Reverse Osmosis For A Fish Tank

The answer, and the number that qualifies it

Yes, a drinking-water reverse osmosis system makes water a tank can use. Whether it is good enough depends on the tank, and the deciding figure is one almost no product page prints.

Reverse osmosis is a high-rejection membrane process, not a zero-solids process. Take the IAPMO performance data sheet for the AquaTru Carafe as a public worked example: against a 750 mg/L TDS challenge, the certified reduction is a minimum of 94.5% and an average of 94.4%, to a permissible product water concentration of 22 mg/L. Certified under-sink systems sit in the same band — the iSpring RCC7AK, for instance, is certified at 97.2% TDS removal.

For drinking water that is excellent. For a reef tank it is the gap that matters. Starting from 300 mg/L tap water, a 95% membrane leaves you around 15 mg/L — and the fraction that passes a membrane preferentially includes the small ions, which is to say the nitrate, phosphate and silicate that feed nuisance algae.

That is the whole case for the fourth stage, and it is why aquarium units are sold as RO/DI rather than RO.

What the DI stage is doing

Deionization is a different mechanism from filtration, which is why it catches what the membrane misses.

A mixed-bed resin exchanges the remaining dissolved ions for hydrogen and hydroxide, which combine into water. It is not a sieve and it has no pore size: it is a chemical trade with a finite number of exchange sites. The Aquatic Life RO Buddie's four stages are sediment, carbon block, the TFC membrane and the deionization cartridge, and retailers describe the division of labor the same way — the membrane removes most of the dissolved solids and the mixed-bed resin takes out the remainder.

Two consequences follow, and both are practical rather than theoretical.

DI resin has no rated gallon figure, and cannot have one. Its life depends entirely on how much the membrane left for it to do. On soft municipal water a cartridge may last for months; on hard water behind a tired membrane it can exhaust in a fortnight. This is why the color-changing resin on the RO Buddie is a genuine feature rather than decoration — it is the only honest indicator available.

The order is load-bearing. DI goes last, after the membrane has removed 94% or more of the work. Put resin in front of a membrane and you will spend a cartridge a week.

Chloramine: the thing that quietly wrecks this setup

This is the failure mode that catches people who did everything else right, and it has nothing to do with the membrane.

Chloramine is harmful to fish for a mechanical reason rather than a toxicological one. The City of High Point's water utility explains that chloramines must be removed from water for fish, amphibians and reptiles because they bind to iron in red blood cell hemoglobin, reducing oxygen delivery. Tampa Bay Water makes the same point: fish absorb monochloramine through their gills, so it has to be removed just as chlorine does.

The trap is that the habits which work for chlorine do not work for chloramine. Letting water stand does not help, and neither does aeration or boiling — chloramine is deliberately more stable than free chlorine, which is why utilities adopted it. A standard 5-micron carbon block in a budget RO/DI unit is not a chloramine filter either; it was specified for chlorine.

What that costs you is resin. Chloramine passes the carbon stage, passes the membrane largely intact, and is then consumed by the DI bed — so the symptom is a DI cartridge that exhausts far faster than it should, and the diagnosis is usually wrong for weeks.

The fixes, in order of how much work they are: find out what your utility doses, which it is required to tell you and which your annual water quality report already states; fit a catalytic carbon prefilter rather than a standard carbon block, which is the stage actually designed for chloramine; and treat with a conditioner that handles ammonia as well as chlorine, since chloramine breaks down into both. The chlorine roundup covers the catalytic carbon distinction in detail.

Freshwater, planted and reef need different answers

"Do I need RO water" has three honest answers depending on what is in the tank, and the hobby tends to give the strictest one to everybody.

The pattern is that the further right you go, the more the point is a reproducible starting point rather than purity for its own sake. A reef keeper is not trying to give fish ultra-pure water; they are trying to make this week's parameters match last week's, and that only works if the water you add salt to is the same every time.

Which also explains the skip box above. A remineralizing stage adds an uncontrolled mineral load back, destroying exactly the property you bought the system for.

If you already have a drinking-water system

This is the cheapest route into tank water and most people miss it, so it is worth spelling out.

A plumbed under-sink system already produces the 94%-plus water above. For a freshwater, planted or shrimp tank, that is very likely all you need — fill buckets from the filtered tap and dose to your target from there. You have bought nothing.

Two cautions. If your system has a remineralizing or alkaline stage, tee off before it, not after. And watch the production rate: an 800 GPD tankless system fills buckets at a reasonable pace, while a 75 GPD tank system refills slowly and you will empty the tank before a water change is done. The roundup's format comparison covers which is which.

For a reef tank, add the deionization stage rather than replacing the system. A dedicated RO/DI unit is cheap enough that running both is usually the right answer anyway — the drinking system stays certified for the kitchen and the tank unit never shares a cartridge schedule with it.

Two honesty notes about aquarium RO/DI units

Since this site judges products on their paperwork, the aquarium tier has to be marked differently and it would be dishonest not to say so.

First, these units carry no NSF/ANSI certification of any kind, and none is expected — they are aquarium equipment, not drinking water treatment units. Every figure on one, including the 50 GPD rating, is a manufacturer specification rather than a certified claim. On this site that normally counts heavily against a product. Here it is simply the state of the category.

Second, the corollary: do not drink from one. Deionised water is the wrong product for a kitchen tap, the unit has no certification to rely on, and the resin is not formulated for it. Keep the certified system for people and the RO/DI for the tank.

The 50 GPD figure also deserves the usual caveat: it is a laboratory number at around 65 PSI and 77 degrees. Cold feed water in February produces materially less, which is true of every membrane and catches every new owner once.