Mother Liquor Recovery: 7 Causes of Yield Loss

Last updated: 2026-10-29

Mother liquor recovery is where crystallizer yield is won or lost: in a typical evaporative circuit, 3–8% of the dissolved product leaves in the purge stream every hour, and without a recovery step that stream is a permanent write-off (typical range, plant-specific). This guide gives process engineers the impurity balance, the bleed rate calculation and the seven field faults that leak yield — starting from our pillar guide to evaporator and crystallizer systems.

Key Takeaways

• Crystallizer yield loss is usually an impurity problem wearing a yield costume.
• The steady-state bleed rate is set by the impurity balance, not by the budget.
• Throttling the purge looks free for weeks — then the impurity steady state doubles.
• Most yield leaks trace to seven recurring, field-diagnosable faults.
• A secondary crystallizer typically returns 85–95% of purged metal (indicative, to be confirmed by testwork).

What Is Mother Liquor Recovery and Why Does Yield Leak Through the Purge?

Mother liquor recovery is a combination of metered purge control and value-recovery unit operations that returns dissolved product, water and reagent value from the waste stream to the crystallization circuit instead of writing them off with the bleed. A bleed stream is a continuous, metered withdrawal that removes dissolved impurities from the circuit at exactly the rate the upstream process introduces them. A secondary crystallizer is a dedicated cooling or evaporative stage that treats the purge stream on its own to recover product crystals before any final disposal. Total organic carbon (TOC) is a surrogate analytical measurement that expresses every dissolved organic in the liquor as equivalent carbon, and in crystallization circuits it is the cheapest early-warning number you can buy.

how MVR Evaporator Selection is decided

The vocabulary matters because each word points to a different engineering decision. A bleed and a purge are the same act seen from two directions: the process engineer bleeds to hold an impurity at its limit; the accountant sees a purge of saleable product. A recovery step is what converts the purge from a cost line into a partial product stream — it changes the economics of the bleed, never the necessity of it. This is where most yield arguments inside a refinery end badly: production throttles the purge to save metal, the impurity balance quietly re-steadies at double the concentration, and the quality lab starts rejecting product weeks later with no visible cause.

Why does the liquor leak yield at all? A crystallizer only makes pure crystals; everything not crystallised stays dissolved in the mother liquor, at saturation, alongside every impurity the feed brought in. Every pathway that carries liquor out of the circuit — the bleed, the centrifuge heel, condensate entrainment, filter cloth wash — carries saturated product with it. Crystal yield loss through the mother liquor circuit therefore has two faces: dissolved product discarded in the purge, and crystals or liquor reporting to the wrong place (condensate, wash filtrate, product rejects).

One boundary note: this is a troubleshooting article for the circuit around the crystallizer, not a vessel design guide — sizing, agitation and heat-balance design are covered in our crystallizer design resource, and worked mass-balance calculations for a specific salt system live in our sodium sulfate crystallization guide.

How Do You Calculate the Bleed Rate That Holds Impurities at Steady State?

The bleed rate calculation starts from one line of algebra. For a well-mixed liquor volume V with impurity ingress I (kg/h) and bleed flow Q_b (m³/h), the accumulation term is d(C·V)/dt = I − Q_b·C. At steady state, C = I/Q_b: the liquor concentration sits exactly where the ingress and the bleed agree it should. That gives the design rule directly — the minimum bleed for each impurity i is Q_b,i = I_i / C_i,limit, and the circuit bleed is the maximum across all impurities:

Q_b = max ( I_Cl/C_Cl , I_Mg/C_Mg , I_Na/C_Na , I_SO4/C_SO4 , I_TOC/C_TOC )

Worked example (stated basis, indicative): a nickel sulfate circuit ingresses 60 kg/h of chloride from feed brine and acid make-up, and the corrosion- and product-driven limit on mother liquor chloride is 40 g/L. The chloride duty alone demands Q_b ≥ 60/40 = 1.5 m³/h. Magnesium, ingressing 8 kg/h with a 15 g/L liquor limit, demands 0.53 m³/h; sodium dominates neither. Chloride is the controlling impurity, so 1.5 m³/h is the circuit bleed — not a rounder, smaller number chosen for looks.

Two consequences follow, and both are field-verified repeatedly. First, if the purge is cut to 0.75 m³/h, chloride does not stay put — it drifts to a new steady state of 80 g/L, exactly double, on a time constant of roughly V/Q_b. With 900 m³ of liquor inventory and a 0.75 m³/h bleed, that is a 50-day exponential approach: the throttle looks free for six or seven weeks, which is why the fault survives every quarterly review. Second, any mother liquor recovery step that returns dissolved impurity to the circuit raises the required bleed. Crystals are pure — a secondary crystallizer can recover 92% of the nickel in a purge and leave essentially all of the magnesium behind in a much smaller final discard — but any route that returns liquor rather than crystals (ion exchange eluates, membrane permeate recycling decisions) must be counted in the impurity balance honestly.

The table below gives the impurity stack a process engineer should trend weekly. Bands are indicative for nickel and cobalt sulfate circuits; establish your own limits from product specification and corrosion allowances.

Every impurity in that table enters at a rate you can measure and leaves at a rate you control. The seven faults below are the standard ways that control fails — each one is a yield leak with a symptom you can see in routine plant data.

7 Faults That Leak Yield Through the Mother Liquor Circuit

Run the diagnosis in order; the faults are ranked by how often they explain mother liquor recovery losses in base-metal sulfate plants, and the first one is implicated more often than the other six combined.

1. Purge rate throttled to save product. Symptom: weekly metal-in-product reconciliation improves for three to six weeks, then off-spec batches and re-dissolved rejects begin; liquor impurity trend doubles against baseline. Root cause: impurity ingress is unchanged while impurity removal is halved or worse, so the circuit re-steadies at C = I/Q_b with the new, smaller Q_b — impurity accumulation in the crystallizer is arithmetic, not bad luck. Corrective action: meter the bleed against the impurity balance (analyser or density trend), never against the month’s metal accounting; recover the purged value in a secondary crystallizer instead of hoping the stream will behave.

2. Fines recirculation without dissolution. Symptom: particle size distribution widens, mother liquor turns turbid, product filtration slows, and dissolved loss in the purge creeps up as fine crystals escape classification. Root cause: the fines loop is designed to dissolve fine crystals by raising the return-stream temperature above saturation; when the fines dissolver runs cold or short of residence time, fines survive the loop, deplete the crystal surface area, and the smallest leave with classified purge flows. Corrective action: verify dissolver outlet temperature at least 5 K above saturation, hold fines residence at 10–15 minutes minimum (typical design basis), and check the hydrocyclone or elutriator cut size against commissioning records.

3. Crystal carryover into the condensate. Symptom: condensate conductivity climbs, the condensate tank glitters under a torch, and the product balance goes short by 0.5–2% (typical entrainment range). Root cause: in evaporative crystallizers, vapor velocity above design — from fouled calandrias, higher duty or foaming — lifts droplets of saturated liquor past a demister that is blinded, collapsed or overdue for inspection. Corrective action: audit vapor velocities against design, inspect and wash the demister on a fixed schedule, and trend condensate conductivity daily — it is the cheapest entrainment alarm on the plant.

4. Mother liquor density control drifting. Symptom: liquor density reads 0.02–0.05 g/cm³ below setpoint at unchanged steam duty, yield per pass sags, and operators compensate by adding more water, making it worse. Root cause: radiometric or Coriolis density sensors foul with gypsum and silica scale in hot, saturated liquor and drift low, so the control system genuinely believes the liquor is lean and dilutes it further — a control loop actively converting setpoint into dilution. Corrective action: verify density weekly against a laboratory sample at 25 °C, clean the sensor on a schedule that matches your scaling rate, and treat the fouling mechanism itself — our evaporator scaling control guide covers the antiscalant and suppression choices.

5. Excessive wash water diluting the circuit. Symptom: the water balance will not close, evaporation duty creeps upward month on month, and the mother liquor cannot reach setpoint density at design steam flow. Root cause: centrifuge or pusher wash water run well above the 0.1–0.3 kg water per kg crystal that good cake washing needs (indicative), with the excess draining into the mother liquor tank as load the evaporator must remove again — sometimes cold, which doubles the damage as sensible-heat duty. Corrective action: meter wash water, count it explicitly in the circuit balance, wash with hot condensate, and tune wash to cake-moisture targets rather than to “until the cake looks clean”.

6. Cooling-crystallizer ΔT collapse. Symptom: in a cooling stage, outlet liquor temperature approaches inlet temperature, yield per pass drops 30–50%, and the secondary circuit runs hotter every week. Root cause: heat-transfer surfaces foul with gypsum, silica or crystallised product itself, so the working ΔT collapses from a healthy 8–12 K toward 2–4 K (typical field progression); the stage still circulates, but it no longer desupersaturates the liquor, and unsucked yield rides out in the purge. Corrective action: clean heat exchangers on a measured fouling trend rather than a calendar, control supersaturation with seeding where the system allows it, and verify the cooling medium supply temperature before blaming the exchanger.

7. Organic contamination from flocculant or solvent carryover. Symptom: TOC trends upward, the evaporator foams at falling levels, crystal habit shifts toward needles or thin plates, and product filtration slows. Root cause: thickener flocculant overdosed upstream and solvent extraction organics entrained into the liquor adsorb on growing crystal faces, changing habit and occluding mother liquor inclusions — how an impurity at milligrams per liter drives a purity reject measured in grams per kilogram. Corrective action: audit upstream reagent dosing against actual feed solids, install or verify activated-carbon polishing on the liquor, and put TOC on the weekly lab schedule so the trend is visible before the habit change is.

Which Mother Liquor Recovery Technology Fits Your Circuit?

Four recovery routes dominate industrial practice, and the choice is set by purge volume, product value and what the impurities in the purge will tolerate. The comparison below is indicative — efficiencies and energies are typical ranges to be confirmed by testwork on your own liquor.

For most metal-sulfate circuits the secondary crystallizer wins, and it is worth being explicit about why: crystals are the only recovery product that is simultaneously pure and in saleable form. IX returns nickel into an eluate that must be polished and re-crystallised anyway; membranes return a concentrate that re-enters the circuit and re-loads the impurity balance; dryness produces a mixed salt your quality department cannot sell into a battery chain. The staged logic is the same one that governs lithium carbonate circuits, where the purge carries boron and magnesium that must be rejected before the lithium is re-concentrated. Whether the secondary stage is a batch pot or a continuous forced-circulation unit matters less than the ΔT you can hold across it — see our continuous versus batch crystallizer comparison. Where the driver is disposal rather than metal value, the sizing logic shifts to brine concentrator sizing and selection.

One design rule ties the table together: recovery efficiency is not the deciding variable. A 92% recovery of a high-value metal beats a 99% recovery of a mixed salt that sells at disposal prices. Rank the options by recovered value per unit energy, not by percentage alone.

Put as two decision sentences: for a purge of 2–8% of circulating liquor that carries a saleable metal sulfate, a secondary crystallizer is preferred because it returns product at the specification the main circuit already sells. For a dilute purge below roughly 1 g/L metal with low suspended solids, ion exchange is preferred because the thermal duty of crystallizing that stream would exceed the value of the metal it recovers.

How Do You Build the Payback Case for a Recovery Step?

The economic model is four lines, and every plant can compute it from its own data before calling an equipment supplier. Annual recovered value R = Q_b (m³/h) × C_M (kg/m³) × η (recovery fraction) × H (operating hours per year) × P (metal price). Against it: energy cost, reagent and consumable cost, and an O&M allowance. Payback = installed capital cost ÷ (R − annual costs). The honest version uses your own metal price assumption and states the recovery efficiency as a testwork claim, not a brochure number — the energy side is developed further in our evaporator energy consumption guide.

MVR vs Multiple Effect Evaporator in practice

MVR vs Multi-Effect Evaporation, step by step

The punchline for metal circuits is that specific energy almost never decides the payback. Take the worked case in the table below: evaporating one cubic meter of purge water at an MVR duty of 30 kWh/m³ costs roughly USD 2.40 per cubic meter at USD 80/MWh, while that same cubic meter — in a nickel circuit at 80 kg/m³ — carries about USD 1,280 of dissolved nickel. Energy is two-tenths of one percent of the value passing through the pipe. In nickel, cobalt and battery-chain sulfates, the payback is killed only by disposal logistics and product-purity risk, never by the electricity meter; energy economics start to matter when the purge carries sodium or magnesium sulfate rather than metal.

State the assumptions when you present this table: metal price, operating hours, recovery efficiency and energy tariff are all site variables. The structural conclusion travels well — for any liquor carrying more than a few tens of kilograms of metal per cubic meter, the value recovered in a year buys the plant in a fraction of it.

Field Case: 120 t/d Nickel Sulfate, Rejects from 6% to 1.5%

A forced-circulation nickel sulfate hexahydrate crystallizer at a third-party hydrometallurgical refinery (anonymised; figures indicative of a real circuit review we supported, to be confirmed by testwork) shows the whole fault family in one campaign. The plant produced 120 t/d of NiSO₄·6H₂O — roughly 24 t/d of contained nickel — with about 30 m³/h of mother liquor recirculating and a bleed set at 4% of recycle flow, 1.2 m³/h, discharged raw to the effluent plant. Battery grade sulfate purity was the sales specification, with magnesium in crystal limited to 20 mg/kg (plant spec, indicative).

The symptom arrived as rejects: 6% of production re-dissolved each month for magnesium at crystal. Mother liquor Mg had re-steadied near 40 g/L; the purge carried 80 g/L of dissolved nickel straight to waste — about 2.3 t/d of nickel, worth well over USD 13 M/y at then-current prices, leaving through a 50 mm pipe. The counter-pressure was predictable: the operating team had throttled the design bleed twice in the preceding year to protect the monthly metal accounting, and each throttle bought weeks of quiet before the impurity balance re-steadied higher.

The fix was a bleed rate calculation done openly, followed by a 2.7 m³/h bleed — 9% of recycle — feeding a small secondary cooling crystallizer sized on the purge alone, the recovery step the circuit had been missing since commissioning. The secondary stage recovered 92% of purged nickel as saleable crystal (testwork-confirmed on site liquor); net nickel loss fell from 2.3 t/d to about 0.4 t/d, and magnesium re-steadied near 19 g/L. Rejects fell from 6% to 1.5% of production within two months of the new steady state and held. The plant’s own metallurgist summarised it in one line: the bleed was never a cost — discarding it unrecovered was. Black-mass and battery-recycling refineries run into the identical impurity stack-up, which we treat in our battery recycling application notes.

Frequently Asked Questions

What is the difference between a bleed and a mother liquor recovery step?

A bleed is the deliberate, metered withdrawal of liquor that holds every impurity at its steady-state limit; it is a control action, and it is mandatory in any circuit with impurity ingress. A recovery step is a set of downstream operations — usually a secondary crystallizer — that processes the bleed to recover dissolved product before final disposal. The bleed protects purity; the recovery protects economics. Circuits need both, and neither substitutes for the other.

How do I size the minimum bleed rate for my crystallizer circuit?

Measure the ingress rate of every significant impurity — chloride, magnesium, sodium, sulfate acidity, TOC — in kilograms per hour from feed analyses and reagent additions. Divide each by its allowable mother liquor concentration in kilograms per cubic meter. The largest result is the circuit bleed, because one uncontrolled impurity ruins the product regardless of how well the others are managed. Re-run the calculation whenever the feed source, neutralant or upstream flowsheet changes.

How much metal value sits in a typical mother liquor purge?

For a nickel or cobalt sulfate circuit at 60–90 g/L dissolved metal, every cubic meter purged carries roughly USD 1,000–1,500 of metal at recent prices (indicative, price-dependent). A 1–3 m³/h purge therefore discards USD 25,000–100,000 of value per operating day if sent unrecovered to effluent. This is why a purge stream should be sampled, assayed and priced like a product stream — because that is what it is.

Can ion exchange replace a secondary crystallizer in mother liquor recovery?

Sometimes, but rarely as a replacement. Ion exchange recovers 90–98% of dissolved metal from dilute streams and returns it as a regenerant eluate, which must then be polished and crystallised anyway. IX suits purges with low suspended solids and modest flow where the metal value is high and the liquor is too dilute to crystallise economically. For saturated sulfate liquor at tens of grams per liter, direct secondary crystallization is simpler and returns saleable crystal directly.

What impurity limits matter most for battery-grade nickel sulfate?

The binding limits are usually magnesium, calcium and chloride in the product crystal, with typical specifications at the low milligrams-per-kilogram level (plant- and customer-specific). Because impurities co-crystallise or occlude in proportion to their mother liquor concentration, the liquor-side control band is what actually protects the product spec. Trend liquor Mg and Cl weekly against a limit derived from your own partition data rather than from generic literature values.

Is crystallizer condensate safe to reuse as wash water?

Usually yes, and it is often the best available source — hot, soft and low in dissolved salts, which suits both cake washing and liquor dilution duties. The caveat is entrainment: condensate can carry fine crystals and dissolved product when demisters degrade, so trend its conductivity before reusing it on product washing. Contaminated condensate puts your yield leak back into the circuit as an impurity and a loss at the same time.

What energy does a mother liquor recovery step consume?

A secondary cooling crystallizer is modest: 15–40 kWh per tonne of liquor circulated for pumping and agitation (indicative). An evaporative secondary stage runs 0.35–0.6 t of steam per tonne of water evaporated, or 25–45 kWh/m³ with mechanical vapor recompression. Against a purge carrying 60–90 g/L of nickel or cobalt, the energy cost is typically under one percent of the recovered value — which is why payback cases for metal-bearing purges close on value, not on energy.

If your purge numbers resemble fault one above, the next step is arithmetic, not procurement: send us your liquor analysis, purge flow and impurity ingress, and the Lexin-tech engineering team will run the impurity balance and recovery duty with you — request a quote to start that review. The supporting references for evaporation and crystallization practice are collected in our resource library.

Conclusion: Treat the Bleed as a Product Stream

Mother liquor recovery is the control system that lets a crystallizer run the impurity balance you actually designed instead of the one the monthly budget prefers: size the bleed from the impurity balance, protect it from throttling, and put a secondary crystallizer on the purge so the yield leak becomes a second product stream. Audit your circuit against the seven faults this week — most plants find that fixing fault one and fault four pays for the review before the next quarter closes.

References

1. Perry, R.H. and Green, D.W., Perry’s Chemical Engineers’ Handbook, 9th edition, McGraw-Hill Education, 2019 — crystallization and evaporation sections.
2. Myerson, A.S. (ed.), Handbook of Industrial Crystallization, 3rd edition, Cambridge University Press, 2019.
3. Mersmann, A. (ed.), Crystallization Technology Handbook, 2nd edition, Marcel Dekker, 2001.
4. U.S. Geological Survey, Mineral Commodity Summaries 2025 — Nickel, https://pubs.usgs.gov/periodicals/mcs2025/mcs2025-nickel.pdf
5. Society for Mining, Metallurgy & Exploration (SME), SME Mineral Processing Handbook, Society for Mining, Metallurgy & Exploration, Littleton, Colorado.
6. Kirk-Othmer, Encyclopedia of Chemical Technology, 5th edition, John Wiley & Sons — entry “Crystallization”.

Changelog

• 2026-10-29: Initial publication — impurity balance, bleed rate calculation, seven fault diagnoses, recovery technology comparison, payback model and anonymised nickel sulfate case.

Related reading

• Crystallizer Design