Surface Tension Induced Instabilities in Reduced Gravity: the Benard Problem

A Benard convection experiment has been set up, and the onset of convection in shallow layers of silicone oil two millimeters or less deep has been studied. The onset has been observed visually or has been determined by the break in the heat transfer curve which accompanies the onset of convection. The outcome of these experiments has been very surprising, from the point of view of theoretical expectations. The onset of convection at temperature differences far below the critical value for fluid depths smaller than 2mm was observed. The discrepancy between experiments and theory increases with decreasing fluid depth. According to theoretical considerations, the effects of surface tension become more important as the fluid depth is decreased. Actually, one observes that the onset of convection tales place in two stages. There is first an apparently surface tension driven instability, occuring at subcritical temperature differences according to conventional theory. If then the temperature difference is increased, a second instability occurs which transform the first pattern into conventional strong hexagonal Benard cells. The second instability is in agreement with the critical temperature gradients predicted by Nield.