They've developed a statistical model of the deterioration and simulated it on computer, using what may seem like an unlikely analogy: a cracking brick wall.
What they've found could one day help scientists better understand this kind of corrosion, and also explain corrosion in other types of alloys.
Although the alloy, called 2024-T3, is strong and resistant to corrosion in general, it is vulnerable to intergranular corrosion -- when tiny pits on the surface grow into crack-like fissures that snake down into a part, weakening the structure, explained Gerald Frankel, professor of materials science and engineering at Ohio State.
Frankel had long experimented with 2024-T3 in the lab, and he suspected that a good way to model this corrosion on computer might be to use the analogy of a brick wall -- with the fissures between the grains simulated as cracks spreading through the mortar between bricks. But modeling the complex microstructure of an aluminum alloy is very difficult.
He approached Doug Wolfe, professor and chair of statistics at Ohio State, and together with their students, they developed a statistical model that depends on the probabilities of a fissure turning various different ways within the alloy.
In the December issue of the Journal of Statistical Planning and Inference, the researchers report that their brick wall model simulated intergranular corrosion in the alloy with near-perfect accuracy.
The fissures seem to follow a random path between grains of the alloy, Frankel said. Metal alloys are made up of many individual grains, which are regions where the atoms are aligned in a particular direction.
Large grains, like the ones on the surface of a new galvanized garbage can, are visible to the naked eye, Frankel said. In the high-strength aluminum alloys used in airplanes, grains are microscopic.
Corrosion often follows the region between the grains, called grain boundaries, which are more susceptible to attack. Sometimes a fissure will turn back on itself and re-emerge harmlessly at the surface, and other times it will progress through a part completely. When the part is on an aircraft skin, it has to be replaced quickly, and at great cost. The United States Air Force spends almost $1 billion each year fighting corrosion, Frankel pointed out.
Wolfe and doctoral student Shiling Ruan simulated a situation in which many fissures started at the top surface of an alloy sheet. They permitted the fissures to turn in all directions -- up, down, left, and right -- until one fissure completely penetrated the sheet and emerged on the other side. Based on thousands of simulations, they estimated the ratio of the length of the path followed by this first key fissure to the width of the sheet.
Their findings mirrored Frankel's experiments in the lab, where he and doctoral student Weilong Zhang had measured the time for intergranular corrosion fissures to propagate through sheets of varying width. They calculated a ratio of fissure length to sheet width of 4.29 -- meaning that for a given sheet width, the average length of the first fissure to zigzag through it would be 4.29 times this width.
In their simulations, Wolfe and Ruan obtained ratios as close as 4.25.
It is important to understand how such fissures propagate, Frankel said. Stresses on an aircraft can cause fissures to turn into large cracks that could lead to disaster if they are not repaired in time.
So knowing how fissure propagation depends on the material microstructure could one day allow a prediction of when preventive maintenance is needed.
To Wolfe's surprise, the factor that affected path length the most was not the direction a particular fissure would turn, but rather what happened when the fissure came to an intersection between paths -- a point in the alloy where three grain edges met. There a fissure could turn, or continue to follow a straight path and effectively "skip" the intersection.
"At first, we conjectured that the most important factor was the probability that a fissure would turn downward, so that it was headed toward the back of the sheet. We didn't expect that skipping was the real culprit in extending fissure lengths," Wolfe said.
Fissures with few skips found short paths through the sheet. Fissures with many skips were more likely to get caught up snaking sideways between the grains, and less likely to ever make it through.
In the future, engineers could potentially design a material's microstructure to increase the probability of skips, if they could determine what exactly caused a skip. Frankel suspects that changes in the chemistry between grains or the angles of atoms in the grains may play an important role. But he also said this work is just beginning, and any such application is a long way off.
"The problem is, we don't know how these probabilities change form plate to plate or alloy to alloy," he said.
Future work would have to assess how fissures travel through other alloys. And since the brick wall model is currently only two-dimensional, the researchers would have to expand it to three dimensions for more realistic results.
This work was funded by the National Science Foundation and the United States Air Force through a grant from S&K Technologies.
Written by Pam Frost Gorder, 614-292-9475; Gorder.email@example.com