Filtration and Separation: How Filters Are Rated and Tested
Every filtration decision comes down to two questions: what has to be removed, and how do you prove the filter actually removed it. This guide covers the fundamentals: surface filtration versus depth filtration, how a micron rating is actually earned, and how filter makers test what they sell.
What a filter actually separates, and why the mechanism matters
Filtration separates a mixture because one part of it is more valuable, or more harmful, on its own. Sometimes the fluid is the product and solids are the contaminant, as in drinking water treatment. Sometimes the reverse is true, as when a valuable solid is recovered from a process liquor. The particles get captured by one of a handful of mechanisms: they are simply too large to pass through the pores (straining), they settle out under gravity, their momentum carries them into the media rather than around it (inertial impaction), they follow the fluid streamline into a pore they cannot pass (interception), or an oppositely charged media surface pulls them out of the stream (electrostatic deposition). The media itself takes several physical forms: perforated plates and membranes, woven fabric, non-woven felt or spun fibre, stacked particle beds such as sand or gravel, and sintered structures where particles are fused together under pressure into a porous solid.
Surface filtration or depth filtration: the difference that decides cost
Surface filtration compared with depth filtration
| Criteria | Surface filtration | Depth filtration |
|---|---|---|
| What it captures | Particles larger than the pore stay on the outside of the medium | Particles are trapped throughout the thickness of the medium |
| Typical service life | Shorter: the surface cakes over and blocks flow | Longer: capacity is spread through the depth of the media |
| Typical cost | Lower per element | Higher per element |
| Mechanical strength | Lower, except stainless steel media | Higher |
| Example media | Pleated cartridges, cellulose filters | String wound cartridges, ceramic and sintered filters |
Surface filters stop particles at the outer layer of the media; the medium does not have to be thick, so unit cost is usually lower, but the surface cakes over as solids build up and the element needs cleaning or replacement sooner. Depth filters use a medium with pores graded from coarse to fine through its thickness, so particles of different sizes are captured at different depths. That spreads the dirt-holding capacity through the whole element instead of concentrating it on one surface, which is why a depth cartridge typically runs longer than a surface cartridge on the same dirty stream, at a higher unit cost. Choosing between them is a cost-per-year question, not a cost-per-piece one: a cheaper surface element that needs changing weekly can cost more annually than a depth element that lasts a month.
Why two filters "rated at 10 micron" can perform completely differently
A micron rating only means something once you know how it was measured. An absolute rating is a hard cut-off: the medium has a sufficiently uniform pore structure that no particle above the stated size gets through, full stop. A nominal rating reports a percentage of retention instead, commonly 90, 95 or 98 percent at the stated size, because media such as felt, paper and woven fabric do not have a uniform enough pore structure to claim an absolute cut-off. A filter with a 10 micron nominal rating at 90 percent retains 90 percent of particles at that size, not all of them, and the remaining 10 percent passes straight through. Mean filter rating averages the pore sizes across the element instead of quoting a single cut-off, which better reflects media that is not uniform to begin with. Beta ratio goes further, expressing the ratio of particles of a given size upstream to particles of that same size downstream: a beta ratio of 1,000 at 99.9 percent efficiency means 1,000 of every 1,000,000 particles that size get through. None of these systems is wrong; they answer different questions, and comparing two cartridges on the micron number alone, without checking which system quoted it, is the single most common specification mistake we see.
How a rating gets tested before it ships
A rating claim is only as good as the test behind it. The bead challenge test passes glass beads of a known, controlled size through the media and examines the filtrate, establishing the absolute cut-off. The single-pass test runs a constant feed through once and measures particle counts in the feed and the filtrate to calculate removal efficiency; the multi-pass test recirculates the fluid through a housing, continuously topping up contaminant, to simulate a hydraulic or lubrication loop. The bubble point test wets the media fully, then reads pore size from the pressure needed to force a gas bubble through it, since that pressure is inversely proportional to pore diameter. A dirt capacity test plots pressure drop against accumulated contaminant to establish how much solids loading an element can take before it needs changing, which is the number that actually predicts service life. Fatigue testing cycles the filter through repeated pressure swings, then rechecks the bubble point to confirm the media has not been damaged. Bacterial and fungal removal is reported separately as a log reduction value (LRV), used where bioburden control matters as much as particle size.
Direct answers
What is the difference between an absolute rating and a nominal rating?
An absolute rating is a hard cut-off: no particle above that micron size passes the medium. A nominal rating is a percentage: a filter with a 10 micron nominal rating at 90 percent retains 90 percent of particles at that size, not all of them, so two filters quoted at the same micron number can leave very different amounts of contamination behind.
What do filtrate, retentate and permeate actually mean on a data sheet?
Filtrate is the liquid that has passed through the filter. Retentate is the fraction of the feed a cross-flow membrane holds back and does not pass. Permeate is the fluid a membrane allows through. The three terms describe where in the process a sample was taken, and mixing them up on a test report misreads the result.
How does a filter maker prove a micron rating before it ships?
Common tests include the bead challenge test (glass beads of known size confirm the absolute cut-off), the single-pass and multi-pass tests (particle counts before and after the filter, run once or recirculated), the bubble point test (pore size read from the pressure needed to force a gas bubble through a wetted medium), and a dirt capacity test that plots pressure drop against contaminant loading to establish service life.
Read the rating basis before you read the price
A quote that states only a micron number, with no rating basis, no dirt capacity figure and no test method, has told you the least useful thing on the sheet. Send us your stream, your target micron size and the rating basis you were quoted, and we will confirm what that number actually means for your process before you specify against it.