OSLO Crystallizer: Design, Sizing and Selection

Last updated: 2026-11-01

An OSLO crystallizer is the design of choice when a specification calls for coarse, low-fines crystals — typically a d50 of 0.5-3 mm (0.02-0.12 in), several times coarser than forced circulation delivers. Supersaturation is generated away from the bed and released gently in an upward liquor flow, so crystals grow rather than nucleate. This guide covers the working principle, the sizing math, and the selection verdicts process engineers need.

If you are still mapping the wider equipment landscape, start with our evaporator-crystallizer overview; this article goes one level deeper on the fluidized-bed type.

Key Takeaways

• This design grows coarse, low-fines crystals by fluidizing a classified bed in upward-flowing clear liquor.
• Supersaturation forms in a separate evaporative or cooling section and is consumed gently across the bed.
• Expect a d50 of 0.5-3 mm (typical), versus 0.2-0.5 mm from forced circulation designs.
• Bed sizing follows a simple law: crystal holdup equals production rate times crystal residence time.
• Pairing with MVR or multi-effect heating keeps steam consumption competitive for large duties.

Why Crystal Size Distribution Decides Crystallizer Choice

In crystallization, the crystal size distribution (CSD) is rarely a cosmetic detail. It decides how the cake filters, how well it washes, whether it dusts in the dryer, and whether it cakes in the bag. For battery-grade nickel sulfate, an off-spec CSD can cost more to fix downstream than the crystallizer itself costs to buy.

The d50 is a statistical measure that reports the particle diameter at which half of the crystal mass is finer and half is coarser. Buyers specify it because it correlates directly with centrifuge cycle time, wash efficiency, and bulk density. Moving d50 from 0.3 mm to 1.0 mm (0.012 to 0.04 in) can double or triple centrifuge throughput on the same filter area, and coarse crystals also retain less mother liquor, which cuts drying load and improves assay.

Fines are the second half of the story. Fine crystals carry high surface area, occlude more impurity-laden mother liquor, wash poorly, and generate dust. They also burden the plant twice: every fines particle that leaves the machine represents solute that was already crystallized once and now has to be redissolved and sent back through the mother liquor recovery loop.

This is why crystallizer selection is really an exercise in nucleation management. A forced-circulation machine nucleates heavily because crystals travel through the heater with the liquor and see high supersaturation and shear; the result is a compact, forgiving unit that makes small crystals. A design that keeps crystals out of the hot zone and lets supersaturation decay across a suspended bed does the opposite: growth dominates, nucleation is throttled, and the product stays coarse. That is the design intent behind the OSLO crystallizer and its modern relatives.

Before going deeper into the hardware, fix one more term. Throughout this article, “coarse” means a d50 above roughly 0.5 mm with a narrow enough spread that screening is unnecessary. If your target is a fine, free-flowing commodity crystal, the calculus changes, and the comparison later in this article will point you to a different machine.

How an OSLO Crystallizer Works: The Fluidized Bed Principle

An OSLO crystallizer is a fluidized-bed crystallizer that suspends a growing bed of crystals in upward-flowing, essentially crystal-free liquor while supersaturation is generated in a separate evaporative or cooling section, typically at superficial liquor velocities of 20-60 mm/s (0.07-0.2 ft/s). The concept dates to the Krystal process developed in the 1920s and 1930s, which is why the machine is still called a Krystal crystallizer or Oslo-type crystallizer in much of the industry.

Evaporator Crystallizer Systems for Mining Brines, step by step

The hardware is arranged as a tall column with an attached circulation loop, and each part of that loop has one job.

The bed zone. Clear, supersaturated liquor enters below a bed of crystals and rises through it. Crystals whose terminal settling velocity exceeds the liquor velocity stay in the column and arrange themselves classified: the largest at the bottom, where supersaturation is freshest, and the smaller ones toward the top. Each crystal is suspended in liquor that is only slightly supersaturated, so it grows layer by layer instead of spawning new nuclei.

The circulation loop. Liquor that has surrendered its supersaturation leaves the top of the bed essentially clear, is pumped through an external heat exchanger — or a surface cooler on vacuum-cooling duties — and returns to the vapor head, where evaporation or flashing regenerates supersaturation. Because the liquor carries no crystals at this point, the exchanger sees clean walls and a low fouling rate. The whole machine lives or dies by keeping it that way.

The fines path. Nuclei form wherever supersaturation peaks, at the vapor head. Crystals too small to settle are carried out with the overflow and stay in the circulation loop, where they either redissolve in the warm leg or keep circulating until they are heavy enough to drop back into the bed. Fines never reach the product, and that single fact gives the OSLO crystallizer its characteristically narrow CSD and low fines content.

Product discharge comes from the bottom of the bed through an elutriation leg, so withdrawal is automatically classified — the coarsest crystals leave first, without screens or cutters.

The net effect is spatial separation of cause and effect: supersaturation is created in one place and consumed in another, with an inventory of growing crystals standing between the two. Mixed-suspension machines blur that separation, which is why they grow faster but dirtier.

Key Design Parameters for an OSLO Crystallizer

The OSLO crystallizer rewards careful parameter selection more than any other crystallizer type, because its performance rests on a balance between liquor velocity, supersaturation, and crystal inventory. The table below lists the parameters and starting ranges we use in practice; all values are typical or indicative and must be confirmed by test work on the actual liquor.

Fluidization velocity and bed height. The circulation pump must lift liquor fast enough to keep the top-of-bed size fraction suspended — typically 20-60 mm/s — yet slow enough that coarse crystals are not swept into the overflow. A bed expansion of 50-100% above the settled height is a workable starting band (indicative). More expansion thins the bed density, shortens residence time, and leaves less margin against feed surges.

Circulation ratio. The ratio of circulating liquor flow to feed flow commonly falls between 30:1 and 100:1 (typical). A high ratio is deliberate: it spreads the supersaturation generated per pass over a large liquor volume, so each pass through the bed removes only a small slice of the driving force and growth, not nucleation, absorbs it.

Supersaturation control. Operating at low Δc — commonly 1-5 g/L, or 20-50% of the measured metastable zone width — is the single most important operating decision. Push Δc too high and nuclei form faster than the bed can absorb them; the fines load spirals, and within hours the machine starts behaving like an agitated forced-circulation vessel with an expensive column on top.

Crystal growth rate. The crystal growth rate G is a kinetic parameter that gives the linear advance of the crystal surface per unit time, typically 10-100 µm/h for inorganic salts at the low Δc values above. G varies with the system, temperature, and trace impurities, and it cannot be predicted reliably from first principles. Measure it — everything in the worked example below flows from this one number.

Residence time. Crystal residence runs from about 12 h on fast systems to 72 h and beyond (hours to days, typical). Because a classified bed drives every crystal toward the same top size, the relation τ ≈ d50/G is the governing equation, not an approximation to apologize for.

Fines destruction. Where the natural loop cannot hold the population down, add a dedicated fines trap or a heated dissolver leg that redissolves overflow fines and returns clear liquor. Options range from a simple heated leg to a settler with an external dissolver tank. The full population-balance treatment, including CSD prediction from these parameters, is developed in our crystallizer design resource.

Sizing Worked Example: From Mass Balance to Bed Volume

The cleanest way to see how an OSLO crystallizer is sized is to carry one number through the chain. Consider 5 t/h of nickel sulfate hexahydrate product (about 11,000 lb/h) with a target d50 of 0.8 mm (800 µm) — a realistic coarse-grade specification. All figures are illustrative; growth rate in particular is to be confirmed by test work.

Assumptions (indicative): crystal growth rate G = 40 µm/h at the design Δc; bed crystal concentration 350 kg/m³ (22 lb/ft³); superficial liquor velocity 50 mm/s (0.16 ft/s).

Step 1 — residence time. For a classified bed, τ ≈ d50/G = 800/40 = 20 h.

Step 2 — crystal holdup. The bed must hold every crystal produced during one residence time: M = 5,000 kg/h × 20 h = 100,000 kg (about 220,000 lb).

Step 3 — bed volume. At 350 kg/m³, V = 100,000/350 ≈ 286 m³ (10,100 ft³). With a 6 m (19.7 ft) internal diameter the cross-section is 28.3 m², so the bed height works out to roughly 10 m (33 ft) before freeboard and the vapor head are added.

Step 4 — circulation. To fluidize at 50 mm/s across 28.3 m², the loop moves 0.05 × 28.3 ≈ 1.4 m³/s, about 5,100 m³/h (22,500 US gpm). That single number explains both the capital signature of a fluidized-bed train — large low-head pumps and generous pipework — and its low Δc per pass.

Step 5 — sanity checks. The nucleation balance must close: fines generated per hour should equal crystals leaving as product, or the bed inventory drifts. The heat balance must close: the evaporative section has to remove the water that carries the solute out of circulation, which is the duty we normally pair with MVR. And the mass balance must close: solute in equals crystal product out plus liquor losses, checked before any of the steps above are trusted.

The arithmetic is deliberately simple — the population balance for a classified bed collapses to τ = d50/G and M = P·τ — but the inputs are not. Two liquor-specific numbers, G and the achievable bed concentration, move the required bed volume by a factor of two in either direction. That is the practical argument for bench and pilot work before freezing a design: the equation is trivial, the kinetics are not.

OSLO vs Forced Circulation vs DTB: Which Wins When?

Four machine families compete for most crystallization duties: fluidized-bed (OSLO), forced circulation, draft-tube baffle (DTB), and batch. They differ mainly in how they treat nucleation and where crystals spend their time.

For coarse, low-fines product, the OSLO crystallizer is preferred because the classified bed discharges only crystals that have earned their size, while fines are structurally excluded from the product stream — no screening step is needed to hold the specification.

For heat-sensitive or high-value salts — nickel sulfate, potassium sulfate, and coarse ammonium sulfate grades — OSLO wins because growth at low Δc preserves habit and limits inclusions, and the crystal-free heater minimizes both product degradation and wall encrustation.

For scaling-prone, high-boiling-point-rise brines at large tonnage, forced circulation wins because it tolerates fouling and keeps running where a fluidized bed would demand disciplined supervision; heater selection and scaling behavior are compared in our forced circulation vs falling film analysis.

For medium crystals at high throughput in a compact plot, a DTB machine wins because it integrates fines removal and clear-liquor recycle inside one short vessel without the column height.

For a multiproduct contract plant running short campaigns, batch wins because changeover and cleaning-in-place matter more than CSD polish.

Duty-specific notes: sodium sulfate spans both worlds depending on phase and temperature window, and we treat it separately in our sodium sulfate crystallization resource. Ammonium paratungstate (APT) and lithium carbonate are duties where crystal quality drives product value directly, making classified-bed growth attractive; the lithium case is developed in our lithium carbonate application guide.

Why Do Fluidized Beds Collapse and How Do You Recover?

Fluidized-bed crystallizers fail in characteristic ways, and every failure mode has a signature in the instrumentation. Learn the signatures and recovery becomes routine rather than dramatic.

1. Bed collapse (de-fluidization) A drop in circulation velocity, a surge of fine solids from the vapor head, or a slug of highly supersaturated liquor can sink the bed into a compacted mass. The signature is a falling differential pressure across the column plus a plugged bottom discharge. Recovery: raise circulation to maximum, back-flush the bottom cone with hot unsaturated liquor, and if the bed has fully set, drain it to the dissolver tank and restart from seed. A collapsed bed costs a shift, not a campaign — if you catch the ΔP trend early.

2. Scaling in the heat exchanger Scaling begins the moment crystals enter the heater, which is a violation of the clear-liquor principle. The early warning is a rising temperature approach across the exchanger at constant duty. Prevention is velocity discipline, correct noncondensable venting, and fines management in the loop. Heater-type selection and scaling behavior are compared in our forced circulation vs falling film article; the same logic governs the OSLO loop.

3. Fines carryover in the overflow When the overflow turns turbid, either the liquor velocity is too high for the current crystal population, nucleation has spiked, or the fines trap has lost efficiency. Turbid overflow also contaminates the downstream liquor circuit and starves the bed of inventory. Fix: cut circulation 10-20%, tighten Δc control, and service the fines trap at the next window.

4. Drainline blockage Bottom drains sit at the densest part of the bed and plug with static crystals during shutdowns. Design drains with a flush-water quench, warm them before opening after a standby, and never leave saturated liquor standing in a drain line through a cold soak.

5. Bed-level hunting and slugging An oscillating bed level usually points to gas entrainment from the vapor head or an unstable circulation control loop. Correct the downcomer submergence, retune the level control, and check the vapor head for vortexing before blaming the bed.

What Does an OSLO Crystallizer Cost to Build and Run?

Capital cost for an OSLO crystallizer typically lands above an equivalent forced-circulation train (indicative): the column is tall, the circulation flow is enormous relative to feed, and the vapor head adds platework and structure. Three items dominate the equipment list — the fluidization column, the external heat exchanger, and the low-head, high-flow circulation pumps. Civil works for the extra height are a real line item that buyers sometimes miss.

how MVR vs Multiple Effect Evaporator is decided

how MVR vs Multi-Effect Evaporation is decided

On the operating side, four drivers set the bill:

Pumping power. Moving thousands of cubic meters per hour against a few meters of head is a continuous multi-hundred-kilowatt load and is usually the largest electrical item on the machine.

Heating energy. The evaporative section is the true energy consumer, which is why an OSLO crystallizer is almost always paired with heat-pump or multi-effect heating. Mechanical vapor recompression is the default on power-rich sites — see how MVR works — and a packaged MVR evaporation-crystallization unit holds specific energy in the tens of kWh per ton of water evaporated (typical). Where steam is cheap and power expensive, a triple-effect evaporation-crystallization unit remains competitive.

Cleaning and maintenance. Exchanger washes, pump seals, and drain flushing are the routine. A well-run bed goes months between cleanings; a misrun one needs them monthly, and the difference is the largest controllable cost in the maintenance budget.

Availability. Every unplanned bed collapse costs a full restart with seeding. The recovery procedures in the previous section exist to protect this number, and they repay the training time many times over.

Operating Data From Fluidized Bed Crystallizers

Published guarantees aside, confidence comes from running machines. The figures below are anonymized and indicative — typical industry values, not a guarantee. Every duty is to be confirmed by test work on the actual liquor.

A composite case illustrates the pattern. A nickel sulfate refinery running two aging forced-circulation lines (d50 around 0.35 mm, typical for that technology) replaced them with a single OSLO crystallizer train heated by MVR. Centrifuge cycle time lengthened from roughly 20 to 55 minutes (indicative), wash-water demand fell by about a third, and the coarse product opened a premium market segment the fines could not serve. The price was a new building bay for the column height — the classic OSLO trade of steel and elevation for crystal quality.

Note how the achieved d50 tracks the duty’s growth kinetics: fast-growing ammonium sulfate reaches the top of the range, slower nickel sulfate sits mid-range, and sodium sulfate depends strongly on the stable phase and temperature window selected.

Frequently Asked Questions

What crystal size can an OSLO crystallizer deliver?
Most duties produce a d50 of 0.5-3 mm (0.02-0.12 in), with 1-2 mm common for ammonium and potassium sulfates and 0.5-1 mm for nickel sulfate. The practical ceiling is settling velocity: crystals much above 3 mm need liquor velocities that start to sweep the bed apart. Treat the upper end as indicative and confirm the achievable size with bench or pilot test work on the actual liquor before writing the specification.

How does an OSLO unit differ from a DTB crystallizer?
Both manage fines, but the geometry differs. The OSLO design keeps the entire crystal inventory in a classified fluidized bed and discharges from the bottom, giving the narrowest CSD and the coarsest product. A DTB machine runs a mixed suspension with an internal fines-settling baffle and an elutriation leg, reaching moderate sizes in a shorter, more compact vessel. Choose OSLO for CSD quality and DTB for plot-plan efficiency.

Is the design suitable for heat-sensitive or high-value salts?
Yes, and this is a classic use case. Because liquor is heated without crystals present and supersaturation is kept low, the design suits heat-sensitive and high-value duties such as nickel sulfate, potassium sulfate, and coarse ammonium sulfate. Growth at low Δc preserves crystal habit and reduces inclusions, so the assay stays high. The trade-off is a taller vessel and more disciplined operation, which premium products easily justify.

How is supersaturation controlled in practice?
Through three coupled levers: the circulation ratio, the heat input to the evaporative section, and fines destruction. A high circulation ratio spreads the generated supersaturation across a large liquor flow, keeping Δc per pass low. Heat input is trimmed so the vapor head works within the metastable zone. Fines destruction removes the nuclei that would otherwise consume the driving force through nucleation instead of growth.

Is pilot test work really necessary before sizing?
Yes. Growth rate G, metastable zone width, and achievable bed concentration are liquor-specific and can vary by a factor of two or more with trace impurities. Sizing errors are expensive because they scale the vessel directly. A bench campaign establishes the kinetic inputs, and a pilot loop verifies bed behavior, fines carryover, and drainability. Most credible suppliers will not guarantee a d50 without at least bench data.

Can the unit be combined with MVR heating?
It can, and usually should be. The clear-liquor circulation loop is an ideal match for mechanical vapor recompression: secondary vapor from the evaporative head is compressed and reused as heating steam, cutting specific energy to a fraction of steam-driven operation. The pairing is standard on large nickel sulfate and sodium sulfate duties where energy dominates operating cost.

What does day-to-day operation demand from a crew?
Mainly vigilance on three instruments: bed differential pressure (crystal inventory), heater temperature approach (scaling), and overflow turbidity (fines). Routine work includes pump and seal checks, exchanger cleaning on a schedule set by the ΔT trend, and exercising the flush drains. A trained crew of two to three operators can typically run a large crystallization train around the clock.

Conclusion

The OSLO crystallizer remains the reference design whenever crystal size is a commercial property rather than a byproduct. Its principle is spatial discipline: generate supersaturation in a separate section, keep the heater crystal-free, and let a classified fluidized bed convert that driving force into coarse, narrow-CSD product at a d50 of 0.5-3 mm (typical). Its sizing is honest — τ = d50/G, holdup equal to production times residence, and a bed volume that scales directly with the kinetics you measure. Where the duty is nickel sulfate, potassium sulfate, or another coarse-crystal specification, no competing machine matches that combination of size and uniformity; where it is not, forced circulation or DTB will serve you better for less. If your project fits the fluidized-bed profile, send us the duty, capacity, and target crystal size through our request a quote page, or explore the companion deep-dives in our resources library covering MVR, sodium sulfate, and lithium carbonate crystallization.

References

• Mullin, J.W., Crystallization, 4th ed., Butterworth-Heinemann.
• Randolph, A.D. and Larson, M.A., Theory of Particulate Processes, 2nd ed., Academic Press.
• Mersmann, A. (ed.), Crystallization Technology Handbook, 2nd ed., Marcel Dekker.
• Perry, R.H. and Green, D.W., Perry’s Chemical Engineers’ Handbook, 9th ed., McGraw-Hill, crystallization equipment section.
• Ulrich, J. and Jones, M.J., “Crystallization,” in Ullmann’s Encyclopedia of Industrial Chemistry.
• Nyvlt, J., Industrial Crystallization from Solutions, CRC Press.

Changelog

• 2026-11-01: First publication.

Related reading

• MVR Evaporator Selection