Why Aquarium Carbon Lasts About a Month

This page contains affiliate links. If you buy through one we may earn a commission, at no cost to you.

You cannot, and the reason is more interesting than the answer. Two entirely separate clocks govern a carbon cartridge. One is how much surface it has left. The other is how far into each grain the water has managed to reach. They are unrelated processes, they both run out somewhere inside the same month, and there is no observation available in a home aquarium that tells you which of the two you are looking at.

Which means the monthly replacement interval on the packet turns out to be roughly right, for a reason nobody prints on the packet. It is worth working through, because along the way the arithmetic makes clear that the fraction of your water chemistry carbon can touch is much narrower than most setups assume.

What is actually in a cartridge

Activated carbon is a sponge at molecular scale. Steam or chemical activation blows a pore network through the grain, and good aquarium grade material offers something like 1,000 square metres of internal surface for every gram. A 20 gram cartridge is therefore carrying around 20,000 square metres of surface, which is two hectares, folded into something that fits in your palm.

Adsorption on that surface is physical, not chemical. Dispersion forces hold a molecule against the pore wall, and the walls are hydrophobic, so what sticks is what water does not want: large, uncharged, oily organic molecules. Tannins from driftwood. Dissolved proteins and their breakdown products. Medication after a course of treatment. Odour compounds. That list is the whole job description.

Now the capacity, which is where the first widely quoted number goes wrong. The figure printed on industrial carbon is the iodine number, typically 900 to 1,100 milligrams per gram. That is a probe measurement using iodine in solution to count accessible micropores, and it is not a field capacity for anything you care about. Real loading for dissolved organic matter pulled out of water runs 50 to 200 milligrams per gram, so the iodine number overstates the useful figure by about sevenfold. At 150 milligrams per gram, a 20 gram cartridge holds around 3 grams of organics. That is the whole allowance.

Clock one: the capacity, and why it cannot be pinned down

Three grams sounds small and the question is what eats it. Almost all of the organic load in a closed tank arrives as food. Feed half a gram of dry flake a day, which is a modest community tank, and some fraction of that ends up as dissolved organic matter the carbon can take. The rest is assimilated, or settles as solids the sponge catches, or is mineralised by bacteria into ions.

That fraction is the problem. It depends on the food, the stocking, the bacterial population and how much lands uneaten, and nothing available to a keeper measures it. So the honest thing is to show what it does to the answer. At 10 percent adsorbable the cartridge lasts 60 days. At 30 percent it lasts 20. At 50 percent it lasts 12.

A factor of five, resting entirely on a number that cannot be measured in the room the tank is in. If the capacity clock were the only clock, the correct answer to the question would be a shrug.

Clock two: the water barely touches the carbon

Here is the part that changes the shape of the problem. Water treatment engineers size a carbon bed by empty bed contact time, which is simply the bed volume divided by the flow through it, and the design range for adsorbing organics out of water is 5 to 20 minutes.

Put an aquarium cartridge on that scale. A fishkeeper 4-in-1 submersible internal filter publishes 135 gallons per hour, which is 511 litres an hour, or 0.142 litres a second. A carbon bag in a unit that size might hold 30 millilitres of granules. Divide and the contact time is 0.211 seconds.

Against a design minimum of 300 seconds, that is short by a factor of 1,420. Not undersized. Three orders of magnitude outside the process window. On a single pass a cartridge like that can only strip the very fastest adsorbing fraction off the outside of its grains, and the water leaves before anything else has a chance to happen.

What rescues it is that a tank is a loop. In a 20 gallon tank, 75.7 litres, that 511 litres an hour is 6.75 turnovers every hour, so the same water comes back 162 times a day. Cumulative contact therefore accrues at 6.75 times 24 times 0.211 seconds, which is 34.2 seconds a day. Over 30 days that is 1,026 seconds, 17 minutes, and now you are inside the design window.

Which is why a month is the right answer

Adsorption inside a grain is diffusion limited. A molecule has to work its way down a pore network to reach fresh surface, and the time to penetrate a depth follows the square of that depth over twice the diffusivity. With an intraparticle diffusivity around ten to the minus ten square metres per second, which is typical for an organic solute in carbon, a molecule reaches 10 micrometres in half a second, 100 micrometres in 50 seconds, and the centre of a 1 millimetre granule in 1,250 seconds, just under 21 minutes of contact.

Set that against the 34.2 seconds a day the loop delivers, and the grains are fully used at 1,250 divided by 34.2, which is 36.5 days. A month and a bit.

So the two clocks, which share no physics at all, both expire within the same few weeks. The capacity clock says 12 to 60 days depending on how you feed. The diffusion clock says 36.5 days regardless of how you feed. A keeper watching the tank has no instrument that separates a cartridge which has run out of surface from one whose interior is only now being reached, and that is the real answer to the question as asked.

It also hands you a lever nobody sells. The diffusion time goes as the square of the grain radius, so halving the grain size cuts it to 313 seconds and brings full utilisation forward from 36.5 days to 9.1. Fine carbon in a thin layer outperforms coarse carbon in a fat bag, by a lot. Water treatment specifications always state a mesh size for exactly this reason. No carbon bearing product on this site states one. In fact none of them states a mass, a grain size, an iodine number or a surface area either, and those four numbers are the only route to the answer the question wants. Our roundups list the media as "activated carbon bag" and that is the whole specification.

The flow figure on that filter does not work

While the filter is on the bench, one number in its record needs reporting. It publishes 135 gallons per hour and a tank range of "80 to 300 gallon freshwater or saltwater". A filter's flow has to be read against the tank it is turning over, and the working minimum for a stocked freshwater tank is around four turnovers an hour.

At 80 gallons, 135 gallons per hour is 1.69 turnovers. At 300 gallons it is 0.45. So across its own stated range the unit delivers between two and a half and nine times less circulation than the job needs. Its drawbacks do flag the issue, saying "135GPH is modest for the top of its range", and modest is a generous way to describe being short by a factor of nine.

Our own catalogue contradicts it, too. The canister filters in our aquarium filter roundups publish 265 gallons per hour for tanks up to 75 gallons, which is 3.5 turnovers an hour and a sensible figure. Two listings on the same site are using flow to tank ratios that differ by eight times.

Our guide on choosing an internal aquarium filter gets the method exactly right, saying to compare the gallons per hour figure to your tank volume and remembering that heavily stocked tanks need more turnover than lightly stocked ones. It is right and it stops one step short: it never says what ratio to aim at, and without the number the comparison cannot be made. The number is about four, and this unit is an honest fit for a 20 to 30 gallon tank rather than the range on its box.

The fraction of your water chemistry carbon cannot touch

This is the correction that matters most, because it is the reason a lot of carbon is in a lot of filters doing nothing. Adsorption works by pressing an uncharged molecule against a hydrophobic wall. Ammonium, nitrite and nitrate are small ions, each wrapped in a tightly held shell of water molecules, and prying that shell off to stick the ion to a greasy pore wall is energetically hopeless. The capacity of activated carbon for all three is close to nil.

So if the reason there is carbon in the filter is a general sense of cleaning the water, and the thing actually being watched is one of those three, the carbon has never contributed and replacing it will change nothing. The process that handles them is bacterial, it lives on the sponge, and our page on why a fish tank goes cloudy after cleaning covers what happens when that colony is disturbed.

The measurement side makes the same point from the other direction, and it is visible in our own catalogue. A kactoily digital aquarium water tester with app alerts continuously tracks pH, ORP, EC, TDS, salinity, SG and temperature. Look at that list as a physicist would: every single entry is a property of dissolved ions. Conductivity and total dissolved solids are the same measurement twice, and they respond to ions in proportion to their charge and mobility. Dissolved organic molecules are uncharged and contribute almost nothing to conductivity.

So the most capable monitoring instrument on this site is, by construction, blind to the entire class of compounds the carbon is removing, and sensitive to the class it cannot remove. Its own drawbacks are honest that the "Calibration routine and steps are not described" and that accuracy figures are maker stated, and neither of those is the real limitation. The real one is that there is no hobby instrument for dissolved organic carbon at all, which is why breakthrough is invisible.

If the goal was clearer water, the tool is different

A fair amount of aquarium carbon is in service for cloudiness, and cloudiness is usually particles rather than dissolved anything. Carbon does not remove particles, the sponge does, and if the cloud is suspended algae then neither of them does much. An Aquatic Eden 3W mini UV-C clarifier is the tool for that specific case, rated for 10 to 20 gallons and up to 30 at low bio load, with timer modes of 3, 6, 12 and 24 hours.

Its record contains one line that is more useful than it looks: "Results depend heavily on high-flow placement". That is a statement about dose. UV works on the product of intensity and exposure time, so the flow past the lamp sets how much each organism receives, and it is the only variable the keeper controls. Which leads to the same absence as the carbon: the listing publishes a lamp wattage and timer modes, and no chamber volume and no rated flow, so the dose cannot be computed from it. Our guide to using a UV clarifier covers the scheduling side. It is also right that the unit should go in a filter chamber rather than loose in the tank.

The honest limit, and the move that beats all of it

The limit on this page is clear and it sits on two assumptions: the adsorbable fraction of the food, which sets the capacity clock and which nobody can measure, and the intraparticle diffusivity, which sets the diffusion clock and which depends on a pore structure no listing describes. Move either and both clocks move. What does not move is the conclusion, because the conclusion is that the two clocks overlap and cannot be told apart, and that survives any reasonable choice of either number.

So here is the thing to do instead, and it needs no media at all. A SEAOURA 4-in-1 gravel vacuum siphon runs on nothing but a height difference, with dual airbags to prime it and a basket to keep gravel and fish out of the tube, and its own record is straight about the one drawback, that "Manual pumping takes effort on large tanks".

Put a water change on the same scale as the cartridge. Pull 25 percent out of a 75.7 litre tank and you have removed 18.9 litres, and with it 25 percent of every dissolved substance in the tank at once. Not 25 percent of the organics. Twenty five percent of the organics, and the nitrate, and the phosphate, and every ion the carbon has no capacity for, in about ten minutes, with no clock to guess at and nothing to replace. Our page on doing a gravel vacuum water change covers the routine.

That is not an argument against carbon. It is an argument for using it the way the chemistry actually supports: as a short, deliberate intervention with a specific target. Driftwood has stained the water, or a course of medication has finished and you want it out. Put carbon in, leave it a few days while the gradient is steep and adsorption is fast, then take it out. Used that way the two clocks above never get near expiring, the invisible breakthrough problem never arises, and the question of when to replace it answers itself: you do not replace it, you remove it when the job it was put in for is done.