I was involved in an unfortunately not-so-nice exchange in this thread, but the net result is that I calculated the volume of hydrogen gas bubble production in an SWG and found that even under best-case conditions with perfect kinetics, there simply isn't enough volume of gas generated to remove carbon dioxide (at equilibrium quantities using Henry's Law) at a rate to explain the significant pH rise in SWG pools. Though lowering the TA will reduce the effect of pH rise from carbon dioxide outgassing, in many SWG pools this doesn't help enough and my theory of hydrogen gas bubbles increasing aeration cannot be the primary source of pH rise.
So I looked for other reasons and the most likely candidate is that of chlorine gas outgassing since that is also produced in the SWG and if not fully dissolved into the water it can outgas (once it becomes aqueous chlorine then this combines with water to form hypochlorous acid very quickly and does not outgas very much). This means that there may be other methods that can help slow down the rate of pH rise in SWG pools. For example, pointing the returns downwards may help make the chlorine gas bubbles spend more time in the water and therefore dissolve more completely so less gets outgassed. Using a slower pump speed can help the chlorine spend more time in the pipe, at least in larger pipe and longer pipe runs.
If anyone has an SWG with a significant pH rise (so frequent acid addition) and can experiment with turning down their returns to see if it makes any difference, that would be great. If one has a 2-speed pump and can see if there is any difference running the SWG with the pump at low-speed vs. high-speed, that would also be good.
I copy below some of the relevant computations from that thread just so it's here at TFP.
OUTGASSING VOLUME IN HYDROGEN BUBBLES
As for the volume of hydrogen gas bubble production, a typical SWG for a moderate pool at 20 grams chlorine per hour is ( 20 / (70.9064 g/mole Cl2) ) = 0.28 moles per hour (of both chlorine and hydrogen gas). Using PV=nRT, Volume = (0.28 moles) * (0.082 l-atm/mole-K) * (298 K) / (1 atm) = 6.8 liters of gas per hour. The normal volume of air that we breathe out is about 0.5 liters so the volume of hydrogen gas that is produced is like 14 breaths per hour (about one every 4 minutes). Though that is not a lot of volume, the bubbles are very small (high surface area to volume ratio) so we can calculate a best-case scenario where carbon dioxide fills the bubble to achieve equilibrium. Henry's Law constant for carbon dioxide is [H2CO3]/pCO2(g) = 0.034 (moles/liter)/atm. At 100 ppm TA (with 30 ppm CYA so 90 ppm carbonate alkalinity and a pH of 7.5) [H2CO3] = 1x10-4 so pCO2(g) = 1x10-4/0.034 = 0.0029 or 0.3% of the volume. So that's a removal of carbon dioxide at a rate of 0.28 * 0.3% = 0.00084 moles per hour. A 15,000 gallon pool is about 57,000 liters so that's 1.5x10-8 moles/liter per hour. So that this isn't very much, but a drop in total carbonates of 2% results in a pH rise of 0.12 so that would take 0.02*1x10-4 / 1.5x10-8 = 133 hours of on-time. Nevertheless, we see a direct relationship between the SWG on-time and pH rise so there could be side reactions that produce excess hydroxyl ion (though I am not aware of any that would -- production of oxygen gas would be produce the same quantity of hydrogen ion as its removal at the other plate producing hydrogen gas). Also, the bubbles breaking the surface of the water could influence the rate of surface transfer beyond direct transport in the bubbles themselves. The most likely reason is chlorine outgassing I discuss next.
Assuming stable long-term FC so that chlorine consumption plus chlorine gas outgassing equals SWG generation of chlorine gas, then over one week of 2 ppm FC per day generation/consumption/loss (14 ppm FC),
0% chlorine gas outgassing ==> no change in pH
10% chlorine gas outgassing ==> +0.12 rise in pH
20% chlorine gas outgassing ==> +0.28 rise in pH
30% chlorine gas outgassing ==> +0.46 rise in pH
So clearly outgassing of undissolved chlorine gas could explain the bulk of the pH rise in SWG pools (with a smaller portion explained by carbon dioxide outgassing for which lowering the TA helps).
CHLORINE CONVERSION TO HYPOCHLOROUS ACID
Cl2(aq) + H2O(l) <---> HOCl(aq) + H+(aq) + Cl-(aq)
The equilibrium constant for the above reaction is K = 10-3.3 = [HOCl][H+][Cl-] / [Cl2]. This source gives the equilibrium constant as 4.2x10-4 whose -log10 is 3.4, close to what I used (my number was adjusted for activities based on ionic strength).
Ignoring the effects of Cyanuric Acid (CYA) which significantly reduce HOCl concentration, 3 ppm FC (which by convention is measured as ppm Cl2) at pH 7.5 is equivalent to [HOCl] = 2x10-5 moles/liter (half of the FC is HOCl at pH 7.5 when there is no CYA). A pH of 7.5 is equivalent to [H+] = 3.5x10-8 and a salt pool with 3000 ppm salt (ppm NaCl) is equivalent to [Cl-] = 5x10-2. So let's calculate the equilibrium concentration of aqueous chlorine in this case:
[Cl2] = [HOCl][H+][Cl-] / 10-3.3 = 2x10-5 * 3.5x10-8 * 5x10-2 / 10-3.3 = 7x10-11 moles/liter
The bottom line is that there is very little chlorine (gas in water or aqueous) left -- most of it becomes hypochlorous acid. It's only in more acidic water that chlorine gas doesn't dissolve as readily since the generation of chlorine and the above dissolving reaction makes the water in that half-cell acidic -- but in pools the water is buffered so far more chlorine can dissolve in the water and becomes hypochlorous acid as shown above. Therefore, because this reaction goes essentially to completion, it cannot be ignored yet that is what virtually everyone does when quoting the net result from the SWG. The rate equation for aqueous chlorine to hypochlorous acid is Rate = 28.6*[Cl2] while the reverse rate is 28000*[H+][Cl-][HOCl]. Using the above numbers, the reverse rate is 9.8x10-10 M/sec. The forward rate of aqueous chlorine to hypochorous acid has a half-life (conversion time for half of Cl2) of 0.024 seconds so isn't slow (i.e. 99% of the produced chlorine should be converted to hypochlorous acid in less than 0.2 seconds). I just noticed that the ratio of rate constants is 1x10-3 which is a little (factor of 2) inconsistent with the equilibrium constant of 4.2x10-4 so I'm tracking down the source for the rate constants to see why there is this discrepancy. [EDIT] This link gives the forward rate constant as 20.9 sec-1. [END-EDIT] This link gives a value for the forward rate constant of 15.4 s-1 so still a 99% conversion in about 0.4 seconds (but the source describes the reaction as slow) while this link reports a rapid hydrolysis where the equivalent rate constant is 3.4x1014 x 10(7.5-14) = 1x108. This link gives an equivalent rate constant of 8x108 x 10(7.5-14) = 250 so more in line with the others. Don't you hate it when sources are so terribly inconsistent? These rates are only for aqueous chlorine to hypochlorous acid, but not for gaseous chlorine to aqueous chlorine. So the rate-limiting step could be diffusion and some of the chlorine gas that is generated may outgas from the pool and not get into the water which would of course account for extra pH rise.