Hu and El-Wakil (1974)

Hu and El-Wakil (1974) performed a series of experiments on the nonisothermal vapour transport across cuboids. In order to create a uniform vapour mass fraction at the hot wall, they forced liquid through a slot cut in the top of it. This caused two deviations from ideality. First, the liquified vapour streaming down the hot and cold walls meant that there was a nonzero tangential velocity at the boundary of the vapour phase. Second, the vapour was saturated at these boundaries.

They did account for the slip velocity at the hot and cold walls in their numerical simulations. They demonstrated that it: `strongly influences flow, temperature and concentration, causing strong asymmetry in them'; `may also cause transition from unicellular to tricellular motion'; and `enhances the convective contributions, thus the heat and mass transfer rates'.

While the saturated boundary conditions would have meant that the vapour was saturated in at least part of the interior (Close & Sheridan 1989, fig. 1b; McBain 1995, fig. 18; McBain, Harris, Close & Suehrcke 1998, fig. 4), they ignored the possibility of internal condensation in their numerical solutions. That the vapour was indeed supersaturated can be concluded from the fact that they had to heat the transparent front and back walls in order to be able to see through them--otherwise they were obscured by condensate!

Their two-dimensional numerical model included transpiration ( $\varPhi \neq 0$) but excluded interdiffusion ( $\mbox{\textit{Pr}}_I=0$). This is common to many of the studies reviewed here, but is inconsistent (§§2.3.2, 4.4, 6.3).