Jamming in Sand

We all know that if you tilt a bucket filled with sand, the sand will only start to flow when the tilt reaches a certain angle. The sand immediately flows fast. If you want the sand to flow slowly, you should gently vibrate the bucket (as you do with breakfast cereal). In the lab we have a setup that allows us to do controlled flow experiments on weakly vibrated sand. We have a disk in the sand that is connected to a rheometer (see inset of the Figure (copied from our PRL (2011)). We measure how much torque (T) it takes to rotate the disk at a certain rotation rate (Ω) for different vibration amplitudes (Γ), where we keep the acceleration below 1g. The result is shown in the Figure. In the absence of vibrations (Γ=0), this curve has a large regime with a negative slope dT/dΩ < 0, this suggests that if you want the disk to rotate faster, you have to impose less torque! This is exactly the instability that causes the observation that you cannot let sand flow slowly out of a bucket. For Γ=0.5, there still is a small regime with a negative slope, but for low Ω, slow flow is possible. For Γ=1, the flow curve is monotonic as it would be for a Newtonian fluid such as water – the vibration liquefied the sand. The behavior of this system is very rich which means we can study many different properties of the system. A first example is that we are trying to characterize the phase transitions from no flow and slow flow to fast flow. Addionally, as recently reported in PRE (2014), we have found that if the grains are weakly vibrated, the system can no longer be described by concidering it purely from a frictional point of view.