Air Leakage Degrades the Thermal Performance of Walls
No surprises here — but testing by the Building Science Corporation begins to quantify the problem
For the past five years, researchers at the Building Science Corporation (BSC) in Massachusetts have been testing the thermal performance of a variety of wall assemblies as part of an ambitious project to develop a new metric to replace R-valueMeasure of resistance to heat flow; the higher the R-value, the lower the heat loss. The inverse of U-factor. . (I last reported on the project in my August 2011 article, A Bold Attempt to Slay R-Value.)
BSC researchers recently released information on the multi-year project:
- On August 1, 2012, BSC principal Chris Schumacher gave a presentation, “Thermal Metric Project: A year of progress,” at the Westford Building Science Symposium.
- On August 6, 2012, BSC released a three-page report, “Thermal Metric Research Project.”
The BSC project is financed by a group of five insulation manufacturers and one trade organization — Dow, U.S. Greenfiber, Honeywell, Huntsman Polyurethanes, IcyneneOpen-cell, low-density spray foam insulation that can be used in wall, floor, and roof assemblies. It has an R-value of about 3.6 per inch and a vapor permeability of about 10 perms at 5 inches thick., and the North American Insulation Manufacturers Association — known collectively as the “Thermal Metric consortium.” While BSC hasn’t yet published a peer-reviewed paper reporting on their research project, the consortium has released enough details to draw a few conclusions.
Developing a new metric or label for wall assemblies
The BSC researchers set out to measure the rate of heat flow through a variety of wall assemblies under different conditions — for example, at different rates of air flow. The intent was to compare the performance of a wall when air was leaking through the wall with the performance of the wall without any air leakage. The researchers also hoped to test wall assemblies at different outdoor temperatures and to quantify the effect of thermal bridgingHeat flow that occurs across more conductive components in an otherwise well-insulated material, resulting in disproportionately significant heat loss. For example, steel studs in an insulated wall dramatically reduce the overall energy performance of the wall, because of thermal bridging through the steel. .
Eventually, the BSC researchers hope to propose a new metric for wall assemblies — something akin to the NFRC label now found on windows. “The window industry was in disarray a decade and a half ago, and it was very difficult for practitioners to select windows,” said Joseph Lstiburek, a principle of BSC. “Then the window industry came up with the NFRC ratings for windows, so that today it is fairly straightforward to get information on the three most important characteristics of windows. That labeling came from the industry, and it provides a tremendous example of how the industry can cooperate.”
Since most wall assemblies are site-built, it’s unlikely that you will ever see these new labels attached to wall assemblies in the “Walls” aisle at Home Depot. However, in the future, designers may be able to look up wall performance data on a website, just as they now look up whole-wall thermal performance values on the Oak Ridge National Laboratory website. (The thermal performance measurements made by the BSC researchers differ from the whole-wall R-value measurements made by the Oak Ridge researchers in several important ways. Unlike the Oak Ridge researchers, the BSC team measured thermal performance at a variety of outdoor temperatures. They also measured the walls’ thermal performance while air was flowing through the wall assemblies.)
“We need a big guarded hot box”
Before the BSC researchers could begin testing wall assemblies, they had to build a large guarded hot box that could accommodate wall assemblies measuring 8 ft. by 12 feet. The resulting hot box is capable of testing assemblies at pressure differences of +/-25 Pascals, at air flows of up to 30 cfm, and at a range of simulated outdoor temperatures (from -25°F to 145°F). The actual range of temperatures used to test wall assemblies was -18°F to 144°F.
Building the guarded hot box was no easy task. “We wanted to look at air flow effects, so we had to measure energy and mass flows,” said Schumacher. “The mass flow component is difficult to measure.”
Insulation performs differently at different temperatures
Researchers have long known that insulation materials behave differently at different temperatures. “An R-13 fiberglass batt performs better at low temperatures than at high temperatures,” said Schumacher. “There is an apparent improvement in thermal performance as the batt gets colder. Since there is more radiation at high temperatures, the performance is degraded at high temperatures. The same thing happens with expanded polystyrene.”
Polyisocyanurate bucks the trend: when tested at a mean temperature below 50°F, the performance of polyiso gets worse rather than better. The reason for the declining performance, Schumacher explained, is that “the trapped blowing-agent gases start to condense at cold temperatures.”
Description of the wall assemblies
During the past year, BSC researchers tested the performance of seven wall assemblies. (Confusion alert: the numbering of the wall assemblies discussed in this article begins with Wall 2, not Wall 1.)
- Wall 2: 2x4 studs, 16 inches o.c., with inset-stapled kraft-faced fiberglass batts;
- Wall 3: 2x4 studs, 16 inches o.c., with face-stapled kraft-faced fiberglass batts;
- Wall 4: 2x4 studs, 16 inches o.c., with damp-sprayed cellulose;
- Wall 6: 2x4 studs, 16 inches o.c., with R-13 open-cell spray polyurethane foam;
- Wall 6: 2x4 studs, 16 inches o.c., with R-13 closed-cell spray polyurethane foam;
- Wall 7: 2x4 studs, 16 inches o.c., with R-13 friction-fit fiberglass batts plus 1-inch-thick exterior XPSExtruded polystyrene. Highly insulating, water-resistant rigid foam insulation that is widely used above and below grade, such as on exterior walls and underneath concrete floor slabs. In North America, XPS is made with ozone-depleting HCFC-142b. XPS has higher density and R-value and lower vapor permeability than EPS rigid insulation. rigid foam;
- Wall 8: 2x6 studs, 16 inches o.c., with R-21 friction-fit fiberglass batts.
The walls were built with a few deliberate air leaks. Each tested wall had three electrical boxes (two duplex outlet boxes and one single-gang switch box) with Romex wiring between the boxes.
Shims were used to introduce a 1/32 inch gap between the OSB sheathingMaterial, usually plywood or oriented strand board (OSB), but sometimes wooden boards, installed on the exterior of wall studs, rafters, or roof trusses; siding or roofing installed on the sheathing—sometimes over strapping to create a rainscreen. and the bottom plates and top plates of the walls, as well as between the gypsum wallboard and the bottom plates and top plates of the walls. While this decision may sound odd, the deliberate gaps were included so that the walls would leak about as much air as “typical” unsealed walls. “The 1/32 inch gap is based on work done by Don Onysko of Forintek, and represents the typical gap resulting from expansion and contraction of wood framing due to seasonal wood moisture content changes,” said Schumacher. “The 1/32 inch gap sounds like a lot, but it gave us leakage rates similar to what we are seeing in the field.” According to the BSC report, “All of the assemblies … were more airtight than most real wall assemblies (e.g. 0.05 to 0.20 cfm50/ft2).” The quoted range of 0.05 to 0.20 cfm50/ft2 refers to leakage rates in opaque walls, but not walls that include a band joist, door, or window, Schumacher explained.
Some GBAGreenBuildingAdvisor.com readers — especially those building in cold climates — may be wondering, “Why test an R-13 wall? Who builds an R-13 wall anymore?” The answer to that question, of course, is that the R-13 wall is alive and well in Climate Zones 1 through 4 — for example, in Virginia and Kentucky and states further south.
Each of the seven wall assemblies was tested at five outdoor temperatures (-18°F, 0°F, 35°F, 108°F, and 144°F) without any air leakage through the walls. Then they were re-tested at two outdoor temperatures (0°F and 108°F) with two types of air leakage at each temperature (first with 10 Pascals of infiltration, and then with 10 Pascals of exfiltrationAirflow outward through a wall or building envelope; the opposite of infiltration.). “Air leakage increases when it is cold,” said Schumacher. “When cold, all the walls get leakier.”
Testing a wall assembly at multiple outdoor temperatures and pressure regimes takes a lot of time. According to Schumacher, “It took us a month to test one wall assembly.”
Heat recovery during infiltration and exfiltration
When infiltration and exfiltration occur at multiple leak sites in a wall — imagine a colander — the heat flow through the wall is less than when the infiltration occurs at a single hole and the exfiltration occurs at another single hole. This is because a leaky wall acts like a heat exchangerDevice that transfers heat from one material or medium to another. An air-to-air heat exchanger, or heat-recovery ventilator, transfers heat from one airstream to another. A copper-pipe heat exchanger in a solar water-heater tank transfers heat from the heat-transfer fluid circulating through a solar collector to the potable water in the storage tank..
Here’s how the phenomenon works in a wall with infiltrating air. First, imagine a perfectly tight wall with a single hole in the center of the wall. All infiltration occurs through this one hole; let’s say the hole admits 20 cfm. This type of infiltration does not benefit from the type of heat exchange under discussion.
If the same 20 cfm of infiltration occurs through one hundred tiny colander-sized holes, the energy penalty attributable to the infiltrating air will be less. That’s because the infiltrating air that passes through all those little holes gains heat as it moves through the wall; in fact, the air gains more heat than it would if it passed through the single hole.
Here’s how the phenomenon works in a wall with exfiltrating air. If all of the exfiltration happens at a single hole, you don’t get any benefit from the heat-exchange phenomenon. But if the infiltration occurs through one hundred tiny colander-sized holes, the warm exfiltrating air will raise the temperature of the wall assembly as it exits, reducing the conductive heat flow through the wall.
“Exfiltration reduces heat flow through the wall,” explained Dave Ober, one of the presenters at the Westford Building Science Symposium on August 1, 2012. “There is a performance benefit. The wall acts like a parallel-flow heat exchanger, so you are getting a little heat recovery.”
Ober continued, “The effect of discrete of holes” — that is, holes that do not resemble a colander — “is easy to calculate because there is no interaction. But uniform air infiltration” — that is, air infiltration through colander-like holes — “recaptures some of the heat loss through the wall by heating some of the air flowing through the wall. Uniform air exfiltration also recaptures a small amount of heat — actually it changes the temperature gradient and reduces the heat flow. You are not recapturing all of the heat; it is a net loss. You are reducing the conductive losses somewhat.”
Because of the heat-recovery effect, Schumacher noted, “Walls with fibrous insulations have reduced heat flows compared to sealed walls with an equal airflow through a discrete hole.”
It’s important to remember the heat carried away by escaping air in the winter incurs a greater thermal penalty than any benefit accrued by the heat-exchange that occurs as air flows in or out of a wall. “There are a lot better ways to achieve heat recovery,” said Schumacher. “It’s better to seal the walls.”
Here’s how the BSC researchers summarized the heat-recovery phenomenon:
- Infiltrating air may recover some heat flow through the windward wall, thus reducing the energy impact of the air leakage;
- Exfiltrating air can reduce the heat flow through walls on the leeward side of the building;
- Any air leakage through a wall assembly will result in increased heat flow.
As part of their thermal metric project, the BSC researchers developed a data-analysis methodology that accounted for the heat-exchange process that accompanies infiltration and exfiltration.
Some of the researchers’ findings
First of all, if the walls were sealed and there was no air flow through the walls, all of the R-13 walls behaved the same. As Gertrude Stein might have put it, an R-13 wall is an R-13 wall is an R-13 wall. “When the nominal R-13 walls are sealed and tested, they have the same heat flow, plus or minus 4%,” said Schumacher. “They all perform roughly the same.” (The results represent averages from tests conducted at five outdoor temperatures: -18°F, 0°F, 35°F, 108°F, and 144°F).
These findings are shown in the bar graph reproduced below.
When the walls were tested at a 10 Pascal pressure difference, however — in other words, when there was air flow through the walls — they no longer behaved the same. The walls insulated with fiberglass allowed more cfm of air flow through the walls than the walls insulated with spray polyurethane foam. The walls insulated with cellulose had air leakage levels that fell between those of the walls insulated with fiberglass batts and those of the walls insulated with spray foam. “There is not much difference between inset-stapled fiberglass and and face-stapled fiberglass,” said Schumacher. “But there is much less air leakage through the spray foam walls.”
The findings are shown in the bar graph reproduced below. (The results shown in this graph and the results in all of the subsequent graphs on this page represent averages from tests conducted at only at two outdoor temperatures: 0°F and 108°F).
Of course, infiltration and exfiltration through a wall assembly degrade the thermal performance of the wall. The leakiest walls (those insulated with fiberglass batts) saw a greater degradation in thermal performance than the tighter walls (those insulated with spray foam). As one would expect, the wall insulated with cellulose has thermal performance results that fall between the performance of fiberglass batts and spray foam.
These findings are shown in the bar graph reproduced below.
When tested under a 10 Pascal pressure difference, the heat flow through the fiberglass insulated walls was about 35% higher than when it was tested without any air flow. The spray foam wall performed better. When tested under a 10 Pascal pressure difference, the heat flow through the open-cell spray foam wall was only 16% higher, and the heat flow through the closed-cell spray foam wall was only 23% higher, than when the same walls were tested without any air flow.
These findings are shown in the bar graph reproduced below.
The percentage degradation in thermal performance was higher for the 2x6 walls insulated with fiberglass batts than it was for the 2x4 walls insulated with fiberglass batts. While the increase in heat flow attributable to air leakage was between 34% and 35% for the 2x4 walls, the increase in heat flow attributable to air leakage was between 45% and 48% for the 2x6 walls. “The percentage effect is much larger on high R-value walls because the heat flows were lower to begin with,” said Schumacher. “So with higher R-value walls, it’s more important to take care of air flow.”
These findings are shown in the bar graph reproduced below.
Note that air leakage degraded the thermal performance of every single wall — even the walls insulated with spray polyurethane foam. According to a summary released by BSC, “All wall assemblies experience a loss in thermal performance due to air movement through the assembly. This is true for all of the assemblies tested regardless of the type of insulation material used (e.g. cellulose, fiberglass, ocSPF, ccSPF, XPS). The energy impact of airflow depends on the flow path, the interaction between the air and the solid materials in the assembly, and the installed R-value of the assembly.”
Address the big holes
In a discussion of advice for builders, Lstiburek said, “In houses, the big holes matter. Plain walls don't contain the big holes. I can make a house insulated with fiberglass just as tight as a spray-foam home or a cellulose home if I seal the big holes. Most folks who use cellulose also get the draft-stopping big-hole thing; the same thing for spray polyurethane foam. Most typical builders who do fiberglass do the bare minimum, and typically don't get it. ... Consider that there is not a lot of leakage in a stud cavity, even with the shims, compared to the big holes everywhere else in a typical building.”
When asked to comment on the results of the thermal metric consortium’s work, Lstiburek said, “Air tightness is important to all insulation systems. I know this should be obvious, but it is clearly not, based on what we see being built out in the field. Having good numbers helps make the case for air tightness unassailable.”
Lstiburek is very proud of the work performed by the consortium. “This is a huge first step,” he said. “I don’t know if anyone appreciates how difficult it was to get to here. We’re not done, but we are a long way down the road to giving good advice.”
The way I see it
The findings of the BSC researchers seem to me to be consistent with long-standing advice from insulation experts, including advice that GBA has provided for many years. Here's a sample of some advice from the GBA website that remains valid:
- “Stopping air leaks is just as important as — maybe more important than — adding insulation. Unless builders prevent air from leaking through walls and ceilings, insulation alone won't do much good.”
- “If you choose an insulation that doesn't stop air flow, it's important to install an adjacent air barrier material.”
- “Fiberglass batts are air-permeable — they do a poor job of resisting airflow — so it is essential that they be installed in continuous contact with an impeccable barrier to air movement. For the very best performance, fiberglass batts should be installed in a sealed cavity (for example, a stud or joist bay) with an air barrier on all six sides.”
- “Dense-packed cellulose resists air infiltration better than fiberglass batts, but it's not an air barrier.”
- “Spray polyurethane foam is better than any other type of insulation at reducing air leakage.”
- “Fiberglass-insulated homes are the leakiest. … Bruce Harley, the Conservation Services Group’s technical director for residential energy services, … assembled airtightness data on Energy Star homesA U.S. Environmental Protection Agency (EPA) program to promote the construction of new homes that are at least 15% more energy-efficient than homes that minimally comply with the 2004 International Residential Code. Energy Star Home requirements vary by climate. (including single-family and multifamily homes) completed in 2004 in Massachusetts and Rhode Island. All of the homes were blower-door tested after completion. … Harley found that houses with walls insulated with spray polyurethane foam were significantly tighter than those houses with walls insulated with cellulose, and that houses with walls insulated with cellulose were significantly tighter than those insulated with fiberglass.”
Martin Holladay’s previous blog: “Study Shows That Expensive Windows Yield Meager Energy Returns.”
Sep 28, 2012 8:45 AM ET
Sep 28, 2012 9:05 AM ET
Sep 28, 2012 3:11 PM ET
Sep 28, 2012 6:01 PM ET
Sep 29, 2012 6:12 AM ET
Sep 30, 2012 1:26 PM ET
Oct 3, 2012 10:20 AM ET
Oct 3, 2012 7:29 PM ET
Oct 3, 2012 7:56 PM ET
Oct 3, 2012 8:47 PM ET
Oct 4, 2012 12:54 AM ET
Oct 4, 2012 6:22 AM ET
Oct 4, 2012 7:35 AM ET
Oct 4, 2012 7:39 AM ET
Oct 4, 2012 9:46 AM ET
Oct 4, 2012 9:58 AM ET
Oct 4, 2012 10:19 AM ET
Oct 4, 2012 10:27 AM ET
Oct 4, 2012 11:13 AM ET
Oct 4, 2012 11:48 AM ET
Oct 4, 2012 12:05 PM ET
Oct 4, 2012 6:09 PM ET
Oct 4, 2012 9:19 PM ET
Oct 8, 2012 8:14 PM ET