PAINTSQUARE BLOG

The author scraped away apparently good coating, only to find water pushing up from the surrounding coatings. This occurred at numerous places around the tank.

At first, seeing fairly standard corrosion, we assumed the presence of contaminants on the steel surface prior to coating.

The Cold Wall Effect? Just like a cold beer on a hot day, but inside out.

Comment from David Cerchie, (9/6/2013, 9:19 AM)

This is an important topic to say the least. One thing that continues to bother me, is the mechanics ascribed to the cold wall effect. When describing the "cold wall effect" it is generally posited that something about the temperature differentials play a part in the phenomenon and perhaps this is true. But I can't help but think the actual water vapors (small molecules) are already passing through the coatings and the condensation is where the "cold wall" plays its part. I'm not convinced that the "cold wall" actually acts as a magnet or drawing force responsible for the permeation. As in the beer can example, the vapors are already moving through the tissue but it is the temperature differential that actually causes the condensation and resulting accumulation of water. Where's our science department?

Comment from Bowdie Campbell, (9/6/2013, 9:56 AM)

How does the tank with achieve thermal equilibrium(hot travels to cold)? Is this process forcing a "capillary effect"? I am fascinated with this idea!

Comment from Tom Schwerdt, (9/6/2013, 10:23 AM)

The cold wall effect can be demonstrated effectively with Atlas Cell testing. Many coatings will blister when a temperature gradient is imposed across the Atlas Cell, when the same coating exposed to the same conditions (other than imposed temperature gradient) shows no blistering. However, there is dispute over where the Atlas Cell testing is applicable in the real world - it is easy to set up a much higher temperature gradient with an Atlas Cell than a coating will ever see in the real world, particularly if cold water is used for the chilling medium on the "back" side of the test panels and real world conditions will only have air on outside of a tank (for example.) Water can transfer heat more than 20 times as fast as air (thermal conductivity of 0.58 versus 0.024)

Comment from Warren Brand, (9/6/2013, 11:38 AM)

I've been troubled by this, too. I have heard that water/moisture routinely travels, or permeates, through coatings all the time - picture a water molecule traveling back and forth through the coating at will, and harmlessly bouncing off the steel substrate. That is why there is such a buzz in the industry about SCAT testing for soluble salts (I think that might be a topic for my next blog). I hate to admit this, but in my 25 years as a coating contractor and after lining, literally, thousands of tanks and other structures, we never, ever, testing for soluble salts and our failure rate was almost zero. My understanding is that many folks believe that water will migrate back and forth through the coating - which is not problematic unless one of three things happen: A. the moisture hits soluble salt, causing a corrosive cell, which leads to blistering. B. the moisture condenses (due to the cold wall effect) thus causing a blister. C. the coating is not compatible with the materials being stored, which leads to entirely new issues not relevant to this blog. I have a very close friend who is a physicist at NIST (National Institute of Standards and Technology) in Washington, D.C. I've posed these questions to him. I'll let you know if I hear anything more definitive. And thank you all for your comments. Please continue to share any thoughts you have - and feel free to challenge me on anything. We're all in the same boat - looking for the truth. Respectfully, Warren

Comment from Gerald Burbank, (9/6/2013, 1:38 PM)

It is my understanding that the condensation that occurs on a cool surface in humid conditions is the result of moisture falling out of the cooler air at the surface. As air cools, the vapor pressure changes (cold air holds less moisture at the same barometric pressure as warmer air - which is why we are so concerned with dew points). Condensation is not necessarily related to some "magnetic" property of the steel that draws the moisture into it. It is more like rain droplets that are spontaneously generated at the colder surface as the air cools when it comes into contact. Chemical deposits (salts or otherwise) that were left on the surface of the steel beneath the coating (after preparing the surface and before applying the primer) are more likely the culprits. If there is an electro-chemical differential between one side of the coating to the other, the coating will act like an osmotic membrane. If the surface of the coating is saturated or submerged, and the coating is at all permeable, then the moisture will be drawn through the coating through the process of osmosis. This will continue to occur until electro-chemical potentials on either side of the coating membrane are equalized. Of course, this can never happen, so the end result is blistering and eventual disbondment. So - No, I don't believe that there is a so called "cold wall effect" per se. The only thing that the cold wall will do is to provide the condensate that may eventually be drawn through the coating. That's my theory anyway.

Comment from Fred Salome, (9/6/2013, 11:48 PM)

I have seen the Cold Wall Effect in action, especially where parts of the surface of a vessel have differing external conditions that cause difference in the internal surface temeprature, such as welded lugs or brackets, or other thermal influences. Where such tanks contain warm aqueous content, blistering of the lining occurs more frequently or sometimes exclusively where there is a heat sink or known cold condition on the outside. The way I see it, any lining will become saturated with moisture after some time in service. Water molecules will enter into and move through the spaces that exist between the larger polymers molecules as well as fill any voids present in the coating. Therefore the steel surface will evetually be in contact with a saturated lining. Water molecules can condense on the steel surface, even in the absence of soluble salts and such, if there is some sort of space or void for this to happen, and the colder the surface the greater the tendency for condensation to occur. So it comes down to the permeability and saturated moisture content of the coating, and it true adhesion (ie lack of void at the interface)to the steel surface. I have always thought that Cold Wall Effect only occurs on some linings, especially polyureas and such which often have a semi-cellular structure, and not on others such as the older type multi-coat low build epoxy linings. In other words, Cold Wall Effect does not in itself cause failure, but will point out deficiencies in some coating types. Screening for this should be become part of the overall assessment of coatings intended for use as linings. Does anyone agree with this?

Comment from Christopher Woods, (9/8/2013, 12:52 PM)

Your readers might find Chapter 2.8 authored by Bernard Appleman titled "The Effects of Soluble Salts on Protective Coatings" , of SSPC Painting Manual Volume #1, interesting reading.

Comment from Warren Brand, (9/8/2013, 1:56 PM)

Hi Gerald, Fred and Christopher - and everyone else. Thanks for your comments. I think a consensus if forming in that it really comes down to the permeability of the coating system, and that, perhaps, the cold wall effect may exacerbate or accelerate a mode of failure - that is, if a coating system is going to be permeable, and fail over time, the cold wall effect may reduce the time before failure. Is that a fair statement? Also, if anyone has any idea for a topic, please let me know. Thanks, Warren

Comment from Tom Swan, (9/8/2013, 8:05 PM)

Osmotic blistering and cold wall effect are both related by the same thermodynamic principles. Saturation Vapor pressure. As was previously mentioned, coatings in immersion service have water molecules in the matrix which move from the immersion side to the steel side and back to the water side when everything is in balance. In the case of salts, when the water finds the salts (it is not pulled to the salts) it solubalizes the salt and due to the change in saturation vapor pressure slows down the waters ability to move out of the solution. As a result, more water enters the solution than leave it until the saturation vapor pressures equalizes with the liquid outside the coating. Since the water enters faster than it leaves, a blister forms. With "cold wall effect" instead of the salt effecting the Saturation vapor pressure it is the temperature. The cold temperatures slow down the water leaving the surface(kinetic energy decreases with temperature) so again, the amount of water coming into the solution by the cold wall leaves slower than the water coming in, again causing blistering and delamination. The Kinetic energy of the cooler water is less than the Kenitic energy of the water on the outside of the coating so look at it as a way to balance out the energy between the two sides. It is basic Thermodynamics.

Comment from Warren Brand, (9/9/2013, 9:27 AM)

Hi Tom, It seems that the root issue remains permeation. Are you saying that all coatings are subject to "having water molecules in the matrix?"

Comment from M. Halliwell, (9/9/2013, 10:26 AM)

There might also be one other mechanism associated with "Cold Wall Effect" that would apply for northern climates (Canada, northern Europe, northern US, etc.)...Assuming that the water can cross the coating (even in a low permeability case) and the cold exterior wall causes condensation, a unique property of water would also come into play. As water temperature moves away from 4oC (39oF), where it is most dense, it actually expands...as you go through heat/cool cycles (or freeze-thaw cycles), any condensation between the coating and substrate would end up generating additional voids / blisters. We see the same effect in asphalt and concrete roadways in cold climate...a minor imperfection allows water in and you get potholes. Even without freeze-thaw cycles (which I assume would generate a greater effect), just a cooling-warming cycle could allow this mechanism to come into play and degrade a coating system faster than anticipated. Just one more bit of food for thought on the topic :)

Comment from Jerry Trevino, (9/9/2013, 2:52 PM)

Warren, this is an interesting situation and discussion. What generic type of coating was it, and what was the chemical that was not compatible with the coating? From a practical point of view, one would not spray or apply a relative warm coating onto a relative much colder substrate. Too big of a temperature difference would affect the curing of the coating. If the coating was waterborne, then I can see some water entrapment possible. I can picture very minute water vapor entrapment between the coating and the steel, assuming the humidity inside was present, however, not enough volume to collect, condense, and then drip out as per the picture. If the coating is permeable enough to allow massive amounts of water or water vapor to migrate through the coating and collect at the steel to coating interface, then I would assume the coating was not fit for immersion service. I do recall only one coating manufacturer address Cold Wall Effect, however, I do not normally see it. There is always some type of molecular exchange and a cross migration of substances on a molecular level whenever two substances are adjacent to each other, as per the discussions of polycarbonates with milk in baby bottles. By the way, what kind of beer was in the hypothetical beer can?

Comment from Warren Brand, (9/9/2013, 8:40 PM)

Hi Jerry, Thanks for the comments. As a point of clarification, and I clearly must not have been that clear, given this is a confusing subject, the cold wall effect is relevant after the tank has been coated and returned to service. And, you are, of course, exactly right about the problems with applying a warm coating onto a cold substrate (although we did do exactly that at 20F for a national ice cream producer in a freezer the size of a small warehouse - perhaps a topic for another blog?). In the case study above, I'm somewhat limited in what I can reveal as this is a current client and we are bound by confidentiality agreements. However, in this case there was one specific chemical in very small quantities that was identified, after the coating was installed, to be detrimental to the existing coating. We were not involved in the original coating application or specification, but during our current due diligence in evaluating this coating failure (and repair protocol) we identified this coating incompatibility. The way this incompatible chemical would likely attack the coating (according to a number of sources) would b for it to soften the coating slightly, but then, when the tank was emptied, and the material evaporated out from the coating, the coating would get hard again. It was our hypothesis that this softening of the coating increased the permeation rate of the existing coating allowing the cold wall effect to have a more pronounced effect than it would otherwise have had. The coating was designed for immersion service, however, there was miscommunication which lead to the current situation. Lastly, and perhaps most importantly, to they type of beer to which I was referring: since this is a hypothetical beer, please feel free to use your imagination to choose whatever beer is most compatible with your hypothetical brat, cheeseburger or ribs. Warren

Comment from Tom Swan, (9/10/2013, 10:43 AM)

While paint is treated like a semi permeable membrane and in many ways it acts like one, in reality,all coatings are permeable to a certain degree an in immersion service contain water in their matrix. As long as a balance is maintained on both sides of the coating, this is not an issue but when something occurs to throw off the balance, soluble salts, soluble organics, soluble solvents or decrease in temperature that slow the rate of water moving away from the surface, this is when blisters occur. None of these "pull" the water through the coating and the water moving through the coating does not know what is on the surface until it gets there, but when it does and the rate of the water leaving is slower than the water coming in, problems will occur. It is nature’s way of maintain energy balance between both sides of the coating.

Comment from William slama, (9/11/2013, 10:56 AM)

Wow! A lot of interesting viewpoints about a very complex process(es). First, I want to clarify that the water vapor is not "pulled" into or through the membrane (coating or lining). It is pushed because the partial pressure of water vapor (as a gas) is greater on the hotter side of the membrane. Those actual pressure differences can be visualized by looking at the water vapor partial pressure on the curve vs temperature. Note that at lower temperatures, say 70 too 120 degrees F, the curve is shallow (the difference in temperature doesn't result in a great difference in partial pressure). As the temperature goes higher the curve gets steeper, and of course at 212 deg. F the pressure is about 1 atmosphere ~14.7 psi. One cannot dispute that as being the primary driving force for water vapor transmission through any membrane that is measurably permeable to water. That is what is commonly referred to as the "cold wall effect". Obvious manifestations of this have been doted from the time we started coating (steel) substrates to protect from corrosion. The most frequent is when a coating or lining that is performing well in a vessel at elevated temperature develops blisters or disbondment that is localized to the area of an external stiffener on the vessel. The actual thermodynamics are complicated -thermal conduction/convection/radiation- the only three ways heat can be conveyed. Regarding permeability, almost all of the polymer systems can be correctly defined as "semi permeable membranes". That is that they cal allow penetration of some species but not others, generally related to molecular size vs the pore size within the polymer (over-simplistic explanation). The most obvious example is that if the lining is exposed to a salt solution, water can penetrate the polymer, but not the salt. That is a primary demonstration in Chemistry 101 - That water on the cater side of the membrane travels through the membrane to get to the salt solution. That is, of course, termed "osmotic" pressure and it can easily reach pressures of several atmospheres.

Comment from William slama, (9/11/2013, 10:57 AM)

I have performed, witnessed, and reviewed many, many "permeation" tests over the years. Most were performed by ASTM E96 and ASTM D1653 and significant variations thereof. The reason is that these tests were originally intended for relatively permeable thin film coatings. As such, the temperatures and gradients did not pose a high differential pressure and it was difficult to get good data on thick, low permeability linings. So we took advantage of the steeper pressure slope , using temperatures n the 140 to 180 degree F range. Here another factor (mentioned above) can enter in. That is the physical response of the polymer to the elevated temperature. In simple terms, because of thermal expansion, the "pores" of the polymer will become larger. An addition all of the polymers employed have a critical temperature, referred to as the "glass transition temperature" Tg. Above that temperature the physical properties of the polymer change dramatically. The internal structure becomes weak and that allows the structure to "creep" or "relax" in response to stresses caused either by differential thermal expansion or external physical forces. The relevance to our topic is that the polymer is easier to penetrate (permeate). As discussed above most believe that the one side immersion test ASTM C868, referred to as the "Atlas cell test" is the best (if imperfect) method to simulate the effect of the cold wall effect in the laboratory. It has, of course, been demonstrated to cause the permeation/blistering that is experienced in the field. Whereas even lesser coatings can withstand boiling water in an isothermal (immersed) condition.

Comment from William slama, (9/11/2013, 10:58 AM)

As we all know, a lot of work (as referenced above) has been done to evaluate the effects of varying amounts of soluble salt contamination that could be present on the steel surface prior to coating. It is generally agreed that at ambient temperatures the effect is not severe, but at elevated temperatures the presence of residual soluble salts can be disastrous. We recently saw complete failure of a duct lining at 130 deg. F where the salt concentration under the lining could be easily measured after the fact. The presence of the salt underneath changes the driving force from just partial pressure of water vapor to one that includes osmotic pressures which can be very high. All of the above is somewhat over simplistic in that the specific properties of the lining or coating have much to do with the result. I have seen some data that suggests that the coefficient of thermal expansion of the lining in the plane of the substrate is very important to reduce the stresses at the lining/steel interface. I have also seen that systems with a very high Tg tend to hold up well at elevated temperatures. Note that most coatings suppliers require vessels to be thermally insulated for higher exposure temperatures. Overall, I believe that everyone in the industry is continually trying to understand all of the salient factors to provide the best solutions for corrosion protection

Comment from Gerald Burbank, (9/12/2013, 6:23 AM)

Interesting discussions. I'm not a chemist or a physicist, so I can't comment on many of the hypotheticals here. My suspicion is that there are multiple causes for such failures, which may include osmotic processes, stress induced by differential expansion and contraction between the substrate and coating film, the coatings adhesion/cohesion, and permeability. It's nice to we a thorough examination of a problem as opposed to the vitriol and posturing that iI see on most discussion boards. Highly professional and informative.

Join the Conversation: