Ammonia in water: NH₃ or NH₄⁺ - why the form decides toxicity

Knowledge base · August 8, 2026

A result from a laboratory or probe says: ammonium nitrogen (N-NH₄) = 1.2 mg/l. Is that a lot or a little? The answer is: it depends on pH and temperature - because the same figure can mean water that is safe for fish or water in which organisms begin to suffer. The key is that ammonium nitrogen is the sum of two forms, and only one of them is truly dangerous.

TAN, the sum of two forms

The laboratory result TAN (Total Ammonia Nitrogen) is the sum of:

• the ammonium ion (NH₄⁺) - a form that is in practice of low toxicity,
• un-ionised ammonia (NH₃) - a highly toxic form, because as a neutral molecule it freely passes through the gill membranes.

These two forms remain in chemical equilibrium with each other: NH₄⁺ ⇌ NH₃ + H⁺. Which way this equilibrium is shifted is decided by pH (mainly) and temperature (additionally).

What share of TAN is toxic NH₃

The share of NH₃ follows from the acid-base equilibrium:

NH₃ share = 1 / (1 + 10^(pKa − pH)), where pKa = 0.09018 + 2729.92 / T[K]

This temperature dependence of pKa comes from Emerson et al. (1975) and remains the computational standard. In practice it is enough to read off the table - share of NH₃ in TAN [%]:

Two things are immediately visible. First, pH rules: at pH 7 a fraction of one per cent of TAN is toxic, at pH 9 it is from a dozen to several dozen per cent. Second, temperature acts in the same direction but more weakly: warming the water by 20 °C increases the NH₃ share three- to fourfold (less at high pH), whereas a rise in pH by one unit increases it about tenfold. In summer both factors rise at once, so their effects multiply.

A caveat on the “×10 per pH unit” rule. It holds as long as pH is clearly below pKa (about 9.4 at 20 °C). Near and above pKa the multiplier shrinks - visible in the table: between pH 8.5 and 9.0 at 30 °C the share rises only ~2.2-fold, not ~3.2-fold.

The same result, two different verdicts

Back to the opening example: TAN = 1.2 mg/l.

• at pH 7.0 and 20 °C → NH₃-N ≈ 0.005 mg/l - safe,
• at pH 8.5 and 20 °C → NH₃-N ≈ 0.13 mg/l - more than 25 times higher, already in the elevated-risk zone.

That is the whole point of this article in two numbers: TAN alone is not a result until pH and temperature come with it.

Effects on aquatic organisms

Un-ionised ammonia damages the gill epithelium, impairs gas exchange and weakens immunity - even at concentrations that do not kill immediately (the mechanism is described among others by Randall and Tsui, 2002). Indicative bands for NH₃-N used in aquaculture and monitoring practice:

Where these numbers come from. The ladder above is an editorial convention of this site - a simplification adopted so that a result can be compared against something straight away, consistent with our calculator. It is not a quotation from any standard or criterion. Real criteria are more complex: in the US EPA ammonia criteria (2013) the permissible concentration is not a fixed number but a function of pH and temperature, and additionally depends on the presence of sensitive mussels and snails - for waters with those organisms the thresholds are markedly lower. If you need a value with legal standing (a report, an expert opinion, a dispute), reach for the criterion proper to your jurisdiction and water type, not for this table.

Sensitivity also differs between species - salmonids are far more sensitive than cyprinids - so for a salmonid stock treat these bands as too permissive.

Do not forget nitrite

Ammonia does not disappear - it is oxidised by bacteria in the process of nitrification:

NH₄⁺ → NO₂⁻ (nitrite) → NO₃⁻ (nitrate)

The intermediate step is often the bottleneck. When the nitrifying community is immature (a new pond, a new biofilter) or overloaded (a stocking surge, high temperature, after disinfection), nitrite accumulates. Its toxicity works by a different mechanism - it oxidises haemoglobin to methaemoglobin, which cannot carry oxygen (hence “brown blood disease”) - and, crucially, it does not depend on pH. So it can neither be predicted from the reaction nor mitigated by adjusting it.

In practice this gives a simple rule: if you measure ammonia, measure nitrite too - especially in the first weeks after stocking and after any intervention in the water circuit. Wider context on nitrogen transformations: Nitrogen and phosphorus cycle.

What to do when the result is high

The order of actions at elevated NH₃ is well established:

1. Stop or reduce feeding - the fastest way to cut off the TAN input.
2. Aerate - ammonia most often occurs together with an oxygen deficit, and damaged gills cope with it worse.
3. Exchange or add water - dilution acts on TAN, that is on both forms at once.
4. Do not “rescue” the situation with an abrupt pH correction. Lowering pH does shift the equilibrium towards non-toxic NH₄⁺, but a jump in reaction is a separate stress for fish, and a rebound upwards can reverse the effect instantly. Change pH slowly and only deliberately.

What this assessment does not cover

• Salinity. The formula above applies to fresh waters. In transitional and coastal waters ionic strength lowers the NH₃ share - using the “freshwater” formula there overestimates the risk.
• Exposure over time. A brief excursion to a “warning” value is not the same as a week in that range; the bands describe a state, not a dose.
• Interaction with oxygen. A fish with damaged gills tolerates an oxygen deficit worse - and vice versa. In summer both factors usually deteriorate together.

When the risk really rises

• Aquaculture ponds - high stocking, intensive feeding → high TAN from fish metabolism and feed breakdown,
• Eutrophic ponds and lakes - algal photosynthesis raises pH during the day, sometimes above 9,
• After water exchange or storms - sudden changes in pH/temperature can abruptly shift the NH₃/NH₄⁺ equilibrium even at constant TAN.

That is why in summer, in warm and fertile water bodies, fish kills most often occur and are explained as “lack of oxygen”, though ammonia can be a co-cause.

In practice

A TAN result alone says little without pH and temperature measured at the same moment - pH in a eutrophic pond can change by more than a unit within a day, so a sample from the afternoon and a sample at dawn describe two different toxicological situations. On converting TAN units (e.g. NH₄ ↔ N-NH₄) see Converting concentration units.

🧮 Tool: Ammonia toxicity calculator (NH₃/NH₄⁺)

• enter TAN, pH and temperature and you will get the concentration of toxic NH₃ together with a rating according to the table above. No sign-up.

If you monitor for longer than a single season, the key thing is that TAN, pH and temperature are recorded together, as one sample - otherwise a year later there is no way to reconstruct the conditions in which a given result arose. In LimnoLog you record them in a single measurement session, also offline in the field, and see them on a shared chart.