Imagine a material cracking — now imagine what happens if there are small inclusions in the material. Do they create an obstacle course for the crack to navigate, slowing it down? Or do they act as weak points, helping the crack spread faster?

Historically, most engineers believed the former, using heterogeneities, or differences, in materials to make materials stronger and more resilient. However, research from Georgia Tech is showing that, in some cases, heterogeneities make materials weaker and can even accelerate cracks. 

Led by School of Physics Assistant Professor Itamar Kolvin, the study, “Dual Role for Heterogeneity in Dynamic Fracture,” was published in Physical Review Letters this fall. 

While Kolvin’s work is theoretical, the results of the research are widely applicable. “Predicting this type of toughening effect helps engineers decide how much reinforcement to add to a material, and the best way to do so,” he says. “Cracks are complex — they interact with the material, change shape, and respond dynamically. All of this affects the overall toughness, which impacts safety.”

Building strong materials

The study found that the key to crack behavior starts at the microscopic level where the material’s microscopic structure influences how it resists cracks running at different speeds.

“Cracks propagate by breaking bonds, and that costs energy,” he explains. “On top of this, materials experience extreme deformations close to where the crack runs, which costs additional energy. In some materials, the amount of this energy cost can depend on the crack’s speed because of microscopic friction between molecules.”

Other materials, like window glass, are mostly indifferent to the crack speed. These materials are made of simple molecules, allowing a crack to propagate slowly or quickly using the same amount of energy. The researchers found that including heterogeneities can help strengthen these materials.

Materials made of more complex molecules, like polymer plastics and gels, on the other hand, are velocity dependent: it takes more energy for a crack to propagate faster. In these materials, heterogeneities are less effective at toughening, and if the crack is fast enough, heterogeneities could help it advance. “That’s something we didn’t expect when we started,” Kolvin says.

Disorder versus design

After discovering which types of materials can benefit from heterogeneities, Kolvin wanted to investigate the best way to add them. “Natural materials like rocks are usually very messy and disordered,” he explains, “but in engineering, heterogenous materials tend to be patterned.” For example, imagine a manufactured material: heterogeneities may be added in a grid-like or other patterned way. Now, contrast that with the irregular freckles and inclusions you might see in a rock found in a streambed.

Kolvin’s question was simple: which material was stronger? The results, again, were surprising. The disordered case — similar to what is found in nature — created the toughest material. 

Among the patterned materials the team tested, only one was as tough as the disordered case — and every other pattern tested made the material weaker.

From lab to landscape

At Georgia Tech, Kolvin’s lab focuses on the mechanics of materials — both solid and fluid. “We are using our expertise in physics to explore questions across different fields,” he says. “A common concept is treating materials as continua — zooming out from molecular detail to look at how materials deform and flow at the large scale.”

This current research follows suit with applications ranging from investigating the smallest material microstructures to predicting earthquake fractures. “Earthquake faults are highly disordered, and simulating these ruptures is a major challenge, usually requiring supercomputers to solve crack propagation in three dimensions,” Kolvin says. “But with the tools our study has developed, we can simulate similar conditions and large systems using just a desktop computer.”

“This opens the doors for scientists, engineers, physicists, and geologists to explore problems right from their own computer, allowing more researchers access to more tools,” he adds. “And new tools often lead to new discoveries.”

 

DOI: https://doi.org/10.1103/j4vb-y1ng