By Steven Mufson
Washington Post Staff Writer
Friday, October 29, 2010; 10:22 AM
Hours after a presidential commission raised new questions about the safety of the cement mixture intended to temporarily seal BP's Macondo exploration well, oil field service giant Halliburton said the final version of the mixture did not undergo a foam-stability test.
The presidential commission investigating the mammoth oil spill said Thursday that the cement mixture repeatedly failed lab tests before the April 20 blowout.
Halliburton later issued its own statement, dismissing some of those lab tests as preliminary or irrelevant but also adding - without explanation - that the cement mixture ultimately used on the well did not undergo the complete battery of safety tests.
As early as February, Halliburton was getting poor results in lab tests of the recipe for the cement it was planning to use, according to evidence collected by the National Commission on the BP Deepwater Horizon Oil Spill and Offshore Drilling.
Three separate tests suggested that the mixture would be "unstable," according to a commission staff letter released Thursday. Halliburton notified BP by e-mail about only one of the tests before the well explosion, according to the commission. The two companies went ahead with the cementing job anyway. Its failure became the first in a cascade of factors leading to the accident.
The results of a fourth Halliburton test - the only one indicating that the cement slurry might have been able to contain the high-pressure pool of oil and gas at the bottom of the Macondo well - were not available until the night of April 19 at the earliest and perhaps not until after the cement was poured, the commission staff said.
The oil spill commission is sifting through the events leading to the April 20 explosion, which killed 11 workers, sank the Deepwater Horizon and triggered a massive oil spill in the Gulf of Mexico. The cement at the bottom of the exploratory well was supposed to have provided a seal until a production facility could be built.
The reason for the cement job's failure has been a matter of dispute for months. Halliburton has pointed at BP; BP has challenged Halliburton. Experts still differ.
Halliburton late Thursday night issued a statement disputing the commission staff's letter, calling the February tests "preliminary" and saying that "final well conditions were not known at that time." Halliburton added that the first of the April tests was "irrelevant" because of errors at the lab, and it said that, contrary to the commission letter, BP was informed of that test.
The news spooked shareholders; Halliburton's stock closed at $31.68 a share, down nearly 8 percent. BP said it had no comment.
The commission letter reiterated that the cement was just one contributor to the disaster. "Cementing wells is a complex endeavor, and industry experts inform us that cementing failures are not uncommon even in the best of circumstances," the commission letter says.
But the new details call into question whether Halliburton's recipe - which mixed nitrogen and other additives with ordinary cement to create a foamy mixture - was the right one.
At the commission's request, Chevron recently carried out independent lab tests of a cement slurry that Halliburton said was the same as that used in the Macondo well. The commission staff said Chevron reported that "its lab personnel were unable to generate stable foam cement in the laboratory using the materials provided by Halliburton."
The commission staff said in the letter Thursday that the Halliburton tests before the Macondo well blowout and the new lab tests conducted by Chevron show that "Halliburton (and perhaps BP) should have considered redesigning the foam slurry before pumping it at the Macondo well."
Halliburton said that the Chevron tests were preliminary and what Chevron tested may have been different from the "unique blend of cement and additives" used on the Deepwater Horizon.
The commission letter appears to conflict with statements that Halliburton officials made during other inquiries into the causes of the Deepwater Horizon disaster.
"All of the design work, all of the testing work that was done by Halliburton in advance of this job indicated that the foam system was stable," Thomas Roth, Halliburton's vice president for cementing, told an investigative panel of the National Academy of Engineering at a Sept. 26 hearing.
Halliburton had conducted two lab tests in February that indicated potential cement problems, but it reported only one of them to BP, the commission said. Halliburton sent BP those results in a March 8 e-mail, although commission staff members said it wasn't clear whether BP ever raised any questions.
"Both tests indicated that this foam slurry design was unstable," the commission letter says. Staff members said separately that Halliburton did not use the words "unstable" in its e-mail. "There is no indication that Halliburton highlighted to BP the significance of the foam stability data or that BP personnel raised any questions about it," the commission letter says.
Halliburton conducted two more tests in April with additional information about conditions at the bottom of the well, but the first of those tests on or about April 13 again indicated that the cement slurry would be unstable. "The results of this test were reported internally within Halliburton by at least April 17, though it appears that Halliburton never provided the data to BP," the commission letter says.
The final test altered some of the conditions and showed that the cement job might hold, but it was not clear that the test was complete before the cement was poured at the well, the commission said. The results were only reported to BP on April 26, the commission said.
Halliburton said that the final test in April was performed on a slurry formulation that included eight gallons of retarder per 100 sacks of cement and that the test showed that the foam mixture was stable. Retarders lengthen the time it takes for cement to set and are frequently used in high temperature environments.
Halliburton said that BP subsequently instructed Halliburton to increase the amount of retarder in the slurry formulation from eight to nine gallons per 100 sacks of cement. New tests were then performed on the thickening time and compressive strength on the nine-gallon formulation, but not on its foam stability, Halliburton said.
In a report released Sept. 8, BP's own internal investigators faulted Halliburton for ignoring troublesome signs in the April 12 test data. BP said the lab tests "should have led Halliburton to conduct further testing and to continue working on the slurry design." One aspect of the cement mixture "could have led to difficulties in foam stability," BP said.
Many of the earlier explanations of the failure of the cement job in the Macondo well have pointed to BP's decision to use six instead of 21 centralizers, devices that ensure the drill pipe is centered in the hole.
In testimony Aug. 24 in Houston before an inquiry board of the Coast Guard and Interior Department into the causes of the Gulf of Mexico oil spill, Halliburton technical adviser Jesse Gagliano faulted BP for failing to use more centralizers.Gagliano said he told BP that such a decision could endanger the cement job.
Asked by a BP lawyer whether Halliburton had tested the effects of a mixture on the cement, Gagliano said, "Yes, we had a lab test showing that it was a stable system."
Halliburton President Tim Probert wrote in a June 1 letter to Rep. Edward J. Markey (D-Mass.), chairman of the House subcommittee on energy and the environment: "Halliburton is confident that the cementing work on the Mississippi Canyon 252 well was completed in accordance with the requirements of the well owner's well construction plan."
The oil spill commission's findings turn attention back to Halliburton and the recipe it used for additives and nitrogen that would make the cement light and durable.
"The fact that BP and Halliburton knew this cement job could fail only solidifies their liability and responsibility for this disaster," Markey said. "This is like building a car when you know the brakes could fail, but you sell the cars anyway."
In a separate background briefing, commission staff members said that BP's design for the well might still have been a factor. The use of a long, steel casing might have contributed to concerns about the level of pressure in the space between the pipe and the sides of the well. That limited the amount of cement that could be used, commission staff members said. In other situations, some companies add cement if they are worried about the cement stability.
In addition, commission staffers said, a different well design might have enabled the oil services giant to use another cement mixture.
Robert Bea, a professor and oil industry expert at the University of California at Berkeley, said that drillers will often run one test on a cement mixture, then a second test as a backup. He called four tests "unusual."
"Given that they are running four, that's telling me they were having trouble getting to a stable design," Bea said. "And so they continued to work with this witches' brew until they got on to something they thought was workable."
In general, he said, mixing light, foamy cement with the fragile rock formations at the drilling site was "a recipe for disaster."
Staff writers Joel Achenbach and Joe A. Stephens contributed to this report.