Over the years, concrete can become permeated with road salt from a steady drip of wintry slush.

If salt reaches the inner steel, it's the beginning of the end: corrosion.

"Once you start getting salt brine down in there, it's a killer," said Harold Windisch, a senior construction engineer with the Pennsylvania Department of Transportation.

Engineers couldn't say for sure yesterday what caused the crack, first spotted in 2005, or what widened it to the point that the road was closed Monday night. But the theory was that road salt was a likely culprit, along with the freezing and thawing of water that penetrated the crack.

How salt promotes corrosion is not fully understood, but the result is clear. Once steel starts to rust, reverting to its natural state of iron oxide, it expands. That cracks the concrete, potentially allowing more salt and moisture to enter. "The problem with this is that it accelerates," said engineer William Schutt, president of Matcor Inc. in Doylestown, which sells corrosion-protection systems.

Today, reinforced concrete is often coated with some sort of sealer or membrane to prevent salt intrusion, or the steel itself may be coated with epoxy. But the cracked pier under I-95 was built in the 1960s. State officials said the steel did not appear to be coated in any fashion.

Initially, the pier was protected by the road overhead. But it sits beneath an expansion joint that connects two sections of highway.

A rubber trough carries away moisture, just like a gutter on a house, but Windisch said the trough could have sprung an unseen leak - exposing the pier to a steady, salty drip.

The bridge will be shored up with four prefabricated pillars until the cracked pier can be repaired or rebuilt.

Asked to look at the repair plans, Lou DaSaro, an associate professor of structural engineering at Drexel University, said he could not comment without knowing more about the loads involved.

Generally speaking, he said, "it seems like a reasonable concept."

Could the problem have been prevented? Schutt said yes.

"I think it should never have gotten to this point," he said after looking at dramatic pictures of the crack, which is up to five inches wide and more than six feet long.

Once the crack was discovered, inspectors could have periodically used special voltmeters to tell whether the inner steel had begun to corrode, Schutt said. The pH of the concrete also could be tested; a drop in pH is a sure sign of salt intrusion.

Another option: Tap the concrete with a hammer. A hollow sound means it has begun to delaminate.

Charles Davies, PennDot's assistant district executive for design, said he did not know what specialized equipment had been used, if any, when the cracked pier was last checked in October. The inspection was "primarily visual," he said, but a hammer might have been used.

Whatever happened on I-95, engineers agreed that the corrosive powers of road salt - any of several chloride compounds, such as calcium chloride - required additional study.

When steel bars are used to reinforce concrete, they tend to develop a passive, protective film. But chloride ions are able to penetrate that protective layer, said Jamie Farny, an engineer with the Portland Cement Association, a trade association in Illinois.

Once the salt ions come in contact with steel, they promote the electrochemical reaction that generates rust, said Scott Humphreys, a structural engineer with the Concrete Reinforcing Steel Institute, another industry association.

"If we knew perfectly how it occurred," Humphreys said, "then things could be done to make sure that it didn't."