Prior to venturing into private practice in 1979, Dr. Ting spent 10 years in building product research at H.H. Robertson Co. He is a leading consultant in metal building product applications and consults with major metal building product manufacturers throughout the world.

Dr. Ting can be reached by writing to his company at 505A McKnight Park Drive, Pittsburgh, PA 15237, or calling (412) 367-2808, or faxing
(412) 367-7990.

When the weather turned foul and day turned to night, prehistoric man sought shelter in caves. As civilization advanced, man began to build. Since that time, the quest has been ongoing to improve the methods by which buildings are constructed and develop products capable of outperforming those that have come before. This fact can be well illustrated by looking at the evolution of exterior wall designs for multistory buildings.

Initially, the walls of multistory buildings were load bearing. Required to support the floors and roof, they were constructed of strong, heavy materials such as stone, brick and concrete block. As long as an adequate base was provided, these types of structures were sufficiently able to resist shear force and overturning moments produced by the wind.

Eventually, the original bearing wall design gave way to a more thermally-efficient, double-layer version. While offering no real improvement with regard to the ability of the walls to prevent water penetration, leaks were generally limited to the outer layer, thus masking any indications of their presence from inside the building.

The advent of reinforced concrete and steel framing prompted a dramatic change in the way exterior walls were designed. No longer required to play a structural role, load-bearing exterior walls gave way to a new product, hanging walls. Supported by the building's floors via attachment along their edges, these new curtain wall systems enabled designers to effectively meet their clients' demands for more usable floor space and shorter erection time frames.

Subsequent improvements in curtain wall systems focused on making the products thinner and lighter without sacrificing their ability to meet wind load requirements. These efforts resulted in the development of aluminum extrusion curtain wall systems. With these systems, individual panel faces are supported at their perimeters by aluminum extrusions.

Typical curtain wall structures consist of multiple wall panels joined together to form vertical and horizontal panel joints. Popular facing materials include glass, 1/8"- to 3/16"-thick solid aluminum plate, composite panel material and 1"- to 2"-thick stone panels. Naturally, it is at the joints where problems with water infiltration can occur. The control of such problems has remained a focus of the building industry from the advent of those initial curtain wall systems through the development of subsequent generations of products.

An analysis of curtain wall water leakage problems has shown them to generally be the result of three simultaneously existing elements: an imperfect seal along the wall panel joints, positive differential air pressure (i.e. exterior air pressure being higher than interior air pressure) and water. When all three elements are present, water running along the panel surfaces is pushed through the poorly sealed joint areas by the pressure differential created by the positive wind. Unfortunately, problems related to the sealing of aluminum extrusion curtain wall systems often go undetected in their early stages. When the discovery of a leak or leaks is made, significant damage to wall insulation and/or corrosion of the curtain wall supporting system may have already occurred and corrections may not be easily made. To provide further frustration, it is not out of the realm of possibility that water leakage will continue to occur at different locations or recur at the same points even after attempts to make repairs have been made.

Because two of the three elements required for leaks to occur happen naturally, designers of curtain wall products have always focused their attention toward improving panel system seals as the most effective means of preventing leaks. A comparison of today's systems to those of years ago shows great strides in this area have been made.

The earliest curtain wall panel systems were based on the perfect seal principle. As its name implies, the principle of this generation of seal design is to make a perfect seal. Initial research efforts were directed toward maximizing the compatibility between the sealant and the component material, providing durability against thermal movements, improving the methods of application and retaining elasticity. Research efforts in this area led to the development of silicone caulk.

As Figure 1 illustrates, the perfect seal design requires the use of silicone to perfectly seal all exposed panel joints. Experience has taught that when problems with this type of sealing method occur, they are generally linked to field workmanship and/or sealant degradation. Unfortunately, little can be done to correct these inherent shortcomings because consistent, perfect field workmanship is impossible to attain and the material degradation of sealant is ultimately inevitable.

The primary advantage of using the perfect seal method is that it affords workmen the opportunity to complete any type of required repair work from the outside without damaging the finished interior wall or disrupting inside activities. For that reason, and because water leakage problems have persisted with subsequent generations of curtain wall designs anyway, the perfect seal method continues to be used today. In an effort to limit damage that can be caused by undetected leaks, some building owners have taken to scheduling costly, periodic curtain wall resealing operations as a part of their routine building maintenance programs.

A change in philosophy preceded the introduction of a second generation of curtain wall systems. Rather than preventing leakage, designers decided instead to accept and control it. The ultimate goal of curtain wall systems utilizing the controlled leakage design principle is to ensure penetrating water does not result in interior damage. Typical panel joint details of the second generation seal are shown in Figures 2 and 3. The main feature of these types of systems is their interior gutters.

With curtain wall systems employing the controlled leakage design principle, the vertical and horizontal panel joints are usually formed by interlocking panel perimeter extrusions with gasket sealing material in the joints. As an integral part of the perimeter extrusion, a horizontal gutter is provided behind the sealed horizontal joint. Due to the fact the gutter is behind the sealed joint, the gutter space is subject to interior air pressure and can thus be called an "interior air gutter." Any water that infiltrates through the sealed joint is directed into, and temporarily stored in, the interior air gutter. The accumulated water in the gutter is drained to the outside via exposed drainage holes after the wind diminishes.

One of the advantages of the controlled leakage design principle is that it significantly reduces the number of locations for possible water leakage. However, this type of system does have drawbacks. First, under the effect of differential air pressure, exterior water can be pushed into the interior air gutter through the exposed drainage holes, causing a large quantity of water infiltration. The water head in the gutter is used to balance the differential air pressure. Therefore, the higher the wind forces, the higher the required gutter height. To prevent water leakage through the panel securing screw holes, the panel securing screws must be located above the expected water level in the gutter. This fastening location creates a lever arm on the fastener pull-out force under negative wind conditions. This can lead to serious structural problems in high wind areas such as hurricane zones.

A second problem with this type of design is that the water drainage action is only possible after the wind has diminished. This delayed water drainage behavior can cause water stains on the panel surface below the drainage holes. Also, any necessary repairs cannot be made from the outside. Thus, interior wall damage and the disruption of interior operations are inevitable in the event repair work is required. In addition, the design relies on shop-applied perfect seals between perimeter extrusions and the facing panel material, and field-applied perfect seals between the vertical joint gutter splices and gutters they connect. Even if the shop seals were all perfect as intended, the problem of the sealant's eventual degradation would remain.

A variation of the original second generation design effectively addresses the problems of water infiltration through the drainage holes and the associated staining. In this design, the water drainage mechanism features concealed drainage holes that direct water into a concealed, pressure-equalized water tunnel that discharges into concealed vertical grooves at both ends along the vertical panel joints. This type of system was successfully employed on a terminal and corridors for American Airlines at the Raleigh-Durham Airport in North Carolina.

The third generation design is known as the unitized curtain wall system. It has been the most commonly used type in recent years. It has three major design features which distinguish it from previous systems. First, multiple panels are pre-assembled in the shop allowing most of the locations requiring perfect seals to be installed in an environment where quality control is more reliable. Second, the ends of the horizontal gutters are dammed and sealed to the vertical split mullions, eliminating the need for field installation of the gutter splices. Third, in an attempt to pressure equalize the wall cavities, the exterior panel joints are not sealed.

The wall cavities linking to the exterior air are segmented at the intersections of the vertical and horizontal extrusions in this design. When exterior water running along the exterior wall surfaces blankets over the panel joints, the pressure equalization mechanism for the isolated wall cavity behind the panel joint ceases to function. Therefore, this type of system cannot be claimed to be a pressure-equalized system.

Typical details of the third generation are shown in Figures 4 and 5. The intersection points of vertical and horizontal extrusions within each panelized unit are perfect sealed in the shop. The horizontal and vertical joints created by the attachment of the panelized units to the building itself are then perfect sealed in the field to complete the installation.

To limit the potential for damage to the sealant lines within the panel field from the time of fabrication through the life of the curtain wall system, the framing employed for the construction of the panelized units is rigid. The anchoring system utilized for attachment permits the units to maintain free rigid body movement (except: the effect of dead weight) in case of relative inter-floor displacements either vertically due to floor load or horizontally due to wind or seismic load. In the case of relative horizontal inter-floor displacement, the rigid body movement of each individual unit will produce large relative displacement with both vertical and horizontal components at the four-comer intersections that would ultimately damage the field-applied seals at those locations. The third generation design is a further improvement over the second generation design by increasing the reliability of the sealant lines. However, the water leakage problem has not been solved.

The principle of the fourth generation design is based on eliminating the effects of the natural forces of wind and rain This design was only recently invented and is based on the premise that all seal lines are assumed to be imperfect. Therefore, concerns about workmanship both in the shop and in the field can be put to rest. The same is true of concerns about sealant degradation.

end of part 1


published in Metal Architecture, February, 1997

The primary elements of the fourth generation curtain wall system design harken back to the more sophisticated version of the second generation design discussed last month (January 1997 Metal Architecture). Typical details of the fourth generation design are shown in Figures 6 to 9. In the earlier system, a pressure-equalized outer air loop was created to accomplish a concealed water drainage mechanism. The evolution of that concept was carried out in two steps. The first step was to expand the depth of the outer air loop to the entire depth of the curtain wall panel. Thus, the gutter space is covered by the outer air loop which is pressure equalized to the outside air. The vertical air seal is provided at the contacting surface between the wall panel and the supporting mullion. The horizontal air seal between the panels is provided behind the gutter space and near the inner surface of the wall panel. As a result of the design arrangements, the gutter in this case is considered an exterior air gutter because the gutter space is pressure equalized to the exterior air. Since the gutter space is pressure equalized, there is no build-up of water in the gutter and the drainage mechanism is instantaneous. Due to pressure equalization, there is no differential air pressure in the wall panel joint cavities and the effect is equivalent to having no wind. The air seal lines having differential pressure are shielded from rain water by the instantaneous water drainage system. Therefore, water cannot reach the air seal lines and the effect is equivalent to having no rain.

The second step added a pressure-equalized inner air loop to protect the air seal between each facing panel and panel perimeter extrusions. This inner air loop is pressure equalized by linkage to the outer air loop through air holes located above the exterior air gutter. This inner air loop is formed within the extrusion profile and miter-matched at the corners so that the air seal looping the perimeter extrusions is completely isolated by the inner air loop. The outer seal along the inner air loop is subjected to the exterior running water. However, because it is pressure-equalized, water will not infiltrate into the inner loop even if the seal is imperfect. Therefore, the inner air loop is a dry air loop that is equivalent to having no rain. Unlike the isolation of the panel joint cavities in the third generation design, all panel joint cavities in the fourth generation design are totally interconnected to ensure the effectiveness of pressure equalization in all conditions.

It must be emphasized that the fourth generation design tolerates imperfect seals anywhere in the system without producing water leakage problems. It separates watertightness performance from airtightness performance and eliminates workmanship and sealant material degradation as factors in curtain wall performance. The fourth generation design has been successfully utilized on the Butler County Airport terminal in Butler County, PA.

The fourth generation design could lead to major changes with regard to testing for the performance of watertightness in curtain wall projects. Currently the water test procedures followed by the industry include the ASTM E-331 static test and AAMA 501.1 dynamic test. In the static ASTM E-331 test, the required differential air pressure is normally set at 20% of the design's maximum positive wind pressure for a duration of 15 minutes. In the dynamic AAMA 501.1 test, the wind pressure produced by the dynamic wind is normally equivalent to 12 psf.

Due to the fact the designs of the first, second and third generations fail to separate watertightness from airtightness, the water test protocol normally includes the following steps: (1) static ASTM E-331 test; (2) dynamic AAMA 501.1 test; (3) structural tests including wind load and inter-floor displacements; and (4) static ASTM E-331 test. The reason for conducting a water test after the structural tests is to ensure that the watertightness performance is not affected by the structural movements. Another problem is that in the designs of the prior three generations, perfect seals are always required, and the locations requiring perfect seals vary with the details for fulfilling the building shape created by the architect. Therefore, all major curtain wall projects require a full-scale mock-up water test implementing typical details.

While logic would dictate any curtain wall water tests be conducted prior to awarding bids, cost prevents it. If every bidder on a given job was required to complete a water test prior to bidding, the bid prices themselves would rise significantly to cover costs related to making dies, samples and mock-ups, and having the tests performed. This economic fact of life prompts most building owners to reluctantly agree to allow bidders to forego testing until after they get the job. This practice gives rise to a very serious question. What if the design fails the water test?

Once the bid is awarded, the building owner has little recourse but to give the curtain wall supplier as much leeway as possible to pass the required test. For example, it is usually allowed to conduct a pre-test, adding caulk as necessary until passage is achieved. While that in itself is not so bad, the same measures can be employed to pass the certifying test. If the mock-up requires repairs to pass the test, what realistic expectations can there be for watertightness of the actual curtain wall system in real world conditions?

To a scientist, tests conducted in such a manner cannot be used as a measure of the watertightness of real buildings. They can only be used to determine the weak points of a given system for the purpose of strengthening quality controls either in the field or in the shop. But even if repairs weren't allowed as a means of achieving passage, there are other issues that demand attention. For instance, is the test duration of 15 minutes really adequate for representing all areas of different weather conditions in the world? Does 20% of the maximum positive wind load adequately represent all areas in the world in storm conditions? Is the short-term structural test adequate to represent the long-term cyclic loadings of nature including wind forces, temperature changes and earthquakes? Finally, what is the best way of factoring the sealant material degradation problem into the test protocol? In consideration of these questions, the following water test protocol using ASTM E-331 would be more appropriate.

Hurricane areas:
Test pressure = 80% of maximum design positive wind load.
Test duration = 60 minutes.

2.Pre-test and repairs using caulking:
Not permitted. Changing the system design to re-test is permissible.

3.Simulated Long-Term Performance Test:
Determine the locations of imperfect seals to be introduced and selected by the architect.

Upon successful testing in accordance to the above standards, repeat the test and systematically introduce the imperfect seals by destroying the seals in succession during the test until water leakage occurs or the test duration ends. The test is passed if no water leakage occurs throughout the test duration. If water leakage is observed during the test, immediately stop the water test and, without changing the conditions, conduct the ASTM E-283 air infiltration test. If the air leakage rate is more than twice that of the conditions prior to any water test,this final test is considered as passed. This final test collectively represents the effects of long-term structural movements, the workmanship problem, and the sealant material degradation problem. Therefore, structural testing is not required for judging the watertightness performance.

The fourth generation design has been successfully tested in laboratories in both the United States and Taiwan with the following concluding results:

The watertightness performance of the system is not affected by the intensity of the differential air pressure applied during the test (within the structural safety of the system). The water test duration is irrelevant to the watertightness performance of the system (continuously tested for 60 minutes with no sign of water accumulation in the gutter).The system passed the Simulated Long Term Test (destroying sealant lines failed to produce water leakage).

In summary, if a successful laboratory water test were to be used as a measure of the long-term watertightness performance of a building, the test mock-up must be able to tolerate imperfect seals anywhere in the system without producing water leakage. To be able to tolerate imperfect seals anywhere in the system, the fourth generation design follows a simple design principle of having complete pressure-equalized inner and outer air loops which can be seen in the cross-sectional details. Therefore, the watertightness performance of the fourth generation design can be judged by examining the cross-sectional details without having to conduct the water infiltration tests. With additional test verifications in the future, it is believed that the costly full-scale water infiltration test can eventually be eliminated.