Pool Water Chemistry

RGB

Member
Jan 9, 2013
11
Thanks JasonLion. Maybe the page could say "the range should be 7.5 to 7.8 with 7.7-7.8 being ideal". Then someone would need to offer a suitable pH kit of course: cresol red covers the required range.

Extensive experience is important, but there is probably very little experience (let alone controlled experimental data) about where the pH will stabilise if not repeatedly adjusted (because most members rely on kits that can not read reliably above 7.8), and whether that higher pH is suitable for pool operation and sanitation. From chem geek's graph of outgassing vs pH, and assuming that the idea of outgassing as the primary driver of SWG-linked pH rise is correct, I am speculating that the system might stabilise at a pH (near 8?) that works well while reducing costs (and any risks) of higher CYA (and therefore higher FC) use and HCl use to fight the equilibrium pH. But this is mere speculation, and I agree with you that practical recommendations need to be well within the experimentally tested range.
 

JasonLion

TFP Expert
Platinum Supporter
LifeTime Supporter
May 7, 2007
37,879
Silver Spring, MD
Getting the PH to stabilize around 7.8 is one of the goals of our recommendations, but that isn't always possible. Given enough aeration, say a negative edge, even fairly low TA levels will eventually drive the PH over 8.6, and higher TA levels will do so more quickly even at lower aeration levels. There are various adjustments you can make so that you can still maintain sanitation and CSI levels with PH in that range, but there is no avoiding the metal stains problem at high PH levels. That rules out any "trouble free" solution at higher PH levels.
 

RGB

Member
Jan 9, 2013
11
Thanks again chem geek and JasonLion. Your generosity in providing advice through this site, and your approach of basing that advice on careful analysis of the underlying chemistry plus a wealth of practical experience, are both outstanding. Swimmer safety does not seem to be an issue (ocean surface pH is around 8.2). I will check the article from Ben Powell ('Don't fight your pool: work with it'), try the experiment of pH >8, CYA ~50 ppm, Borate ~50 ppm, FC 3-4 ppm mainly from SWG, over a full year cycle (with a reliable pH meter and an eye open for scale or stains in case I have to fight a little) and report back then if there are any surprises.

After digging a bit deeper into the published science about CYA-FC-pH effects:

Estimates of dissociation constants for the key chloroisocyanurates vary substantially between studies, with substantial effects on downstream calculations. For example, based on measured dissociation constants (with fairly large error terms) and some extrapolations that may not be entirely valid (but are at least specified), Gardiner (1973) concluded that in the typical usage range for swimming pools (CYA >> 3 FC) the equations are independent of pH in the range 7.5-9.0 and
simplify to:
HOCl = 0.13 * FC/CYA, or
required FC = target HOCl * 7.7 CYA

Gardiner’s equations taken with previously documented results for 99% kill of S. feacalis within 1 minute indicate a requirement for at least 2.4 to 6.4 ppm FC, at 30 - 80 ppm CYA (in the absence of common pool contaminants). This bacterial kill rate is below public health targets to avoid person-to-person transmission, but probably adequate for private pools. The FC levels are close to those in Pool School.

Estimated dissociation constants from early studies lead to lower estimates of [Hy] = [HOCl] + [OCl-] than subsequent measurements (Pinsky & Hu 1981). That would add a safety margin to use of the early equations such as Gardiner’s.

The dissociation of Hy from chloroisocyanurates in solution increases with temperature, through differential changes to multiple dissociation constants (Pinsky & Hu 1981). This does not seem to be fully elucidated, so it is probably hard to factor into chem-geek’s spreadsheet.

CYA and chloroisocyanurates only absorb UV radiation below 250 nm (Sancier et al 1964). OCl- also absorbs at longer wavelengths (peak around 300 nm), which are present in sunlight at the earth’s surface (Buxton & Subhani 1972; Pinsky & Hu 1981; Nowell & Hoigne 1992). Chloroisocyanurates are far more photostable than OCl-. The quantum yield for photodegradation of OCl- depends on wavelength and pH (Cooper et al. 2007; Watts & Linden 2007). By comparison of the spectra, it can be seen that compounds which absorb only below 250 nm might give partial protection of chlorine through shielding under artificial UV disinfection; but there is no overlap with solar radiation wavelengths reaching the earth’s surface, even under current ozone depletion (Madronich et al. 1998).

Brady et al. (1963) made the curious observation that the extinction coefficient of OCl- is considerably lower in borate relative to carbonate buffer at the same pH. This has been attributed to reversible formation of a hypochloritoborate complex {B(OH)3OCl-} which does not absorb around 300 nm, and which reaches a maximum concentration at pH 8.3 (Bousher et al. 1987). Formation of this complex is expected to reduce both the photodegradation of OCl- and the HOCl (sanitising chlorine) concentration. Therefore, borate will act as a photoprotectant and buffer of HOCl release from total hypochlorite (Hy above plus complexed OCl-), with a pH sensitivity different from CYA.

Overall, photodegradation by sunlight of FC in CYA solutions is expected to increase at high pH and/or high temperature (because less of the chlorine is complexed with CYA) and the pH effect will be moderated by borate.
Another counter effect at high pH is reduced loss of Hy via chlorine gas:
2 HOCl + 2 H+ + 2 Cl? ? 2 Cl2 + 2 H2O (light, metal oxides) ? 4 HCl + O2
2 ClO? + 4 H+ + 2 Cl? ? 2 Cl2 + 2 H2O (light, metal oxides) ? 4 HCl + O2
4 ClO- + 2 H2O ? 2 Cl2 + 4 OH- + O2

Putting aside the borate and Cl2 effects for a moment: high pH also increases the ratio of OCl- to HOCl. Because OCl- is more light-sensitive (Nowell & Hoigne 1992) and less microbicidal than HOCl, the trade-off between photodegradation and sanitation will be less favourable at high pH, even in the presence of CYA. More chlorine will have to be supplied to replenish losses through photolysis at pH 8 than at pH 7. However, the estimated amounts of this increase vary, over the range 1.3- to 3-fold (Nowell & Hoigne 1992; Fuchs & Lichtman 1961). Whether this ends up costing more than HCl usage to maintain a lower pH in an SWG pool will depend on many environmental and economic factors (day-length, light spectrum, light intensity, temperature, borate concentration, organic load, pool cover type and usage pattern, aeration, SWG cell cost and life, SWG chlorination rate, target FC level -- which depends on CYA level, automatic and manual FC boost thresholds, electricity cost, HCl cost, supplementary bleach cost, time cost for monitoring etc).

Because many factors affect chlorine consumption, it is quite a challenge to design critical experiments to test hypothetical mechanisms or economics. Outliers in preliminary experiments (as noted by mas985 in chlorine-output-from-swg-t8-40.html) are one warning of unintended variables likely to preclude firm conclusions. But still interesting to consider.

Bousher A, Brimblecombe P, Midgley D (1987) Stability constants for hypochloritoborate and hypobromitoborate complex ions in aqueous solution. J Chem Soc - Dalton Trans 943-946
Brady AP, Sirine G, Sancier KM (1963) Equilibria in solutions of cyanuric acid and its chlorinated derivatives. J Am Chem Soc 85:3101-3104
Buxton GV, Subhani MS (1972) Radiation chemistry and photochemistry of oxychlorine ions. 2. Photodecomposition of aqueous solutions of hypochlorite ions. J Chem Soc-Faraday Trans I 68:958-969
Cooper WJ, Jones AC, Whitehead RF, Zika RG (2007) Sunlight-induced photochemical decay of oxidants in natural waters: Implications in ballast water treatment. Environ Sci Technol 41:3728-3733
Fuchs R, Lichtman I (1961) Stabilization of active chlorine containing solutions. USA Patent 29888471.
Gardiner J (1973) Chloroisocyanurates in treatment of swimming pool water. Water Res 7:823-833
Pinsky ML, Hu HC (1981) Evaluation of the chloroisocyanurate hydrolysis constants. Environ Sci Technol 15:423-430
Madronich S, McKenzie RL, Bjorn LO, Caldwell MM (1998) Changes in biologically active ultraviolet radiation reaching the Earth's surface. J Photochem Photobiol B-Biol 46:5-19
Nowell LH, Hoigne J (1992) Photolysis of aqueous chlorine at sunlight and ultraviolet wavelengths.1. Degradation rates. Water Res 26:593-598
Sancier KM, Brady AP, Lee WW (1964) Absorption spectra of solutions of cyanuric acid and its chlorinated derivatives. Spectrochimica Acta 20:397-403
Watts MJ, Linden KG (2007) Chlorine photolysis and subsequent OH radical production during UV treatment of chlorinated water. Water Res 41:2871-2878
Wikipedia Hypochlorous acid and Hypochlorite pages
 

chem geek

TFP Expert
LifeTime Supporter
Mar 28, 2007
12,083
San Rafael, CA USA
I'm not sure how I missed this post from earlier this year, but I just saw it now after looking at this thread so I'll comment on it.

RGB said:
Estimates of dissociation constants for the key chloroisocyanurates vary substantially between studies, with substantial effects on downstream calculations. For example, based on measured dissociation constants (with fairly large error terms) and some extrapolations that may not be entirely valid (but are at least specified), Gardiner (1973) concluded that in the typical usage range for swimming pools (CYA >> 3 FC) the equations are independent of pH in the range 7.5-9.0 and
simplify to:
HOCl = 0.13 * FC/CYA, or
required FC = target HOCl * 7.7 CYA
Gardiner is wrong. There is a pH dependence even with CYA present -- it's just not as strong. With an FC of 3 and a CYA of 30 ppm, the following are HOCl concentrations (in ppm Cl2 units) at various pH:

pH ....... HOCl ....... (FC/CYA)/HOCl
7.5 ...... 0.0422 ...... 2.37
8.0 ...... 0.0361 ...... 2.77
8.5 ...... 0.0312 ...... 3.21
9.0 ...... 0.0234 ...... 4.27

Perhaps Gardiner wasn't using O'Brien's numbers and I don't know what units he was using (I didn't look up his paper yet).

RGB said:
Gardiner’s equations taken with previously documented results for 99% kill of S. feacalis within 1 minute indicate a requirement for at least 2.4 to 6.4 ppm FC, at 30 - 80 ppm CYA (in the absence of common pool contaminants). This bacterial kill rate is below public health targets to avoid person-to-person transmission, but probably adequate for private pools. The FC levels are close to those in Pool School.
The Pool School minimum FC that is 7.5% of the CYA level is roughly equivalent to 0.04 ppm FC with no CYA (at 77ºF). The EPA DIS/TSS-12 laboratory standard is roughly equivalent to 0.4 ppm FC with no CYA, but that's a 6-log kill in 30 seconds or less (for E.coli), but all the chlorinated cyanurate products passed the field test where the CYA level rises substantially over 4-12 months. For commercial/public pools, I'd recommend an FC/CYA ratio of around 20% for the equivalent of 0.2 ppm FC with no CYA that should be very reasonable for disinfection.

RGB said:
Estimated dissociation constants from early studies lead to lower estimates of [Hy] = [HOCl] + [OCl-] than subsequent measurements (Pinsky & Hu 1981). That would add a safety margin to use of the early equations such as Gardiner’s.

The dissociation of Hy from chloroisocyanurates in solution increases with temperature, through differential changes to multiple dissociation constants (Pinsky & Hu 1981). This does not seem to be fully elucidated, so it is probably hard to factor into chem-geek’s spreadsheet.
Pinsky & Hu data is suspect. Read Reeevaluation of Chloroisocyanurate Hydrolysis Constants. The O'Brien paper shows the very meticulous methods that they used and is probably the best data available for the equilibrium constants.

RGB said:
CYA and chloroisocyanurates only absorb UV radiation below 250 nm (Sancier et al 1964). OCl- also absorbs at longer wavelengths (peak around 300 nm), which are present in sunlight at the earth’s surface (Buxton & Subhani 1972; Pinsky & Hu 1981; Nowell & Hoigne 1992). Chloroisocyanurates are far more photostable than OCl-. The quantum yield for photodegradation of OCl- depends on wavelength and pH (Cooper et al. 2007; Watts & Linden 2007). By comparison of the spectra, it can be seen that compounds which absorb only below 250 nm might give partial protection of chlorine through shielding under artificial UV disinfection; but there is no overlap with solar radiation wavelengths reaching the earth’s surface, even under current ozone depletion (Madronich et al. 1998).
The conclusions about CYA absorption are wrong, though not by that much. See Ultraviolet absorption spectra of derivatives of symmetric triazine—II: Oxo-triazines and their acyclic analogs where you can see in Figure 4 that there is absorption above 250 nm and the extinction coefficient is at around 30 at 250 nm, but data a pH 9 was cut short so we don't know how high the absorption goes. It doesn't take much of an extinction coefficient to have a CYA shielding effect so that is still possible. If you can find any paper that shows the actual measured UV absorption above 250 nm at a pH in the 7-8 range (or even 9 would be better than nothing), then please let us know.

RGB said:
Brady et al. (1963) made the curious observation that the extinction coefficient of OCl- is considerably lower in borate relative to carbonate buffer at the same pH. This has been attributed to reversible formation of a hypochloritoborate complex {B(OH)3OCl-} which does not absorb around 300 nm, and which reaches a maximum concentration at pH 8.3 (Bousher et al. 1987). Formation of this complex is expected to reduce both the photodegradation of OCl- and the HOCl (sanitising chlorine) concentration. Therefore, borate will act as a photoprotectant and buffer of HOCl release from total hypochlorite (Hy above plus complexed OCl-), with a pH sensitivity different from CYA.
This paper gives a stability constant of around 102.25 = 178. So this gives 178 * 4.5x10-3 * 4.3x10-7 = 3.4x10-7. If there were no CYA in the water, then this would imply that the chlorine concentration would be cut roughly in half, but with CYA in the water there is effectively no change at all since the chlorine bound to borate basically comes from the ample amount bound to CYA while the amount remaining as hypochlorous acid and hypochlorite ion barely budge.

RGB said:
Because many factors affect chlorine consumption, it is quite a challenge to design critical experiments to test hypothetical mechanisms or economics. Outliers in preliminary experiments (as noted by mas985 in chlorine-output-from-swg-t8-40.html) are one warning of unintended variables likely to preclude firm conclusions. But still interesting to consider.
His experiment wasn't an outlier. We did the experiment after seeing several reports of people getting greater than expected protection of chlorine breakdown at higher CYA levels and since that time this has been seen in SWCG pools to a very great extent. So something is going on even if we don't understand it fully. If it's not CYA shielding, then it's something that accomplishes a similar effect in terms of higher CYA protecting chlorine even at the same FC/CYA ratio so same hypochlorous acid and hypochlorite ion concentrations.
 

RGB

Member
Jan 9, 2013
11
Thanks again chemgeek for the interesting perspectives. I agree about Brady’s good work, and I don’t think he concluded that it provided reliable constants for the very different conditions typical of swimming pools. In response to your request about spectra, see Sancier cited above: “Examination of the spectra up to 5500 A using a 5 cm cell and saturated solutions of cyanuric acid in the pH range 1.3-12 revealed only one other absorption band. This band peaks at about 2800 A with a molar absorptivity e of about 0-2 1. cm-l mole-l.” Note that the UV-C absorbtion peaks for CYA are about 100,000 times higher (e = 10,000 to 30,000), and absorption by hypochlorite at 280 nm is about 1,000 times higher (e ~ 200).

See Madronich cited above for the concept of action spectrum based on integration of the quantum yield for photodegradation and the incident spectral irradiance. “The solar radiation at the top of the Earth’s atmosphere contains a significant amount of radiation of wavelength shorter, and therefore more energetic, than that of visible light (400-700 nm). Wavelengths in the range 100-400 nm constitute the ultraviolet spectral region. The shortest of these wavelengths (UV-C, 100-280 nm) are essentially completely blocked (absorbed) by atmospheric oxygen (02) and ozone ..... A useful measure is the biologically active irradiance ... defined as the area under the spectral overlap function....”

There are experimental data on quantum yields for photodegradation in Cooper et al. 2007 and Watts & Linden 2007 (cited above), some of which seemingly contradict established wisdom in the pool industry (sensitivity of OCL- vs HOCl) and still puzzle me.

See mas985 cited above for the outlying experiment. No offence is intended. It is almost inevitable that there will be a lot of uncontrolled variables at that stage of observation. The preliminary observations are still very interesting. But they are not explained by current scientific knowledge. It remains quite a challenge to design critical experiments to test hypothetical mechanisms or economics.
 

chem geek

TFP Expert
LifeTime Supporter
Mar 28, 2007
12,083
San Rafael, CA USA
See Figure 3 in this paper for some very decent absorption spectra for HOCl and OCl-. I've used that in a model that very accurately predicts chlorine consumption and is consistent with many measurements from other papers (except one that shows faster degradation). I'm well aware of the cutoff of the UV in sunlight and have that in a spreadsheet with that model.

Wavelength (nm): ............. 290 ........ 300 ........ 310 ........ 320 ....... 330 ........ 340 ........ 350 ....... 360 ........ 370 ....... 380
Irradiance (W/cm2-nm): 3.00E-10 2.00E-07 3.00E-06 1.50E-05 2.50E-05 2.70E-05 3.00E-05 3.00E-05 4.00E-05 3.00E-05

O'Brien's paper said nothing about any lack of applicability to swimming pools and we're talking equilibrium chemistry where it doesn't just stop occurring in different environments. Other additional equilibria can exist including ion pairs that can affect apparent solubility, but basic equilibrium chemistry doesn't vary by simply occurring with other chemicals in a pool. It is, of course, possible for there to be other chemicals in a pool that act like CYA forming weak bonds that affect the hypochlorous acid concentration, but in practice this isn't seen and indirect measurements of HOCl via ORP measurements or amperometric sensors generally track O'Brien's equilibrium constants. ORP is, of course, very noisy, but you can see in this post how in the second graph the calculated HOCl in real pools is correlated with ORP that is presumably most affected by it.

Now the one variation from O'Brien's paper that is very real is temperature. There is some info on that in this paper from Wojtowicz (see values for K7 and K9 temperature dependence). For swimming pools, this is a moderate effect, but it shows up more in hot spas. My Pool Equations spreadsheet has this temperature dependence turned off by default (around line 230 "Use Temp. Dependent Cl-CYA") because I'm not confident about the Wojtowicz data and using 77ºF constants is worst-case for pools and spas so is conservative.

As for the Sancier paper, I already had that, but that molar absorptivity of 0.2 peaking at 280 nm is too low though it would only take molar absorptivities of around 10-20 to explain the seen effect. So as I wrote, if it's not CYA shielding, then it's some other effect that acts in the same way and is as yet unexplained. It's absolutely an effect we've seen in many pools, especially the SWCG pools where chlorine introduction rates are much more regular so the "data" is far less noisy. 4 ppm FC at 80 ppm CYA loses less in sunlight than 2 ppm FC at 40 ppm CYA even though both have the same hypochlorous acid and hypochlorite ion concentrations (when the pH is the same in both cases). Don't forget that there also seems to be degradation of chlorine bound to CYA as well since the loss rate during exposure to sunlight far exceeds the theoretical calculated amount from the unbound CYA alone. So in theory, the 4 ppm FC should lose chlorine at almost twice the rate of 2 ppm FC, but in fact it loses less. If you have any explanation for this phenomenon, please let us know since we've never figured this one out.
 

RGB

Member
Jan 9, 2013
11
Pinsky & Hu (cited above) explain why one would like measurements closer to pool conditions (and I doubt that O'Brien would have disagreed). Their motivation was right, though as Wojtowicz explained their new data analysis had problems. Nowell & Hoigne (1992) conclude from convincing analysis that the effective wavelengths for significant photolysis of aqueous chlorine by sunlight are between 320 and 340 nm, so it would be great to see detailed absorption spectra for (chloro)cyanurates in that region. Maybe a forum member with access to a scanning spectrophotometer can help. But Sancier tells us that examination of the spectra up to 550 nm in the pH range 1.3-12 revealed only one other absorption band, with a peak around 280 nm and molar absorptivity of about 0.2. So unless Sancier is mistaken, we are not going to find molar absorptivities in the range you mention. The observations you mention about high CYA are very interesting. The characteristics of pools where the effects are not observed are probably equally important for understanding. For now, these observations are preliminary in a scientific sense, despite their potential practical significance (too many uncontrolled variables to be published as a contribution to scientific understanding). They are not explained by current scientific knowledge. Speculating about potential mechanisms may help to generate competing hypotheses. But it remains quite a challenge to design critical scientific experiments to test hypothetical mechanisms or economics.
 

RGB

Member
Jan 9, 2013
11
Reporting back as promised after running the pH experiment for more than a year.

During this time my test salt-water-chlorinated pool always drifted up in pH, and never equilibrated below my intervention zone between pH 8.2-8.6. Over this period, I added more than 60 litres of 9 N HCl, and no carbonate or bicarbonate. The pool bicarbonate level was in equilibrium with atmospheric CO2 within a few months, and I used boric acid (at about 60 ppm boron) as the main pH buffer. Evidently pH updrift in salt-water-chlorinated pools involves mechanisms beyond outgassing of CO2. Loss of volatile disinfection byproducts is the most likely mechanism, as discussed elsewhere (Birch 2013b).

Borate makes much more sense than bicarbonate as a primary pH buffer in salt-water chlorinated pools, and a critical review of the science behind this use is provided elsewhere (Birch 2013a).

So many factors affect the rate of chlorine consumption that great care is needed to control all of these factors in any experiment to isolate the effect of any single factor (such as CYA concentration). Then the relationship will depend on the levels of all the other factors used in the experiment. Experienced professional chemists have struggled with this problem in experiments testing CYA concentration vs sanitizer consumption (Wojtowicz 2004). One factor with the potential to cause very different experiences in this context between ‘similar’ salt pools is the bromide content of the salt supply, as discussed elsewhere (Birch 2013b).

Summaries can be seen without needing to download the full pdf files, at http://members.iinet.net.au/~jorobbirch/

Birch RG (2013a) Boric acid as a swimming pool buffer. http://members.iinet.net.au/~jorobbirch/Boric_acid.pdf
Birch RG (2013b) BABES: a better method than “BBB” for pools with a salt-water chlorine generator. http://members.iinet.net.au/~jorobbirch/BABES.pdf
Wojtowicz JA (2004) Effect of cyanuric acid on swimming pool maintenance. Journal of the Swimming Pool and Spa Industry 5, 15-19.
 
  • Like
Reactions: tillmac62

chem geek

TFP Expert
LifeTime Supporter
Mar 28, 2007
12,083
San Rafael, CA USA
Thank you for your very detailed write-ups. I briefly skimmed them and will go over them in more detail in the future. As for pH rise in SWG pools, even after lowering the TA and adding borates, it still varies by pool in terms of how much pH rise there is. It would be great to know exactly why. As for Cl2 outgassing, I don't think it's from equilibrium since the level is far too low and instead HOCl is the primary chlorine component that outgases unless the bather (or organic) load is high enough to have substantial chloramines and most residential pools have very little chloramines (and all of them don't outgas anyway -- most are oxidized -- and the amount of nitrogen trichloride that is not oxidized is negligible in such pools). When I talk about Cl2 outgassing, I assume it is from undissolved gas bubbles generated in the SWG but not completely dissolved in the pool. "Gas shooters" (those who add chlorine gas as their method of swimming pool chlorination) know this problem very well and adjust their flow rate accordingly.

My Pool Equations spreadsheet shows that even with no carbon dioxide outgassing at all, that if 10% of the generated chlorine is outgassed, then at 2 ppm FC per day and 70 ppm TA and 80 ppm CYA, the pH rises from 7.50 to 7.63 over a week. If 20% of the chlorine gas outgases, then the pH rises from 7.50 to 7.80. There is some anecdotal evidence that shorter pipe runs and faster flow rates lead to more of a pH rise, but it would be much better to get to the bottom of this. It does seem to be related to the SWG on-time since the use of higher CYA levels that let one turn down the SWG on-time does help reduce the rate of pH rise (of course, there could be something to do with the CYA itself, but since the effect isn't noticeable in non-SWG pools, I attribute the rise to the SWG on-time).
 

chem geek

TFP Expert
LifeTime Supporter
Mar 28, 2007
12,083
San Rafael, CA USA
I've now had a chance to read your articles in more detail. With regard to Boric Acid, I have only a few comments as the article is outstanding. You might give a specific example where boric acid being a weak acid with its addition doesn't generally need pH adjustment in pools. For example, adding 50 ppm Borates (technically 50 ppm Boron as a unit-of-measure) as boric acid to water with a pH of 7.5 and TA of 80 ppm results in a drop in pH to 7.3. That gives one a rough idea of what one may expect.

You mention 100 ppm borates, but the EPA limit is close to 50 ppm. The 360 ppm you quote is "chemical concentration in the water" from Table 5 of the EPA report (360 being the level at an MOE of 100), but that is in ppm sodium tetraborate pentahydrate which is equivalent to 53.4 ppm Boron (I made that same mistake initially and is why I have EDIT sections in the Are Borates Safe to Use? post). There are likely greater algaestatic properties at higher levels around 80 ppm or so, but technically this is beyond what the EPA allows. Nevertheless, I agree with you that even at 100 ppm the quantities of pool water one would need to drink on a regular basis are high in terms of being a risk. The EPA rule comes largely from the Margin of Exposure (MOE) of 100 (factor of 10 for inter-species since human trials were not done and factor of 10 for intra-species variability since the number of tested subjects for any given study was rather small).

Under safety, you might mention that in lower quantities the body efficiently processes it and excretes excess. So toxicity is about exceeding the body's ability to excrete boron and building up higher boron levels in the body. This is sort of implied in what you wrote, but might be better explicitly stated. I thought that this video gave a nice explanation of why borates kill insects but not humans.

As for the effect of calcium and TDS (ionic strength, actually) on the borate equilibrium, these aren't big factors in pool water. Lowering the pH from 7.8 to 7.5 in a pool with 80 ppm TA and 50 ppm Borates and 1600 ppm TDS with no calcium hardness at all requires 17.81 fluid ounces of 31.45% HCl per 10,000 gallons but with 1000 ppm CH it requires 20.40 fluid ounces so a 14-15% increase. Similarly, a pool with no TDS (not really possible) and no CH requires 17.22 fluid ounces while one with 3000 ppm TDS (and still no CH) requires 18.22 fluid ounces so a 6% increase. The reason TDS doesn't have as much of an effect as you might expect is that TDS also affects the carbonate buffer system in the opposite direction. With no borates in the water at all, with 80 ppm TA (and no CH) it takes 6.89 fluid ounces with no TDS while it takes only 6.18 fluid ounces at 3000 ppm TDS. The statement you make about doubling the required acid amount when the pKa changes by 0.2 is probably not a realistic example for actual pool variations in CH and TDS.

As for the saving of acid using borate if the cause of the pH rise is CO2 outgassing, that savings would only be seen if one were to lower the carbonate alkalinity, so generally lowering the TA level, as a result of having another pH buffering system. One can raise the CH level to compensate for the saturation index. You mention that there would be savings if additional bicarbonate addition were avoided, but what really goes on is that the acid addition is lowering the TA level. So I think it's better to mention that.

As for not needing bicarbonate when borates are used, there are two points to make. First is that the carbonates are needed for calcium carbonate saturation if one has plaster surfaces to protect. The second is that the carbonate buffer system is effective at preventing a drop in pH so if one is using net acidic sources of chlorine then the carbonate system is needed and is more effective than the borate system at preventing a "pH crash" (though if CYA is present then that is an additional buffer system). On the other hand, this discussion is in the section on SWC pools so my comments are probably a moot point.

That's interesting about borate ion and hypochlorite ion complexing. I suspect the effect is too small in pools to consider, but as you say it would be good to get better experiments done to know for sure.

I'm not sure about a blue appearance with borates. What we hear and what I've seen in my own pool is sparkle and a definitively flattened meniscus when taking samples (Nakath's report on surface tension notwithstanding) so I presume the effects are due to either an altered refractive index or surface tension or both. The "sparkle" is referring more to light reflection from the water surface especially when there are some small waves in the pool. At least that's my best guess at this point. You should do some experiments of your own regarding looking at the meniscus in samples without and with borates. Maybe the meniscus effect is not a surface tension change but some sort of differing attraction/cleaning effect on the plastic vial though it sure looks like a surface tension effect to me.

I have found the drop-based mannitol borate test to be very consistent and accurate (in terms of what was expected based on measured addition). The test strips have been OK -- as good as one would expect given their large steps and harder-to-read saturation differences.

I'll do a separate post for your other article on "BABES".
 

chem geek

TFP Expert
LifeTime Supporter
Mar 28, 2007
12,083
San Rafael, CA USA
With regard to BABES: a better method than “BBB” for pools with a salt-water chlorine generator, I also have some comments, but overall the article is very good.

You write that "HOCl (through its equilibrium product OCl-) breaks down in sunlight", but HOCl itself also breaks down in sunlight, just not as quickly. This is seen in this link. The net result is that the half-life of OCl- is around 35 minutes while that of HOCl is around 130 minutes. The direct breakdown of HOCl cannot be ignored. So you might add "also through its equilibrium...".

In section 3.2 Sodium hypochlorite, you mention that in the assay many different chemical species have the same oxidizing power per molecule as Cl2, but I wouldn't put it that way. The test is measuring the reserve capacity of chlorine where it is essentially counting chlorine atoms (those in the +1 state that are not in combined chlorine molecules such as chlorourea or chloramines). The way it is written, it makes it sound like chlorine bound to CYA has the same oxidizing power as HOCl, but that is not true. Chlorine bound to CYA is not only a very weak (at least 150 times less than HOCl) disinfectant, but it is also a very weak oxidizer as well.

You should emphasize that Figure 1 is not applicable when CYA is present. You show the correct graph later in Figure 7.

That is very interesting about the salt effect on indicator dye equilibrium including the phenol red pH test. Along with the roughly same order-or-magnitude effect on CSI from TDS (roughly 0.2 lower at 3000 ppm TDS than 500 ppm), this could explain the greater likelihood of plaster issues in some SWC pools. I always thought the 0.2 difference was pretty small, but combined with a second 0.2 difference then this is 0.4 lower which is more substantial.

In section 3.2.3 Salt and bromide, you write that "HOBr gives a less intense colour than HOCl (by a factor of about 2.25 according to Taylor Technologies)", but that factor of 2.25 is just the unit of measurement (molecular weight) difference between Br2 and Cl2. The HOBr on a molecule per molecule basis gives the exact same intensity change as HOCl. The sanitizing effect of HOBr is reasonably and correctly measured in ppm Cl2 units as if you thought it was chlorine. Also, when CYA is present, the HOBr will be far far high in concentration than the HOCl as chlorine binds to CYA while bromine does not (Wojtowicz notwithstanding).

In spite of White's claims that bromamines don't need to be removed by shocking, spa users will disagree. Most bromine spas do need to be periodically shocked, usually with chlorine, in order to get rid of the bromamine smell.

When you were using my spreadsheet at any temperature other than 25ºC (77ºF), it wasn't clear whether you set line 230 "Use Temp. Dependent Cl-CYA" to TRUE. You should if you plan on looking at HOCl concentrations at other temperatures. Setting that line (in columns B and C) to TRUE uses the temperature dependent hydrolysis constants from Wojtowicz.

If you calculate absolute chlorine losses from the increased OCl- at higher pH, it's not very much in comparison to overall FC losses seen on a daily basis. In spite of Wojtowicz saying that chlorine bound to CYA doesn't degrade, there is something breaking down or using up chlorine much faster than explained by photolysis of HOCl and OCl- even at lower pH. At 4 ppm FC with 80 ppm CYA, then at pH 8.0 there is 0.087 ppm OCl- and 0.025 ppm HOCl (at 82ºF) so even using a 30 minute half-life for OCl- and 2 hours for HOCl, then over 8 hours of full-time equivalent noontime sun that would be a loss of 2*8*0.087 + (8/2)*0.025 = 1.5 ppm FC so at high pH this can be explained, but at pH 7.5 we still see large losses. There is some chlorine oxidation of CYA itself, but again that's not enough to explain the usual 2 ppm FC daily losses.

As for CYA helping to reduce Cl2 amounts remaining, remember that the chlorine production amounts overwhelm CYA concentrations, but your point is well taken that downstream as the water gets diluted away from the plates the CYA may help keep the pH and HOCl concentrations in check somewhat to improve Cl2 rates of solubility. Certainly, it binds to HOCl that forms thus buffering it away. However, the point I've made about undissolved chlorine gas outgassing is about chlorine gas that has NEVER dissolved. It is formed at the plate as bubbles and doesn't fully dissolve. So CYA isn't that relevant except that it helps to lower the HOCl concentration and therefore equilibrium Cl2 concentration, but that has never really been much of a limiting factor in the chlorine gas bubble dissolving rate. It's a matter of physical diffusion and even having zero chlorine concentration in the water won't make it diffuse infinitely fast. It is the physical diffusion limit that appears to be at play here and that is a function of things like bubble size which determines the surface area to volume ratio, the length of pipe run that determines the amount of time for dissolving, whether returns are pointed up or down which determines the path length (so time) for the bubbles, and so on.

All your points about reasons why the pH would net rise due to chlorine not getting all used/consumed are valid, but they are mostly all true for non-SWC pools yet we fairly consistently see SWC pools have a greater tendency for the pH to rise than non-SWC pools. So that is where the speculative hypothesis of undissolved chlorine gas outgassing and increased aeration causing greater carbon dioxide outgassing come from.

I didn't see you mention anywhere borates positive effect on reducing scaling in the SWC cell. I think that is very important. Because the borates are strong buffers against a rise in pH, the effectively cut down the amount of pH rise at the hydrogen gas generation plate about in half and that helps to prevent calcium carbonate scaling as well as calcium phosphate scaling.

I would not say that > 50 ppm CYA is too much in a residential SWC pool. 80 ppm seems to work very well in many such pools and the higher CYA seems to help reduce the SWC on-time with its associated benefits of slower pH rise.

As for pH rise in borate-buffered SWC pools, there is still CO2 outgassing unless the TA is very low and/or the pH quite high (above 8.0 unless the TA is low). The other factors you mention regarding not all chlorine getting used/consumed are of course still relevant.

You wrote "But Langelier’s brief was to reduce pitting in cast-iron water pipes, a far cry from the challenge of ideal solute composition for a SWC swimming pool!" However, if you looked at the bottom of my Pool Equations spreadsheet, you would find that the derivation of the Langelier Saturation Index has absolutely nothing to do with boilers, closed systems, cast iron corrosion, or any other such nonsense. It is a chemical equilibrium equation solution for the saturation point of calcium carbonate, nothing more. So please stop perpetuating this falsehood that gets repeated over and over again by many in the industry that somehow because Langelier wasn't looking at pools that his equation somehow doesn't apply. That is simply not true at all.

I also would not say that "There is no strong indication from experience that any calcium saturation index is important in modern pool concrete etching." onBalance has done tank experiments with plaster coupons that show that the index plays a strong role in etching of poorly made plaster but that even plaster that is will made will etch at more extreme saturation indices such as those in the -0.7 to -1.0 range. I would say that low pH is more detrimental than a low calcium level. Even at zero calcium levels in the water, there is a limit to dissolving/pitting rates and those rates are more of a function of pH and temperature.

You write "concrete is damaged by the formation of CaCO3, through reaction of carbonic acid with bonding agents including calcium silicate hydrates." That doesn't make sense since pool plaster (concrete) curing requires the conversion of calcium hydroxide produced from that curing to calcium carbonate. The bicarb startup works best to form the strongest pool plaster surface.

You write "If the protection from ‘balanced water’ is primarily from a protective surface coating of CaCO3, then MgCO3 or other low-solubility salts may be similarly effective." This is again simply not true. The saturation of calcium carbonate is much lower than that of magnesium carbonate so trying to trade off by having more magnesium instead of calcium simply does not work. Calcium carbonate in pool plaster dissolves into the water and you can't practically have enough magnesium concentration to form magnesium carbonate in its place for an equally strong surface. Total Hardness is therefore pretty much irrelevant. It is Calcium Hardness (CH) that matters since it is calcium carbonate saturation that matters.

Your Box 2 approach of low calcium and high pH isn't bad but remember the risk of metal staining at higher pH. If you've got metal ions in the water such as iron, copper, manganese, etc. then you risk staining at higher pH levels. You referenced Ben Powell's high pH poolsolutions.com page so that talks about the pros and cons of the approach.

The catalytic breakdown of chlorine from some metal ions is something I don't think applies practically in pools due to the much lower concentration of chlorine in the pools. Such problems with metal ion impurities are a problem for chlorinating liquid but the auto-breakdown of chlorine (not photolytic) at pool concentrations should be negligible.

You wrote "Polyquat algaecides are sometimes claimed not to deplete pool chlorine, without reference to independent scientific tests." I don't know where you heard that claim. In fact, we have most definitely seen where the two react with each other where high concentrations of Polyquat most certainly deplete chlorine. What has been said by observation of many pools is that the chlorine reaction with Polyquat is slower than that of linear quats. At recommended Polyquat levels, the increased chlorine demand is relatively low but when using linear quats it is much more noticeable. Specific quantification of these observations has not been done.

As for copper stains, they may be from copper carbonate formed in the pool plaster itself by copper displacing calcium so forming copper carbonate where calcium carbonate existed.

If there is the risk of salt splash-out problems such as damage to concrete or other surfaces, I would not recommend running at the highest salt levels the manufacturers recommend but rather the lowest levels that still give reasonable SWC efficiency.

Your 30-50 ppm CYA recommendation may not be enough for SWC pools in a lot of sunlight or in desert areas, at least not unless the SWC cell is sized higher, but then you have the other problems of higher SWC output including faster pH rise.

Again, overall, a very good article -- the only reason I had so many comments is that there was a lot of content in this article on which to comment.
 

chem geek

TFP Expert
LifeTime Supporter
Mar 28, 2007
12,083
San Rafael, CA USA
I recently ran into Extended experimental investigation: The effect of sunlight on the chlorine levels in pools and noticed that Graphs 1 and 2 are much more consistent with chemical theory where Cyanuric Acid (CYA) hugely prevents chlorine degradation. The pH was not monitored (though a pH buffer was used), but we can still calculate relative rates of decline at different CYA levels to see if it's consistent with what we see in pools. We can calculate a decay rate constant based on C = C0*e-k*t. So, k = 1.16 for no CYA (though the 0.63 at 1 hour should be 0.78 for consistency with the data at other times), and using the 3 hour data we have k = 0.13 for 25 ppm CYA, k = 0.093 for 50 ppm CYA, k = 0.78 for 100 ppm CYA. This shows that there is chlorine loss beyond that from unbound chlorine and that higher CYA levels provide additional protection though not proportional to the CYA level.

That link was to a school's chemistry assignment assessment guide and student response example, but it looks like it was some student's actual data (it's too close to reality to be a fake response).

It would be very interesting to see what happens to the losses when the initial FC/CYA ratio is held constant. If no other mechanisms of protection were occurring, then increasing the FC proportional to the CYA level would result in noticeably higher absolute losses but we usually see the opposite. We presumed this was due to some sort of CYA (or chlorine bound to CYA) shielding effect, but that is a speculative guess. It remains one of the mysteries for which we do not have a definitive explanation.
 

chem geek

TFP Expert
LifeTime Supporter
Mar 28, 2007
12,083
San Rafael, CA USA
I'm not sure about a blue appearance with borates. What we hear and what I've seen in my own pool is sparkle and a definitively flattened meniscus when taking samples (Nakath's report on surface tension notwithstanding) so I presume the effects are due to either an altered refractive index or surface tension or both. The "sparkle" is referring more to light reflection from the water surface especially when there are some small waves in the pool. At least that's my best guess at this point. You should do some experiments of your own regarding looking at the meniscus in samples without and with borates. Maybe the meniscus effect is not a surface tension change but some sort of differing attraction/cleaning effect on the plastic vial though it sure looks like a surface tension effect to me.
Finally someone did an experiment with borates though it was with tap water vs. pool water (so not quite the controlled experiment we would like to see). In this post a 13% reduction of surface tension was seen at 30 ppm Borates so extrapolating to 50 ppm Borates it might be a 22% reduction. I'd still like to see a "same water before and after borates" experiment, but until then at least we've now got something a little more quantitative.
 

jv92red

LifeTime Supporter
Mar 7, 2013
117
SoCal
Finally someone did an experiment with borates though it was with tap water vs. pool water (so not quite the controlled experiment we would like to see). In this post a 13% reduction of surface tension was seen at 30 ppm Borates so extrapolating to 50 ppm Borates it might be a 22% reduction. I'd still like to see a "same water before and after borates" experiment, but until then at least we've now got something a little more quantitative.
chem geek - I will be doing Borates in my pool in a few days, let me know I can set up an experiment on this. So just take pictures of water samples before and after to compare the meniscus? That would be easy if so. If not then let me know what I need to do. I know my borate level now is at zero since I maintain it myself and the pool is new as of August 2013.
 

SunnyOptimism

Well-known member
Jun 20, 2014
860
Tucson, AZ
chem geek - I will be doing Borates in my pool in a few days, let me know I can set up an experiment on this. So just take pictures of water samples before and after to compare the meniscus? That would be easy if so. If not then let me know what I need to do. I know my borate level now is at zero since I maintain it myself and the pool is new as of August 2013.
If only we had one if these -

http://www.kruss.de/products/tensiometers/force-tensiometer-k6/

Or even better -

http://www.kruss.de/products/tensiometers/bubble-pressure-tensiometer-bp50/


Sent from my iPhone using Tapatalk
 

chem geek

TFP Expert
LifeTime Supporter
Mar 28, 2007
12,083
San Rafael, CA USA
chem geek - I will be doing Borates in my pool in a few days, let me know I can set up an experiment on this. So just take pictures of water samples before and after to compare the meniscus? That would be easy if so. If not then let me know what I need to do. I know my borate level now is at zero since I maintain it myself and the pool is new as of August 2013.
You'd need special instrumentation to quantitatively measure the effect, but yes you could take photos of your pool water in a sample tube before and after addition of the borates (after some mixing, say an hour of circulation). It wouldn't be quantitative, but would at least possibly show some difference.