Studying the Brazil-Nut effect is studying the interaction of granular media. People have been studying the segregation that happens in mixed granular media for a long time now, however there have been many theories replacing older theories and none of them seems to be quite correct.

    One of the main theories involves a kind of convection effect that causes grains to move up in the center and down at the sides of a shaken vessel. The large grains get trapped at the top, because they are too large to fit within the narrow descending stream. Some researchers have pointed to a sieving effect, where small grains fall through the gaps between larger ones. Other researches have found that with large amplitudes of shaking the larger intruder particles get tossed upward a bit and the smaller particles quickly fall into the empty space left behind. One of the newest theories takes into account the vibration frequency where the particles are said to undergo condensation [1]. In this state the particles are moving in free space but cannot exchange places with their neighbors. Since different size and density particles have different frequencies of condensation they will have different motions while together. With this kind of model the researcher found that there should be a reverse Brazil-Nut effect. If a sphere A is twice the mass and diameter of a sphere B, the larger particles float. But if they are six times heavier and twice the diameter, the larger ones would sink. This would be the reverse Brazil-Nut effect. This would be something worth testing for in our experiment.

    The inspiration for our research project comes mainly from an article that describes a relationship between pressure of the air in the shaking chamber and the movement of the large intruder particle to the top [2]. In this experiment a single larger intruder particle was put into a chamber full of smaller particles. It was vertically shaken and the pressure was varied, as was the density ratio of the two kinds of particles. The effect the pressure had was to lower the peak rising time of the particle. They found that when the pressure was lowered to about 1 torr the intruder particle did not rise at all. We would like to further explore the effect of pressure on the granular segregation dynamics.