracked cone, or a leaking access door or flange — disturbs the vortex and re-entrains already-separated powder back into the air stream. The result is a sudden, often large, loss of recovery: product that the cyclone had captured is swept downstream as if the cyclone were not there. Because such leaks are easy to miss, they are a common and costly cause of unexplained fines losses and emission excursions.

Bag filters (baghouses)

A bag filter passes the exhaust air through fabric filter bags that capture the fine powder leaving the cyclones. It does two jobs. First, it routinely captures the very finest fraction that no cyclone can separate — which is why bag filters are effectively required to meet modern recovery and emission targets. Second, and just as important, it acts as the safety net for cyclone problems: when a cyclone underperforms or air leaks into a discharge, the escaping powder is caught by the bag filter rather than lost up the stack, and the fault shows up as rising bag-filter differential pressure instead of as a product-loss and emissions event. Bag filters are used both to maximise product recovery and to meet local dust-emission requirements,^[10] with collection efficiency reported between 99.5% and 99.98%.^[4] Modern dairy bag filters use food-grade filter media and an automatic pulse-jet cleaning system, in which a short reverse pulse of compressed air periodically de-dusts each bag so that captured powder falls away and the filter does not blind. Venturi designs above the bags can significantly reduce the compressed-air demand of pulsing,^[11] and modern controls pulse “as much as necessary but as little as possible” to cut energy and emissions.

An important development is the CIP-able (washable) baghouse, which can be cleaned in place. Because it can be cleaned to the same hygiene standard as the chamber, the fines it collects retain the microbiological quality of the main powder and can be returned to product rather than downgraded — directly increasing yield. Collecting all fines in a single washable stage also simplifies the plant, reduces the number of unit operations, and is a gentler form of collection than cyclones, which suits the high-fat and high-protein powders common in dairy.^[12]

Keeping bag filters working: cleaner bags, more recovery

A bag filter only delivers its rated recovery and emission performance if the bags are kept in good condition and the cleaning system works correctly. The practical issues that erode performance are well known:

• Blinding — bags progressively clog with fine, and especially with sticky or fatty, powder, raising differential pressure and reducing airflow and recovery. High differential pressure across washable baghouses is a recognised cause of operational interruption.^[13]
• Pulse-cleaning faults — failed or mistimed pulse valves, low compressed-air pressure, or blocked venturis leave bags un-cleaned, so differential pressure climbs and emissions rise.
• Bag damage — pinholes, seam failures or detached bags let fines straight through, causing a sudden rise in emissions and a loss of recovered product.
• Condensation and dewpoint — if the exhaust cools below dewpoint, moisture binds powder to the bags, accelerating blinding and risking microbial growth in non-CIP systems.

Monitoring differential pressure across the filter is the single most useful indicator: a rising trend signals blinding or cleaning faults before they become an emissions or yield problem. On plants with washable baghouses, a disciplined CIP regime keeps the fines product-grade and the bags performing. The pay-off for keeping bags clean is direct: more recovered product, lower emissions, and stable airflow that keeps the whole dryer running on specification.

Plants Without a Bag Filter: Reducing Emissions

Some older or smaller plants run on cyclones alone, with no bag-filter stage. On these plants the finest fraction that the cyclones cannot catch goes to atmosphere, which is both a loss of product and a particulate emission. Fines must be removed from the exhaust air for economic, safety and emission reasons.^[3] In the UK and EU, particulate emissions from milk powder plants are controlled under environmental permitting, with site-specific limits informed by best-available-techniques (BAT) guidance; the trend in limits is downward, and a cyclone-only plant can find itself unable to meet a tightened limit.

Options for a plant without a bag filter, in rough order of effectiveness:

• Reduce fines generation at source — the levers in the table above (higher feed solids, correct atomisation, avoiding over-drying) cut the quantity of fines reaching the exhaust in the first place, helping both emissions and yield without capital expenditure.
• Optimise the existing cyclones — correct cyclone sizing, sealing, and the elimination of in-leakage restore the recovery the cyclones were designed to deliver; worn or poorly sealed cyclones underperform badly.
• Add a bag filter — the definitive solution. Retrofitting a bag filter downstream of the cyclones lifts recovery from ~90% to well above 99.5% and brings the plant into line with modern emission limits, with the recovered fines paying back part of the investment.
• Wet scrubbing — where a dry bag filter is impractical, a wet scrubber can reduce particulate emissions, though it converts a dry recoverable product into an effluent load and is generally less attractive than a bag filter for a food powder.

Worked Example: The Value of Recovered Fines

The value at stake is easy to under-estimate. Consider a plant producing 5,000 kg/h of milk powder, where 15% of production — 750 kg/h — leaves the tower as fines in the exhaust air:

• Cyclones alone at 90% recovery return 675 kg/h and lose 75 kg/h of finished powder to the exhaust.
• Cyclones plus a bag filter at 99.9% combined recovery lose around 0.75 kg/h.
• The difference — roughly 74 kg/h of recovered product — at a powder value of, say, £2,500/tonne and 6,000 operating hours per year, is in the order of £1.1 million per year of product that would otherwise have gone up the stack.

The exact figures depend on your product value, fines fraction and recovery efficiencies — but the order of magnitude is why bag filters are standard on modern plants, and why keeping them working is not a maintenance afterthought but a core profitability issue. (Powder value changes continually and varies by product and market; use your own current figures.)

Fines and Dust-Explosion Safety

A dust explosion requires five things together — combustible dust, dispersion in air at an explosible concentration, an ignition source, confinement, and oxygen. All of these are naturally present in spray dryers and the equipment connected to them. The drying chamber, cyclones, bag filters, fluidised beds and conveying systems are the highest-risk locations in a dairy plant, and because they are interconnected, an explosion starting in one vessel can propagate to all of them.^[14]

The hazard is governed by the dust’s explosibility constants — K_st and P_max — which are product-specific and must be measured by testing, not calculated from process conditions. Protection is engineered through a combination of measures: ignition-source control, explosion venting (vent areas sized to recognised standards such as NFPA 68 or EN 14491), explosion suppression, and isolation devices to stop a flame or pressure front passing between connected vessels.^[15] Dairy spray-dryer explosions are not theoretical: documented incidents include a destructive milk spray-dryer explosion at a plant in north-west Spain.

The practical message for fines management is this: a high fines fraction is not only a yield and emissions issue, it is a safety-sensitivity issue. The correct response to high fines is never to ignore the dust risk but to confirm that the dryer, cyclones and bag filter have explosion protection designed to current standards, that the powder’s K_st and P_max have been tested, and that isolation between vessels is in place. This is specialist territory and must be handled by competent explosion-safety engineers.

Frequently Asked Questions

What counts as “fines” in milk powder?

Fines are the smallest particles in the powder, commonly taken as those below about 90 micrometres, though the exact cut-off depends on the product and the measurement method.^[1] They are fully dried product, but are light enough to be carried out of the chamber in the exhaust air.

How much of the powder becomes fines?

It varies widely with product and plant. Typically 10–20% of production leaves the tower as fines in the exhaust,^[3] and the fines loading reaching the bag filter can be anywhere from about 40% to 90% of the powder produced.^[4] Pilot work has shown up to 35% of powder becoming fines under some conditions.^[2]

Does concentrating the feed more reduce fines?

Yes, within limits. Higher feed total solids forms more robust particles and shifts the size distribution coarser, reducing fines,^[6] and it saves energy because removing water in the evaporator is cheaper than in the dryer. The limit is viscosity and stickiness — too high a concentration causes its own problems.

Why use both cyclones and a bag filter?

Cyclones (one or several, in parallel or series) do the primary bulk recovery and bring fines losses low in normal running, but they cannot separate the very finest fraction and they lose efficiency if air leaks into the discharge. A bag filter then captures that finest fraction at 99.5–99.98% efficiency, meets emission limits, and acts as the safety net that catches the powder when a cyclone leaks or underperforms.^[4][9][10]

Why do bag filters lose performance?

Mainly blinding (bags clogging with fine or sticky powder, raising differential pressure), pulse-cleaning faults, bag damage, and condensation below dewpoint.^[13] Monitoring differential pressure across the filter is the best early-warning indicator.

Are fines a dust-explosion risk?

Yes. Finer, drier powder is more dust-explosion-sensitive, and the dryer, cyclones, bag filter and connected equipment are the highest-risk locations on the plant.^[14] Explosion protection must be engineered to standards such as NFPA 68 or EN 14491 and assessed by qualified specialists.^[15]