A Critique of the PBL Research Report

"An Assessment towards the Phenomenon of Discoloration and Deterioration of Plaster Surface Materials within the Recreational Water industry & The Carbonate System"

Discussion given by Don Hafer at the National Plasterers Council Conference 2001

By Que Hales, Doug Latta, & Kim Skinner

INTRODUCTION

PBL has released a 1996 study on plaster deterioration which concluded that "the most probable cause" of discoloration, delamination, nodules, spot etching, corrosion and overall degradation to pool cement (plaster) finishes, is high levels of aggressive carbon dioxide. The author and senior analyst of PBL, Don Hafer, made this conclusion after observing some plaster coupons (which were placed in five water tanks with five different balance programs) reportedly suffer these abnormalities.

Essentially, Mr. Hafer claims that when balancing pool water using the "carbonate system" (adding strong acids and strong bases), the amount of carbon dioxide (expressed somehow as carbonic acid) increases and will often exceed the appropriate compensating levels necessary to stabilize the "carbonate equilibrium" (calcium bicarbonate). This, he claims, results in compromising the pool plaster finish and creating the above mentioned detrimental results. Additionally, he suggests that even if water is balanced according to the Langelier Index, this condition of high carbon dioxide (CO2) levels could still result and lead to overall plaster degradation.

EXAMINATION OF CO2 LEVELS

We first examined whether PBL accurately obtained the high levels of carbon dioxide indicated in its report. Immediately noted was that a significantly high content of CO2 was claimed with a pH reading above 8.3. This is in contradiction to proven chemical science. The following is an example of a chemical analysis by PBL: the tap water used in the experiment was recorded to have a pH of 8.47, total alkalinity 120 ppm, calcium hardness 320 ppm, temperature 19.7 degrees C, TDS 544ppm, and a CO2 content of 40 ppm.

According to Standard Methods 18th Edition, and Water Chemistry by Snoeyink & Jenkins, (both of which PBL cites as references), and the NALCO Water Handbook 2nd Edition, the end-point for CO2 is at a pH of 8.3. In other words, CO2 ceases to exist in measurable quantities above a pH of 8.3. When using the titration method (as PBL claims to have used), the above tap water would result in a reading of "0" ppm for CO2. Using the reliable nomograph method as outlined in Standard Methods, it would be less than 1 ppm of CO2. This contradicts the claim of 40 ppm of CO2 by PBL. Interestingly, on page 26 of the PBL report, a reference indicates that "tritrating to a phenolphthalein end point pH of 8.3 corresponds to the neutralization of carbonic acid to bicarbonate. " It should also be pointed out that a titration test kit for CO2 should not be used for swimming pools due to several chemical interferences (as listed in Standard Methods and the Hach manual) that are normally contained in pool water.

Every CO2 analysis performed by PBL has the ppm amounts of CO2 far exceeding the results that would be given by the nomograph or calculation methods. One example from each tank is as follows: The listed PBL readings are followed by the nomograph reading and calculation result for CO2.

Tank Number	Date	PBL pH	PBL TA	PBL CO2	Nomograph CO2	Calculated CO2
1	12/18	7.31	60	29	5.5	5.8
2	12/14	8.03	130	60	2.6	2.4
3	1/3	7.90	80	40	1.8	2.0
4	1/15	7.59	150	22	7.5	7.7
5	1/15	7.98	120	30	2.4	2.5

As one can see, PBL's claimed results for CO2 do not coincide with the results given with the reliable nomograph and calculation methods, and are in fact, many times higher than would be in actuality. This claim by PBL is simply not plausible.

EXAMINATION OF OBSERVED RESULTS

The next aspect of the report we examined was the observations and apparent conclusions drawn from the "plaster coupons" in the various water tanks. PBL states that each of their five water tanks had a different water balance maintenance program. They are listed as aggressive, alkali, neutral, control, and zero degree hardness. Placed in each tank were 40 plaster coupons for approximately 45 days. Summaries of the observed results are as follows:

• The "aggressive" tank #1 had discoloration on all coupons, and nodules formed on four separate coupons. No coupons showed any spot etching.
• The "alkali" tank #2 indicated that all coupons discolored, and six different coupons showed signs of spot etching, one of which also had a nodule forming.
• The "neutral" tank #3 was noted to have discoloration on all coupons, thirteen plaster coupons spot etched, while no nodules formed.
• The "control" tank #4 showed only a slight appearance of discoloration on all coupons, and slight spot etching on four plaster coupon samples. No nodules formed.
• The "Zero Degree Hardness" tank #5 did not produce any discoloration, spot etching, nodules, or any degradation over the forty-five day period.
  

Pictures of plaster coupons were shown by PBL at the 2001 National Plasterers Council Conference in Reno. Mr. Hafer directed our attention to certain areas on these photographs claimed to be spot etching. The images we observed did not appear to be any type of spot etching we are familiar with. Additionally, these photographs were supposedly performed with a 400X magnification, which means that the pictures should contain an area about the size of a grain of sand. Thus, we are uncertain whether any coupon "spot etched" at all.
From these observations, PBL concluded that the high levels of aggressive CO2 (again, for some reason, expressed as carbonic acid) in tanks numbers 1, 2, & 3, led to discoloration, spot etching, nodules, and overall degradation of these affected plaster coupons. PBL observed that the slightly lower levels of CO2 in tank #4 resulted in only slight discoloration on all coupons and slight spot etching on four coupons. Finally, PBL observed that its "special" non-chemical program in tank #5 led to low CO2 amounts where no discoloration or degradation was observed on any plaster coupon.

ASSUMING HIGH LEVELS OF CO2 DID EXIST
Although it has been demonstrated that excessively high levels of CO2 are not plausible in any of the PBL water tanks, let us accept for a moment for the sake of argument that those high levels of CO2 did exist in these water tanks. We still question, why water chemistry balance (including high CO2 levels) would be assumed to cause the above results? For example, why did only 30 coupons out of 160 (the total coupons in tanks 1, 2, 3, & 4) result in spot etching and/or nodule growth? If CO2 in water is homogenous and consistent throughout its volume (and this was circulating water), logic would dictate that all the coupons would be affected equally. Not only should water that contains high CO2 affect all of the plaster coupons equally and similarly, but also affect each coupon surface equally and uniformly (and result in uniform etching), not in individual spots or locations. PBL failed to explain this inconsistency.
As mentioned above, the coupons in tank #5, reportedly did not exhibit any plaster phenomenon such as spot etch, nodules, or discoloration, and PBL claimed the reason for this was the low amount of CO2 in tank #5. However, the chart on page 76 for tank #5 indicates a period of time that the CO2 reading was at 30ppm (according to PBL's test method). It appeared that this condition could have existed for as long as one month. By PBL own standards and claimed readings, this amount of CO2 is too high and aggressive.
Tank #4 had recorded results for CO2 of 22 ppm or less for a period of 50 days, with the exception of only one recorded reading of 40 ppm, which lasted only 4 days. Yet, these coupons suffered some discoloration and some spot etching, while tank #5 coupons did not. PBL does not address this discrepancy.
The PBL study also claimed that the increase in calcium hardness levels in tank numbers 1,2, &3 was due to high CO2 levels above the compensating amount, which dissolves calcium from the plaster finish. PBL does correctly state that when the carbonate hardness of water is low, the compensating amount of CO2 is also low. On page 67, PBL claims that only about 3 ppm of CO2 is necessary for about 100 ppm of CaCO3, and when the calcium hardness is lower than 100 ppm, then even less CO2 would be required for maintaining the calcium bicarbonate equilibrium. However, the chart on page 76 (under the data for the "zero degree hardness" tank 5), shows PBL's own listed readings of CO2 always exceeding 3 mg/l - several fold - the highest being 30 ppm as mentioned above. The calcium hardness in tank #5 started out at zero and eventually increased to 100 ppm.
Additionally, on page 67, PBL writes, "It would appear that the use of soft water without the use of chemicals normally prescribed by the recreational water industry was beneficial to these plaster coupons." But tank #5 had an increase in the calcium level of 100 ppm without any acknowledgment or explanation by PBL. Where did the calcium come from if not from the plaster coupons? Additionally, the above statement by PBL contradicts studies, which demonstrate that soft water (low calcium, low alkalinity) is aggressive against cementitious products.

DISREGARDING CLAIMED HIGH CO2 LEVELS AND USING LANGELIER INDEX
As is stated above, it is not plausible for the CO2 levels to have been as high as claimed by PBL. So, we will now disregard the claimed high CO2 levels and compare the results of five different water tanks using the Langelier Saturation Index. Tank #1 water was to be maintained with a negative index, but only ended up negative temporarily. Interestingly, these coupons apparently did not exhibit any spot etching, although some in the swimming pool industry claim that aggressive water is the actual cause of such. Tank #2 water was maintained with a positive index (base, alkali, or scale forming water), and reportedly resulted in six coupons spot etching. PBL attempted to maintain Tank #3 with a balanced index, but was actually slightly positive overall. These coupons apparently suffered the worst, supposedly resulting in 13 coupons spot etching. Tank #4 also was maintained with a slight positive index, but had higher alkalinity levels than tank #3. These coupons apparently had minimal spot etching on four coupons. Tank #5 water was maintained with similar alkalinity levels as tanks #2 & #4, but with a slightly higher pH, and started out with a zero calcium hardness content, resulting in a slightly negative Index also temporarily. Curiously, these coupons apparently showed no visible signs of degradation or other effects.
Interestingly, an analysis of these particular reported results seem to have indicated that the most aggressive water (as in tank #1 & 5) had the least aggressive effect observed on plaster coupons. This is opposite of what the Saturation Index would predict. The coupons in the alkali tank #2 showed signs of etching, also opposite of Index predictions. The neutral tank #3's coupon seemed to show the most problems. Finally, tank #4's coupons, with higher alkalinity levels than tank #3 (but similar to tank #2) showed very little effects of carbonation and spottiness. Essentially, these results are inconsistent and generally in opposition to the Langelier Saturation Index predictions.

EXAMINATION OF THE CAUSES OF DISCOLORATION & NODULES
In regards to the claimed discoloration caused by reportedly high CO2 levels of all coupons observed in the first four tanks, it appears that PBL overlooked an obvious factor as to the cause of discoloration. PBL's own test results indicated that copper was present in the tap water used to fill the first four water tanks, but unfortunately they did not test for iron. Well water, which is used to supplement Lake Mead's water for the Las Vegas area (where PBL is located), is likely to contain iron. It is far more probable that the discoloration noted on coupons was caused by the presence of these minerals, rather than supposedly high CO2 levels. Furthermore, as for the coupons in tank #5 that did not discolor, PBL filtered and softened the tap water that entered this tank. PBL performed this process to remove the calcium for its "zero degree hardness" program. However, this softening process is also known to remove other minerals such as copper and iron! Therefore, no copper or iron was available to discolor the tank #5 coupons.
Despite these facts, PBL concluded that the plaster coupons discolored due to aggressive CO2, and interestingly, did not explain the mechanism or process of how CO2 would cause plaster coupons to turn gray or brown. In contrast, it is known that copper can precipitate out as copper carbonate, copper sulfate, or copper oxide and stain plaster finishes. Iron can be oxidized and precipitate out as iron oxide onto the plaster as a brown deposit.
PBL infers that the nodules formed on a few coupons were also the result of aggressive CO2. This claim does not follow proper chemical rationale. The formation of a nodule is the process of precipitating calcium carbonate, or in other words, calcium becomes insoluble. When CO2 levels are high enough to be aggressive towards calcium carbonate, then these conditions (high CO2) would actually prevent calcium carbonate nodules from forming, and instead, make and keep calcium in a soluble form. It is already well known that delamination of plaster from the underlying base can cause calcium nodules to form. PBL purposely created conditions whereby water could seep down the side of a few petri dishes (used to form the plaster coupons), and create a void area simulating delaminations. Therefore, it is not surprising that some nodules formed.

EXAMINATION OF CHEMICAL FORMULAS
Other discrepancies in the report include PBL's assertion that adding carbonates increases the CO2 and/or carbonic acid in water. PBL provided a chemical formula to advance this concept: CO3 + H20 = H2CO3 + O. Obviously, the above claim is erroneous and does not follow proper chemical rationale. Instead, adding a carbonate (such as soda ash) to water does not increase the content of CO2/H2CO3 in water; in fact it will decrease the amount of CO2/H2CO3 in water if it is present. The following formula illustrates: Na2CO3 + H2CO3 = 2NaHCO3. Sodium carbonate plus carbonic acid equals sodium bicarbonate. Carbonic acid has been neutralized and the amount decreased.
PBL additionally asserts that adding acid to water to neutralize carbonates (i.e., extreme high pH) would also increase the content of carbonic acid. PBL provided the following formula to illustrate: H + CO3 + H2O = H2CO3 + H. Again, this formula is erroneous. The correct formula is H + CO3 + H2O = HCO3 + H2O. As one can see, the acid added reacts with carbonate to form bicarbonate. No carbonic acid is created in this reaction.

OMISSIONS BY PBL
In laying the foundation that high levels of CO2 are aggressive towards plaster (cement), PBL quoted from the sixth section in the third chapter of Concrete Corrosion & Concrete Protection by Biczok, 1964 (although PBL listed it as Biczolis, 1967). Mr. Hafer transferred most of the Biczok section into his work, but without quotation marks, indents, or other such standard indicators to cue the reader to transitions from Biczok's words to Hafer's. In the process, he also omitted critical, key paragraphs, which would have shown Hafer's own positions to be incorrect. For instance: "It will be readily perceived now that water containing carbonic acid is not aggressive to concrete unless two conditions are fulfilled simultaneously: pH value lower than 7.0 and presence of aggressive carbonic acid." PBL also did not refer to the information preceding the above quote, which laid the groundwork for the conclusion made by Biczok. The main significance of this is that PBL cited water balance problems even though the pH of the water in all of their tanks was consistently above 7.3. (There was one test in one tank that resulted in a pH of 6.34 for a few days, but those coupons in tank #1 weren't observed as having any etching problems.)
Finally, we question how PBL could rule out the effects of application and finishing techniques (troweling) on coupons when only 30 out of 160 coupons suffered some degradation. If troweling is performed on coupons individually, (whereas water is universal) and the PBL study indicated "that various finishing techniques were used to fabricate imperfections on plaster surfaces," how then can finishing techniques be ruled out as a possible cause to the few affected coupons? Also, if Hafer is going to refer to concrete industry texts as authoritative on the subject of pool plaster deterioration, why does he omit key sections and disregard material from those sources that cite high porosity, and other improper practices and conditions as causes of corrosion, leaching, and excess carbonation?

ADDITIONAL COMMENTS
The PBL 1996 research report seemed to promote a chemical program consisting of sanitation by Hydroxyl and Oxygen Radicals as a preferred method. At Mr. Hafer's presentation for the NPC, the concept of Hydrogen Peroxide as an oxidizer and PHMB as a residual sanitizer was recommended as the best method as to not cause degradation to swimming pool plaster finishes. In both cases, sufficient problems with the report, methodology, and testing exist to cause doubt that any accurate conclusions could be reached. What is certain is that the promoted chemical regimens have not been demonstrated to be either superior or inferior to other methods relative to etching, staining, discoloration, etc.

SUMMARY

It is well documented in concrete and water chemistry textbooks that a high content of carbon dioxide (CO2) can be aggressive towards cement products. However, it is also well documented that high levels of CO2 generally do not exist when the pH is above 7.0, and that CO2 (or even carbonic acid) levels are not considered aggressive when the pH is above 7.0. PBL cited water balance problems even though the pH of the water in their tanks was consistently above 7.3. The Langelier Saturation Index is often used in the swimming pool industry to predict if water is aggressive towards cement or is scale forming. The observed results by PBL however, are in contrast to Langelier Saturation Index predictions. Some chemical formulas and concepts provided by PBL are fundamentally incorrect, which can lead to such false conclusions.

In recognition of this, it becomes obvious that the PBL study is fundamentally flawed. The CO2 levels as claimed by PBL could not have been as high as were recorded, and were in fact, a fraction of the recorded amounts. The assumptions that carbonic acid in the water led to the plaster phenomena problems on a few affected plaster coupons leap beyond reason, logic, and the limitations of the experimental data. As mentioned in the body of our critique, there is serious doubt that any plaster coupons actually spot etched. Notwithstanding, the supposed results from the effects of the different water conditions on the plaster coupons did not follow a consistent or logical pattern. Some results were similar where they should have been different, and some results that were different should have been similar.
Not only did PBL not prove or document spot etching causes, they did not describe how or why spots would appear as etched areas as opposed to uniform etching or why some areas of the plaster coupons were unaffected. There is no legitimate basis given for discarding "finishing techniques" as possible contributing factors to its study, especially when concrete chemistry books cite that high porosity can lead to corrosion, leaching, and excess carbonation of the cement surface.
Additionally, the presence of copper and likely presence of iron in the tap water was ignored as probable contributors to any noted discoloration on coupons.
Therefore, the authors of this critique paper reject the PBL study and suggest that the conclusions promoted by PBL are without foundation, unwarranted, and have no basis in fact.