Downsides to salt pools

[waste]

Richard, (as often happens when I try to keep my 'word count' down), I misrepresented what I was trying to say (go figure ). Though I specificly said I'd look at the SWCG pool areas, it was my own 'shorthand' for 'looking' at them compared to pools of the same age and ~ usage which diddn't have the SWCGs. This also raises the issue of my not being 'blind' (ie 'double-blind experiment') - undoubtably, I'll report some more or less corrosion to the pools in question. However, since I work in a relatively small area, many other factors will be removed from my observations. (in my 'miss-spent' youth, I studied psychology, and as it's a 'soft science', they made sure to 'pound-in' the necessity of using the 'hard science' methodology)

[Waterworks]

I just found another potential problem, or at least a very good source for testing. A residential customer of ours with an indoor salt pool called and said his ladders and light were corroding, so I went to check it out with my service man. The ladders were so corroded that they were falling apart. His pool is full-tile and his original white grout lines are now black, his original blue tiles are now greenish. The light ring is black also. We went and checked his cartridge filters and they were black, caked with rust. The pump basket was also black. The ladder rails that are above water are rusting, as you'll see in the pictures, but nowhere near as bad as the underwater parts.
Outside the pool there is a slide, metal deck drains, metal exhaust vents, tile decking, tile safety grip coping and a flagstone deck section with grout. The slide posts showed no corrosion, but I think they are aluminum. One of the metal deck drains is showing some rust, one is just fine. the metal exhaust vents show no corrosion. The flagstone and tile deck sections show no corrosion.

We installed the SWG in May 2005. The pool was installed around 2000. Before the SWG came along there was no corrosion. We went on a service call in March of last year to clear up the pool, as it was cloudy. At that point, the generator had been turned all the way down. We had not had any complaint of corrosion, and did not actively look for it at this point. My service man turned it up to 30% and told the customer to keep checking the chlorine levels and to keep them high until the pool was clear again, then turn the SWG down. The customer has not brought in a water sample since, and while I was speaking with him today he did not know where his test kit was. When I looked at the SWG today, it was still set on 30%. The pool room had absolutely no chlorine smell and was not hard on my eyes. His kids use the pool almost every day and have never complained about anything.

Here are the results of my water test:
Chlorine - I used my FAS-DPD test kit. I filled the sample to 10ml, then added two scoops of powder. As the powder hit the water it turned bright pink, but then went back to clear. As I shook the sample, it would periodically turn to pink, but then back to clear. I'd seen this before so I added my R0871 and after 25 drops the sample had turned pink and stayed that way. I kept adding drops until I had done 110. Then I mixed 1 part pool water with 3 parts distilled water and it still took 100 drops in a 10 mL test, meaning a chlorine reading of somewhere between 55-100 ppm. I did an OTO test and the sample turned bright red.
pH - phenol red test was off the charts so I used my Hannah meter and got 8.4
TDS (Our city water is about 650ppm) - 2850
Salt - 3200 on the Aquarite machine and 3000 on my Taylor Kit
Alkalinity - 80
Calcium - 140
CYA - 0

It seems to me that this info coincides with what I already thought. Most SWG's create WAY too much chlorine for indoor pools. This excess chlorine causes corrosion. Since this corrosion ties in with the installation of the SWG, the SWG takes the blame.

I brought one of his ladders back to my shop so I could try to do some kind of test on it. I was thinking of setting up 2 buckets, one with high chlorine and one with high salt and setting each leg of the ladder (because they are evenly corroded at this point) in it's own bucket. I would take pictures before and after and compare the effects. I would also test the water every day or two and keep it balanced. I wish I had a way to test a third part so I could do high chlorine with CYA. Does anyone think this test would be worthwhile, and does anyone have any tips to make the test better?

If anyone can tell me how to post a picture, I will post a couple pics of the ladder.

Brad

[chem geek]

I brought one of his ladders back to my shop so I could try to do some kind of test on it. I was thinking of setting up 2 buckets, one with high chlorine and one with high salt and setting each leg of the ladder (because they are evenly corroded at this point) in it's own bucket. I would take pictures before and after and compare the effects. I would also test the water every day or two and keep it balanced. I wish I had a way to test a third part so I could do high chlorine with CYA. Does anyone think this test would be worthwhile, and does anyone have any tips to make the test better?

If anyone can tell me how to post a picture, I will post a couple pics of the ladder.

Brad, I think that this forum is now limiting uploading of pictures, but if you send them to me (via E-mail -- click on my name on any post and there will be a link to E-mail to me) then I can host them on the website I use for the other pool stuff I have and then your post can link to that.

Sorry to hear about this customer, but man, that much chlorine with no CYA is horribly corrosive. As for what test to do vis-a-vis salt, part of what you want to compare is moderate chlorine levels (say, 5 ppm FC with no CYA) with salt at 3000 ppm vs. the same chlorine levels with salt at around 300 ppm. We already know and understand the effects of CYA on chlorine so that isn't as useful. What we really want to know is whether the salt in the presence of chlorine accelerates the corrosion so you want the same chlorine level in both buckets with the only difference between the two being the salt level. To make this a reasonably fair test, go ahead and add baking soda to raise the TA and maybe even some calcium chloride for CH. That would make it very much like pool water (so try and keep the pH around 7.5 as well). Whenever you test or add chemicals, be sure and do the same sort of mixing (circulation) in the two buckets as that also affects corrosion (if there is no circulation, then the buildup of "rust" products probably slows down additional corrosion).

This story really emphasizes the need for customer education about SWG and especially about testing your water regularly. I can hear Carl's words now...

Stay ahead of your water!
Take 2 to 5 minutes every day for pool maintenance!

Richard

[chem geek]

I thought of another possible reason why SWG manufacturers recommend high levels of CYA (typically 70-80 with 80 "best"). In this thread (starting at around post #30) I said that the reason for high CYA in SWG pools is that this is more efficient at producing chlorine. Though this appears to be true (based on pool owner's experiences with runtime), I didn't say what happens when there is a buildup of chlorine at the plate. In addition to production of chlorine, there is also a competing reaction of producing oxygen. If the reaction producing chlorine slows down, then the reaction rate of producing oxygen can speed up in its place, though the production of oxygen is normally less favored than that of chlorine (technically, the production of oxygen is more likely to occur in an equilibrium sense, but it has a much higher activation energy than the production of chlorine so results in an overvoltage such that chlorine is more favored than oxygen).

It is interesting to note that different SWG systems use a different voltage to drive their cells. Some operate in the range of 6-9 volts, but others are in the range of 22-30 volts. Though the voltage level required to only produce chlorine and not oxygen is quite low (perhaps between 1.16V and 1.45V at normal pool concentrations of pH and chlorine), higher voltages will probably produce both chlorine and oxygen with chlorine being produced more. I do not know if higher voltages change this relationship, but suspect that it might, with the oxygen proportion increasing somewhat at higher voltages (though still produced in lower quantity than chlorine). At any rate, having lower CYA levels may very well increase the oxygen proportion and this can be dangerous since the other plate is producing hydrogen. Though hydrogen by itself is not explosive and in water does essentially nothing, the combination of oxygen and hydrogen can be dangerous if this combination of gasses were to accumulate (if the pump shuts off but the SWG cell keeps running, for example).

The blog link at the first post in this thread has been reporting some explosions of SWG systems and I wonder if those pools were using lower CYA levels than recommended and produced more oxygen that could have been at explosive levels with hydrogen. It is also possible that the explosions weren't actual ignitions of hydrogen and oxygen, but high pressure bursts from the buildup of gas pressure. It should be understood that under normal operating conditions when the SWG only operates when the pump is running, that the risk of explosion is virtually nil. I'm just trying to sort out some unusual circumstances (where the salt cell didn't shut off when the water flow stopped) to better understand what is going on.

[EDIT] A further note: Even the most dangerous levels of hydrogen gas and oxygen gas in a 2:1 ratio will not spontaneously explode unless the pressures are very high (much higher than what pipes would handle, so bursting would occur before explosion) or the temperatures are very high (at least 400C) so generally some sort of spark or flame is required to trigger a literal thermal explosion. In other words, the risk of oxygen production and explosion is low; the risk of a pressure burst is high if the salt cell operates in a closed valve environment (which would only occur if some component failed to operate properly). [END-EDIT]

Richard

[giroup01]

At any rate, having lower CYA levels may very well increase the oxygen proportion and this can be dangerous since the other plate is producing hydrogen.

Richard,

Can you expand a bit on the chemistry behind low CYA levels favoring O2 production ?

Thanks,

[chem geek]

Can you expand a bit on the chemistry behind low CYA levels favoring O2 production ?

Sure.

First let's take a look at the only reaction that appears to be reasonably possible at the cathode (the negatively charged plate). The other reactions that can occur at this plate include the plating of metals (iron, copper) that are in the water. Also, the [EDIT] high (sorry about saying "low" at first -- I was "thinking" the right thing but "wrote" the wrong word) [END-EDIT] pH in the region of this plate make scaling of calcium carbonate much more likely.

2H+ + 2e- --> H2(g) ...................... Eo = 0V

In other words, this is the generation of hydrogen gas. It can also be written as coming from water to produce hydroxyl ion:

2H2O + 2e- --> H2(g) + 2OH- .......... Eo = -0.8277V

The different Eo readings reflect the fact that Eo is defined to be at standard conditions which means that each of the species dissolved in water have a concentration of 1 mole/liter and each gaseous component has a partial pressure of 1 atmosphere.

Now let's look at the two different competing reactions that appear to be reasonably possible at the anode (the positively charged plate):

2H2O --> O2(g) + 4H+ + 4e- ........... Eo = -1.229V

2Cl- --> Cl2(g) + 2e- ....................... Eo = -1.35827V
Cl2(g) + H2O --> HOCl + H+ + Cl-
----------------------------------------
Cl- + H2O --> HOCl + H+ + 2e- ......... Eo = -1.482V

The first reaction is where oxygen is produced from water. The other reactions are the production of chlorine where I show the additional reaction of chlorine dissolving in water to produce hypochlorous acid. You can see that the Eo value is more negative after chlorine gas dissolves in water to make hypochlorous acid and that would seem to imply that the reaction is not favored, but again this is only because the Eo is defined for standard conditions (which are theoretical and may not always be actually achievable) of 1 atmosphere pressure of chlorine gas with 1 mole/liter concentration of hypochlorous acid AND 1 mole/liter concentration of hydrogen ion (i.e. a pH of 0), AND 1 mole/liter concentration of chloride ion.

For the cathode, I listed the reduction potentials as oxidation potentials (i.e. with negative numbers) to indicate that a voltage needs to be applied to make the reaction occur. Note that it takes a lower (absolute) voltage to produce oxygen than chlorine. This represents a difference of about 25 kilojoule per mole in favor of oxygen formation over chlorine formation. HOWEVER, chemical reactions have what is known as an activation energy which is an energy barrier that must be overcome before the reaction can happen. In electrochemistry, this barrier is called overvoltage which represents the additional electrical potential needed to get an electrochemical reaction started. Oxygen has a very high overvoltage of 700 mV while chlorine has an overvoltage of only 50 mV.

So if we add the overvoltages to the standard potentials, we see that to produce oxygen it takes 1.229+0.7 = 1.929V while to produce chlorine it takes 1.35827+0.05 = 1.40827V. Or we can look at producing hypochlorous acid directly as 1.482+0.05 = 1.532V. Now these voltages do not account for the actual concentrations of the various species in pool water as I noted above.

If we want to really know about the true voltages required and which reactions are favored, we need to convert from standard conditions to actual concentrations. There is a formula for doing that called the Nernst equation as follows:

E = Eo - RT/nF * ln(K)

where "K" is the ratio of product concentrations (raised to powers corresponding to their stoichiometric quantity in the formula) to reactant concentrations. "R" is the gas constant, "T" is the temperature in Kelvin, "n" is the number of electrons in the equation and "F" is the Faraday constant. Sometimes this Nernst equation is written with a "+" and the "K" value is inverted to have reactants divided by products.

Anyway, if I assume reasonable initial conditions of a pH of 7.5 to get [H+] = 3.7x10^(-8), normal dissolved oxygen levels (based on 21% oxygen in the air and using Henry's Law) to get [O2] = 2.6x10^(-4) moles/liter and converting to 0.0065 atmospheres (partial pressure), using normal chlorine levels of 3 ppm with 30 ppm CYA to have [HOCl] = 5.8x10^(-7) moles/liter and chloride concentration of 3000 ppm to have [Cl-] = 0.052 then this adjusts the oxygen equation by (0.02585/4)*ln(0.0065*(3.7x10^(-8))^4) = -0.47V so E = -0.75V for oxygen. The chlorine equation is adjusted by (0.02585/2)*ln(5.8x10^(-7) * 3.7x10^(-8) / 0.052) = -0.37V so E = -1.11V for chlorine. I'm not certain on the handling of the oxygen quantity in terms of molarity vs. partial pressure, (I believe I did it correctly) but it turns out it doesn't matter much in this case so let's keep going.

Applying the overvoltages again gives is E = -1.45V for oxygen while we get E = -1.16V for chlorine. So at least initially, the production of chlorine is favored over the production of oxygen and if the voltage applied to the salt cell is set to be BETWEEN 1.16V and 1.45V, then only chlorine can be generated and oxygen cannot. [EDIT] I neglected to do the concentration adjustment calculation for the cathode for the production of hydrogen, but that would modify the overall voltage by the same amount regardless of whether oxygen or chlorine were produced at the anode so it's not that relevant to this discussion. [END-EDIT]

If higher voltages are used, then both chlorine and oxygen will be produced with chlorine favored over oxygen. If you have ever done the "home" experiment of taking two carbon cores from D-cell batteries and connecting them to a 6V transformer (as I have), then putting this into salt water produces both chlorine and oxygen (you can smell the chlorine and you can "pop" the gasses with a flame where the oxygen and hydrogen combine) and can somewhat vary the ratio of the two depending on the salt concentration. If you use sodium bicarbonate as the electrolyte, then you get only hydrogen and oxygen and not chlorine.

The salt cells appear to use voltages that are much higher than these minimum amounts that would ensure that only chlorine gets generated. I'm sure that this is partly due to the rather low chloride concentration in pool water (especially today, since early salt generators wanted 6000 ppm instead of 3000 ppm). Otherwise, you would need very large generating plates to get decent generation rates. So with voltages of 6-9 volts for some manufacturers and 22-30 volts for others, it would seem that both chlorine and oxygen could get produced. My guess is that the higher voltage units have a greater tendency to produce more oxygen proportionately than the lower voltage units, but that's just a guess.

Now what I do not know is whether the materials (coatings) used for the generating plates in a salt cell are designed to somehow inhibit the oxygen generation, but I suspect that is not the case. (If anyone knows about this, please let us know).

As the electrolysis proceeds, the product concentrations build up at the generating plates (though there is water flow to help keep them lower) and this slows down the generating reactions. Note that the reaction with oxygen produces hydrogen ion so is acidic and that the production of chlorine AND dissolving of that chlorine in water also produces hydrogen ion so is also acidic. However, the oxygen reaction produces one hydrogen for each electron while the chlorine reaction produces half a hydrogen for each electron. This means that as the pH gets lowered near the plate, the chlorine reaction will become more favored, assuming that the chlorine gas is able to dissolve in water rapidly. Having CYA in the water helps make this happen because it combines with the hypochlorous acid that is produced so that the reaction of chlorine dissolving in water can continue at a rapid pace. So the lower buildup of products in the chlorine reaction, due to CYA binding with the generated hypochlorous acid, means that this reaction becomes even more favored compared to the oxygen generation.

This can be netted out as follows:

2Cl- --> Cl2(g) + 2e-
Cl2(g) + H2O --> HOCl + H+ + Cl-
HOCl + CYA --> Cl-CYA + H2O
----------------------------------------
Cl- + CYA --> Cl-CYA + H+ + 2e-

Each succeeding reaction above moves the products into another form and that takes them away from being able to go "backwards" to create reactants again (or to inhibit or slow down the creation of more products). It is almost as if one turned up the water flow with respect to sweeping away chlorine gas, but that this is done chemically.

Richard

[giroup01]

Sure.

Thanks! ...I'll have to chew on that, but yes, thank you!

When you say:

Having CYA in the water helps make this happen because it combines with the hypochlorous acid that is produced so that the reaction of chlorine dissolving in water can continue at a rapid pace. So the lower buildup of products in the chlorine reaction, due to CYA binding with the generated hypochlorous acid, means that this reaction becomes even more favored compared to the oxygen generation.

Would this explain why SWG manufacturers seem to specify a somewhat higher than normal CYA? To increase the ratio of chlorine to oxygen?

[Gunslinger]

Richard, your explanations of SWG chemistry are truly awesome, but I don't believe for NY second that the reason for the manufacturers' 70-80ppm CYA recommendation (I've actually seen 80-100ppm recommended) has anything to do with real-world chemistry. Rather, I suspect it has more to do with a slightly different interpretation of the acronym "CYA", rooted in the early days of the SWG evolution when reliable operation was not up to par with current devices. In a loss-of-power event, for example, high CYA levels would help to ensure that all your residual FC doesn't disappear on a sunny day before you've noticed something is amiss. I believe there is no longer any rationale justification for this recommendation other than my suspicion that the manufacturers have no interest in pursuing the matter and are content to leave it unchanged.
Personally, I barely manage to keep my CYA level above 25ppm, and have experienced no consequences from that during the four seasons I have operated my Pool Pilot SWG -- although I do test my FC and Ph quite often.

[chem geek]

Gunslinger,

You may be right, but users of this forum, including Evan (waterbear) noticed that the output efficiency of the salt cell fell dramatically at lower CYA levels. That is, you needed to increase the time and/or power of the salt cell to achieve the same FC level when the CYA was below the 70-80 recommended range and this effect was MUCH larger than the sunlight protection effect of the additional CYA (which is quite minimal at higher CYA -- see this graph) . Now this specific CYA level probably varies by manufacturer, but this effect is absolutely real as it was reported by multiple users. Maybe your salt cell is one of the lucky ones that operates well at lower CYA, or maybe you never got to the 80 ppm CYA where the efficiency dramatically improves for some salt cells.

So at least some SWG manufacturers may be recommending higher CYA in order to have higher efficiency. The issue then becomes whether efficiency of chlorine generation is the only factor or whether the proportion of oxygen generation is also affected. I don't know the answer to that and was just speculating based on chemistry -- we don't have real-world data on SWG output (i.e. measuring the gasses) to know for certain.

giroup01,

I believe the high CYA recommendation is primarily for SWG cell efficiency as mentioned above. I do not know if the oxygen effect is real (there might be coatings on the plates to inhibit oxygen generation, if that's even possible) or if the manufacturers know about it if it is or if they recommend higher CYA partly for that reason. Basically, I don't know a lot. But if the chlorine generation rate goes down at lower CYA but the conductivity stays the same and there is the alternate reaction of generating oxygen available, then it seems reasonable to assume that more oxygen is generated at lower CYA. The extreme example of this occurs if you remove all of the chloride and just have a different electrolyte (sulfates, carbonates, etc.) in which case you get pretty much ONLY oxygen and hydrogen. I had fun as a kid making such concoctions and carefully exploding them (this is definitely a "don't try this at home" scenario where you must not use glassware and should wear glasses).

Richard

[Waterworks]

I was just speaking with my customer that has the indoor Stainless Steel pool. (This is not the same pool that I spoke of last week with the massively rusted ladders and +100 ppm chlorine.) Once his pool was up and running with the SWG he quickly noticed corrosion issues with the walls and floor of the pool. He is now using pucks, and I had him add a little bit of extra CYA. The build up of CYA over time won't be an issue because he drains the pool periodically. He has not noticed any corrosion since getting rid of the SWG and switching to pucks.

I told him that I thought that high chlorine was the cause of the corrosion and that I thought that by adding some CYA and more closely monitoring chlorine levels we could avoid the corrosion while allowing him to use his SWG again. The pool is completely 316L grade Stainless and is grounded properly. When the initial corrosion happened he brought a metallurgist out to look at the problems. The metallurgist told him that 316L stainless could not handle sodium chloride in concentrations above 500 ppm. This is vastly different from everything I've read about 316L steel.

Let me know what you think.

Brad