Buoyancy effects

Bird et al. (1960, p. 527) assumed that buoyancy forces would only be important if the vapour was lighter than the gas ($\zeta>0$). This follows from the one-dimensional assumption (§2.6.2). If, however, the no-slip condition gives rise to radial concentration gradients at finite mass transfer rates, then buoyancy forces will always affect the velocity and vapour flux fields (Markham & Rosenberger 1980; §3.2.3).

For a light vapour, questions of convective stability are foremost. C. Y. Lee and Wilke (1954) pointed out that convection could be eliminated by making the tube thin enough. Sparrow and coworkers (Sparrow, Nunez & Prata 1985; Nunez-Testa 1986; Nunez and Sparrow 1988; Sparrow and Nunez 1988) studied the postcritical flow under the assumption of axisymmetry. Suehrcke and Harris (1995; Suehrcke, Harris & McBain 1996) investigated the transition criterion under the same restriction. It has recently been shown that the least stable modes for such a system are not axisymmetric, and that the assumption of axisymmetry leads to an estimate of the critical density gradient (Rayleigh number) that is seven times too high (McBain, Suehrcke and Harris, in press).