Philadelphia dodged several potential catastrophes during a dramatic June 21 refinery blast, which released about 5,239 pounds of a deadly chemical and launched pieces of shrapnel as large as a truck hurtling across the 1,300-acre refinery complex, according to federal findings released Wednesday.
The disaster at the Philadelphia Energy Solutions complex began with the early morning failure of an elbow section of pipe that had corroded to half the thickness of a credit card, according to investigators for the U.S. Chemical Safety and Hazard Investigation Board (CSB). A fire triggered three successive explosions, the largest of which blew a fuel tank into massive projectiles, including one weighing 19 tons that traveled 2,100 feet and landed on the opposite bank of the Schuylkill.
The failure of a section of pipe was similar to a 2012 accident at a Chevron refinery in Richmond, Calif., which prompted the CSB to recommend that refinery operators inspect all components of piping systems. That was not done at PES, said Kristen Kulinowski, the CSB’s interim executive.
The agency’s report is the first to confirm the release of toxic hydrofluoric acid, a material used as a catalyst in the alkylation unit that was destroyed in the blast. Since 2015, the CSB has investigated two other refinery accidents involving HF-related equipment, but neither caused a release of the material.
“The board remains concerned that the next time there is a major explosion at a refinery that uses HF for alkylation, workers and those living nearby will not be so lucky,” Kulinowski said.
The acid can destroy tissue and bone, and has caused fatalities with skin exposures to as little as 2.5% of body surface area, the CSB said. “If inhaled, HF can cause severe lung injury and pulmonary edema — fluid in the lungs — which can result in death,” the report said.
Despite the release of hydrofluoric acid, the incident caused no serious injuries — only five refinery workers experienced minor injuries that required first aid treatment, CSB said. The agency said it was unaware of any offsite or onsite health impact from the hydrofluoric acid release.
“It’s likely, though we have no way of measuring this, that it just dissipated into the atmosphere,” Kulinowski said. “There were no people in the vicinity, or we have not heard of any reports of human exposure.”
The large pieces of steel equipment that were blasted through the air miraculously landed harmlessly in the refinery, missing processing equipment and millions of barrels of crude oil and fuel stored in tanks. Penrose Avenue and the Platt Bridge were also in the range of the flying missiles.
“I think this was catastrophic enough as it was,” Kulinowski said. “We’re focusing on what could have happened had it been worse, but in fact, it was bad, just as it was. We need to focus on making sure that this kind of explosion doesn’t happen anymore because it’s just a matter of time before the facts are just a little bit different.”
The Chemical Safety Board has investigated about a dozen refinery incidents in the last 20 years, including those resulting in fatalities, serious injury, and property damage. The agency has no enforcement powers and can only make recommendations. Other agencies with the power to penalize, including the Occupational Safety and Health Administration, are also investigating.
The CSB’s 10-page “factual update” does not attempt to affix blame or fully explain the circumstances leading to the fire, which shut down the largest oil refinery on the East Coast. The report does not suggest whether the refinery’s longstanding financial problems or its maintenance practices contributed to the accident.
The refining company and its workforce appear to be the biggest casualty of the accident. Its owners announced the plant’s closure on June 26, throwing most of the 1,100 workers out of jobs, and filed for Chapter 11 bankruptcy protection in July. The refinery remains shut down as the U.S. Bankruptcy Court in Delaware determines whether its creditors will take ownership, or the complex will be sold.
PES estimated that about 676,000 pounds of hydrocarbons — mostly butane, butylene, and isobutane used in the alkylation processing unit — were released during the event, of which 608,000 pounds were combusted, or burned up.
Thirty seconds after the fire started, a control room operator emptied most of the hydrofluoric acid in the alkylation unit into a “rapid acid deinventory drum,” a system the refinery’s previous owner, Sunoco Inc., had installed more than 10 years ago to reduce the chances of a catastrophic release of HF in such an accident. The alkylation unit itself remained intact.
But a “low concentration” of HF remained in some of the equipment that failed during the event, and PES estimated that 5,239 pounds was released. About 1,968 pounds of the released hydrofluoric acid was contained by a water spray in the unit that was installed to reduce the spread of an inadvertent release. About 3,271 pounds was released to the atmosphere. In a gaseous state, HF is called hydrogen fluoride.
The CSB says that 48 of the nation’s 150 refineries operate alkylation units that use hydrofluoric acid. The board in April called on the Environmental Protection Agency to revisit the effectiveness of existing regulations for HF and whether safer technologies are available for alkylation, which produces a fuel-blending agent that increases the octane levels of gasoline.
The refinery previously reported to federal regulators that a worst-case release of 71 tons of HF could cause a cloud that would extend more than seven miles, potentially exposing a population of more than one million people. The amount released into the atmosphere on June 21 was less than two tons.
But the proximity and concentration of the HF exposure are more of a measure of the risk. The refinery’s previous owner, Sunoco, was cited by OSHA in 2009 after 13 contract workers were sent to the hospital for “precautionary” medical treatment following a release of only 22 pounds of HF. The workers were installing upgrades to the alkylation unit, including the equipment that allowed PES to rapidly empty the acid on June 21 after the fire broke out.
At the time of the June 21 fire, a shelter-in-place order was issued for residents near the refinery, but no evacuation ordered. The Philadelphia Health Department measured an “elevated” level of hydrogen fluoride gas outside the South Philadelphia refinery during the accident, but the reading was dismissed as a “false positive.”
Peter DeCarlo, a Drexel environmental engineering professor, said the CSB report should prompt the city to revisit his call for a more robust air-monitoring system to alert the public about dangerous releases. “If all that material had been released at once at ground level, we would certainly know about it,” he said.
The disaster unfolded in a matter of minutes. CSB said the alkylation unit was operating normally, and then at 4:00:16 a.m., a sudden loss of containment in the piping system in the unit caused a combustible mixture of process fluid and hydrofluoric acid to release, forming a “ground-hugging vapor cloud.”
Less than two minutes later, the cloud ignited. About 30 seconds later, the control room operator emptied the hydrofluoric acid from the unit into the safety storage drum.
At 4:15 a.m., 13 minutes after the fire began, the first explosion occurred, followed by a second at 4:19 a.m. At 4:22 a.m., the third explosion occurred when a large drum containing butylene, isobutane, and butane ruptured.
That explosion, captured by remote television cameras, sent three large fragments of the drum flying in different directions. A 38,000-pound piece traveled four-tenths of a mile and landed on the opposite bank of the Schuylkill, near the refinery’s tank farm. Two other pieces, one weighing 23,000 pounds and another weighing 15,500 pounds, landed in the refinery.
The reason for the pipe failure remains a metallurgical mystery. One of the CSB’s remaining challenges appears to be trying to resolve why one elbow section of pipe in the doomed alkylation unit failed while others that had been inspected recently showed no sign of excessive corrosion.
The piping circuit in the alkylation unit that contained the ruptured elbow was installed in about 1973, and CSB said it appears to be original piping. The pipes are subject to regular ultrasonic thickness measurements at designated “condition monitoring locations” as part of a PES inspection program to monitor metal losses due to corrosion.
The most recent measurements did not indicate a thin pipe — PES recorded wall thicknesses ranging from 0.229 to 0.345 inches in previous tests, above the minimum thickness of 0.18 inches.
The elbow joint that failed, however, was among several that were not subject to the measurements. CSB measured the pipe after the incident and said that at its thinnest point, it measured a mere 0.012 inches thick, 7% of the minimum thickness allowed, or about half the thickness of a credit card.
The piping was installed before new standards were adopted in 1995 that set maximum levels of metals such as nickel, chromium, and copper, whose presence in a steel alloy can cause greater corrosion when it comes in contact with HF.
The ruptured steel pipe elbow contained a much higher percentage of nickel and copper alloys than would be permitted in a new installation. Curiously, an adjacent elbow joint, 18 inches away from the ruptured pipe, showed no sign of accelerated corrosion. It has the same markings as the ruptured pipe and was also original equipment from 1973, said Lauren Grim, the CSB’s supervisory investigator.
“This was just kind of a rogue piece because of its metallurgical composition that corroded faster,” said Grim.
Grim said it was not unusual for pipe sections that come in contact with hydrofluoric acid to experience variable corrosion rates, which is why the agency called for refinery operators to test 100% of critical pipe sections.