Some steel anchor bolts produced in 2010 for the new San Francisco-Oakland Bay Bridge failed the same quality tests as bolts manufactured in 2008, which broke last month.
Results released last week contradicted earlier statements by California Department of Transportation officials that only the 2008 bolts had failed tests.
Both sets of bolts, provided by Ohio-based Dyson Corp., were designed to secure vital seismic safety equipment on the eastern pier of the new suspension span.
The two failed tests for the 2010 bolts, conducted by Caltrans' materials lab, involved stretching bolts to ensure that they attained a required 14 percent elongation before rupturing. Other samples had passed the same tests conducted by a contractor.
Digital Access for only $0.99
For the most comprehensive local coverage, subscribe today.
Caltrans spokesman Will Shuck said that there were "no elongation test failures" because some samples exceeded the 14 percent standard, even though others fell short.
"Elongation results are based on an averaging of samples," he said. "Those averages are within desired specifications."
Thomas Devine, a UC Berkeley professor in the department of materials science and engineering, called such averaging "not a good practice," and "not appropriate" because it could result in the use of bolts that did not meet minimum standards.
"What you are interested in is not the average performance. You are interested in the complete performance," he said.
Accepting parts that failed to meet one quality standard takes on greater significance, Devine said, given other warning signs involving quality control.
The Bee previously reported that the supplier, Ohio-based Dyson Corp., and some of its subcontractors, did not pass Caltrans quality audits prior to the bolts being produced. Dyson also had a higher number and rate of quality test failures for its parts than any other supplier for the new suspension span.
Separate tensile strength tests showed that the east pier bolts were particularly susceptible to becoming brittle if contaminated by hydrogen, experts have said. That vulnerability might have worsened after they were dipped in molten zinc to guard against corrosion. The zinc process, called galvanization, causes changes in the surface of the steel that sometimes leads to cracks that provide pathways for hydrogen atoms.
A tensile strength result above 145,000 or 150,000 pounds per square inch suggests vulnerability to hydrogen contamination, experts said. The higher the strength number, the greater the likelihood that harmful hydrogen can enter the steel.
The suspect Bay Bridge bolts produced in 2010 showed readings from about 153,000 to 165,000 pounds per square inch – levels that should have prompted some attention to possible hydrogen contamination, said Joe Payer, chief scientist of the national corrosion center at Ohio's University of Akron. The bolts made in 2008, 32 of which failed, were rated up to 171,000 pounds per square inch.
Payer said that using high-strength bolts is not unusual or problematic if they are properly manufactured and protected from corrosion.
But their use in marine environments, particularly when coated by molten zinc, can be problematic, according to the U.S. Navy's technical manual for ships. The Navy prohibits such steel fasteners with a strength rating of 150,000 pounds per square inch or higher if they might contact water or heavy condensation.
For the 2008 bridge bolts, irregularities in the production process and corrosion from water exposure after installation might have combined to cause hydrogen intrusion that led to their failure when tightened last month. Those factors could likewise eventually affect the 2010 bolts, which so far have not ruptured, experts said.
UC Berkeley's Devine said that if hydrogen contamination took place during manufacturing, it might have been discovered with a test that was not used, which stresses the part for a day or longer at a higher level than it would experience in service.
That test could not have foreseen the bolts becoming brittle due to the corrosive effects of water once installed. But Devine called a failure after only a short period of water exposure "almost bizarre."
"Here you have a part that is expected to function in a marine environment, obviously exposed to rain and to fog, and is expected to last 150 years," he said. "The bottom line is that this is not something that should have been a possibility. It should have been considered from the beginning."
The combination of quality control problems, worrisome tests results, zinc-related cracking and water-related corrosion all add up, Devine said.
"That's typically the case when you have failure of parts in service," he said. "Generally speaking, it's not one particular shortcoming, but an accumulation of issues."
Letters released by Caltrans last week also show that the agency ordered the prime contractor, a joint venture of American Bridge Co. and Fluor Corp., to conduct magnetic particle beam tests – a way to check for small cracks – on eight kinds of parts. These included anchor bolts for the span's main cable, as well as the bolts that later failed on the eastern pier.
But in a letter to Caltrans resident engineer Gary Pursell, American Bridge/ Fluor project director Michael D. Flowers confirmed "informal communications" with the agency that Flowers said allowed the contractor to skip tests of the anchor bolts on the eastern pier, which had already been installed.
A fix for the bolts has not yet been selected, and will take months to build, test and install, placing the new span's projected opening date of Sept. 3 in doubt.
The Bee's Charles Piller, (916) 321-1113. Follow him on Twitter @cpiller.