Specific Cake Resistance: Filter Sizing From Lab Tests

Last updated: 2026-11-03

Specific cake resistance, alpha, is the resistance that a unit mass of deposited filter cake offers to filtrate flow, expressed in m/kg. Crystalline solids such as gypsum typically sit near 1e10-1e12 m/kg; gelatinous hydroxide sludges can exceed 1e13. Measured with a leaf test and inserted into the Ruth filtration equation, it is the single number that converts a bench result into plant filter area within solid-liquid separation design.

Key Takeaways

• Specific cake resistance (alpha, m/kg) is the one cake property that transfers from lab to plant.
• The Ruth filtration equation turns a leaf-test t/V plot directly into required filter area.
• Compressibility index n tells you how much area you lose when pressure drops.
• Practical designs apply an indicative 1.2-1.6x area margin over the leaf-test result.
• When alpha is too high, conditioning and chemistry beat buying more filtration area.

What Is Specific Cake Resistance?

Specific cake resistance is a lumped material property that expresses how strongly a unit mass (1 kg) of deposited cake impedes filtrate flow, and it is reported in meters per kilogram (m/kg) — units that look odd until you see where they come from in Darcy’s law.

Darcy’s law for flow through a porous bed states that the pressure drop dP equals viscosity mu, times superficial velocity, times a hydraulic resistance per unit bed depth. A filter is a shrinking bed: as filtrate passes, solids accumulate and the cake mass per unit area grows. Ruth fused Darcy’s law with that growth into the constant-pressure filtration equation, and it is worth writing in plain text:

t/V = (mu * alpha * c) / (2 * dP * A^2) * V + (mu * Rm) / (dP * A)

where t is filtration time (s), V is cumulative filtrate volume (m3), mu is filtrate viscosity (Pa.s), dP is applied pressure drop (Pa), A is filter area (m2), c is the mass of dry cake solids deposited per unit volume of filtrate (kg/m3), Rm is the medium resistance (1/m), and alpha is the specific cake resistance (m/kg). In this form the equation is a straight line: plot t/V against V from a leaf test, and the slope b gives alpha = 2 * dP * A^2 * b / (mu * c), while the intercept gives Rm.

The physical meaning of alpha is worth internalizing. It is the pressure drop required to drive filtrate of unit viscosity at unit superficial velocity through enough cake to contain one kilogram of dry solids. Coarse, rigid, well-formed crystals offer little resistance per kilogram; fine, platy, gelatinous particles offer enormous resistance per kilogram, because the same mass packs into far more pore-space obstacles.

Why is it the one number that transfers from lab to plant? Because alpha consolidates particle size, shape, packing, tortuosity, and porosity into a single hydraulic term that does not depend on the machine. Leaf test filtration on 0.01 m2 and a 300 m2 press see the same cake, provided slurry conditioning, pressure, and temperature are reproduced. Medium resistance and machine hydraulics change with scale; alpha does not. That is the entire logic of filter area sizing, and everything else in this article is bookkeeping around it.

Typical Values by Material and What Compressibility Means

Alpha spans five orders of magnitude across common hydrometallurgical solids, and the compressibility index n spans a further decision: whether pressure alone can do the job. The compressibility index is an exponent that describes how much alpha rises as filtration pressure rises, defined by alpha = alpha0 * dP^n, where alpha0 is the resistance at unit pressure and dP is the pressure drop. Read the bands below as typical and indicative — confirm every value by test work on your own slurry, because crystal habit and particle size distribution move numbers by a factor of two or more within each class.

The n bands carry the practical message. Below 0.3, the cake is essentially incompressible: doubling pressure nearly doubles the rate, so a cheap pressure increase substitutes for area. Between 0.3 and 0.6 — where most hydrometallurgical residues sit — returns are real but diminishing: doubling pressure buys roughly 40-70% more cake per pass (indicative), not double. Above 0.6, the cake compresses under its own pressure gradient; extra squeeze mostly consolidates the cake and lowers moisture rather than raising rate, which is why membrane presses exist. One caution the table cannot show: n itself is only valid across the pressure range it was measured over. Extrapolating a leaf-test n from 200 kPa up to 2 MPa has burned more than one sizing study. Measuring n properly takes at least three pressures in the intended operating band, each regressed separately, with the log-log slope of alpha against dP taken as the index.

How Do You Measure It? 7-Step Leaf Test Procedure

A pressure leaf test is a small-area constant-pressure filtration that reproduces plant slurry and pressure on a laboratory scale — the workhorse behind leaf test filtration and every t/V plot filtration analysis. A laboratory pressure leaf filter with a 0.01 m2 leaf (roughly 110 mm diameter) and instrumented filtrate collection is the standard tool. The procedure:

1. Prepare and condition the sample. Take a representative, agitated slurry sample; hold solids content and flocculant dosage at the values intended for the plant; measure and record solids content by drying a subsample.
2. Control temperature. Run the test in a constant-temperature bath at the expected plant temperature, because viscosity enters the slope directly and a 20 degree Celsius swing moves mu by roughly a factor of two.
3. Mount cloth and measure area. Fit the same cloth (or media permeability) intended for the plant filter; record the exposed leaf area A to better than 1%.
4. Run pressure steps. Filter at 100, 300, and 600 kPa in sequence with fresh slurry each step (typical); each pressure needs its own run because the slope changes with dP and the intercept changes with both dP and cloth history.
5. Record time versus filtrate volume. Log t and V continuously from the first drop; discard nothing — the early points carry the intercept.
6. Plot t/V against V and regress. For one run, suppose the regression of t/V (s/m3) on V (m3) gives slope b = 1.2e9 s/m6 and intercept a = 2.2e4 s/m3 at 600 kPa with A = 0.01 m2, mu = 1.3e-3 Pa.s, and c = 441 kg/m3. Then alpha = 2 * dP * A^2 * b / (mu * c) = (2 * 600000 * 0.0001 * 1.2e9) / (1.3e-3 * 441) = 2.5e11 m/kg. Sanity-check that alpha rises, and the slope steepens, at the lower pressures.
7. Extract medium resistance and correct for viscosity. Rm = a * dP * A / mu = (2.2e4 * 600000 * 0.01) / 1.3e-3 = 1.0e11 1/m, typical of a fresh cloth. Finally, restate alpha and Rm at the design temperature: divide the slope-equivalent resistance by the viscosity ratio, or simply re-run step 4 at design temperature if it differs from the lab by more than 10 degrees Celsius.

Worked Example: Sizing a Filter From One Leaf Test

Here is a fully redoable sizing from the numbers above. Duty: 10 t/h of dry solids from a neutralized slurry fed at 20% w/w solids. The leaf test gave alpha = 2.5e11 m/kg at 600 kPa and n = 0.45, cake moisture 22% w/w (wet basis), and the plant wants a 4 h cycle with 2.0 h of cake formation (typical for a washed cycle; confirm the split by cycle optimization).

Mass balance first. At 20% w/w feed, 10 t/h of solids arrive in 50 t/h of slurry, so 40 t/h of water enters. A 22% w/w cake on 10 t/h of solids weighs 12.8 t/h and carries 2.8 t/h of moisture, leaving 27.2 t/h of filtrate; at 1.2 t/m3 that is 22.7 m3/h, and c = 441 kg/m3. Note the assumption: liquor density 1.2 t/m3 is realistic for sulfate liquors; using 1.0 would change c and every downstream number, so measure it.

Sizing next. With t = 7200 s of formation, the Ruth equation at constant pressure gives (V/A)^2 = 2 * dP * t / (mu * alpha * c) = (2 * 600000 * 7200) / (1.3e-3 * 2.5e11 * 441) = 0.0602, so V/A = 0.245 m3/m2. Each square meter deposits 0.245 * 441 = 108 kg of dry solids per cycle, and the duty needs 40 t per cycle: 40,000 / 108 = 369 m2 of active area. Applying a 1.25 scale-up margin (justified in the next section) gives roughly 460 m2 installed, which rounds to two presses of common frame size — a layout worth checking against the economics of a filter press for mining service.

Now read the fine print that makes this a real design. Cake thickness is 108 / 0.78 / 1900 = 73 mm, thicker than the 30-50 mm chambers of a typical press. The honest responses: specify deeper chambers where available, or stop formation at chamber depth and shorten the cycle (forced into a 40 mm chamber, formation ends near 0.6 h and the area climbs to roughly 675 m2), or revisit the filter press cycle optimization case entirely. Moisture, too, is an assumption: 22% w/w is what the leaf said, and plant cake usually runs 2-4 points wetter (indicative), which shifts the mass balance — track it against measured filter cake moisture data.

Finally, the pressure penalty. With n = 0.45, alpha at 300 kPa is 2.5e11 * (300/600)^0.45 = 1.8e11 m/kg. Repeating the sizing at the same 7200 s gives V/A = 0.203 m3/m2, 89.5 kg of solids per square meter per cycle, and 447 m2 required — a 21% area penalty (and a 40% power saving). Half the pressure, roughly a fifth more filter: that trade is exactly what n quantifies.

Why Does Laboratory Alpha Over-Predict Plant Area?

Despite everything above, a plant built on a raw leaf-test area usually underperforms. The mechanisms are known and worth pricing in:

• Cloth resistance in service. The fresh-cloth Rm from step 7 is a fiction within weeks; blinding and pore logging raise effective medium resistance severalfold. Factor 1.05-1.15 (indicative); audit with the filter cloth blinding checklist.
• Flocculant breakdown under pump shear. The flocs measured in the lab are destroyed by the centrifugal feed pump; effective alpha at the press can be 1.5-3x the leaf value (indicative) when dosing strategy is fragile.
• Temperature and viscosity drift. A slurry sized at 40 degrees Celsius but filtering at 25 degrees loses roughly a third of its rate through viscosity alone.
• Full-scale cake compression. Long feed headers and uneven flow let part of the cake form at lower pressure than the leaf test saw, lengthening the effective cycle.
• Air breakthrough and cake cracking. Cracks short-circuit vacuum and reset the formation clock on coarse cakes.
• Uneven feed distribution. Real plates do not fill equally; the slowest chamber sets the cycle time.

No single factor dominates all plants, which is why practice applies a composite area margin. The indicative bands below assume leaf testing was done properly; if the lab skipped viscosity control or used unrepresentative slurry, the margin must be larger.

Treat 1.2-1.6 as a range, not a license: a well-instrumented pilot press on the real slurry can justify the low end, while a cold-climate plant with an aggressive feed pump has no business claiming it. Record which factor drove your margin when you commission it — the first year of plant data will either validate the split or tell you exactly which assumption to fix before the next duty is sized.

When Specific Cake Resistance Is Too High, What Do You Change?

When specific cake resistance is too high, change the slurry before you change the filter — and rank the levers by cost. Each verdict below assumes test work confirms it on your own material.

1. Flocculant and conditioning. Cheapest by far. Agglomerating fines can cut alpha by a factor of 2-5 (typical) for grams per tonne of polymer, but gains evaporate under shear, so dose point and mixing energy decide success. First lever, always tested.
2. Feed dilution. Raising liquor fraction dilutes c and thins the cake, buying rate at the cost of more filtrate to handle per tonne. Worthwhile when the thickener upstream is underloaded; a wash-ratio test on a bench-scale CCD wash unit tells you the cost.
3. Temperature. Every 10 degrees Celsius cut in viscosity is nearly free capacity if waste heat exists; heating a large slurry stream rarely pays unless the circuit is hot already.
4. Filter aid / body feed. Precoat and body feed rescue gelatinous sludges, but the aid is consumed per tonne forever; it is a rental solution and should be priced as one.
5. Switch to pressure filtration. If the current duty is vacuum, moving to 300-600 kPa typically halves area demand for moderately compressible cakes (n below 0.6). For coarse, fast-draining solids the reverse holds — check vacuum belt filter selection logic before over-buying pressure.
6. Change the upstream chemistry. Most expensive, sometimes decisive. In zinc plants, a jarosite residue filters several times better than a goethite residue of the same duty (typical), and in neutralization duty, gypsum crystal habit can be engineered by seeding and pH staging. Verdict: only justified when the filtration saving is large and the flowsheet is young enough to absorb the change.

The engineering verdict: work down this list in order. Levers 1-3 usually deliver an order-of-magnitude more capacity per dollar than lever 5, and lever 6 is a flowsheet decision, not a filtration decision.

Field Data: Three Filter Upgrades Traced to Alpha

Three anonymized cases from our test records, each with the measurement method stated so you can judge the basis.

• Zinc plant goethite residue. Leaf tests on plant slurry gave alpha = 2.8e11 m/kg at 500 kPa (multi-pressure leaf, 0.01 m2 leaf, plant cloth). Plant presses over 30 days of tally ran at throughput consistent with alpha near 4.0e11 — a scale factor of about 1.4, dominated by flocculant shear. After moving polymer addition upstream of the feed pump and cutting its speed, press cycle time fell from 4.2 to 3.1 h and installed area demand fell 28% at constant duty (plant mass balance, 30-day average).
• Copper smelter effluent neutralization. Switching from single-stage lime to staged limestone-lime neutralization changed gypsum habit; leaf tests at 300 kPa showed alpha falling from 8.5e11 to 3.2e11 m/kg and cake moisture dropping from 26% to 19% w/w (oven drying method). The existing vacuum drum was replaced by a press with one-third less area at doubled duty (leaf test series plus plant trial, 72 h mass balance).
• Nickel hydroxide precipitate. A dilute-then-dose conditioning change cut effective alpha from 6.0e11 to 3.8e11 m/kg (weekly leaf tests on plant samples over 6 weeks) and lifted dryer-feed solids, with cake moisture falling from 30% to 25% w/w (oven drying, weekly composite). Area was unchanged; capacity rose 22% (plant throughput tally).

The pattern across all three: the biggest wins came from chemistry and conditioning, not from metal.

Frequently Asked Questions

What is the difference between specific cake resistance and cake permeability?

Permeability describes flow through a unit volume of porous bed; specific cake resistance describes flow per unit mass of cake solids. They are inversely related through cake bulk density, but alpha is what filtration equations need, because filters accumulate mass, not volume. Use permeability in packed-bed work and alpha in filtration sizing; the two are convertible only if cake density is measured.

Can I size a plant filter from one leaf test at one pressure?

You can, but you are gambling on compressibility. A single pressure gives one alpha point, so any deviation in plant pressure or slurry conditioning is uncorrected. The standard practice is a multi-pressure series — typically 100, 300, and 600 kPa — which yields both alpha and the compressibility index n. The extra half-day of lab work is the cheapest risk reduction in the whole sizing chain.

How many leaf test runs are enough for design?

At minimum: three pressures, each in duplicate, plus runs on aged slurry and re-flocculated slurry if the plant will see both. That is 6-10 runs on one sample (typical). If feed varies seasonally or between ore sources, repeat the series per feed case. Design numbers come from the worst realistic case, not the average, and margins come from the scale-up factors discussed above.

Does filter cloth choice change alpha?

No — cloth changes medium resistance Rm, not the cake. But cloth choice changes the practical result: a blinded or poorly matched cloth lengthens the cycle and raises cake moisture, which then changes the slurry the next cake forms from. Size with alpha from the leaf test and cloth resistance from a service-aged cloth sample, not a fresh one.

Why is plant cake wetter than the leaf test cake?

Usually several causes stack up: shorter effective drainage time in the real cycle, higher feed solids variability, air breakthrough or cracks bypassing dewatering, and temperature differences that leave more viscous mother liquor in the pores. Plan for plant moisture 2-4 percentage points above the leaf result (indicative) and verify with early plant sampling.

What alpha is too high for pressure filtration?

There is no hard wall, but above roughly 1e13 m/kg (typical, at 300-600 kPa) pressure filtration needs conditioning help — body feed, precoat, or membrane squeeze — and area margins grow fast. Above that, question whether mechanical dewatering is the right duty at all; thickening plus safe pumping to a tailings facility sometimes beats owning hundreds of square meters of filter for a gel.

How does temperature affect filter area sizing?

Directly and strongly, because viscosity multiplies both terms of the Ruth equation. Rate scales roughly with the inverse of viscosity, so a slurry sized at 20 degrees Celsius needs notably more area than the same slurry at 60 degrees Celsius. Always report test temperature, measure filtrate viscosity at that temperature, and restate alpha at plant temperature before sizing.

Conclusion

Specific cake resistance is the number that makes filtration design a calculation instead of a guess: measure it with a disciplined leaf test, correct it with an honest scale-up margin, and it will tell you the filter area, the cake thickness, and the price of running at lower pressure before any vendor quote arrives. Just as important, it tells you where to spend — most plants get more capacity from conditioning and chemistry than from steel. If you want this procedure run on your own slurry, with sizing you can defend in a design review, request a quote for a bench filtration program, or browse the rest of our calculation guides first.

References

• Wills, B.A. and Finch, J.A., Wills’ Mineral Processing Technology, 8th ed., Butterworth-Heinemann.
• SME Mineral Processing and Extractive Metallurgy Handbook, Society for Mining, Metallurgy and Exploration, Englewood, CO.
• Rushton, A., Ward, A.S. and Holdich, R.G., Solid-Liquid Filtration and Separation Technology, 2nd ed., Wiley-VCH.
• Svarovsky, L., Solid-Liquid Separation, 4th ed., Butterworth-Heinemann.
• Wakeman, R.J. and Tarleton, E.S., Solid/Liquid Separation: Principles of Industrial Filtration, Elsevier.
• Purchas, D.B. and Wakeman, R.J., Solid/Liquid Separation Equipment Scale-Up, Filtration & Separation Studies.
• Ruth, B.F., “Studies in Filtration III. Derivation of General Filtration Equations,” Industrial & Engineering Chemistry.
• Tiller, F.M., “The Role of Porosity in Filtration,” AIChE Journal.

Changelog

• 2026-11-03: First publication.