Filter media: what actually works
Sponge, ceramic, bio-balls, K1 — every forum has the same argument. Here is what the biology and the evidence actually say, and where the marketing numbers fall apart.
The filter media war
Ask which biological filter media is best and you will start a fight. One camp swears by sintered glass at £30 a litre. Another says ceramic rings are all you need. Someone insists plastic K1 changed their tank forever, and a quiet contrarian in the corner runs stainless pot scrubbers and has the clearest water in the club.
Underneath the argument is a simple, testable question: how much biological filtration do you actually get per litre of each medium? Manufacturers answer with enormous surface-area figures — 700, 1000, even 1500 square metres per litre. Those numbers are the heart of the marketing, and they are also where the science and the sales copy part company.
The honest summary is this: for the job most media are sold to do — converting ammonia and nitrite — the type of medium matters far less than the industry implies. What matters is accessible surface, oxygenated water actually passing through it, and time. But there are real, secondary differences worth understanding, and that is where the nuance lives.
What biological filtration actually is
Biological filtration is not the media. It is a film of living bacteria that grows on the media. The medium is just real estate. Understanding the tenants explains everything about what makes good real estate.
The organisms doing the work are nitrifying bacteria and archaea: ammonia oxidisers that turn toxic ammonia (NH3) into nitrite (NO2−), and nitrite oxidisers that turn nitrite into far-less-toxic nitrate (NO3−). In many mature tanks a single organism — complete-ammonia-oxidising comammox Nitrospira — does both steps at once. This is the same chain covered in our nitrogen cycle guide.
Three facts about these bacteria decide which media work:
- They are strictly aerobic. Nitrification is an oxygen-hungry process — roughly 4.3 mg of oxygen is consumed per mg of ammonia-nitrogen oxidised. No oxygen, no nitrification. This single fact demolishes most surface-area claims, as we will see.
- They live in a thin biofilm, not deep inside pores. Nitrifiers colonise surfaces as a film only a few cell-layers thick, fed by the water flowing past. A bacterium is roughly 1–2 microns across; a pore has to be open, wet and oxygenated for it to be usable.
- They are slow-growing and self-limiting. Their population grows to match the food supply — the ammonia your stock produces — and no further. This is the fact that quietly settles most of the argument.
The surface-area numbers are mostly fiction
Here is the sleight of hand. When a bag of porous media claims "700 m² per litre," that figure typically comes from BET-style analysis — nitrogen-gas adsorption, a lab technique that measures every nook a nitrogen molecule can reach, right down to nanometre-scale pores (where a figure is not gas-adsorption it is usually a geometric estimate, which runs into the same accessibility problem).
A nitrogen molecule is about 0.3 nanometres wide. A nitrifying bacterium is around 1–2 micrometres — roughly a thousand times larger. The overwhelming majority of the surface area in that headline number sits inside pores no bacterium could ever enter, let alone a bacterium that also needs a supply of oxygenated water delivered to it.
The number you are sold is not the number bacteria use
Manufacturer surface area is typically total pore area measured by gas adsorption. The area that actually hosts nitrifiers is the accessible, oxygenated, flow-exposed surface — a small fraction of the headline figure, and far more similar between media than the marketing suggests.
It gets worse for the big numbers. Even where pores are large enough to colonise, water and oxygen have to reach the bacteria living in them. A few millimetres into a dense porous block, flow stops, oxygen is consumed by the outer biofilm, and the interior goes anaerobic. Anaerobic zones do not do nitrification; at best they host slow denitrification, which is a different job entirely. So the deep internal area a manufacturer counts is, for ammonia control, dead space.
This is not hobbyist speculation — it is why wastewater engineers abandoned raw surface area decades ago. Moving-bed biofilm reactors (MBBRs), developed by Hallvard Ødegaard's group at the Norwegian University of Science and Technology, are rated not by total surface but by effective specific surface area: the protected, accessible area that actually grows a working biofilm. Engineering estimates put that effective figure for typical carriers on the order of tens to a few hundred square metres per cubic metre — a fraction of a square metre per litre — against the hundreds of square metres per litre printed on aquarium packaging. The engineers measure what bacteria can reach; the marketing measures what a nitrogen molecule can reach.
And the number that actually matters is not area at all — it is the areal nitrification rate: how much ammonia a given area of biofilm processes per day. Decades of recirculating-aquaculture research put this at roughly 0.2–1.0 grams of ammonia-nitrogen per square metre of media per day at 15–20 °C, rising to around 1–2 grams at tropical temperatures. That rate is governed by temperature, oxygen, pH and the ammonia on offer — not by the brand of plastic or ceramic the biofilm happens to be sitting on.
"The bacteria don't read the packaging. They colonise the accessible, oxygenated surface — and there is far less difference between good media there than the price tags imply."
What actually determines effectiveness
Strip away the marketing and three things decide how much real biological capacity a filter delivers — none of which is the brand on the bag.
1. Accessible surface with flow through it, not around it. Bacteria only work where oxygenated water touches them. Media packed so tightly that water channels around the outside, or a canister so clogged that flow collapses, wastes most of its surface. Open, evenly-flushed media beats a denser medium that water skips past. This is why flow and filter maintenance matter more than the media choice.
2. Oxygen. Because nitrification burns oxygen, media held in well-oxygenated, moving water outperforms the same media sitting in a low-flow dead zone. It is also why trickle (wet/dry) filters and fluidised moving-bed media punch above their raw surface area — the bacteria get all the oxygen they can use. If your tank runs low on dissolved oxygen overnight, your biofilter feels it too.
3. Maturity. Nitrifiers double slowly — on the order of a day, not minutes. A biofilter is only as strong as the colony it has grown, which takes weeks. A brand-new bag of premium media is biologically inert; an old, ugly sponge full of established bacteria is worth more than any fresh product on the shelf. Never sterilise mature media chasing a cleaner look — rinse it gently in tank water, never under the tap.
Notice what is missing from that list: the medium itself. Given enough accessible surface, flow and time, the population grows to match your bioload and stops. Which brings us to the fact that ends most arguments.
Bacteria scale to food, not to media volume
Your biofilter grows to process the ammonia your fish produce — and no more, because there is no more food. Doubling your media does not double your "capacity" if the ammonia supply is unchanged; the extra surface simply sits underpopulated. Filtration capacity is set by your stocking and feeding far more than by your media brand.
The media, compared
With all that in mind, here is an honest, relative comparison of the common biological media. The columns that matter are not the fantasy surface-area figures but the usable surface, whether water reliably flows through, and how well the medium keeps working over time.
| Media | Usable bio-surface per litre | Flow-through | Clog resistance | Best role | Value |
|---|---|---|---|---|---|
| Sponge / foam (20–45 ppi) | High | Good until it clogs | Low — needs regular rinsing | All-round bio + mechanical | Excellent |
| Ceramic rings | Low–medium | Excellent (large gaps) | High | Flow-friendly filler, pre-filter | Poor (overpriced for area) |
| Sintered glass | Medium (accessible « claimed) | Good | Medium (can break down) | Compact canister bio | Poor (premium price) |
| Plastic bio-balls | Low submerged / high in trickle | Excellent | Very high | Wet/dry & trickle filters | Good in the right filter |
| Moving-bed (K1 / K3) | Medium–high (protected, always flushed) | Excellent (fluidised) | Self-cleaning | Sumps, dedicated bio-chambers | Good |
| Porous rock (pumice / lava / branded) | Medium (much claimed area inaccessible) | Good | High | Bulk bio media | Excellent as raw pumice/lava |
Where each medium genuinely differs
Sponge and foam. Cheap, high accessible surface, and it doubles as mechanical filtration — which is also its weakness. Trapping debris is what clogs it, and a clogged sponge loses the flow its bacteria depend on. Kept rinsed (in tank water), coarse-to-fine foam is one of the most cost-effective biological media there is. The pot-scrubber crowd are, essentially, running open foam with even better flow.
Ceramic rings. The classic default, and quietly one of the weakest per litre for actual bacterial surface — those big central holes and wide gaps are mostly open space. What they are good at is refusing to clog and letting water pass freely, which is why they work as a stable filler or coarse stage. Useful, but you are paying a premium for surface area that mostly is not there.
Sintered glass. The headline surface-area champion on paper, and the biggest gap between claim and reality. Its fine internal porosity inflates the BET figure enormously, but most of it is too small or too deep to host aerobic nitrifiers. It performs perfectly well — just not proportionally to its price or its printed numbers. Some grades also soften and break down over a year or two.
Plastic bio-balls. Context is everything here. Submerged, a bio-ball is mostly empty space and has poor surface per litre. In the trickle or wet/dry filter they were designed for — constantly wetted and exposed to air — that open structure delivers superb oxygenation and they perform far above their raw area. Right tool, right filter.
Moving-bed media (K1, K3). The genuinely clever option. Fluidised in a stream of water or air, the pieces tumble constantly: biofilm is kept thin and active, the protected internal surfaces are always flushed with oxygenated water, and clogging is a non-issue because the tumbling media scours itself (provided the chamber is aerated or flowing enough to keep it moving). Its usable surface is high precisely because flow and oxygen are guaranteed — the two things surface area is worthless without. It needs a chamber that lets it move, which is its only real constraint.
Porous rock — pumice, lava, and the branded versions. This is the clearest place to save money. Branded porous rock media quote the same giant, mostly-inaccessible surface-area figures as everything else, at many times the price of the raw mineral they are made from. Plain aquarium-safe pumice or lava rock offers comparable usable surface for a fraction of the cost. As with all porous media, the accessible fraction is what counts, and it is far lower than the label.
Does more media mean more filtration?
Up to a point, and then no. You need enough accessible, flowing surface to grow a colony that keeps pace with your ammonia production, plus a sensible margin for feeding spikes and new stock. Beyond that margin, extra media does nothing, because — to repeat the fact that matters — the bacteria have no extra food to grow on.
This is why "how many litres of media" is the wrong first question and "how much livestock, fed how much" is the right one. A lightly-stocked planted tank needs surprisingly little biological media, because the plants themselves take up ammonia directly. A heavily-fed community of large fish needs a lot — not because of the media brand, but because of the ammonia load. Our filter capacity calculator works from exactly this direction: biological surface area and fish load, rather than manufacturer hype.
What the science can and can’t settle
It is worth being honest about where the evidence is strong and where it runs out. Almost everything rigorous we know about biological filter media comes from two well-studied fields: municipal wastewater treatment, where MBBR and trickling-filter performance has been measured for decades, and recirculating aquaculture (RAS), where commercial fish farms live or die by biofilter design. Both point the same way: nitrification is governed by effective surface area, oxygen, temperature and flow, and areal conversion rates converge across media types once those conditions are equal.
What is almost entirely missing is controlled, peer-reviewed testing of the specific hobby media brands people argue about, under ornamental-aquarium conditions. To date there is no published head-to-head trial establishing that ceramic rings out-nitrify sponge, or that premium sintered glass beats plain pumice per litre, in a home tank. The closest peer-reviewed aquarium work (for example a study by Scagnelli and colleagues in the Journal of Exotic Pet Medicine, which found four of five “quick-start” bottled-bacteria products failed to control ammonia) has examined starter bacteria rather than media. The strongest hobby-side evidence — large, uncontrolled comparisons such as those documented at aquariumscience.org — is genuinely useful and points firmly at “media type barely matters”, but it is not peer-reviewed science, and it is fair to say so.
What genuinely remains unknown
Nobody can put a confident number on exactly how much of a given medium’s surface a mature biofilm truly occupies; on how far comammox Nitrospira — only described in 2015, in two Nature papers — redistribute the work at the very low ammonia levels of a planted tank; or on how much your particular filter channels water past the media rather than through it. The honest verdict is not “brand X wins” — it is that the variables science can actually measure are not the ones printed on the packaging.
So what should you actually buy?
Can the science settle the war? Largely, yes — just not in the direction the premium brands would like:
- Media type is a minor variable. Given accessible surface, oxygenated flow and time, most biological media converge on similar real-world performance. The colossal surface-area figures do not translate into proportional filtration, because bacteria cannot use most of that area.
- Buy on the things that actually differ: flow-through and clog resistance, oxygenation, longevity, and cost. Coarse-to-fine sponge, plain pumice or lava, or moving-bed media give you the most usable biology per pound. Ceramic rings and sintered glass are fine but overpriced for what the bacteria can reach.
- Match the medium to the filter. Bio-balls belong in trickle filters; K1 needs room to tumble; sponge suits almost anything but must be kept rinsed.
- Protect what you have. A mature, established colony is worth more than any upgrade. Never rinse media in tap water, never replace it all at once, and keep flow up.
The uncomfortable truth for the marketing department is that the cheapest options — open foam, raw pumice, a tumbling handful of K1 — are, per litre of usable surface, about as good as it gets. The expensive argument was never really about the bacteria.