(405a) Sublimation As a Function of Diffusion

In the sublimation of organic light emitting diodes (OLEDs), OLED vapor is transported from high temperature to colder temperature. The pressure at high temperature is close to the equilibrium vapor pressure of the material being sublimed, and the vapor pressure in the colder region equals or exceeds the vapor pressure at the wall temperature. These results are consistent with sublimation controlled by diffusion. If diffusion is rate limiting, the reciprocal of the flux should vary linearly with the length. If diffusion is rate limiting, the total amount sublimed should vary with the square of the tube diameter.  These expectations are supported by experiment.  Predictions for convective flow and Knudsen diffusion are not supported by experiment.  Diffusion does seem the key.

  However, the success of these predictions does not mean that the sublimation mechanism results from intermolecular diffusion or even from self-diffusion. In intermolecular diffusion, there are two solutes; here, there is one. In self-diffusion, there is one solute, but a portion of this solute can be separately identified by radioactivity, activation, or even optical rotation; here, no such identification is made. In diffusion in porous catalysts, the second species for diffusion is sometimes identified as the catalyst itself, in what is called a “dusty gas model.”  Here, because the mean free path is much less than the tube diameter, this dusty gas model is inappropriate. Moreover, intermolecular diffusion must be measured relative to some form of velocity, like a mass average or a volume average velocity. In many intermolecular diffusion experiments, this velocity is zero. In sublimations like those studied here, there is a real non-zero velocity, which is the sublimation rate. While diffusion is a quick label for what is occurring, diffusion in the conventional sense is not involved.

 This paper discusses the causes of this behavior.  They may arise from a mass balance, from inertia, or from bulk viscosity.  Which of these mechanisms is most likely is discussed.