A Review of Research Conducted to Determine the Effec tiveness of Air Barrier Installations Laverne Dal gleish Air Barrier Association of America, Inc. 1030 15th Street N.W., Suite 460, Washington, DC 20005 Phone: 617-838-8819 • Fax: 866-956-5819 • E-mail: ldalgleish@airbarrier.org S y m p o s i u m o n B u i l d i n g E n v e l o p e T e c h n o l o g y • No v e m be r 2 0 1 3 D a l g l e i s h • 3 7 3 8 • Da l g l e i s h S y m p o s i u m o n B u i l d i n g E n v e l o p e T e c h n o l o g y • No v e m be r 2 0 1 3 ABSTR ACT Air barrier use goes beyond simply what the building code requires. The Air Barrier Association of America (ABAA), in conjunction with the U.S. Department of Energy’s Oak Ridge National Laboratory (ORNL), and Syracuse University (SU), is in the final phase of a five-year research project on the energy performance of buildings. This paper covers some of the results of that project, which shows how air barriers performed when installed in a test building and was monitored for a two-year period during which the walls were tested on a regular basis for air leakage. At the end of the monitoring period, the walls were dismantled and the air leakage paths were examined. Results showed the need for addressing details during installation to achieve optimum performance. A second round of testing was conducted, during which different types of air barrier were installed with three different air leakage rates. The energy flows were then monitored to answer the question of whether it makes any difference in how tight a building should be. The results of this research can be incorporated by the design professional into actual buildings. SPEAKER Laverne Dalgleish – Air Barrier Association of America, Inc. Laverne Dalgleish has over 30 years of experience in the building industry, with 20 of those years focused on energy-efficient building envelopes. He has been involved in standards development on an international and national level. Dalgleish has developed site quality assurance programs for air barrier, spray polyurethane foam, exterior insulation finish systems (EIFS), groundwater heat pumps, insulation, windows, and doors. He has managed a number of research projects for the industry, including research on air barrier performance, wall energy ratings, and wall drainage systems. S y m p o s i u m o n B u i l d i n g E n v e l o p e T e c h n o l o g y • No v e m be r 2 0 1 3 D a l g l e i s h • 3 9 Background In 2001, when the Air Barrier Association of America (ABAA) was established, documents had been circulating that suggested that air leakage in and out of a building could account for 30-40% of the energy used to heat and cool the building. At that time, this was a generally accepted principle. As the use of air barriers in buildings across the United States grew, proposals were made to the International Codes Council (ICC) to add requirements for buildings to include an air barrier system for energy conservation and energy efficiency. During these code hearings, the assertion of 30-40% energy savings was challenged. As ABAA dug into the history to determine where this figure came from, they found that although it was generally accepted, there was limited data to substantiate this claim. Simultaneously, the National Institute for Science and Technology (NIST) undertook an investigation of the impact of commercial building envelope airtightness on HVAC energy use. Its investigation showed that air leakage accounted for up to 40% in heating climates and up to 15% in cooling climates.1 In 2004, ABAA started a discussion with Oak Ridge National Laboratory (ORNL) regarding the development of a research project in which the energy impact of including an air barrier system in a building could be quantified. ABAA worked with Dr. André Desjarlais and Dr. Achilles Karagiozis of ORNL to develop a research plan to address this issue. This research plan included: • Project administration • Material property characterization • Subsystem and wall characterization • Laboratory wall testing • Advanced moisture engineering modeling • Exterior field testing of air barrier assemblies • Wall optimization • Information technology transfer Exterior field-testing of air barrier assemblies was included to correlate the performance in a laboratory setting with real-world performance. Exterior Field Testing Facility To be able to correlate the laboratory testing with real-world performance, it was envisioned that a test facility should be located where wall assemblies that included an air barrier could be installed and, ultimately, the performance of these walls tracked. The first step was to determine the preferred geographic location. The location needed to have the desired weather conditions necessary to induce loads on the air barrier in the wall assembly. It was determined that a northern climate would provide the best weather conditions for testing. Locations in the upper central and upper eastern regions of the U.S. were then considered. Another criterion for the location of the test facility was easy access for an individual to be able to check on the building operation from time to time in order to troubleshoot any problems. After discussions with the University of Illinois and Penn State did not result in an agreement, discussion began with Syracuse University (SU). It was quickly realized that none of the desired locations had a facility that could be used for testing purposes—the test facility would have to be constructed. The funds to construct the test facility were raised from participating industry manufacturer members. The test facility was constructed on the Syracuse University property so that A Review of Research Conducted to Determine the Effec tiveness of Air Barrier Installations Photo 1 – BEST building at Syracuse University. the facility could be used by the university or by the industry even after the project had finished. Once the location was secured, a design for the testing facility was developed and agreed upon. The design allowed for 34 wall panel openings approximately 4 x 9 ft. in size. The building itself had to be constructed as tightly as possible so that air leakage would be directed to the test specimen walls rather than to the building shell itself. The foundation was a continuous slab-on-grade foundation and, therefore, airtight. The walls and ceiling were sprayed with mediumdensity closed-cell spray polyurethane foam to assist in making the building shell airtight. As the spray foam was installed between the studs, additional sealing was installed in all the gaps and cracks where spray foam was not installed. Construction of the test facility began in the fall of 2008 and was completed in January 2009 (see Photo 1). A fan pressurization/depressurization test was then performed to determine the air leakage rate of the whole building and to confirm that the whole building was as airtight as possible. Once the test facility was completed, ORNL wired the building and installed a data transfer board so that all of the data regarding the performance of the individual wall assemblies could be collected. This data was then transmitted to ORNL for analysis. A complete weather station was installed on the test facility as well, so that all weather data could also be collected. Five miles of wiring was installed to connect the various sensors in the test specimen walls to the data transfer station. Plastic eave troughs were installed on the interior walls to act as raceways for the wiring. Exterior Field Testing of Air Barrier Assemblies In the spring of 2009, the test facility was ready for the test specimen walls to be installed. The original intent was to provide two openings to each participating manufacturer. The manufacturers would construct their own wall specimens and install their air barrier material. It was determined, however, that it would be more cost-effective for ABAA to arrange for all of the base walls to be constructed first in preparation for the manufacturers installing their air barriers. Doing this ensured that all the base walls would be exactly the same, and the only difference would be the specific manufacturer’s air barrier installation. The ABAA worked with ORNL to design three base walls. One was wood-framed, one was steel stud-framed, and the third was a concrete masonry unit (CMU) wall. A carpenter who was also an air barrier installer was hired to construct and install all the walls to ensure that the air barrier substrates were suitable for air barrier installation. During both the construction phase and the installation phase of each 4 0 • Da l g l e i s h S y m p o s i u m o n B u i l d i n g E n v e l o p e T e c h n o l o g y • No v e m be r 2 0 1 3 Photo 2 – Pressure-testing a wall specimen before installing. test specimen wall, ORNL was present to install the sensors within the wall and subsequently make all of the connections to the wiring running to the data transfer station. The first challenge was to design a test specimen wall that would be consistent with normal construction practices and could also be reproduced. The test specimen wall needed to include the normal gaps and cracks that would allow air to leak in and out. The test specimen wall identified in ASTM E2357, Standard Test Method for Determining Air Leakage of Air Barrier Assemblies, was used as the base for designing the test specimen wall. As the test specimen walls could only be 4 ft. wide, the window opening in the ASTM E2357 test specimen wall was dropped. To more accurately represent a typical residential wall, simulated floor and roof assemblies were added to the ASTM E2357 test specimen wall. Spacers were added between framing members to simulate the normal twisting and bowing that occurs as framing material dries out on the jobsite. Electrical outlets—both interior and exterior—were included, as they are a normal feature of construction and would also ensure that air leakage paths were incorporated into the test specimen wall. When the base test specimen walls were completed, they were tested for air leakage to confirm that the air leakage of each test specimen wall was not greater than the ability of the test equipment to measure the air leakage. Once the baseline test specimen walls were constructed, the manufacturers were invited to install their particular air barrier material on a baseline test specimen wall (Photo 2). They could choose the wood-framed, the steel stud-framed, or the CMU. Each manufacturer sent a representative of its choosing for the installation of its proprietary air barrier system. ABAA and ORNL worked with the manufacturers in preparing the test specimen S y m p o s i u m o n B u i l d i n g E n v e l o p e T e c h n o l o g y • No v e m be r 2 0 1 3 D a l g l e i s h • 4 1 Photo 3 – Wall specimens installed into openings with cladding being installed. Photo 4 – Sensors and wiring installed in a wall specimen. wall but did not interfere with the air barrier installation process. A wide range of individuals came to install the air barrier systems for the manufacturers, including local contractors, experts hired for this installation, technical people directly from the manufacturer, and finally, some who seemed to have limited installation experience. The length of time that these installers spent installing their air barriers on two 4- x 9-ft. test specimen walls varied from a couple of hours to four days. Some of the installers installed the air barrier material only and asked ABAA’s carpenter to install a cladding system, whereas other installers also installed the cladding system (e.g., veneer stone and EIFS). Once each test specimen wall was completed, it was installed into a specific opening in the building (Photo 3). ABAA and ORNL used a random selection system to determine on which wall the test specimen would be installed. Once installed, ORNL completed the final connections of the wiring and checked to ensure that all sensors were operating properly (Photo 4). The test specimens were then sealed into place using sealant foam around the perimeter between the test panel and the wall opening. Once all the test specimens were installed, all of the data that were being generated by the sensors were collected (Photo 5). This data collection period lasted one year to allow for the weather conditions from all four seasons to impact the test specimen walls’ performance. In addition to data being collected on moisture, temperature, and air pressures, SU conducted pressurization tests on the test specimens. Air Leakage Initial ResultsSyracuse University first conducted air leakage tests on the test specimen walls in August 2010. Each of the test specimen walls was expected to leak air when a pressure differential was applied to it, but the air leakage of the test specimen walls was expected to be very low. The initial results were somewhat surprising. The air leakage ranged from 0.41 to 1.95 L/(s·m2) at a pressure difference of 75 Pascals (Pa).2 In all of these cases, the values exceeded the maximum air leakage rate of 0.20 L/(s·m2) at a pressure difference of 75 Pa set by ABAA for air barrier assemblies. The obvious question became, “Why were the values so high when all of the materials were well-known and widely accepted air barrier materials?” See Photo 5. Syracuse University examined the test specimen walls to determine the reason for the high leakage rates. One of the reasons 4 2 • Da l g l e i s h S y m p o s i u m o n B u i l d i n g E n v e l o p e T e c h n o l o g y • No v e m be r 2 0 1 3 Photo 5 – Data acquisition board that relays the data to ORNL. Photo 6 – Pressure testing apparatus installed inside and outside. they identified was that the extraneous air leakage of the test apparatus had not yet been subtracted from the results. However, the whole test facility had been pressurized to match the specimen test pressure specifically to reduce the extraneous air leakage of the test apparatus. As SU dug deeper into where the air leakage paths were located, it became apparent that some of the air leakage was going around, rather than through, the test specimen walls. Sealant foam—specifically, low-density polyurethane foam applied from cans—had been used to seal between the test specimen walls and the test facility opening and was proving to be an ineffective seal. Sealant foam, when installed, generally results in large cells in the material. The low density, combined with irregular cell structure, produces a material that has a higher air permeance rate than anticipated. See Photo 6. The university then drafted a plan to reseal the test specimen walls to the test facility wall using caulking. In addition, polyethylene film was installed between the test frame and the edge of the test specimen wall to potentially eliminate any air leakage between the test specimen wall and the test facility wall itself. The test specimen walls were then retested. The air leakage rates for the modified test specimen walls now ranged from 0.0285 to 0.0002 L/(s·m2) at a pressure difference of 75 Pa.3 (See Table 1.) Forensic Review of Test Specimen Walls Even though the test specimen walls’ air leakage rates were now within the expected range, a forensic examination of the actual air leakage paths of the wall specimens was conducted to determine the specific air leakage paths of each test specimen wall. The test specimen walls were deconstructed and examined at the end of the year of data collection. Some of the more excessive air leakage paths identified were (Photos 7 through 10): • Voids in the sealant foam • Joints in the framing that allowed air to leak around the air barrier material • Electrical outlets • Stud/plate interface • Areas around tubing used to measure pressures in test specimen wall • Penetrations not properly detailed • Transition membrane not fully S y m p o s i u m o n B u i l d i n g E n v e l o p e T e c h n o l o g y • No v e m be r 2 0 1 3 D a l g l e i s h • 4 3 Table 1 – Air leakage test results before and after extraneous leakage reduced. Wall Type Leakage (L/min) Leakage (L/min) % Flow (New Procedure) (August Testing) Decrease 10 Pa 25 Pa 50 Pa 75 Pa 50 Pa 75 Pa 75 Pa Wall AB13 47.3 96.35 165.04 225.11 256.9 311.6 27.76 NA 7.96 14.56 22.97 30 106.8 116.2 74.18 Wall AB27 7.99 13.6 20.33 25.75 120.6 137.6 81.29 Wall AB16 11.47 20.47 31.72 40.99 110.7 126.2 67.52 Wall AB19 57.24 100.54 153.98 197.58 266.8 320.5 38.35 Wall AB09 1.75 2.99 5.29 7.32 92.8 98.6 92.58 Wall AB23 21.25 42.15 70.2 92.92 203.4 252.5 63.20 Wall AB22 ** ** ** ** ** ** Wall AB06 15 32 47 58 200.4 229.1 74.68 Wall AB24 27.47 54.64 92.96 126.04 237.9 258.3 51.20 Wall NA 231.84 ** ** ** ** ** Wall AB20 0.52 1.15 2 2.73 108.1 124 97.80 Wall AB21 81.82 130.75 186.4 229.38 293.1 331.7 30.85 Wall AB02 34.01 58.56 88.33 112.35 199.8 251.3 55.29 Wall AB25 75.09 123.55 180.05 224.43 228.9 264.8 15.25 Wall AB13 165 246 ** ** 381.1 457.9 Wall AB23 36.61 61.07 89.93 112.77 263.1 334.2 66.26 Wall AB10 19.08 31.31 45.5 56.7 157.7 202.7 72.03 Wall AB16 20.35 33.31 48.36 60.15 104.9 131 54.08 Wall AB07 20.87 35.4 52.9 66.99 232.5 252 73.42 NA 37.12 58.55 82.65 101.12 ** ** Wall AB26 6.74 9.52 12.35 14.39 148.1 156.8 90.82 Wall AB04 22.29 37.91 56.65 71.65 74.2 96.1 25.44 Wall AB08 53.38 99.25 158.66 208.76 217.1 274.3 23.89 Wall AB08 4.05 5.9 7.84 9.26 175.9 224.3 95.87 Wall AB11 48.19 81.91 122.35 154.72 119.7 159.7 3.12 Wall AB13 0.246 0.501 0.858 1.171 1.336 1.620 27.757 NA 0.041 0.076 0.119 0.156 0.555 0.604 74.182 Wall AB27 0.042 0.071 0.106 0.134 0.627 0.716 81.286 Wall AB16 0.060 0.106 0.165 0.213 0.576 0.656 67.520 Wall AB19 0.298 0.523 0.801 1.027 1.387 1.667 38.353 Wall AB09 0.009 0.016 0.028 0.038 0.483 0.513 92.576 Wall AB23 0.111 0.219 0.363 0.483 1.058 1.313 63.200 Wall AB22 ** ** ** ** ** ** Wall AB06 0.078 0.166 0.244 0.302 1.042 1.191 74.684 Wall AB24 0.143 0.284 0.483 0.655 1.237 1.343 51.204 NA 1.206 ** ** ** ** ** Wall AB20 0.003 0.006 0.010 0.014 0.562 0.645 97.798 Wall AB21 0.425 0.680 0.969 1.193 1.524 1.725 30.847 Wall AB02 0.177 0.305 0.459 0.584 1.039 1.307 55.292 Wall AB25 0.390 0.642 0.936 1.167 1.190 1.377 15.245 Wall AB13 0.858 1.279 ** ** 1.982 2.381 Wall AB23 0.190 0.318 0.468 0.586 1.368 1.738 66.257 Wall AB10 0.099 0.163 0.237 0.295 0.820 1.054 72.028 Wall AB16 0.106 0.173 0.251 0.313 0.546 0.681 54.084 Wall AB07 0.109 0.184 0.275 0.348 1.209 1.310 73.417 NA 0.193 0.304 0.403 0.526 ** ** Wall AB26 0.035 0.050 0.064 0.075 0.770 0.815 90.823 Wall AB04 0.116 0.197 0.295 0.373 0.386 0.500 25.442 Wall AB08 0.278 0.516 0.825 1.086 1.129 1.426 23.894 Wall AB11 0.251 0.426 0.636 0.805 0.622 0.830 3.118 adhering to substrate • Holes in fluid-applied materials • Blisters in fluid-applied material • Shrinkage of spray polyurethane foam • Joints in substrate Conclusions Many conclusions can be noted from the data collection and analysis that took place as a result of this research project. First, the baseline walls used were typical of the construction practices commonly seen across North America; therefore, it should be expected that the air barriers installed in the test specimen walls would provide an airtight seal for these typical walls in any building. Second, many of the installation details currently available do not address the air sealing of electrical boxes. (Note: Airtight electrical boxes that are available in Canada proved to be impossible to obtain locally.) Third, attention to details will greatly reduce the air leakage of a building assembly, often by simply using common materials. In addition to these conclusions, it was also noted that when transition mem- 4 4 • Da l g l e i s h S y m p o s i u m o n B u i l d i n g E n v e l o p e T e c h n o l o g y • No v e m be r 2 0 1 3 Photo 8 – Holes in fluid-applied membranes. Photo 7 – Adhesion of deficiencies. Photo 9 – Reason why sealant foam does not seal. Photo 10 – Reason why selfadhered did not adhere. branes are used in conjunction with spray polyurethane foam, these membranes have to be properly adhered to the substrate; otherwise, they will be pulled off and prove ineffective. Second, all fluid-applied membranes need to be free of any holes caused by installation practices. Third, sealant foam, which is used widely to reduce air leakage, does not provide the performance of an air barrier material. One general conclusion was that all of the air barrier materials themselves generally performed as specified by the manufacturer. However, air barrier system performance was directly related to the knowledge, skills, and abilities of the installer. From this study, it was determined that all identified air leakage paths were attributed to installation issues related to the installers’ various degrees of skill and attention to detail. This was evident in the measured air leakage performance of the various test specimen walls. It was clear that simply installing an air barrier material is not enough to produce the expected energy savings. The air barrier material needs to be combined with air barrier components and air barrier accessories to produce an air barrier assembly that is relatively airtight. These air barrier assemblies, all of which are building assemblies, need to be combined to produce the air barrier system, which is the whole building. The designer needs to provide the details of the air barrier system, and the installer on site needs to turn these details into an effectively installed air barrier system. “The devil is in the details” adage certainly applies here. In the case of the installed performance of air barrier assemblies, it can be said, “The devil is in the detailing of the air barrier assembly.” When it came to the installation of the air barrier assemblies, it was expected that manufacturers would provide the best installer available, as their material was part of an international research project. The “best-ofthe- best” installation resulted in acceptable to high performance of the test specimen’s walls (as far as air leakage). The technical knowledge of the way to install the product properly, updated with the findings identified by this part of the research project, needs to be transferred to the industry and the workers who install air barriers on a daily basis. The Air Barrier Association of America is continually updating its installer training and certification program to transfer this information out to the projects in the field. The installer training and certification program is part of the Site Quality Assurance Program that also includes contractor accreditation and site audits. Reallife situations are fed back to the ABAA office, which uses this feedback to update and enhance the program. Footnotes 1. Steven J. Emmerich, Building and Fire Research Laboratory; Timothy P. McDowell, TESS, Inc.; Wagdy Anis, Shepley Bulfinch Richardson and Abbott. NISTIR 7238, “Investigation of the Impact of Commercial Building Envelope Airtightness on HVAC Energy Use.” Prepared for U.S. Department of Energy — Office of Building Technologies. June 2005. 2. Thomas Thorsell, Korbaga Funtu Woldidan, Denis Pradhan, and Jensen Zhang (PI). “Air Leakage Measurements of 25 Wall Assemblies With Different Types of Air Barriers.” Syracuse University Building Energy and Environmental Systems Laboratory (BEESL). Progress Report #1 to ABAA. August 25, 2010. 3. Denis Pradhan, Jensen Zhang (PI), Thomas Thorsell, Robin Mocarski. “Air Leakage Measurements of 25 Wall Assemblies With Different Types of Air Barriers.” Syracuse University Building Energy and Environmental Systems Laboratory (BEESL) Progress Report #3 to ABAA. January 27, 2011. S y m p o s i u m o n B u i l d i n g E n v e l o p e T e c h n o l o g y • No v e m be r 2 0 1 3 D a l g l e i s h • 4 5
