CYA and Lifetime of Chlorine

After accumulating multiple pieces of conflicting evidence, I think it's about time to discuss and investigate the mechanism of how CYA protects chlorine from sunlight. The starting point for the theory, that I'm starting to think is only partially correct and needs to be enhanced, is that the chlorine that is in the form of hypochlorous acid or hypochlorite ion breaks down in direct noontime sunlight with a half-life of around 35 minutes while chlorine that is attached to Cyanuric Acid (CYA), also known as chlorinated isocyanurates, breaks down from the sun with a half-life of around 8.4 hours.

This graph shows the net result. The conclusion from this graph is that a little CYA provides a lot of protection of chlorine and that there are diminishing returns for using high CYA levels. There are two pieces of evidence that are in conflict with this theory:

1) Some users, most notably Janet (user name Aylad), report that in their non-SWG pools using high levels of CYA shows dramatic improvement in chlorine's staying power. In Janet's case, with a CYA below 60 she found that the FC would go from 7-8 to 2-3 in one day (5 ppm FC per day) while with a CYA of 80-90 the FC would go from 8-9 and take 3 days to go below 5 (about 1.2 ppm FC per day). That is a huge improvement that is wholly inconsistent with the graph.

2) Several users of SWG pools have found that raising the CYA to higher levels, especially approaching 70-80 that some manufacturers recommend, has a dramatic increase in FC levels at the same SWG output. Though one theory is that the CYA makes the SWG cell more efficient by combining with the generated chlorine in the cell "hiding" it from the plates in terms of equilibrium (thus making the generation proceed more quickly), an alternative explanation proposed by some is that the higher CYA levels simply protect the chlorine from destruction from sunlight at a rate faster than the baseline theory outlined at the start of this post.

I've been thinking of mechanisms that might explain the above data and that could be added to the theory to make it predict more accurately. One possibility is that CYA itself is able to absorb UV radiation and possibly re-radiate it as non-UV radiation at lower energy, with the rest of the energy becoming kinetic (i.e. heat or temperature increase). This link shows that indeed CYA does absorb UV at the pH found in pools, though it absorbs even more in more basic/alkaline solutions.

If one adds direct CYA absorption and essentially shielding of UV from lower depths of the pool, then the "CYA shielding chlorine" description would in fact be accurate for this mechanism (while it is not accurate to describe the chlorinated isocyanurates which do not "shield" chlorine but are distinctly different molecules with different absorption rates and affect disinfecting chlorine levels). The net effect of this new mechanism would be to have higher CYA levels reduce chlorine loss at a greater rate than shown in the graph I linked to at the top of the post.

So how can we prove that this new mechanism exists (or is likely) and explains what is being reported in (1) and (2) above? Let's start with the easier of the two, namely the second item of whether CYA improves SWG cell efficiency. This can readily be determined by comparing SWG FC output at different CYA levels, BUT with no sunlight shining on the pool (i.e. either at night or with an opaque cover or with an indoor pool). To the degree that CYA increases the SWG cell output to generate higher FC levels, then this leads credence to the efficiency theory; if not, then the protection from degradation from sunlight is more likely.

As for whether CYA "shields" chlorine through absorption of UV (clearly it does absorb some UV, but the question is more one of whether this is a significant mechanism in quantity), this should be a function of the depth of the pool. The presence of higher concentrations of CYA essentially lower the density of UV radiation reaching lower depths in a pool. So this protective effect of CYA should show up more in deeper pools where a significant fraction of the water is at greater depths and should be less effective in shallower pools. The chlorinated isocyanurates, on the other hand, do not have this same effect since they do in fact degrade (the chlorine attached to them degrades to chloride ion) and are in fact less likely to interact with sunlight in this way so light is more likely to continue to lower depths (i.e. it doesn't act as a shield and even if it does, it's a much smaller concentration than unbound CYA itself).

The experiment would be harder and would require measuring the difference in the destruction of chlorine in waters of different depths at varying CYA levels. The half-life of the chlorinated isocyanurate would be the dominant factor in the shallowest basin of water while CYA's "shielding" effect would be a greater factor in the deepest basin of water.

Richard

One factor you forgot is the penetration of UV into the water itself. UV effects will be most pronounced at the upper water level and less at the depths.

I looked up the absorption coefficient of light in water and I knew that blue penetrated more deeply (which is why water looks blue -- the longer wavelengths get absorbed at shallower depths and they also scatter more). The absorption rate may be seen in a graph at this link where the absorption in the 235-240 nm range that is the peak area where hypochlorous acid (and hypochlorite ion) breaks down from UV is 0.0005 cm^-1 = 0.05% per cm = 1.5% per foot so there is very little absorption of UV from the water in pool depths. Even at 8 feet the light intensity is exp(-0.015*8)=88.7% so only an 11% loss. For red light the coefficient is 0.01 cm^-1 = 30% per foot so at 8 feet the light intensity is exp(-0.30*8)=9.1% so nearly 91% of red light is absorbed.

[EDIT] It is true that the shorter wavelengths near 200 nm are similar to the red wavelengths in that water strongly absorbs such wavelengths. So the real calculation is to take the spectra of sunlight and multiply it by the absorption (actually, the photolysis) spectra of hypochlorous acid and hypochlorite ion vs. wavelength and multiply that by the absorption by water vs. wavelength to see the net effect. I can't find the photolytic absorption and breakdown of chlorine vs. wavelength anywhere, so if anyone can find that then this would be helpful. [END-EDIT]

The bigger problem with the CYA absorption theory is that the absorption spectrum in the link in the first post in this thread shows that the CYA absorption is at lower wavelengths and doesn't appear to be near 235-240 nm. [EDIT] However, later calculations in posts below show that it only takes a relatively low absorption to have a significant effect and that chlorine itself shields lower depths essentially through sacrificial breakdown. [END-EDIT]

Richard

Richard,

What are the assumptions behind the half life curve? For example, does it assume a constant UV exposure? Over the course of a day, I would assume that UV exposure varies quite a bit depending on the individual's pool. Factors such as sun angle, shade and pool covers will vary the results substantially. The evedence sited could be due to other factors and not necessarily the CYA level although that would contribute to it as well. So to correlate CYA levels to protection of chlorine other factors must be taken into consideration:

1) CYA level - probably is dominate factor
2) Pool latitude and time of year
3) Shade and what time of day
4) Pool cover usage
5) How often pool is used
6) How much organics are dropped in the pool

Although items 2-6 may be secondary, I am not sure they can be ignored. I suspect that this is why there are such varying results for different levels of CYA.

The curve is based on half-life numbers that are for direct noontime directly overhead sun (i.e. summer in L.A. and Florida latitudes). However, there are some inconsistencies depending on the source. For unbound chlorine (i.e. no CYA) the half-life is usually given as 35 minutes based on average pool depths but I found one source that found 11.6 minutes but that was for water in a quartz (transparent to UV) tube. For the bound chlorine (i.e. chlorine attached to CYA), this was calculated from the pool study curves but again I found two such sets of curves, one implying 8.4 hours and the other implying 6 hours. I used 35 minutes and 8.4 hours for the graphs I made. However, changing these numbers even by factors of 2 or so doesn't get even close to Janet's results.

It is this latter assumption that the protection of chlorine from sunlight is primarily due to the bound chlorine having a longer half-life that I am questioning. Not that this may not be happening, but that there may be an additional factor from CYA by itself acting as a "shield" of sorts protecting lower depths.

It is true that Janet's reports may have been influenced by other factors such as those that you listed, especially those that could have lowered chlorine demand after adding more CYA such as lighter bather load or fewer organics or time of year, etc., but it sounded like there wasn't much change in these other parameters. In any event, having a net chlorine loss go from 5 ppm FC per day to around 1.2 ppm FC per day is huge and this latter loss at FC levels between 5 and 9 ppm FC is far lower than predicted by direct sunlight which hits her pool. So even if there were other factors partially contributing, the only known change was the increase in CYA so I thought it worth questioning the prevailing theory.

Richard

Do you think a small experiment using a bucket of water, < 3 Gallons, and small amounts of CYA and chlorine would be valid and applicable to a swimming pool? To experiment with a pool is difficult to control and can be expensive as well so I though perhaps a reduced size experiment might actually measure the results you presented. One could measure the CL loss over a 24 hour period for varing amouts of CYA. To me, this would be fairly definitive and perhaps show a significant increase in retention going from 60 to 80 for which current understanding has not accounted for.

However, it wouldn't actually tell us much about SWG operation but if people without SWGs are seeing improvements when raising CYA from 60 to 80, then perhaps what is happening has less to do with SWGs and more to do with binding properties of CYA.

The bucket test would be useful for determining effects that are not dependent upon depth. So if there is some sort of chemical or physical protection not yet accounted for (i.e. more than chlorinated isocyanurates), then that would be seen in such a test. A separate test with something deeper (and wider, so the sun can shine in to full depth) would be needed to test the "shield" effect from direct CYA absorption of UV. Probably a garbage can might be OK for an hour of overhead sun, but wouldn't work for more than that unless the barrel could be tilted to point towards the sun. Ideally, multiple bucket or cans with differing CYA concentrations and the same starting FC would be used at the same time -- that helps eliminate daily variations in sunlight, etc.

Another experiment that might work in a pool that already has a high CYA in it would be to turn off the circulation pump (and do not use the pool and have it be a calm windless day) and after some length of exposure to the sun, then measure the Free Chlorine level at different depths. The problem is that diffusion will tend to reduce differing concentrations, but some sort of gradient should be apparent if the "shield" effect is strong enough. If there were a measuring device (or something sensitive to UV) that could be made waterproof (e.g. inside a quartz glass container) and could measure UV, then that would be pretty definitive by measuring UV levels at various depths and various concentrations of CYA.

Remember that for a given FC level it is still expected to see the chlorine last longer at higher CYA levels. It's the specific amount of protection vs. CYA level that is what we are looking for.

Richard

The formula for determining the light intensity passing through a solution is as follows:

I/Io = exp(-µ*l)

where µ (mu) is the absorption coefficient and is a function of wavelength (so the above formula is for a specific wavelength). "l" is the path length which for the units in the links I referred to is in centimeters.

There is also a formula for absorbance defined as follows:

A = -log10(I/Io)

and there is a molar extinction coefficient defined by the following equation:

A = ε*c*l

where ε (epsilon) is the molar extinction coefficient, c is the molar concentration in moles/liter, and l is the path length in centimeters.

so ε = µ / (log10(e) * c) = µ / (2.303 * c)

3 feet is about 91 centimeters and the molar concentration of 50 ppm CYA is 0.00039 moles/liter. So,

µ = 2.303 * c * ε = 0.00090 * ε

So to get any reasonable absorption from CYA (so that µ is near 1/91 so at 91 cm we have µ*l = 1) we need ε to be over 10. This source gives an extinction coefficient (for gaseous HOCl) of 123 M^(-1)cm^(-1) with an absorption peak at 220 nm. It looks like this may happen from the absorption spectra graph in the range of breakdown of hypochlorous acid. This link provides interesting detailed information about chlorine (and bromine and chlorine dioxide) in terms of half-life at various depths (no CYA present). Interestingly, there is quite a difference in half-life by depth at higher concentrations of chlorine indicating that the chlorine at shallower depths acts as a sacrificial shield to chlorine at lower depths and that this effect is more pronounced at higher concentrations. This makes sense, though it's something I hadn't thought of until I saw this data. It implies that the overall half-life averaged over all depths is longer at higher chlorine concentrations (but remember this is with no CYA at all). The difference in half-life by depth in summer at 0.53 ppm FC is 11 minutes at 0 meters, 26 minutes at 1 meter, 48 minutes at 2 meters, and 71 minutes at 3 meters (implying an extinction coefficient of 500) while at 17.6 ppm the half-life by depth was 9 minutes at 0 meters, 10.3 hours (not minutes) at 1 meter, 20.7 hours at 2 meters, and 31 hours at 3 meters. (implying an extinction coefficient of 74 in the first hour; later hours don't make sense). Thus, the chlorine is most depleted from water near the surface so having good circulation is essential in order to keep chlorine levels more uniform throughout the pool. It also appears, from the pH dependence, that perhaps hypochlorous acid (HOCl) is less susceptible to breakdown from sunlight than hypochlorite ion (OCl-). This implies that having a pool at lower pH results not only in more disinfecting chlorine, but has the chlorine last longer (though the effect may not be very strong from, say, 7.8 to 7.2).

Also, note that there is a non-linear effect from the concentration of whatever protective agent is present at the shallower depths (be it hypochlorous acid itself or CYA). So if I use a molar extinction coefficient of 10 and 50, then I would get the following for I/Io at 3 foot depth:

CYA (ppm) ... I/Io (10) .. I/Io (20) .. I/Io (50)
0 ................. 1.00 ........ 1.00 ........ 1.00
10 ................ 0.85 ........ 0.72 ........ 0.44
20 ................ 0.72 ........ 0.52 ........ 0.20
30 ................ 0.61 ........ 0.38 ........ 0.09
40 ................ 0.52 ........ 0.27 ........ 0.039
50 ................ 0.44 ........ 0.20 ........ 0.017
60 ................ 0.38 ........ 0.14 ........ 0.0077
70 ................ 0.32 ........ 0.10 ........ 0.0034
80 ................ 0.27 ........ 0.074 ........ 0.0015
90 ................ 0.23 ........ 0.054 ........ 0.00067
100 .............. 0.20 ........ 0.039 ........ 0.00030

So to see the dramatic change seen from higher CYA levels, the CYA shielding effect has to be strong enough to be the predominant effect. The shielding effect would "shield" not only unbound chlorine, but also chlorine bound to CYA. Note that using an extinction coefficient of 20 in the above table one finds the difference between 50 and 90 ppm CYA being a factor of 3.7 which is not far off from the factor of 4.2 that Janet was seeing. So perhaps adding an additional protection factor similar to the "20" column in the above table might be the thing to do. This link indicates that the chlorinated isocyanurates are unstable in sunlight, but it is unclear how much of that is due to breakdown from the equilibrium hypochlorous acid vs. direct breakdown itself. The study just shows that CYA is itself stable in sunlight. If the CYA absorption effect is really this strong, then deeper pools should be more protected at the same CYA level since more of their water volume will be at deeper depths "shielded" from the UV.

An experiment using shallow depth water with different levels of CYA will help isolate the two effects. If the CYA "shielding" or absorption is the main effect, then there should be little protection of chlorine in shallow water. If instead the chlorine combined with CYA has a longer half-life and that is the main effect, then higher CYA levels even in shallow depths should show significant protection and should roughly follow the curve in this graph. I suspect that there will be a some of both processes going on.

The original CYA patent by Fuchs may be seen at this link. There were interesting laboratory tests that appear to have been made at shallow depths and only show a small amount of the "depth" variation one sees with higher chlorine levels. The UV lamp they used appeared to have 1 ppm FC drop to 0.5 ppm FC in 1.7 hours so was not as strong as sunlight. The rate of chlorine loss seemed to track the amount of unbound chlorine, but with diminishing returns starting at a rate of 0.29 per hour at no CYA, 0.16 per hour with 1 ppm CYA, 0.13 per hour with 2 ppm CYA, 0.092 per hour with 5 ppm CYA, 0.071 per hour with 50 ppm CYA and an actual increased loss of 0.088 per hour at 100 ppm CYA. This is somewhat consistent with the original theory of a 35 minute half-life in direct sunlight with no CYA and an 8.4 hour half-hour limit when bound with CYA. This is probably where the industry got its original data for its tables. Note that CYA also has a protective effect on chlorine loss from oxidation of iron and copper. Though the patent speculates CYA may coat metals, it appears that the effect is explained by the reduction in disinfecting chlorine and therefore the rate of corrosion based on its concentration. It should be noted that in the patent "real pools" showed the greater protection effect of higher CYA levels by about a factor of 2 at 10 ppm CYA and over a factor of 3 at 50 ppm CYA. Thus there does appear to be a "shielding" depth factor for CYA protection separate from that explained solely by Cl and Cl-CYA breakdown. The fact that the chlorine levels were the same and only the CYA level increased, yet had a greater effect in a real pool with "depth" is very strong evidence.

The good news with this new information is that at sufficiently high CYA levels using a higher FC (to compensate for disinfection and prevention of algae) should not result in larger losses. Going from 30 ppm to 90 ppm requires about triple the FC level, but the loss rate may be cut down by a factor of 7 for a net overall savings of over a factor of 2. If we can validate this, then it should be possible to run a high CYA pool with high FC levels economically, especially in deeper pools.

Richard

I got a response from Janet saying that the depth of her pool is 3.5 feet in the shallow end to 8.5 feet in the deep end (hopper shape with a long gradual slope). This is deeper than many pools with a 3 foot shallow end and a 6 foot deep end so could help explain why the higher CYA shows such a dramatic shielding effect in her pool. It should show this effect in the 6 foot pools as well, but would be more pronounced in a deeper pool. Her pH is usually 7.4 but tends to drift up towards 7.8 (I asked that since chlorine is more protected at lower pH). I think we're getting somewhere!

So now we just need to see if the SWG effect has anything to do with efficiency of the SWG cell or if it's just an issue of increased protection of chlorine from sunlight. That's not an easy test to do unless someone hasn't yet added high CYA levels to their SWG pool and is willing to increase those levels (and measure the SWG output rate overnight before and after the increase in CYA).

Richard

I was planning on slowly raising my CYA from 30 ppm to 80 ppm but to take advantage of solar, I am running my pump during the day.

One thought though is that pools with solar covers should have a considerable sheilding effect of UV. Experiencing an increase in production with a solar cover may indicate the same thing as doing the test overnight.

Another way to perform a shorter test would be to put the SWG into spa mode and run for a measured period of time. This would make it easier to test the chlorine levels since they should be higher and the test could be done in a much shorter period of time.

Mark,

Agreed. The spa mode sounds like a much faster way to test for the SWG efficiency effect (either at night or with a UV-shielding cover) since the volume of water is much smaller. I'd hate for you to increase your CYA level in your pool only to find out that you may want to dilute it, but we already know that most high-CYA (70-80 ppm) SWG pools can run fine at lower FC (most are at 3 ppm) though some need to run closer to 5 ppm.

Running the SWG at night or with a cover should have the FC rise (rather than be maintained) so it will be any difference in the rate of rise (if any) that is what we are looking for.

Thanks,
Richard