Doug Coe, a normally confident engineering manager for the new east span of the San Francisco-Oakland Bay Bridge, walked into the nearby Oakland project office looking as if he were fighting back tears. Joel Sayre, then a bridge spokesman who worked there, remembers tensing in alarm.
Engineers had discovered an alarming corrosion problem with the "post-tension" tendons, and were pumping gallons of rusty water from the ducts that held them, Sayre said Coe told him. "Oh my god," he recalled Coe saying that afternoon in late spring of 2006. "What are we going to do?"
Coe, whom the California Department of Transportation would not permit to answer questions, was talking about thousands of steel tendons in the skyway section of the new span the elevated roadway that runs from the Oakland footing to the suspension bridge near Yerba Buena Island. Ducts containing the tendons, crucial to structural integrity, had been left unsealed. Rainfall and water used to cure concrete, tainted by construction debris exposed to salty bay mist, had entered many of them.
The bridge was billions of dollars over budget and years behind schedule. Rusty water meant tendons had corroded. Sayre said Coe, lead skyway engineer, described a potential nightmare that could stop construction cold.
Such concerns led Caltrans to examine hundreds of tendons. The agency found little significant corrosion a false alarm. It laid out that conclusion in a report and moved on.
Because of the concerns of Sayre and others, The Bee compared that Caltrans study against about 115,000 pages of construction and inspection records and found the conclusions were based on wrong information. The records show that the agency misstated in its report the extent of water contamination and its own inspection efforts. Conclusions that corrosion caused no harm were based partly on underestimates about how long tendons were left exposed and vulnerable, and on suspect testing methods.
In March, anchor bolts meant to secure seismic equipment on the new bridge broke an event attributed partly to water-induced corrosion. Caltrans similarly failed to take basic precautions to protect many skyway tendons from water even after the agency completed its study. Those lapses and others, said nine leading experts in the corrosion of bridge tendons, introduced uncertainty about the durability of the skyway.
Experts said that while a total collapse seemed unlikely, if Caltrans miscalculated corrosion estimates a major quake could cripple sections of the skyway. The span's construction violated universally accepted techniques meant to prevent corrosion, they said, and could result in costly inspection and maintenance headaches nearly unheard of for similar modern bridges.
Among The Bee's findings:
Beginning more than two years before Coe's discovery, inspectors frequently warned about water leaks and corrosion. Bridge spokesman Andrew Gordon could not say why officials failed to address the problems without significant research.
Experts blamed water problems on design or construction errors. Leaks of grout a cement-based filler that normally prevents or halts corrosion between hundreds of ducts forced long construction delays that left tendons exposed. They said the errors made further, unseen corrosion of tendons likely.
Caltrans used the wrong tests for corrosion, resulting in "essentially useless" findings, said UC Berkeley engineering professor Thomas Devine, an internationally known authority on corrosion-caused cracking in metals. He called the agency's research "woefully inadequate" and "meaningless" for detecting "environmentally assisted cracking," which can worsen as tendons fatigue under stress, and can ultimately cause breaks.
University of South Florida professor Alberto A. Sagüés, sole independent evaluator of the Caltrans study, gave it a vote of confidence based on faulty assumptions provided by Caltrans. Sagüés declined to comment.
Caltrans declined Bee requests to interview toll bridge manager Tony Anziano and chief Bay Bridge engineer Brian Maroney about the tendons.
In written responses to questions, Gordon said "Corrosion has been an issue since the first steel bridge was built, and what we learned from this particular challenge six years ago is that we're never going to be done fighting the battle against corrosion in the San Francisco Bay."
Neil Hawkins, engineering professor emeritus at the University of Illinois, reviewed the Caltrans study and other documents for The Bee. He said in an email that the agency probably had faced a grim dilemma when the problems came to light.
"The alternatives may have been proceeding as they have done or tearing down completely the existing construction," Hawkins said. "The issue then becomes both technical, economic and political. Who bears the blame and why did this irregularity occur?"
Duct tape defense
On a cool, overcast day in November 2004, during a routine check of the bridge deck, Caltrans inspector Laura Rubalcaba stumbled on a surprise: Water was leaking through damaged vents that led into tendon ducts.
Those galvanized steel ducts run inside the concrete segments that make up much of the bridge. Once tendons are inserted into the ducts and anchored at one end, hydraulic jacks stretch or "stress" the tendons, compressing segments to boost strength. Builders then fill ducts with grout to protect the tendons against corrosion.
Such methods have been commonplace, worldwide, since the 1950s. Until recent decades, water leaks that caused tendons to corrode also were common. Corroded tendons can crack and rupture, causing maintenance headaches or in extreme cases, bridge failures. In 1992, after some structures collapsed, United Kingdom authorities declared a four-year moratorium on post-tensioned bridge construction and a longer moratorium for using the technique on bridges that, like the skyway, are built from precast segments. In 1999 and 2000, major Florida bridges failed due to tendon corrosion.
Strict federal and state rules were developed to ensure that salt, water and construction debris can't enter ducts and corrode tendons. Those rules and the skyway contract required grouting within 10 days of tendon installation; 30 days if the builder blows a rust-inhibiting powder into the ducts.
Rubalcaba, like dozens of other inspectors, policed contractors who built the bridge. Where she stood, according to construction diaries, many tendons in ungrouted ducts had been installed more than two months earlier.
"(T)he top of the grout injection/vent hoses were not sealed against the rain," she wrote in her daily inspection diary, referring to vents used to insert grout or let air escape during grouting.
"I found many instances where it was obviuos (sic) that rain water was already in the ducts with the stresssed (sic) tendons," she wrote. Caltrans did not suspend work, as required by the contract. Instead, in the style of an overgrown home-improvement project, "I duct taped over the tops of the tubes myself," she noted.
Duct tape, America's beloved fix-it solution, thereby became one Caltrans inspector's first line of defense against a potentially major problem on an iconic $6.4 billion bridge. That day occurred near the genesis of grouting delays that left some tendons unprotected for nearly 17 months.
Six months later, Rubalcaba reported damaged grout vents along a third-of-a-mile stretch of the bridge roadway. Other inspectors made dozens of similar warnings, going back years before skyway chief Coe's revelation.
Inspector Art Pannu noted water entering at a joint between concrete sections.
Mehdi Bassiri recorded that "almost 25-30 feet of tendon fell into water," apparently referring to bay water. "They let it dry up and pulled it into the duct two kinks could be seen." Saltwater and kinked strands often lead to corrosion and rupture.
Supervisors signed off on each warning. Each described a violation of the contract, yet work was not stopped. Water continued to flow into ducts. When Coe flagged the problem, no one knew how many had been contaminated or how many tendons had corroded.
They would soon learn the answer: thousands. Vents had been leaking along the entire length of the bridge.
Devil in the data
Caltrans officials began a lengthy study of the problem. While their analysis found that the corrosion was minor, leading experts have challenged the agency's methodology and conclusions.
Caltrans first examined 18 tendons by threading a borescope a camera attached to a flexible tube inside ducts, according to the agency's report. A few "worst case" examples that had been left ungrouted for 14 months were encrusted with rust. Several strands twisted wire cables that are bundled into tendons were extracted for closer scrutiny.
Agency experts judged those strands "moderately" corroded. They showed a buildup of rust that suggested the formation of pits. These are indentations that form in corroding steel and precede cracking, the rupture of wires, then strands and finally, entire tendons. If tendons snap, a bridge might become unstable under the everyday strain of thousands of cars and trucks, or in a large quake.
Eventually, Caltrans viewed about 1,600 of the more than 5,600 tendons in the bridge. The agency said in its report that this represented almost 80 percent of those not yet grouted and therefore visible using borescope cameras. Nearly half displayed minor rust, but just a fraction suffered moderate corrosion and researchers found little salt.
A few samples extracted for lab tests by Caltrans and the Federal Highway Administration revealed shallow pits and rare cracks.
Sagüés, the professor hired to review the report's approach, offered a qualified blessing. "(B)ased on the information available to date," he wrote, "the corrosion damage in the tendons with delayed grouting appears to have had little impact on integrity."
Long-term effects, Sagüés said, would be similarly slight "if subsequent grouting was performed promptly and properly," and ducts dried prior to grouting as required by the contract.
Yet the construction record shows that Caltrans often ignored that advice.
Many tendons examined by Caltrans remained ungrouted in wet ducts months after the agency completed its examinations. Nor did Caltrans use special grout designed for reducing corrosion risks in wet ducts.
Many vents at the road surface were not properly sealed until at least December 2006, six months after the problem was discovered. Five to nine inches of rain fell locally during that period, depending on the measurement location, so water and debris from the surface flowed into ducts through the unsealed vents. Caltrans also failed to clear water from the ducts prior to grouting in more than 250 cases after the June 2006 discovery of the problem, inspection diaries noted.
Sagüés cautioned in his assessment for Caltrans that "uncertainty remains" because "direct inspection was not possible in most tendons in the Eastbound bridge," referring to one of the span's two parallel roadways.
The report said that Caltrans almost exclusively tested tendons in the westbound portion of the skyway structures because nearly all tendons in the eastbound span had been encased in grout by the time the problem was found.
Because exposures to water for tendons in each span were nearly identical, Caltrans concluded, any corrosion in the untested eastbound span would be no worse than that found in the westbound span, which had been given a clean bill of health.
Yet the record shows that about 1,000 eastbound-span tendons still sat in ungrouted ducts when Caltrans began to address the problem in June 2006. Gordon said in his written response that Caltrans could not explain the failure to examine those tendons without conducting additional research.
Also contrary to the agency's reporting, eastbound tendons were left ungrouted 73 days longer, on average, than their westbound counterparts, and had many more cases of extreme exposure. Nearly 90 percent of the more than 500 tendons grouted at least a year after installation are in the eastbound span.
Asked if further examinations should be made of the eastbound tendons, Gordon wrote, "(A)s we maintain this bridge for the next century and a half, we will continue to use the latest technology."
Rather than providing a final picture of corrosion, the report offered an incomplete, misleading snapshot, independent experts said.
John Broomfield, a globally respected British consultant on corrosion of steel within concrete, and recipient of the distinguished lectureship award at this year's annual meeting of the U.S. Transportation Research Board, called Caltrans' report "strongly challengeable."
"As Prof. Sagüés has pointed out in his comments," Broomfield said, "obviously the conclusions of the study can't be validated if they are based on incorrect information."
Experts said a few extra months of exposure would not necessarily cause dramatically more corrosion as the corrosion rate is affected by the presence of water, air and salt. Continuous exposure to all three factors would damage tendons most quickly.
In the skyway, that rate varied, according to the Caltrans study. "Moderate" corrosion occurred after 11 months of exposure in some tendons; in others, equal corrosion took just over four months.
Broomfield and other experts said that some untested tendons had undoubtedly corroded at the faster rate. Likewise, he said, the interval between the corrosion inspections and final grouting sometimes several months could have allowed corrosion to worsen in tendons that Caltrans said were only moderately damaged.
Richard E. Weyers, emeritus professor of engineering at Virginia Tech and an expert on corrosion of steel in concrete, said pitting and stress often accelerate corrosion.
UC Berkeley's Devine, a metallurgist, examined the Caltrans study and related materials. In interviews and a written analysis prepared for The Bee, he said that Caltrans made fundamental testing and interpretation errors that cast doubt on its conclusions.
Devine said Caltrans misinterpreted its own data, incorrectly ruling out potentially harmful concentrations of sulfur and salt.
Caltrans also incorrectly assumed harmful cracks would be accompanied by visible pitting, Devine said. Microscopic pits also can lead to deep cracking, and can occur "in a wire that otherwise exhibits very little visually observable corrosion."
The agency based its study largely on 1992 research and photographs by corrosion expert Augusto S. Sason.
Sason tested tendons exposed to moist air with a sharply different basic chemistry than the standing water in the skyway ducts. The skyway conditions likely caused greater and different corrosion, according to Devine.
"It points to the inappropriateness of using Sason's photographs as way of fingerprinting (skyway) corrosion," Devine said.
Caltrans relied on a test used by Sason to measure tensile strength stretching wires until they break to rule out vulnerability to cracks related to environmental conditions.
Yet, Sason noted that heavily corroded wires often pass strength tests, but even wires with slight visible corrosion show clear signs of microscopic pitting and cracks in bending fatigue tests. Caltrans neither conducted bend tests nor mentioned Sason's fatigue findings.
"This was so ironic, that Caltrans could rely markedly, heavily on (Sason's) report, and seemingly ignore the most important point being made," Devine said.
"You can conduct tests to show that there's not a problem, or that it's not as bad as it seems," Devine said. "They did the tests that were the least sensitive to the presence of pits, and lo and behold, the test says that there is no problem."
'War on Corrosion'
A few weeks after The Bee requested the corrosion study last year, Caltrans posted a video to YouTube "War on Corrosion." The video described the agency's handling of the tendon matter as effective and responsible.
Construction records reveal a more complex story.
Grout migrated between ducts through leaks at the joints of concrete sections. Leaking grout blocked adjacent ducts, preventing the insertion of tendons. So in each skyway section, tendons for all adjacent ducts had to be installed and stretched before any could be grouted leading to long delays.
Thomas C. Janssen, a spokesman for the lead builder, Kiewit Corp., defended the company's approach.
"At Caltrans' request, we protected the tendons with Vapor Phase Inhibitor Powder," he said in a written statement.
Records show that the builder applied the rust inhibitor in only about one in four ducts even though nearly all tendons experienced delays and almost never used the powder as required by the contract to ensure effectiveness.
The video did not mention that about 1,100 ducts were so leaky that they had to be grouted together in "families" of two to 21 ducts an unorthodox practice that shocked numerous experts.
Significant grout leaks and water intrusion used to crop up in the 1980s and 1990s, said Michael Kreger, a professor of civil engineering at Purdue University who has co-authored numerous scholarly papers on post-tensioned bridges and corrosion. Even then, such serious problems were uncommon, he said. He called the skyway experience "horribly unfortunate."
The combination of water in ducts and grout leaks between ducts, experts said, tends to encourage air and water pockets "voids" that can cause further corrosion.
"It seems to me almost certain that you can't avoid large air pockets with a duct-filling method of that kind," said Nick Buenfeld, who heads the Department of Civil and Environmental Engineering at Imperial College London, one of the world's top engineering schools. "If the joints between the sections allow grout to flow through them, I would expect them also to allow moisture and oxygen to flow through."
European bridges have failed where joints allowed water and salt to enter, Buenfeld said. "Sea spray does carry chloride a long way," he said.
"Grouting, though delayed, provided the protection," for such joint leaks, bridge spokesman Gordon said. Asked about the likelihood that grouting in families caused voids and other problems, Gordon said he "cannot make that speculation."
James Thompson, a teaching professor at Carnegie Mellon University, Buenfeld and Broomfield, among others, blamed the problems on design or construction errors.
TY Lin International, lead skyway designer, referred questions to Caltrans.
Kiewit spokesman Janssen said the ducts "were not designed to be watertight during construction," and that the company had worked according to contract requirements.
Gordon said that the agency has not assigned blame for the tendon issues.
"(W)e remained focused on problem-solving," he wrote.
Experts agreed that strong skyway foundations and piers, plus the factor of safety "10 percent extra tendons," according to Caltrans make a disastrous collapse of the bridge improbable, even in a devastating quake.
But if tendons are more corroded than Caltrans' study indicates, said Hawkins of the University of Illinois, a massive temblor might render one or more sections of the skyway unusable.
Another looming question, experts said, is whether taxpayers have purchased the $6.4 billion bridge they were promised, or a structure that will require costly repairs relatively soon.
"It's reasonable to suggest, strongly suggest, that a number of experts who are completely independent, knowledgeable in these areas, sit down and review what has been done," said Weyers, of Virginia Tech, " and give suggestions about what would need to be done."
Construction and testing doubts pose "a very large question" about the span's long-term performance, said Merrill Walstad, a structural engineer and technical adviser to the Post-Tensioning Institute, an industry research group. He added: "The least that should be done is an investigation."
Several experts said that radar or X-ray examinations might provide clues about the state of the tendons. But tendon excavations might be warranted.
Broomfield suggested close review of bridge sections that experienced extreme conditions: long grouting delays, widespread water contamination and many ducts grouted together in families.
One expert contacted by The Bee said Caltrans deserves the benefit of any doubts about the skyway.
The bridge contains "future ducts" available to insert extra tendons to boost capacity. Those ducts, said Maher Tadros, professor emeritus of civil engineering at the University of Nebraska, reinforced his view "that Caltrans, the designers and builders have followed the required standard of care," because they offer the opportunity for a "belt and suspenders" approach to structural integrity.
Asked if the future ducts could be used to compensate for possible corrosion, Gordon said, "This bridge is built to serve the Bay Area for a century and a half, and it is only prudent to provide opportunities to enhance the bridge during that time."
Samuel I. Schwartz, chief executive of a New York engineering company and former chief engineer for highways and bridges in New York City, said Californians face disheartening questions about quality control for the lifeline between Oakland and San Francisco.
"Based on the degree of corrosion, as an owner why should I accept this job?" he said, given the increased potential for future problems that might require expensive maintenance. "Does it mean the bridge's structural elements would last 40 years instead of 150?"
Call The Bee's Charles Piller, (916) 321-1113. Follow him on Twitter @cpiller. Dan Hill contributed to this report.