Page B1.2 . 04 April 2001                     
ArchitectureWeek - Building Department



When Bad Things Happen to Good Buildings


The Advent of Curtain Walls

Curtain walls proved to be the answer. The transition to curtain walls began in the late 1800s and continued until the middle of the 1900s. In the early part of the transition, walls were still massive, but they no longer supported the structure's floor loads. In the latter part of the transition, lightweight walls were "hung" on the structural frame.

It soon became apparent that the lack of mass and lack of inherent resistance to the effects of water exposure required new ways of managing water penetration and protecting vulnerable materials such as cavity drainage systems, internal waterproofing elements, and durable flashing materials. But they haven't evolved without a few bumps in the road.

Why is it that water problems are a bigger issue now for buildings than they have ever been? The weather hasn't changed significantly.

What has changed, however, is demonstrated by three recent trends: over-reliance on sealants to do the job of waterproofing; the push to make buildings air-tight to reduce energy costs; and the widespread use of moisture-sensitive materials in wall construction.

Instead of providing redundancies to serve as fail-safe protection against water penetration, designers and contractors began to rely solely on surface-sealed barrier walls.

Metal flashings that once were soldered are now lapped and "sealed." We ask more of sealant performance than we have had reason to expect. The result has been too many walls that leak immediately after construction.

The Problem of Waterproofing

Even small amounts of water penetration can have serious consequences. Improvement in air tightness can paradoxically create problems in moisture retention, because the lack of air flow slows drying.

Water that might have penetrated and then evaporated within a few days may now require weeks to dry, during which time a building might be exposed to additional rainstorms.

The net result is accumulation of moisture, prolonged high humidity, and even saturation within wall cavities, creating an environment ripe for rapid deterioration and mold growth.

Despite the relatively high moisture absorption of the transitional wall systems of the early 20th century, many of these walls survived well with minimal attention for many decades. The transitional walls were primarily constructed of stone and mortar that could remain wet without rapid deterioration.

Compare this with walls assembled today from soluble gypsum sheathing boards, corrosion-prone light-gauge metal studs, and insulating glass with water-degradable edge seals.

The transition from masonry load-bearing walls in many cases meant a transition to glass, now one of the most common cladding materials. The quintessentially brittle material, glass has introduced its own set of challenges. Glass is usually installed in metal framing systems, which means that thermal movement is inevitable: Aluminum frames subjected to a change in temperature can move about 2.5 times more than glass subjected to the same change. Breakage can result if the design does not accommodate this differential movement.

From Falling Water...

Bad things happen to some good, even great, buildings. Sometimes good buildings fail because their designers have pushed the limits of technology in order to create something new. As an example, Frank Lloyd Wright's Fallingwater is now undergoing structural repair to its famous cantilevers.

Sometimes good buildings fail because of a defect in a common material or component. Twelve years after its construction, the building at 303 Congress Street in Boston suddenly settled six inches (15 centimeters) due to errors in the production of the concrete mix that was used in its precast piles; the building was eventually demolished.

And sometimes buildings fail through a combination of these scenarios when familiar materials and technologies are used in new ways.

...To Falling Glass

Bostonians are familiar with one of the most famous examples: the John Hancock Tower, which was clad in more than an acre of plywood after its mirror-glass windows began to fracture in 1972 and 1973.

A "gag order" imposed on the parties to the resulting legal dispute prevented the release of the facts regarding the cause of the breakage, giving rise to many theories and myths, some of which exist to this day.

Initially, many design professionals thought the reason for the breakage lay in the fact that the tower swayed excessively in the wind. Although it was indeed swaying substantially, this was not the reason for the glass breakage.

Another hypothesis was that wind forces at "hot spots," which resulted from the rhomboid shape of the tower, caused overstressing of the glass. Substantial "hot spots" did exist, but only a small percentage of the glass was subject to anything near the load for which it had been designed.

Still another myth was that the windows broke because of the stress they endured from the settlement of the tower's foundation.

The Devil in the Detail

But in fact, extraordinary external forces and the building's structural design were not the cause of the failure. The problem actually lay in the insulating glass itself.

The insulating-glass units that made up the facade were fabricated with a thin lead-tape spacer to separate the two panes of glass. The tape was soldered to the glass after the edge of the glass was coated with a film of copper to make it more receptive to the solder.

This created a tenacious bond between the spacer and the glass, which constituted the product's greatest strength as well as the source of its demise.

The lead-tape seal insulating unit was the premier product of the time. It was expensive, but it performed very well with relatively small sheets of clear glass, the typical application in the 1940s through the early 1960s that is still performing well.

However, by the late '60s, large sheets of glass with tints and reflective coatings became popular. The large sizes and increased thermal loads associated with the tints and coatings caused substantial differential movement and increased stress along the glass-to-tape bond, and eventually, the bond began to separate.

The bond, however, was so strong in some areas that the tape ripped microscopically small pieces of glass from the glass surface. These sites concentrated stress from wind loads and ultimately proved catastrophic.

Recovering Face

Forty years ago, the window and curtain wall industry faced a predicament typical of emerging building technologies: rapid development and distribution of innovative products of varying performance and durability; product promotion based on unproven track records; and consequential disarray in the architectural profession responsible for specifying these products.

The industry began to clean its own house, providing specification guides and performance criteria, and developing objective product testing.

Not everything we did in years past was good. For example, the one-acre (4000-square-meter) copper roof on Grand Central Terminal, built in 1913, was doomed from the start due to lack of proper detailing to accommodate thermal movement and to avoid metal fatigue.

Not everything we do now is bad. But innovation is and should be relentless. And with innovation comes reduced predictability and increased risk. To meet the challenge that innovation presents, we must use the lessons of our history, coupled with sound technical fundamentals and a healthy dose of common sense.

Thomas A. Schwartz, PE, is president of Simpson Gumpertz & Heger Inc. in Arlington, Massachusetts. He was a principal investigator of the glass breakage at the John Hancock Tower and is a frequent lecturer and author on issues of building envelope performance.

This article originally appeared in the Winter 2000 issue of ArchitectureBoston.



ArchWeek Photo

Wind tunnel studies displayed on all four sides of the John Hancock Tower contradict the previously held notion that wind pressure simply increased with building height.
Image: Simpson Gumpertz & Heger Inc.

ArchWeek Photo

Glass and edge deflections associated with environmental loads on an insulating glass unit.
Image: Simpson Gumpertz & Heger Inc.

ArchWeek Photo

Cross section of "Bondermetic" insulating glass showing 1/4 inch (6.4-millimeter) thick lites of glass and 1/2 inch (12.7-millimeter) wide lead-tape spacer with soldered connection to glass.
Photo: Simpson Gumpertz & Heger Inc.

ArchWeek Photo

The copper roof on Grand Central Terminal was destined to fail because of lack of proper detailing to accommodate thermal movement and to avoid metal fatigue.
Photo: Simpson Gumpertz & Heger Inc.

ArchWeek Photo

John Hancock Tower with much of its curtain wall glazing temporarily replaced with plywood.
Photo: Michael Shellenbarger/ University of Oregon Library

ArchWeek Photo

Near ground level, most of the John Hancock Tower curtain wall was replaced with plywood.
Photo: Michael Shellenbarger/ University of Oregon Library


Click on thumbnail images
to view full-size pictures.

< Prev Page Next Page > Send this to a friend       Subscribe       Contribute       Advertise       Privacy       Comments
GREAT BUILDINGS   |   DISCUSSION   |   SCRAPBOOK   |   COMMUNITY   |   BOOKS   |   FREE 3D   |   ARTIFICE   |   SEARCH © 2001 Artifice, Inc. - All Rights Reserved