Archive for June, 2013

CYA as a hypochlorous acid buffer

This post will describe the effects of using Cyanuric Acid (CYA) to lower active chlorine concentration in terms of breakpoint chlorination and the net resulting chloramine production. The basic breakpoint reaction was described by Griffin (1939), but the first reasonable detailed model was proposed by Wei & Morris (1974 — in Chapter 13 in the same “Chemistry of Water Supply, Treatment and Distribution” book that has the O’Brien paper on the chlorine/CYA equilibrium constants). Subsequent improvements were made to the model by Saunier & Selleck (1976) and most recently by Jafvert & Valentine (1992) and Vikesland, Ozekin and Valentine (2000) which should be considered to be the best model to-date. The following is the paper (a link to be able to purchase the most recent two online is here and here).

Chad T. Jafvert and Richard L. Valentine, “Reaction Scheme for the Chlorination of Ammoniacal Water”, Environ. Sci. Technol., Vol. 26, No. 3, 1992, pp. 577-585.

Though the model lists 14 reactions, including both forward and reverse reaction rates, the dominant reactions are the following:

(1) HOCl + NH3 —> NH2Cl + H2O
Hypochlorous Acid + Ammonia —> Monochloramine + Water

(2) HOCl + NH2Cl —> NHCl2 + H2O
Hypochlorous Acid + Monochloramine —> Dichloramine + Water

(3) HOCl + NHCl2 —> NCl3 + H2O
Hypochlorous Acid + Dichloramine —> Nitrogen Trichloride + Water

(4) NHCl2 + NCl3 + 2H2O —> 2HOCl + N2(g) + 3H+ + 3Cl
Dichloramine + Nitrogen Trichloride + Water —> Hypochlorous Acid + Nitrogen Gas + Hydrogen Ion + Chloride Ion

The first reaction producing monochloramine is by far the fastest. It is over 95% complete in one minute when the FC is around 10% of the CYA and the ammonia is much less than the chlorine so that the chlorine level remains fairly constant. With no CYA, the reaction is mostly complete in a couple of seconds. The subsequent reactions are far slower.

You can then see that hypochlorous acid participates in two reactions (after initially producing monochloramine quickly), one producing dichloramine and another producing nitrogen trichloride so the net reaction varies as the square of the hypochlorous acid concentration. You can see that nitrogen trichloride is broken down by dichloramine and the latter is produced with a reaction rate that varies linearly with hypochlorous acid concentration. So in the steady state, the amount of nitrogen trichloride is linearly dependent on the hypochlorous acid concentration. This can also be seen by the following rate reaction balance at steady state.

k3*[HOCl]*[NHCl2] = k4*[NHCl2]*[NCl3]
Rate of formation of Nitrogen Trichloride = Rate of destruction of Nitrogen Trichloride

so, k3*[HOCl] = k4*[NCl3]

The nitrogen trichloride concentration in the steady state is linearly proportional to the hypochlorous acid concentration. Since nitrogen trichloride is very volatile, this implies that the rate of outgassing of nitrogen trichloride may be proportional to the hypochlorous acid concentration since the outgassing rate is likely to be proportional to its concentration in the water.

A similar rate reaction balance for dichloramine gives the following.

k2*[HOCl]*[NH2Cl] = k3*[HOCl]*[NHCl2] + k4*[NHCl2]*[NCl3]
Rate of formation of Dichloramine = Rate of destruction of Dichloramine

and substituting the earlier steady-state equation we have

k2*[HOCl]*[NH2Cl] = k3*[HOCl]*[NHCl2] + k3*[HOCl]*[NHCl2]

which reduces to

k2*[NH2Cl] = 2*k3*[NHCl2]

So the ratio of monochloramine to dichloramine is constant and independent of hypochlorous acid concentration.

We can look at the steady-state for monochloramine assuming a constant introduction of ammonia into the water.

k1*[HOCl]*[NH3] = k2*[HOCl]*[NH2Cl]
Rate of formation of Monochloramine = Rate of destruction of Monochloramine

so the ratio of ammonia to monochloramine is constant and independent of hypochlorous acid concentration. Finally, we can look at the steady-state for ammonia.

k = k1*[HOCl]*[NH3]
Rate of formation of Ammonia = Rate of destruction of Ammonia

which says that for a constant rate of introduction of ammonia, the amount of ammonia, and therefore monochloramine and dichloramine (from above), are inversely proportional to the hypochlorous acid concentration.

Earlier models had reactions forming an intermediate, and the Jafvert & Valentine model has this as well, but it is not the dominant reaction in that model. The following shows the intermediate reactions such as found with Wei & Morris.

(5) NHCl2 + H2O —> NOH + 2H+ + 2Cl
Dichloramine + Water —> Intermediate + Hydrogen Ion + Chloride Ion

(6) NOH + NH2Cl —> N2(g) + H2O + H+ + Cl
Intermediate + Monochloramine —> Nitrogen Gas + Water + Hydrogen Ion + Chloride Ion

(7) NOH + NHCl2 —> N2(g) + HOCl + H+ + Cl
Intermediate + Dichloramine —> Nitrogen Gas + Hypochlorous Acid + Hydrogen Ion + Chloride Ion

In the Wei & Morris model, there is no destruction of nitrogen trichloride, so it’s rate of production is the product of the hypochlorous acid concentration and the dichloramine concentration. In the above, reaction (7) is more dominant than reaction (6). The formation of the intermediate NOH is a rate limiting step so dichloramine is built up and therefore the rate of production of nitrogen trichloride is linearly dependent on the hypochlorous acid concentration.

For a realistic example, consider a pool with 3 ppm FC and no CYA vs. a pool with 3 ppm FC and 30 ppm CYA. Both are at a pH of 7.5 (there is far more nitrogen trichloride produced at lower pH) and the temperature is 77F. If it is assumed that the chlorine level is maintained at a constant level and that there is a constant introduction of ammonia in the water at a rate of 0.1 ppm N per hour, then we have the following steady state amounts (using Jafvert & Valentine in a spreadsheet I made here):

SPECIES ……………………… NO CYA ………… 30 ppm CYA
Monochloramine …………. 0.02 ppm ……….. 0.70 ppm
Dichloramine ……………… 2.97 ppb ………… 85.42 ppb
Nitrogen Trichloride …….. 70.96 ppb ………. 2.35 ppb

You can see from the above that with no CYA in the water, there is less monochloramine and dichloramine but more nitrogen trichloride compared to having CYA in the water. The differences are roughly a factor of the CYA level because that is roughly the difference in the hypochlorous acid concentration (the breakpoint chlorination spreadsheet assumes 3 ppm FC with 30 ppm CYA results in about 0.05 ppm hypochlorous acid at pH 7.5 — the actual amount is closer to 0.042 ppm). Nitrogen trichloride is the most volatile and irritating. The monochloramine odor threshold is 0.65 ppm; for dichloramine it is 100 ppb; for nitrogen trichloride it is 20 ppb. The equilibrium concentrations in air for monochloramine and dichloramine are somewhat lower than that in water, but nitrogen trichloride is extremely volatile so will not saturate the air before becoming extremely noticeable and irritating.

The above is just for breakpoint chlorination of ammonia. As seen in Table 4.1 on document page 62 (PDF page 85) of this link, urea has 68% of the nitrogen in sweat compared to 18% for ammonia while in urine it’s 84% vs. 5%. There is no definitive model for oxidation of urea by chlorine, though some mechanisms have been proposed (by Wojtowicz) including the slow formation of a quad-chloro urea followed by rapid breakdown to dichloramine and nitrogen trichloride. If I repeat the above analysis using an 80%/20% split of urea to ammonia and assume a steady state buildup, then I get the following results.

SPECIES ……………………… NO CYA ………… 30 ppm CYA
Monochloramine …………. 0.01 ppm ……….. 0.28 ppm
Dichloramine ……………… 1.19 ppb ………… 34.17 ppb
Nitrogen Trichloride …….. 70.84 ppb ………. 2.35 ppb

You can see that the resulting nitrogen trichloride is the same as before, but that there is lower monochloramine and dichloramine by a factor of 2.5.

The rate of ammonia/urea introduction of 0.1 ppm N per hour is for heavy bather loads since it represents a chlorine usage of nearly 1 ppm FC per hour. One swimmer may produce around 0.1 ppm N per hour in 1000 gallons so only a pool with many people being active would have this sort of usage. Of course, having children in the water that urinate would provide a very high load. If a child urinates 100 ml (3.4 fluid ounces), then in 1000 gallons this is about 0.3 ppm N.

Note that the urea model assumes no interactions between the chloramines and chloroureas or related species and that’s probably not realistic, but there are no studies or models I can find analyzing such interactions.

Since the UV in sunlight breaks down nitrogen trichloride fairly quickly and since air circulation is also good outdoors, the current recommendations for FC as a % of CYA are reasonable for outdoor pools. The slower breakpoint is not generally a problem unless the bather load is high. For commercial/public pools with higher bather loads, an FC that is 20% of the CYA level may be more appropriate. For indoor pools, the slower breakpoint might be more of an issue so perhaps an FC that is 20% of the CYA level may be better even when there is not high bather load in such pools. From the models, not using any CYA at all in any pool (indoor or outdoor) can result in far higher irritating nitrogen trichloride concentrations and also has the chlorine level be too strong for corrosion and oxidizing swimsuits, skin and hair. Since it is not practical to maintain 0.2 ppm FC everywhere in an indoor pool due to local usage and imperfect circulation, using CYA as a hypochlorous acid buffer makes sense, but should not be overdone.