Slow Progress on New Blowing Agents for Polyiso

musingsheader image

Slow Progress on New Blowing Agents for Polyiso

At least one manufacturer is now researching ways to improve the cold-weather performance of polyisocyanurate

Posted on Jan 20 2017 by Martin Holladay

R-valueMeasure of resistance to heat flow; the higher the R-value, the lower the heat loss. The inverse of U-factor. tests by the National Roofing Contractors Association and the Building Science Corporation have revealed that the thermal performance of polyisocyanurate is greatly reduced at cold temperatures. While the R-value of polyiso at a mean temperature of 75°F might be about R-5.7 per inch, the effective R-value of the polyiso drops to about R-4.8 per inch at a mean temperature of 25°F.

Over the past three years, has published several articles on this problem. (See, for example, In Cold Climates, R-5 Foam Beats R-6 and Cold-Weather Performance of Polyisocyanurate.) As builders have learned that polyiso wasn’t doing as good an insulating job at cold temperatures as they had thought, they’ve started to ask questions, including:

  • How long have polyiso manufacturers known about this problem?
  • Are polyiso manufacturers doing anything to try to solve the problem?

In response to such questions, it’s fair to say that polyiso manufacturers have been tight-lipped. John Straube, a professor of Building EnvelopeExterior components of a house that provide protection from colder (and warmer) outdoor temperatures and precipitation; includes the house foundation, framed exterior walls, roof or ceiling, and insulation, and air sealing materials. Science at the University of Waterloo in Ontario, has had several conversations with representatives of polyiso manufacturers. In a 2015 interview, Straube told me, “Maybe the manufacturers … are not telling us the truth. But in apparently honest conversations with several polyiso manufacturers, they’ve told me that they do not fully understand why this is happening. PIMA [the Polyisocyanurate Insulation Manufacturers Association] doesn’t even believe the phenomenon is real. Some of their members … say, ‘I am not 100% sure it is a real effect.’ Well, it is real. … Here’s the thing about polyiso manufacturers: They do not share their proprietary formulations. Manufacturer A may not know what is in Manufacturer B’s foam. So now, when we discuss cold weather performance, they are kind of going, ‘I’m really surprised.’ But it’s possible that they are jiving me.”

Like Straube, I’d like to hear more straight talk from polyiso manufacturers. That’s why my interest was piqued when I got a chance for a brief glimpse behind the curtain at a recent building envelope conference, where John Letts, the technical director for insulations at Firestone Building Products, a manufacturer of polyiso, presented a research paper called “Improving Polyiso Thermal Performance at Low Temperature.”

Letts presented his paper on December 5, 2016 at the Thermal Performance of the Exterior Envelopes of Whole Buildings XIII International Conference in Clearwater Beach, Florida. The paper’s co-authors were Jennifer Yao and Michael J. Hubbard.

A history of the problem

John Letts explained how manufacturers got to this point. Before 1993, manufacturers of polyiso used CFC-11 (Freon) as a blowing agent. Polyiso made with CFC-11 performed better at cold temperatures than today’s polyiso. At a mean temperature of 40°F, for example, this pre-1993 polysio actually performed a little better than it did at a mean temperature of 75°F.

In 2000, most manufacturers of polyiso switched from HCFC 141B (the blowing agent used between 1993 and 2000) to a mixture of pentane and carbon dioxide. “Because of the ozone depletion potentialAmount of damage to the ozone layer a given chemical can cause compared to trichlorofluoromethane (CFC-11), which is given a value of 1.0 on this relative scale. of our blowing agent, we migrated in 2000 to pentane,” Letts said. “Pentane has a low global warming potential and zero ozone depletion potential.”

After the switch from HCFC 141B to pentane, the cold-weather performance of polyiso took a nosedive. According to Letts, however, manufacturers were unaware of this problem until relatively recently. In his presentation, Letts implied that Firestone didn’t realize that polyiso performs poorly at cold temperatures until the NRCA published their R-value test results. Discussing the thermal performance of polyiso, Letts said, “The NCRA research shows that it drops down at low temperature. It’s puzzling data. Pentane doesn’t behave like other blowing agents.”

Letts admits that Firestone has a problem. He said, “The R-value at 40°F mean temperatures are unexpectedly low for the current products.” And in the published paper, the researchers wrote, “The surprisingly poor low mean temperature (40°F) R-value of polyiso produced with pentanes has been confirmed in our studies.”

Changing the definition of R-value

In addition to looking at possible solutions to the problem of poor polyiso performance at cold temperatures, Letts’s paper addressed a secondary topic: Is it possible to change the legal definition of R-value?

The researchers wrote, “Historically, R-values have been measured at a mean temperature of 75°F, typically with at least a 40°F temperature difference between the high and low temperature plates. ASTMAmerican Society for Testing and Materials. Not-for-profit international standards organization that provides a forum for the development and publication of voluntary technical standards for materials, products, systems, and services. Originally the American Society for Testing and Materials. standards for insulations all have mandatory R-values at a mean temperature of 75°F, as does the Federal Trade Commission (FTC). The FTC mandated data be reported at 75°F, because it realized that testing at different temperatures would yield different results and by mandating a test temperature they could eliminate some gamesmanship with R-value reporting. But why report at a mean temperature of 75°F?”

This question is of intense interest to many insulation manufacturers, and Letts's paper discusses it at length. Ultimately, however, this is a political rather than a scientific question. Suffice it to say that any changes to the definition of R-value would create winners and losers, setting the stage for for intense lobbying and political infighting.

What’s going on?

According to Letts, the traditional explanation for the cold-weather drop-off in polyiso performance — that the phenomenon is caused by the condensation of the blowing agent at low temperatures — doesn’t hold up to close examination. Something else is going on.

“It has been noted that the R-value versus mean temperature curve is odd for polyiso, especially around mean temperatures of 40°F and lower,” the researchers wrote. “Whereas some condensation of the blowing agent would be expected since n-pentane’s boiling point is 36°C and isopentane’s boiling point is 28°C, HCFC-141b’s boiling point is 32°C at 1 atm, which is right in between these two pentanes but it does not have this oddity in its R-value versus mean temperature curve.”

At the Florida conference, Letts explained, “I always like to think of things at a fundamental molecular level. The polymer network is key. Past blowing agents have been polar whereas the present blowing agents are relatively nonpolar.”

The researchers wrote, “During our examination of the low mean temperature R-value issue, we tried to think of the problem at the molecular polymer level. If the pentanes are absorbing into the polymer network in the struts and cell walls of the foam much like moisture condenses on glass on a high humidity day, and if these absorption sites or surface areas are less available for the pentanes to interact with the polymer domain, then the pentanes may not condense as readily. A variety of solvents have proved surprisingly very effective in increasing the R-value at a mean temperature of 40°F relative to the R-value at a mean temperature of 75°F. Most of these solvents have boiling points that are higher than the pentanes, but, still, their 40°F R-values are better than the control without them. These solvents tie up the sites for either absorption or adsorption, because they are better solvents than the pentanes (isopentane and n-pentane), which free these pentanes to remain in the gaseous phase.”

Developing a new type of polyiso

Firestone has embarked on a research program to look for solutions to the problem. For Letts, the goal is to find a blowing agent that performs as well as polyiso performed back in the 1980s. Letts told the audience in Florida, “What we want to do is move the product back to the way the product performed with CFC-11.”

While Firestone is fixated on improving the performance of polyiso at a mean temperature of 40°F, it could be argued that cold-climate builders also care about polyiso performance at even colder mean temperatures — at 25°F, for instance. (At a mean temperature of 25°F, polyiso performs significantly worse than it does at a mean temperature of 40°F.) Nevertheless, if the performance of polyiso can be improved at a mean temperature of 40°F, it would at least represent an improvement over the current situation — even if the solution still leaves cold-climate builders grumbling.

Every type of polyiso gets a score

To score the performance of various types of polyiso, Letts uses a numerical ratio he developed. The ratio has as its numerator the effective R-value per inch of polyiso at a mean temperature of 40°F, and has as its denominator the R-value per inch of polyiso at a mean temperature of 75°F.

In their paper, the researchers wrote, “One method to measure the low temperature R-value performance of insulation materials is to divide their R-value at a mean temperature of 40°F by their R-value at a mean temperature of 75°F, which yields a low mean temperature ratio (henceforth to be designated as the ratio). A ratio above 1.0 means that the R-value at 40°F is higher than the R-value at 75°F and vice versa. … Historically, polyiso R-value as a function of mean temperatures has increased at mean temperatures below 75°F to at least 40°F, or this ratio is greater than 1.0. A review of published R-values at mean temperatures of 75°F and 40°F for polyiso expanded with CFC-11 or HCFC-141b gave a ratio percentage of 1.042 to 1.170, with an average of 1.084 for CFC-11…”

According to Letts, the ratio for the polyiso that Firestone sells today is only 0.90. (To see a table listing examples of Letts's ratio for older samples of polyiso blown with CFC-11 and HCFC-141b, see Image #2, below.)

Exploring two possible solutions

Firestone is looking at two different approaches to solving the cold-weather performance problem. Approach #1 is to develop a form of polyiso made with a new blowing agent, code-named “Solution A,” offered by one of Firestone’s suppliers.

Approach #2 is to add a little bit of butane to the pentane blowing agent that Firestone now uses.

Solution A is closely guarded secret

Firestone has performed experiments on polyiso blown with “Solution A.” This approach is promising, but the company is a long way from being able to bring this new type of polyiso to the market.

The researchers wrote, “… [W]e have demonstrated the efficacy of Solution A; R-value performance at a mean temperature of 40°F is nearly equal to that at a mean temperature of 75°F. It is interesting to note that even with 30 mole % methyl formate, the addition of Solution A with the polyol 2 yielded the highest ratio (12%), and the compressive strength was still very good. However, much more work needs to be done to optimize this technology.”

At the Florida conference, Letts said that when the new blowing agent is used, “We’re getting R-values approximately the same at 40°F mean temperature as at 75°F mean. Plant trials are imminent. We expect to commercialize this.” However, polyiso blown with Solution A probably won’t be available for several years.

Approach #2: Add butane

Letts was disappointed with the results of the experiments that involved adding butane to the standard pentane blowing agent.

The researchers wrote, “A concern with present day polyiso blown with iso/n-pentane is condensation of the pentanes as the temperature decreases. Hence, addition of 15% by weight of either n-butane (boiling point 32°F) or isobutane (boiling point 10.9°F) to the pentane mixture in the laboratory was explored to determine if the ratio 40°F to 75°F could be increased above 1.00. The vapor pressure at 40°F, for example, should be higher than for the iso/npentane, especially for the isobutane blend. When we used the same formulation in the previous work except the iso/n-pentane was modified with n-butane or isobutane, the R-values at 75°F were similar, but the 40°F R-values were better than the control. However, the R-value at 40°F were still below the R-value at a mean temperature of 75°F. Their ratios were 0.967 and 0.976 for n-butane and isobutane, respectively. Although they were better, they were still 3.4% and 2.2% lower than R-values at 75°F.”

At the Florida conference, Letts explained, “We looked at the addition of butanes to the standard pentane [blowing agent] formulations. At 40 degrees, the performance is not very good — it’s still not back where we wanted to be, as we were in prior years [with CFC-11].”

Questions from the audience

At the Florida conference, questions from the audience revealed that Firestone is still a long ways away from a solution to the polyiso problem.

One audience member commented, “There is a huge reduction in polyiso performance when the mean temperature is 25 degrees. It’s not just a problem at 40 degrees. Just when the cold starts going, you pull the plug. The outdoor temperature is well below 45 degrees for much of the winter in Minnesota.”

Letts answered, “You are absolutely correct. You need insulation when it is coldest. But we need to get the performance of polyiso back where it was when we used CFC-11. That’s a good place, even if maybe it’s not the best place.”

Another audience member asked Letts, “What is the global warming potential of the new blowing agents you are studying?”

Letts answered, “I don’t know.”

Martin Holladay’s previous blog: “Building Science Information for Builders.”

Click here to follow Martin Holladay on Twitter.

Tags: , ,

Image Credits:

  1. Image #1: L&W Supply
  2. Image #2: John Letts, Jennifer Yao, and Michael Hubbard
  3. Image #3: Building Science Corporation

Jan 20, 2017 10:43 AM ET

Is the GWP really unknown?
by Charlie Sullivan

It's astounding to think Letts wouldn't know the GWP of the secret ingredient. Perhaps it's some newly proposed gas that hasn't been analyzed for GWP, but would they really invest in studying the performance vs. temperature before figuring that out? More likely he is bluffing, knowing that if he said a number, someone could cross reference it and find out what they are using. Or perhaps they made him undergo a selective memory purge procedure before allowing him to give this talk.

Thanks for the good reporting and the appropriate emphasis on the fact that 25 F is a bigger problem than 40 F. In the paper, there's a nod to the fact that the location of the insulation in the assembly matters, but the paper still concludes that looking a colder temperatures is not necessary. That's seems like a contradiction.

Another point I'd like to see better addressed by the manufactures is how this settles out after 10 to 50 years of aging. Maybe we can simply design based on that eventual settled performance and consider any short-term improved performance a bonus.

The discussion of how the different blowing agents interact with the cells walls is very interesting.

Small correction: I believe the date was Dec. 2016.

Jan 21, 2017 3:34 AM ET

Still better than EPS?
by Adam Liberman

It seems that despite the problem, the R-value per inch at 25 deg F is still significantly better than EPS. So if I want to maximize the benefit of 2" of rigid non-permiable insulation, and it rarely goes below 20 deg F, then Polyiso would still be the most effective of the low GWP choices. Correct?

Jan 21, 2017 10:52 AM ET

Edited Jan 21, 2017 10:53 AM ET.

More R-values
by Armando Cobo

Maybe, manufacturers should provide a chart with REAL temperature R-values that Designers and Builders can rely on... start at 0°F to 100°F in 10°F or 20°F intervals, in that way, folks in AK and CAN can use those values that are appropriate for them, as well as folks in AZ or TX to use theirs, and everyone in between.
I'm sure providing truthful and open information would be helpful for most of us.

Jan 21, 2017 12:31 PM ET

EPS is better at 25 F
by Charlie Sullivan

Adam, the thermal conductivity of EPS varies with temperature too, and in fact it become a better insulator at low temperature. You can see that this plot of thermal conductivity (lower values = more thermal resistance)

Below about 7.5 C (45 F) the EPS beats the sample of polyiso shown in that chart. By the time you get down to 25 F, the EPS has twice the thermal resistance of the polyiso!

The sample shown there may be a particularly bad one, so that may be a little pessimistic. But they key point is that as polyiso is getting worse at low temperatures, EPS is getting better.

Jan 21, 2017 12:58 PM ET

That's 25F MEAN temperature through the foam. @ Charlie Sullivan
by D Dorsett

Those performance tests are done with at least a 50F temperature difference between the warm & cold sides of the foam. So in an insulating sheathing option with other insulation on the interior, that would be 0F outdoors, 50F on the interior side of the foam to achieve a 25F mean temp. When it's 25F outdoors, the mean temp through foam sheathing would be significantly warmer.

The mean temp through the foam layer depends on the stackup of other materials (including insulation layers), the indoor temp, and the outdoor temp, and in some assemblies, the thermal mass in each layer if it's enough to induce a significant time delay.

It's silly to look at the performance at the 99.6th percentile temperature bin- most of the time you're more interested in the average seasonal performance, or the average performance during the coldest 5-10 weeks of a typical year. If your January binned hourly mean outdoor temperature is 25F, (a typical US climate zone 5 location, like Chicago: ) estimate the mean temp through the foam layer based on the proportional R values for when it's 25F outside. Even though the foam performance may crash (or rise) at lower temperatures, the total number of hours in the heating season that it's lower performance are small compared to the total heating season, or the fraction of heating degree-days.

Register for a free account and join the conversation

Get a free account and join the conversation!
Become a GBA PRO!