September 3, 2026. Brasília. The Senate said yes. Brazil’s critical-minerals bill — a legal framework for strategic minerals, a credit fund for mining projects, tax incentives for processing at home — cleared its final congressional vote and went to President Luiz Inácio Lula da Silva for signature. The lower house had passed it back in May. Behind the bill sits the fact the whole market works from: the world’s second-largest rare-earth reserves, behind only China’s. And one producer already at commercial scale — Serra Verde in Goiás, fresh off a deal worth around USD 2.8 billion with USA Rare Earth to build a magnet supply chain.
Producing at home means precipitation circuits, and precipitation circuits live or die at the wash. Wash harder and the cake gets cleaner; that part is free. The bill arrives in the filtrate — dissolved metal carried out with the wash water, a diluted stream, more liquid for the recovery circuit to swallow. The arithmetic is plain. One tonne of precipitate cake holding 150 litres of mother liquor at 40 g/L carries 6 kg of dissolved rare earths in that liquor. Displace 80 percent of it and 4.8 kg reports to the wash filtrate. Push to 95 percent and 5.7 kg reports there. The last 0.9 kg is the expensive one: it drinks a disproportionate share of the wash water and pays it back at a weaker grade than the stream it joins. Purity and recovery pull the same stage from opposite ends.
Decide which result the wash stage must protect
Start with the product specification
Fix the product limit first. Which dissolved impurity governs acceptance — residual salts, acidity, one named element the next step cannot tolerate? Until that number exists, “cleaner” is a direction, not a duty. The limit comes from the process route and the buyer’s specification. The filter supplier does not set it. A carbonate feeding calcination and separation lives under one limit; an oxalate on a different flowsheet lives under another. The wash stage serves the route it actually feeds.
With the limit fixed, the second boundary appears. Washing displaces mother liquor, and the liquor carries dissolved rare earths. That metal is recoverable on one condition: somebody collects the stream, assays it, routes it back to a step that can take it. Unmeasured, it leaves. A precipitate can pass its purity test and still bleed money through a wash stream nobody sampled.
So the decisions fall in a fixed order. One: the cake quality the product needs. Two: how much dissolved metal the wash may carry away. Three: where that liquid goes next. Wash-water volume adjusts last, against those three answers. Adjust it first and the plant tunes blind.
Measure cake and filtrate as separate outputs
Sample the two outputs as what they are. Cake: residual impurity against the product limit, wet mass, moisture. Liquids: feed filtrate and wash filtrate apart, each assayed for dissolved metal. Put the results side by side in the review — cake quality on one line, metal balance on the next.
One residual number cannot carry both stories. The cake improves because more liquor got displaced. The displaced liquor went somewhere, at some concentration, in some volume. On the opening example, the run from 80 to 95 percent displacement moved 0.9 kg more metal per tonne of cake and grew a bigger, weaker stream doing it. A review showing only the cake number hides both halves. Processing economics at home in Brazil will be built on numbers like those. Cake colour decides none of it.
Use the press cycle to separate washing from dewatering
Washing sits between two pressure stages
On our HDLY Automatic Vertical Press Filter, the full cycle runs six stages — Filtering, 1st Squeeze, Washing, 2nd Squeeze, Drying, Discharging. Washing gets its own stage, after the first squeeze, before the second. The placement is practical. The wash result reads on its own, before squeeze two and the air drying change the moisture and muddy the comparison.
Wash liquid enters through the same manifold pipe as the slurry feed. It crosses cake and cloth, then leaves by the discharge pipe. The cake is part of the washing equipment. On the HDLY the cloth lies horizontal, cake forming evenly to 45 mm, so liquid faces a similar path across the whole chamber. It cannot hunt for thin spots. The chamber offers none.
Two nameplate figures frame the duty. Maximum pressure 1.6 MPa. Cake moisture down to 10 percent or less. Neither predicts how a given precipitate washes. A coarse, well-aged oxalate drains like a different material from a fine, gelatinous carbonate precipitated cold and fast. The press figures stay still. The result moves.
Keep the comparison tied to the same precipitate
Hold the material steady while wash settings change. Precipitation pH, reagent dose, aging time, solids content — all four move cake structure and the paths liquid takes through it. Let them drift between runs and the difference lands in the wash column of the results table. It does not belong there.
Compare few conditions. Record each one fully: cake impurity, wash filtrate volume and metal content, moisture, the time every stage took. The deliverable is an operating range, inside which this precipitate meets its product limit while metal leaves in a stream concentrated enough to be worth recovering. A universal recipe does not exist. A trial promising one is selling something.
Filter cloth is part of the purity and recovery balance
Select for the actual precipitate and liquor
The cloth holds two jobs that fight each other. Hold back fine precipitate. Let liquid through fast enough to keep the cycle economic. Our vertical-press range spans 5 to 50 μm in filtration fineness and 5 to 300 L/m²·s in air permeability. Fineness speaks to retention. Permeability speaks to flow.
The catalogue spread is real. V11P05HC, multifilament PP, filters at 5 μm and breathes at 5 L/m²·s. The polyester V12S025H runs 20 μm and 25, for strong-acid and weak-alkali duty. Choose on fineness alone and the plant can land clear filtrate with a dead cycle time. Choose on permeability alone and solids turn up in the filtrate. A gelatinous precipitate sharpens the choice further — it blinds a tight cloth within cycles, while a coarse crystalline one slips an open cloth’s pores before the cake bridge even forms.
Material is the third leg. We supply PP, polyester and nylon. PP multifilament rates “Excellent” against acid and alkali. Rare-earth liquors are aggressive and specific — hot chloride and sulfate streams, acid liquors, ammonia-bearing washes. Match the material to the actual stream and its temperature. Run the candidates against the real precipitate before anyone fixes a specification. The full NHD filter cloth range is set out with its materials and duty data, vertical filter presses among the equipment it serves.
Check the filter cake release between cycles
What stays on the cloth after discharge joins the next cycle uninvited. A retained layer narrows the working surface, shifts the flow, quietly rewrites the wash result the team believes it is repeating.
On the vertical press the cloth cleans on both sides in turn as it travels. In our plant the fabric is calendered — that is what improves cake release and holds filtration precision across the width — and every finished cloth is tested before it leaves: weight, thickness, density, filtration precision, tear strength, air permeability.
In a trial series the cloth is a recorded variable. Condition and discharge behaviour go in every run’s notes. A precipitate releasing cleanly on cycle one and sticking by cycle six is reporting blinding that no single-run test ever shows. And a cloth swapped mid-series to test a candidate is its own experiment. It stays out of the wash-water comparison.
Route the displaced metal before increasing wash demand
Confirm the destination for wash filtrate
Before anyone reaches for more wash water, answer a plumbing question. Where does the wash filtrate go — back to an upstream recovery step, into another liquor stream, out to separate treatment? Each answer carries capacity: tanks, pumps, downstream load.
A dilute stream hauling the same metal at twice the volume behaves like a different stream. The recovery circuit will treat it as one. Returned to precipitation, its volume inflates the next batch. Sent to solvent extraction, its acidity and impurity load arrive with it. None of this is philosophy. It is pipework and tankage.
This is where the balance stops being a laboratory question. The displacement that shaves the last impurity off the cake can cost more in recovery load than it earns in product grade. Cake quality and filtrate composition measured on the same runs put that trade in numbers. Arguments about principle become unnecessary.
Put the trade-off into the equipment duty
Write the duty the way the press will meet it. Precipitate feed and its variation. Cake quality limit. Metal accounting boundary. The wash stage required. The route the displaced liquid takes.
Against that duty, filtration area, chamber sequence, cloth and automation get discussed as one machine doing one job. The HDLY runs its cycle from a control panel, self-diagnosis and automatic alarms included, so the sequence repeats. Targets and stream routing stay with the plant, where they belong.
A stated duty protects the comparison of offers, too. One proposal prices six stages, washing included. Another quietly prices four, without it. Those are two machines, not two prices. Requiring the sequence in writing surfaces the difference before the contract — not after commissioning, when it surfaces as a claim.
Conclusion
Purity first, recovery accounted, wash water last. A precipitate circuit tuned in that order stops trading product grade against metal losses it never measured, and the argument inside the plant changes: nobody defends a wash setting with cake colour when the filtrate numbers are sitting on the same page. The press and the cloth then do the work they were specified for, inside a duty written down before prices were compared.
Plant teams setting that duty for a new circuit can work it against the NHD range before committing to a flowsheet. Send the precipitate data and the product limit by email to sales@chinanhd.com, and the wash stage gets sized against both numbers — the cake limit and the metal the filtrate may carry — rather than against cake appearance alone. For a running conversation instead of an email thread, the team answers on WhatsApp at +86 136 6732 4277.
FAQs
Q1: Should cake purity or dissolved-metal recovery come first?
A1: Purity first — the product limit defines the duty. Then measure the metal in the wash filtrate and settle its recovery route. A setting that meets the limit while sending unmeasured metal downstream has done half the job. On a producing line, the unmeasured half is rarely the cheap half.
Q2: Does more wash water always improve precipitate quality?
A2: Up to a point. It improves displacement; past the point it mostly dilutes. On the worked example, the step from 80 to 95 percent displacement moved 0.9 kg more metal per tonne of cake and drank a disproportionate share of the wash water doing it. Cake quality and filtrate composition, measured on the same runs, show where the point sits for a given precipitate.
Q3: Can the press pressure rating predict rare-earth washing performance?
A3: No. The 1.6 MPa figure is an equipment ceiling. Performance comes out of the precipitate, the cake it forms, the cloth and the cycle settings. A coarse oxalate and a gelatinous carbonate prove it on the same machine — same pressure, different wash curves. It gets proven on the material itself, or it does not get proven.