Experiment probes pattern formation during debonding of viscoelastic adhesives

Poke your finger into a jar of honey, and you feel a drag force that depends on how fast you’re moving your finger. Remove it, and some honey clings to your finger and stretches into a long fibril that eventually breaks. Those properties are characteristic of a viscous liquid. When you push against a piece of soft rubber, on the other hand, the force you feel is proportional not to speed but to displacement, and when you pull away, the rubber remains in one piece; such are the properties of an elastic solid. Typical commercial adhesives have both viscous and elastic properties—that is, they are viscoelastic.

When a film of viscous, elastic, or viscoelastic material is sandwiched between two smooth, rigid surfaces that are then pulled apart, patterns form that include “fingers” of air penetrating the gap between the surfaces. Fingering patterns in the viscous and elastic cases have both been studied before. But pattern formation in the viscoelastic regime, which is most relevant to adhesive failure and thus to adhesive performance, has remained unexplored. Now, Julia Nase, Anke Lindner, and Costantino Creton of the École Supérieure de Physique et de Chimie Industrielles in Paris have taken a first look at fingering patterns in adhesive materials with a range of properties, spanning the continuum between viscous liquid and elastic solid. ¹

The researchers’ first task was to find a good model material whose viscous and elastic properties they could tune reproducibly. They settled on a commercially available kit consisting of a polymer fluid and a curing agent that forms cross-links among the polymers. Typically, the fluid and curing agent are mixed in a 10:1 ratio to form a clear elastic solid that can be used for microfluidics devices. But the researchers found that smaller amounts of curing agent produced materials with the desired range of viscoelastic properties. A mixture with 2–3% curing agent forms a softer solid, similar to the protective film that comes on a new computer monitor. A mixture with 1% curing agent forms a material at its gel point: the minimum density of cross-links needed to connect all the polymers in an infinite network. That material feels sticky to the touch and forms visible fibrils, like the adhesive on a piece of packing tape. (For more on polymers, viscoelasticity, and stickiness, see the Physics Today articles by Tom McLeish, August 2008, page 40 , and by Cyprien Gay and Ludwik Leibler, November 1999, page 48 .)

To study the fingering patterns, the researchers coated a glass slide with a layer of the polymer adhesive between 50 μm and 500 μm thick. They brought a 6-mm-diameter steel probe in contact with the material from below and pulled it away, while watching the air fingers using a camera mounted above the slide. In particular, they looked at the initial fingering wavelength, the distance from one finger to the next at the moment the fingers started to form.

In the case of a viscous liquid layer, the fingers penetrate the bulk of the material, as shown in figure 1(a), and the wavelength is known to depend on the thickness of the layer, the speed of the probe, and the material properties of the liquid. That result is described by the so-called Saffman–Taylor instability, which occurs when a less viscous fluid (air, in this case) enters a more viscous fluid in a confined geometry. For an elastic solid layer, the fingers enter at the interface between the layer and one of the surfaces, as shown in figure 1(b), and the wavelength depends only on the layer’s thickness.

The researchers found that for every amount of curing agent, the initial wavelength they observed was described by either the viscous model or the elastic model. Example data are shown in figure 2. (As the debonding progressed, though, the sizes and shapes of the fingers deviated from either of the two models.) The sharp transition between viscous and elastic behavior occurred at around 1.5% curing agent—meaning that even some materials with cross-link densities well above the gel point behaved like viscous liquids. That seemingly strange result occurs because the debonding behavior depends not only on the properties of the adhesive layer but also on how well the probe sticks to the adhesive. If the energy needed to deform the adhesive material is less than the energy needed to create a crack between the adhesive and the probe, then the debonding will follow the bulkdeformation mechanism of viscous liquids. Otherwise, it will follow the interfacial mechanism of elastic solids.

To test that explanation, the researchers ran a new experiment on material with 2% curing agent. But instead of a steel probe, they used a glass probe that had been treated with plasma. The plasma opens up sites for chemical bonds that enhance the adhesion between the probe and the polymer material. Sure enough, the material that displayed the elastic mechanism when the steel probe was used followed the viscous mechanism in the new experiment with the treated glass probe.

There’s a lot left to understand about pattern formation in the viscoelastic regime. In particular, the shapes of the fingers and their behavior in the later stages of debonding remain unexplained. Says Lindner, “We would be happy if our results stimulate theoretical investigations and help answer these questions.”