Ammonia vs Ammonium: How pH and Temperature Set the Toxic Fraction
The NH3/NH4+ equilibrium explained with a fraction table by pH and temperature, why test kits read total ammonia nitrogen, and what this means for water changes.
Two forms of one compound
Ammonia is the main end product of protein breakdown in fish. Fish excrete it through the gills and in faeces, and more enters the water as bacteria decompose uneaten food, dead plants and other organic matter; the digestion of food eaten by the fish is the primary source, so the load rises with feeding rate. In water this nitrogen exists in two forms linked by a reversible reaction: un-ionized ammonia (NH3) and the ionized ammonium ion (NH4+), written NH3 + H+ ⇌ NH4+. The sum of both is called total ammonia nitrogen (TAN).
The two forms differ sharply in toxicity. University of Florida IFAS describes un-ionized ammonia as roughly 100 times more toxic to fish than ammonium, and the Southern Regional Aquaculture Center (SRAC) refers to NH3 as the toxic form and NH4+ as the non-toxic form. Ammonium has a pKa of 9.25: below that pH the protonated ammonium ion predominates, above it NH3 does. In low-pH water the equilibrium shifts towards ammonium; in high-pH water it shifts towards ammonia.
The effect of pH and temperature
The share of TAN present as NH3 rises with both pH and temperature, and falls slightly as salinity increases. pH has the stronger effect: according to SRAC, each increase of one pH unit raises the amount of un-ionized ammonia about tenfold. The table below reproduces part of SRAC Publication 463, Table 1, which was calculated from the equilibrium data of Emerson et al. (1975, Journal of the Fisheries Research Board of Canada 32:2379–2383). Each value is the fraction of TAN in the un-ionized form; the original table also covers 6–14 °C and pH up to 10.2, and does not list values below pH 7.0.
| pH | 16 °C | 20 °C | 24 °C | 26 °C | 28 °C | 30 °C |
|---|---|---|---|---|---|---|
| 7.0 | 0.0029 | 0.0039 | 0.0052 | 0.0060 | 0.0069 | 0.0080 |
| 7.2 | 0.0046 | 0.0062 | 0.0083 | 0.0096 | 0.0110 | 0.0126 |
| 7.4 | 0.0073 | 0.0098 | 0.0131 | 0.0150 | 0.0173 | 0.0198 |
| 7.6 | 0.0116 | 0.0155 | 0.0206 | 0.0236 | 0.0271 | 0.0310 |
| 7.8 | 0.0182 | 0.0244 | 0.0322 | 0.0370 | 0.0423 | 0.0482 |
| 8.0 | 0.0286 | 0.0381 | 0.0502 | 0.0574 | 0.0654 | 0.0743 |
| 8.2 | 0.0445 | 0.0590 | 0.0772 | 0.0880 | 0.0998 | 0.1129 |
| 8.4 | 0.0688 | 0.0904 | 0.1171 | 0.1326 | 0.1495 | 0.1678 |
| 8.6 | 0.1048 | 0.1361 | 0.1737 | 0.1950 | 0.2178 | 0.2422 |
| 8.8 | 0.1566 | 0.1998 | 0.2500 | 0.2774 | 0.3062 | 0.3362 |
| 9.0 | 0.2273 | 0.2836 | 0.3456 | 0.3783 | 0.4116 | 0.4453 |
To estimate un-ionized ammonia, multiply the measured TAN by the fraction for the tank's pH and temperature. SRAC's worked example: at pH 8.6, 30 °C and 3 mg/L TAN, 0.2422 × 3 gives about 0.73 mg/L NH3. Applying the same method to the table, 1 mg/L TAN at 26 °C corresponds to about 0.006 mg/L NH3 at pH 7.0 but about 0.057 mg/L at pH 8.0 — the same test reading carries roughly ten times the toxic load in the harder, more alkaline tank.
What a test kit actually reads
Common ammonia test kits do not measure NH3 directly. According to UF/IFAS, both common chemistries — the Nessler method and the salicylate method — measure total ammonia nitrogen in mg/L (ppm), so the un-ionized share has to be worked out from pH and temperature as shown above. A reading is therefore only interpretable together with a pH reading and a thermometer reading taken at the same time.
- Nessler-type kits can give falsely high readings when ammonia-binding products or many water conditioners are present, and for 24–72 hours after formalin treatment.
- Salicylate-type kits are not affected by ammonia-binding products or formalin.
- Both report TAN, not the toxic fraction; the fraction table converts one into the other.
Implications for water changes
UF/IFAS notes that a 25–50 % water change will remove some ammonia, provided the incoming water itself contains none, and that this is practical in tanks and very small ponds; SRAC adds that maintaining high dissolved oxygen slightly reduces the toxic effect of un-ionized ammonia. Because the toxic share depends only on pH and temperature, any step that raises pH while ammonia is still present increases the NH3 fraction of what remains. This applies to replacement water that is more alkaline or warmer than the tank, and to raising pH or alkalinity in a tank that still has measurable TAN; SRAC similarly warns against liming waters with high ammonia because high pH increases ammonia toxicity. Conversely, SRAC observes that the low pH which follows an algae die-off reduces the toxic proportion of the ammonia it releases.
The underlying problem is removed by nitrification, in which bacteria oxidise ammonium first to nitrite and then to nitrate. UF/IFAS states that nitrifying bacteria function poorly at total alkalinity below 20 mg/L, that a new biofilter needs six to eight weeks to build sufficient bacteria, and that nitrification produces CO2 and hydrogen ions that lower pH unless enough alkalinity is present. Overfeeding is a major cause of high TAN, and a temporary reduction in feeding is the standard aquaculture response; UF/IFAS considers withholding food for one to three days to have no negative effect on the fish.