Burning down a freeway is not easy.

But set off enough high-octane fuel at the wrong place and even multiton supporting steel girders turn wobbly as a chocolate bar in the sun. That is what a speeding gasoline trucker managed to do before dawn Sunday to the busiest interchange in Northern California.

Authorities still are piecing together how one of the 18-wheelers traversing the Bay daily crashed and erupted into a fireball that collapsed one freeway onto another at untold disruption to regional traffic.

Miraculously, no one plunged into the inferno, the void it left behind or the debris below. But some experts think the incident calls for a review of heat and blast vulnerability for critical interstate spans.

"This time we were lucky," said UC Berkeley civil engineering professor and steel bridge expert Abolhassan Astaneh-Asl. "This is a wake-up call for major bridges."

State officials pinned blame on driver James Mosqueda, 51, and his employer, and they suggested the state was moving rapidly to rebuild the damaged portions of Interstates 580 and 880 more or less as they were built 50 years ago.

The episode was a freakish "anomaly," according to Will Kempton, head of Caltrans.

But fiery tanker crashes have collapsed freeways in Philadelphia and Boston, and now homeland-security analysts fear copy-cat terrorist attacks.

The burning tanker at the MacArthur Maze released over three hours about the same energy as the split-second detonation of 200 tons of TNT, equal to an extremely low-yield atomic bomb.

"It certainly is a message of something we should be concerned about, and it doesn't take a rocket scientist to figure it out," said David McCallen, a senior executive at Lawrence Livermore Lab's nonproliferation and homeland-security directorate.

The light, flexible steel spans of the MacArthur Maze survived the 1989 Loma Prieta quake that pancaked the Cypress Freeway, which was made of stiff concrete.

But at 3:41 a.m. Sunday, Mosqueda flipped and ruptured his tanker at almost precisely the Achilles' heel of the arching skyways -- the underside of the pier where thin, supporting steel girders are unprotected by concrete or anything else, according to UC Berkeley's Astaneh-Asl.

"I think this was really the perfect fire, tragically," said Astaneh-Asl, who studied the MacArthur Maze intensely after the earthquake.

The extent of Mosqueda's fuel load was unclear Sunday. But at least 8,600 gallons of unleaded gasoline ignited in a continuous roar -- more fuel than burned inside the World Trade Center towers on Sept. 11, 2001 -- and turned the 20-foot space between I-880, where Mosqueda crashed into a guard rail, and the I-580 overhead into an oven that roasted the exposed steel girders to more than 2,000 degrees.

At 1,000 degrees Fahrenheit, steel in girders and bolts goes soft, said Astaneh-Asl, who studied the collapse of World Trade Center towers for the National Science Foundation.

"When steel gets that warm, it loses its strength and cannot carry its load any more," he said. "It's not to say the steel melted. Some portions may have melted, but the steel got soft, like rubber."

Other experts say structural failure can come at even lower temperatures.

"The bottom line is this kind of thing can happen," said Forman Williams, a pre-eminent combustion expert at UC San Diego, who was tapped by the federal government to explain the World Trade Center collapse. "It is a rare event, but all of these are rare events, and the more of these you can protect against, the better off you are. ... You don't know until you do the study."

Caltrans officials dismissed the notion Sunday of fireproofing freeway stilts and bridges. Astaneh-Asl said roadways really cannot be fireproofed like buildings, partly because of vibration.

But he suggested that some of California's most critical bridges and interchanges might deserve the extra expense of coating exposed steel girders in a few inches of concrete.

That coating might cost 5 percent to 10 percent of the construction cost for the bridge but be enough to delay collapse for two hours, perhaps enough time to cool or extinguish a fuel fire.

"From an accidental standpoint, these are definitely rare events," said Livermore's McCallen. "But I think this will probably elicit some additional debate within the (engineering) community about whether something more for fire is needed.

"It's certainly something that people who worry about homeland security should keep at the forefront of what they're thinking about."

Reach Ian Hoffman of the Oakland Tribune at 510-208-6458 or ihoffman@angnewspapers.com.