Evaporator Crystallizer Systems for Mining Brines | Guide

Last updated: 2026-10-18 · Reviewed by: Senior Metallurgical Engineer, {BRAND}

A mechanical vapour recompression evaporator crystallizer typically evaporates water at 15–30 kWh of electricity per tonne of water, whereas a four-effect steam train consumes 0.35–0.6 t of steam per tonne evaporated (typical ranges; basis: 5–10 t/h water duty, brine boiling point rise 5–8 K, saturated steam at 6 bar(g), grid power). Which train wins is decided by energy prices, boiling point rise and the crystal size specification.

Key Takeaways

• An evaporator crystallizer train splits into cheap concentration duty and expensive crystal production duty.
• Boiling point rise, not flow rate, sets the compressor and effect count more than any other parameter.
• MVR wins on cheap power and moderate BPR; multiple-effect wins where waste steam already exists.
• Forced circulation is the default when solids must be made; OSLO buys size, DTB buys throughput.
• Material choice and fouling control decide run life; both must be settled before CAPEX is compared.

What Does an Evaporator Crystallizer Actually Do?

The phrase covers two distinct duties that are often confused in tenders. The first is concentration: removing water from a dilute process liquor to shrink downstream volume, recover reagents or raise tenor ahead of a precipitation or crystallization step. The second is crystal production: taking the liquor past saturation and growing a saleable or disposable solid product with a defined crystal size distribution. An evaporator crystallizer is a train that performs both duties in one thermal circuit — concentrate, then supersaturate, then suspend and grow crystals, then discharge a pumpable slurry. In a zero liquid discharge scheme the same train also has a third, quiet job: delivering a mother liquor clean enough, or a salt dry enough, that the site can stop building ponds.

Three definitions carry the technical weight of this page, and vendors use all three loosely.

Mechanical vapour recompression (MVR) is a vapour-side heat pump that compresses the boil-off vapour a few degrees above its saturation temperature and returns it as the heating steam, so that the energy entering the system is shaft power rather than fresh steam. The working principle, compressor types and their limits are set out in how MVR works; the equipment family is detailed on the MVR evaporator page.

Boiling point rise (BPR) is a colligative property that lifts the boiling temperature of a brine above that of pure water at the same pressure, and it grows steeply as the liquor approaches saturation. BPR is the single most under-rated number in a quotation: it sets the effective temperature difference available to drive heat transfer, and it caps how far a single compressor stage can lift the vapour.

A forced circulation crystallizer is a slurry-recirculating vessel that pumps the magma through an external heat exchanger at high velocity and flashes it in a low-pressure volume, keeping the tubes swept clear of solids while supersaturation is released in the flash rather than on the heat-transfer surface.

One more term, because it appears on every data sheet: steam economy is a performance ratio that expresses kilograms of water evaporated per kilogram of live steam consumed. The duty split matters commercially because the two halves of the train are priced differently: clarification, preheating and falling-film pre-concentration are cheap per tonne of water, while producing crystals with a controlled size is not. The last 10% of the water in a ZLD application typically costs more per tonne than the first 60%, and the train should be split accordingly rather than forced through one machine.

Which Evaporator Crystallizer Equipment Families Cover the Duty?

No single machine covers the whole curve from dilute feed to dry salt. Every evaporator crystallizer train in practice is an assembly of families, each doing what it does best, and the selection question is where to draw the boundaries. The families and their typical roles are:

• Falling film evaporator — clarified liquor flows as a film down the inside of long tubes. Highest heat-transfer coefficient and lowest residence time, so it is the standard pre-concentration stage; it is not suitable where crystals would bridge the tubes, so it stops short of saturation.
• Multiple-effect evaporator (MEE) — several vessels in series, each boiling at a lower pressure than the previous one, so the same kilogram of steam is reused effect by effect. Mature, power-light, and the default where site steam is cheap or surplus. The trade-offs against MVR are worked through in MVR vs multiple-effect evaporation.
• MVR evaporator — one or two large vessels with a mechanical compressor; buys energy independence from the steam system at the price of electrical infrastructure and compressor maintenance.
• Forced circulation evaporator / crystallizer — the workhorse for making solids in chloride and sulphate brines; robust to scaling, modest crystal size. Equipment detail is on the forced circulation evaporator page.
• DTB crystallizer (draft tube baffle) — an internal circulator with a baffle zone for fines removal and destruction; higher throughput and tighter control than simple forced circulation.
• OSLO crystallizer — a fluidised-bed design where crystals grow in a gentle bed and classify on size; the choice when a coarse, narrow crystal size distribution justifies the cost. Seeding, fines destruction and internals practice are covered in crystallizer design fundamentals.
• Brine concentrator — a vertical-tube, high-recirculation MVR machine pushed to near-saturation on high-BPR brines; common in lithium and coal-chemistry duty.
• ZLD train — the full assembly: pre-treatment and hardness removal, falling-film pre-concentration, the crystallizer itself, then thickening, centrifuge or filter, and drying. The overall architecture is described on the zero liquid discharge applications page, and supporting notes live in the resource library.

Table 1 puts indicative numbers against the families so the boundaries can be seen in one view.

Table 1 — Indicative equipment comparison (basis: clarified chloride/sulphate brines, water duty 5–50 t/h, typical vendor practice; confirm by quotation)

The CAPEX column explains most engineering arguments in this field: a falling film stage removes water at roughly half the cost per tonne of a forced circulation vessel, so every tonne of duty that can be done upstream of the crystallizer is a tonne bought cheaply.

Which Design Parameters Decide the Outcome?

Eight parameters decide whether an evaporator crystallizer quotation will perform, and they interact more than tender documents admit.

MVR Evaporator Selection

MVR vs Multi-Effect Evaporation: selection rules

Boiling point rise. BPR at the boiling surface — not in the feed tank — sets the driving force. A brine that reads 3 K in the feed tank at 10% solids can read 9 K in the crystallizer at 28% solids, and a single-stage MVR compressor needs roughly 4–6 K of lift just to overcome BPR plus a temperature margin for heat transfer. Two-stage or twin-screw compressors extend this, at cost. The practical judgement from operating plants: BPR should be measured on clarified mother liquor from test work, not calculated from feed analysis, because concentration polarisation in a circulating tube adds 1–2 K that feed-based numbers never see.

Evaporation duty. Expressed in t/h of water, this sets the number of vessels rather than their sophistication. Above roughly 40–60 t/h, trains are split into multiple lines because single-vessel diameters and compressor sizes stop scaling economically.

Specific energy. kWh per tonne of water evaporated for MVR; tonnes of steam per tonne of water for MEE. These two are only comparable after converting steam to its equivalent energy cost — the comparison that decides most projects, covered under the selection logic below.

Steam economy. Nameplate economy for a four-effect train is often quoted at 3.5–4.0 kg of water per kg of steam; installed consumption including preheaters, stripping, ejectors, start-up and fouling margins typically lands at 0.35–0.6 t of steam per tonne of water. Ask which number a quotation is using.

Crystal size distribution (CSD). Specified as d50 and coefficient of variation. The crystallizer type fixes what is achievable: OSLO produces the coarsest and narrowest bed, DTB is intermediate, simple forced circulation is the widest and finest. Nucleation control (fines destruction, seeding rate) moves the result within each family by roughly ±30%.

Residence time. Crystal growth needs time; typical magma residence is 1–3 h in forced circulation and DTB service, and short-circuiting the discharge faster than that produces fines at the centrifuge no matter what the vendor curve says.

Slurry density. Typically 10–30 wt% suspended solids in the crystallizer body; below about 10 wt% secondary nucleation dominates and CSD degrades, above about 35–40 wt% the circulating pump and the centrifuge both struggle. This is the parameter most often left out of vendor heat-and-material balances and the one that decides whether the centrifuge downstream will actually work.

Table 2 — Key design parameters with typical ranges (basis: chloride/sulphate mining brines, typical industry practice; verify with bench and pilot test work)

How Do Chloride Brines and Materials of Construction Interact?

Mining brines concentrate everything the ore released, and the wetted alloy sees the peak of that concentration at the hottest, most saturated point of the train. The material question in evaporation service is therefore simpler but sharper than elsewhere in the plant: it is almost entirely a chloride-versus-temperature-versus-pH decision.

Titanium is the default for hot, oxidising, near-neutral to mildly acidic chloride duty — the boiling surface of a lithium or magnesium-bearing brine concentrator at 90–110 °C is classical titanium territory, usually Grade 2, occasionally Grade 12 where crevice corrosion is a concern. Duplex and super duplex stainless steels (2205, 2507) carry most of the pre-concentration duty at moderate chloride and temperature and cost far less than titanium per vessel. Where pH is depressed and chloride is high at the same time — acid mine drainage concentrates, some rare-earth liquors — nickel alloys enter the shortlist and the cost of the crystallizer body can double. Vapour spaces and condensate sides are not exempt: vapour from a chloride brine carries entrained droplets, and condensate corrodes carbon steel condensers quickly unless the demisters and washing arrangements are right.

Two cautions earned in service. First, the alloy that is correct at the crystallizer is often over-specified for the pre-concentration train and under-specified at the salt discharge handling — specify per zone, not per train. Second, galvanic couples between titanium shells and duplex internals concentrate attack on the smaller member; design the internals with the couple in mind. The full treatment of alloys and corrosion allowances is a separate discipline and is covered in the pressure vessel materials article on this site; for scaling — the other half of the wetted-surface problem — see evaporator scaling and fouling control.

How Is Scaling and Fouling Controlled Over the Run Life?

Scaling — not corrosion — ends most evaporator crystallizer campaigns early. The deposit chemistry is predictable: calcium sulphate, calcium carbonate and silica account for the large majority of fouling in mining brine service, and each has a different lever. A run-life programme that works looks like this:

1. Map the supersaturation drivers. Tag where in the train each scale-forming species crosses its solubility limit — usually the flash surface of the last effect and the heat-transfer surface of the crystallizer heater. Fouling that appears somewhere else is usually transport of fines or organics, not in-situ growth.
2. Intercept hardness upstream. Lime–soda softening, ion exchange or nanofiltration ahead of the train removes the problem at a fraction of the cost of removing deposits from titanium tubes. Every mg/L of calcium removed upstream is run life bought cheaply.
3. Seed where supersaturation cannot be avoided. Recycling fine crystal seed into the circulating slurry gives calcium sulphate a place to grow that is not the tube wall. This is standard practice on gypsum-saturated brines and the reason many forced circulation trains run six to twelve months between washes.
4. Keep tube velocity above the deposition threshold. In forced circulation heaters, 2.0–2.5 m/s is usually the floor; dropping below it during turndown converts a swept surface into a nucleation site within weeks. Interlock the minimum circulation flow to the heat input, not to operator attention.
5. Design the wash, not just the operation. An on-line boil-out with dilute acid or chelant, valved and piped from day one, converts a three-day shutdown into a four-hour wash. Trains that were not piped for cleaning are cleaned late and cleaned badly.
6. Watch the mother liquor, not just the tubes. Silica and sulphate accumulate in closed recirculation; a weekly liquor analysis catches the drift months before the first ΔT loss shows on the DCS.

The economics are one-sided: fouling control hardware and test work typically cost 1–3% of train CAPEX and routinely return their cost in the first avoided wash outage, because a forced outage on a 20 t/h train costs production and steam simultaneously.

Which Evaporator Crystallizer Train Wins in Practice?

Conditional conclusions, because the honest answer is a decision tree — but a decision tree with definite leaves.

Choose MVR when electricity is cheap or self-generated (mine power, behind-the-fibre renewables at USD 0.05–0.08/kWh), steam is expensive or absent, the brine BPR stays below about 8–10 K, and the duty is continuous and steady. MVR also wins on emissions accounting where the site is decarbonising, and it de-risks a project that would otherwise need a new boiler house. It loses when the crystallization duty is large relative to the concentration duty, because compressors do not tolerate high-BPR, high-slurry service.

Choose multiple-effect when site steam already exists (a smelter, a SART circuit, a coal-fired power neighbour), the duty is large, and the power supply is weak or unreliable. MEE is the lowest-technology-risk option, tolerant of dirty steam and repairable by any mechanical contractor. Its floor is the steam price: at more than about USD 25–30/t of steam, the four-effect train usually loses the life-cycle cost argument to MVR.

Choose steam-driven forced circulation in multi-effect arrangement when BPR is high (10–15 K), solids production dominates, and the feed is dirty. This is the classic salt-from-brine configuration: no compressor to foul with entrained liquor, every effect is a crystallizer, and CSD requirements are modest.

Between OSLO and DTB: choose OSLO when the product specification demands coarse, narrow crystals (potash, some sulphate salts sold on size), throughput is moderate, and operators are available to manage a fluidised bed. Choose DTB when throughput and availability matter more than crystal perfection, when fines destruction must be built in, and when the feed chemistry varies — DTB forgives more and produces more.

Brine concentrators enter when the train must reach near-saturation on a high-BPR brine and crystallization is handled separately, which is common in lithium spodumene and DLE circuits. The full comparison framework is developed in MVR vs multiple-effect evaporation, and the individual machine pages — MVR evaporator and forced circulation evaporator — carry the device-level detail.

What Do Energy and Cost Ranges Look Like, and on What Basis?

Ranges, clearly labelled, because absolute numbers without a basis are marketing.

Energy. MVR: 15–30 kWh per tonne of water evaporated, typical for a two-stage centrifugal or twin-screw compressor with 6–12 K total lift (BPR plus heat-transfer margin). Four-effect MEE: 0.35–0.6 t of steam per tonne of water, installed basis. At USD 0.07/kWh electricity and USD 20/t steam, that is roughly USD 1.1–2.1/t of water for MVR against USD 7–12/t for the steam train — the reason MVR has taken most of the new-build market in power-rich regions, and the reason it has not in steam-rich ones.

CAPEX. Indicative FOB equipment ranges: USD 0.8–2.0 million per t/h of installed water duty for multi-effect concentration trains, and USD 1.2–2.5 million per t/h where crystallization duty and exotic alloys (titanium, super duplex) are included. Basis: chloride brine, FOB ex-works, excludes civils, installation, high-voltage supply and spares; treat as ±30% screening figures to be replaced by vendor quotations, never as budget authorization.

Operating cost. As a rule of thumb from operating plants, total OPEX per tonne of water evaporated (energy, labour, consumables, maintenance) runs 1.6–2.2× the energy line alone; membrane pre-treatment and upstream softening shift this ratio down by removing wash chemistry.

Where the crossover sits. With electricity at E USD/kWh and steam at S USD/t, MVR beats four-effect MEE on energy cost when E × 22.5 (mid-range kWh/t) is less than S × 0.48 (mid-range t/t) — in plain terms, MVR wins when electricity costs less than about USD 0.045 per kWh per USD 2/t of steam, all other things equal. BPR above ~10 K and crystallization duty above ~30% of train CAPEX move the crossover against MVR.

FAQ

Q1. When does MVR beat multiple-effect evaporation?
MVR wins when electricity is cheap relative to steam, typically below USD 0.05–0.07/kWh against steam above USD 20/t, when brine boiling point rise stays under about 8–10 K, and when the duty is continuous. It also wins where a site has no boiler house and building one would dominate CAPEX. Multiple-effect keeps the advantage where surplus steam already exists, where power is unreliable, and where the crystallization duty is too large for compressor-based machines.

Q2. What boiling point rise makes MVR uneconomic?
Above roughly 10–12 K of boiling point rise at the crystallizer, a single-stage compressor cannot deliver a useful temperature difference and a two-stage machine buys lift at falling efficiency. Near-saturated sodium or magnesium chloride brines with 12–15 K BPR usually push the design to steam-driven forced circulation or a dedicated brine concentrator. Always measure BPR on clarified mother liquor from test work rather than calculating it from feed analysis.

Q3. When should OSLO be chosen over DTB?
Choose OSLO when the crystal size specification is the commercial constraint — coarse, narrow distributions with d50 above roughly 1.5 mm, as required for some potash and sulphate products — and throughput is moderate. Choose DTB when availability and tonnage dominate, when the feed chemistry drifts seasonally, or when fines destruction must be integral to the machine. OSLO costs more per vessel and needs more skilled operation in exchange for the size advantage.

Q4. What slurry density should a forced circulation crystallizer run at?
Typically 10–30 wt% suspended solids. Below about 10 wt%, secondary nucleation dominates and the crystal size distribution degrades; above 35–40 wt%, the circulating pump loses head and the centrifuge cycles badly. The number must appear in the vendor’s heat-and-material balance, not just the guarantee — it is the parameter most often discovered missing when the downstream centrifuge fails to meet capacity.

Q5. How much water must a zero liquid discharge train recover?
Functionally all of it: ZLD means no liquid leaves site except product, so the train must recover 98–99.9% of feed water depending on pond losses and blowdown philosophy. The economics concentrate in the tail: the first 60% of water is cheap falling-film duty, while the last 10% passes through the crystallizer at several times the cost per tonne. Softening and pre-concentration capacity should be sized generously to keep the crystallizer small.

Q6. Which materials suit hot saturated chloride brine service?
Titanium Grade 2 is the default for the boiling surfaces of hot, near-neutral chloride brines, with Grade 12 where crevice corrosion is a concern. Duplex 2205 covers most moderate-temperature pre-concentration duty, and super duplex 2507 extends the envelope. Acidic high-chloride liquors push toward nickel alloys and a step change in cost. Specify alloy per zone, verify with coupon testing in the actual liquor, and design out galvanic couples between titanium and stainless internals.

Q7. What data is needed before sizing an evaporator crystallizer?
Minimum: full feed chemistry with seasonal variation, target production rate and crystal specification, site energy prices and availability of both power and steam, water balance, and boiling point rise measured on mother liquor at the expected terminal concentration. With those, a credible vendor can commit to an evaporator crystallizer configuration within weeks. Without them, any quotation is a placeholder — budget and schedule both benefit from funding a bench-scale evaporation and crystallization test programme first.

Conclusion

An evaporator crystallizer project is won or lost on three decisions made before any equipment is priced: how the duty is split between cheap pre-concentration and expensive crystallization, which energy system — compressor or steam — matches the site’s price reality, and whether the crystal specification really requires an OSLO or whether a well-run forced circulation or DTB machine will do. Get those three right and the train will run for decades; get one wrong and no vendor guarantee repairs it. Measure boiling point rise on real mother liquor, put slurry density in the heat balance, and pipe the wash before commissioning day.

If you are assembling a complete train — softening, evaporation, crystallization, solid-liquid separation and drying — and want it engineered, procured and commissioned as one responsibility, review our turnkey hydrometallurgical plants service and send us the brine analysis and water balance for a configuration proposal.

References

1. USGS, Mineral Commodity Summaries 2025 — lithium, potash and salt production context for brine processing demand. usgs.gov
2. Wills, B.A. & Finch, J.A., Wills’ Mineral Processing Technology, 8th ed., Elsevier — chapters on hydrometallurgical circuits and dewatering.
3. Perry, R.H. & Green, D.W., Perry’s Chemical Engineers’ Handbook, 9th ed., McGraw-Hill — sections 12 (evaporators) and 18 (crystallization).
4. ALTA Metallurgical Services, ALTA 2024 Proceedings — uranium/REE/nickel-cobalt sessions; brine concentration and crystallization practice papers.
5. MetSoc, Conference of Metallurgists (COM) Proceedings — lithium brine processing and ZLD case studies.
6. CSIRO Mineral Resources — brine chemistry, scaling and hydrometallurgical process testing. csiro.au

Changelog

• 2026-10-18 — Initial pillar refresh: equipment family comparison, design parameter tables, selection decision logic, energy and cost ranges with stated basis.

{
“@context”: “https://schema.org”,
“@graph”: [
{
“@type”: “TechArticle”,
“headline”: “Evaporator Crystallizer Systems for Mining Brines | Guide”,
“about”: [“evaporator crystallizer”, “mechanical vapour recompression”, “multiple-effect evaporation”, “forced circulation crystallizer”, “zero liquid discharge”],
“datePublished”: “2026-10-18”,
“dateModified”: “2026-10-18”,
“author”: {“@type”: “Person”, “name”: “Process Metallurgist, {BRAND}”},
“publisher”: {“@type”: “Organization”, “name”: “{BRAND}”}
},
{
“@type”: “FAQPage”,
“mainEntity”: [
{“@type”: “Question”, “name”: “When does MVR beat multiple-effect evaporation?”, “acceptedAnswer”: {“@type”: “Answer”, “text”: “MVR wins when electricity is cheap relative to steam, brine boiling point rise stays under about 8–10 K, and the duty is continuous; multiple-effect keeps the advantage where surplus steam exists or power is unreliable.”}},
{“@type”: “Question”, “name”: “What boiling point rise makes MVR uneconomic?”, “acceptedAnswer”: {“@type”: “Answer”, “text”: “Above roughly 10–12 K at the crystallizer, a single-stage compressor cannot deliver a useful temperature difference and two-stage machines lose efficiency; near-saturated brines usually force steam-driven forced circulation or a brine concentrator.”}},
{“@type”: “Question”, “name”: “When should OSLO be chosen over DTB?”, “acceptedAnswer”: {“@type”: “Answer”, “text”: “Choose OSLO when coarse, narrow crystal size distributions (d50 above roughly 1.5 mm) are the commercial constraint and throughput is moderate; choose DTB when availability, tonnage and feed variability dominate.”}},
{“@type”: “Question”, “name”: “What slurry density should a forced circulation crystallizer run at?”, “acceptedAnswer”: {“@type”: “Answer”, “text”: “Typically 10–30 wt% suspended solids; below 10 wt% secondary nucleation degrades CSD, above 35–40 wt% pump and centrifuge performance suffer.”}},
{“@type”: “Question”, “name”: “How much water must a zero liquid discharge train recover?”, “acceptedAnswer”: {“@type”: “Answer”, “text”: “Functionally all of it: 98–99.9% of feed water depending on losses, with the last 10% handled in the crystallizer at several times the cost per tonne of the first 60%.”}},
{“@type”: “Question”, “name”: “Which materials suit hot saturated chloride brine service?”, “acceptedAnswer”: {“@type”: “Answer”, “text”: “Titanium Grade 2 for hot boiling surfaces, Grade 12 where crevice corrosion is a concern, duplex 2205 and super duplex 2507 for moderate-temperature pre-concentration, nickel alloys for acidic high-chloride liquors.”}},
{“@type”: “Question”, “name”: “What data is needed before sizing an evaporator crystallizer?”, “acceptedAnswer”: {“@type”: “Answer”, “text”: “Feed chemistry with seasonal variation, production and crystal specification, site energy prices, water balance, and boiling point rise measured on mother liquor at terminal concentration.”}}
]
},
{
“@type”: “BreadcrumbList”,
“itemListElement”: [
{“@type”: “ListItem”, “position”: 1, “name”: “Home”, “item”: “/”},
{“@type”: “ListItem”, “position”: 2, “name”: “Evaporation & Crystallization”, “item”: “/evaporation-crystallization/”},
{“@type”: “ListItem”, “position”: 3, “name”: “Evaporator Crystallizer Systems for Mining Brines”, “item”: “/evaporation-crystallization/”}
]
}
]
}

Related reading

• Crystallization