Performance of gas saturators in the presence of exit stream temperature gradients and implications for chemical vapor deposition saturator design

Abstract

When a carrier gas is passed through a gas saturator at temperature T 0 containing liquid or solid reagent under conditions leading to its saturation with an equilibrium vapor pressure p∘ of the reagent, and then into a tube where the exit stream temperature is increased to T ∞, the downstream partial pressure of reagent, p ∞, may be less than p∘ due to thermal diffusion of the reagent driven by the temperature gradient. The mass transfer process has been modeled for two cases: (1) a linear temperature gradient over a tube length L 2 and (2) a tube length L 1 of constant temperature T 0 followed by a tube length L 2 with a linear temperature gradient. Solution of case (1) is defined by the dimensionless Peclet number Pe 2 = υ L 2/ D and a dimensionless “thermal diffusion number” Td = α ln ( T ∞/ T 0), where υ is average gas velocity in the tube, D is the reagent gas—carrier gas binary diffusion coefficient, and α is the reagent gas—carrier gas thermal diffusion factor. Solution of case (2) is characterized by Td, Pe 2, and Pe 1 = υ L 1/ D. Chemical vapor deposition processes used in the fabrication of electronic and optoelectronic semiconductor devices require the reproducible control of reagent partial pressures to better than ± 0.4%. Gas saturators for these processes are commonly operated under conditions where p ∞ is depedent on both flow rate and the downstream temperature profile, and where p ∞ can be as much as 10% lower than p∘. The reproducible control of reagent partial pressure is best brought about by designing the saturator system so that Pe 1 > 10, conditions leading to p ∞ ≅ p∘ independent of flow rate or downstream temperature.