CHEMICAL VAPOR DEPOSITION OF THIN FILM MATERIALS FOR ...

However, there is sizeable run-to-run growth rate variation for films grown at a specific
temperature, particularly at higher deposition temperatures (up to 8%). The growth rate is
sensitive to the position of the substrate in the reaction zone. This is likely related to the
temperature and precursor concentration distribution in our hot wall tube reactor, which can be
approximated by a plug flow reactor (PFR). From the vaporizer gas outlet to the substrate surface,
the tube temperature gradually increases to the reaction zone temperature due to thermal
conductance while the precursor concentration decreases because of thermal decomposition and
deposition on the surrounding tube wall. The highest growth rate can be achieved over a
particular region of the reactor because of a favorable combination of temperature and precursor
concentration. By inspecting the color of the tube walls after each experiment, we found that
most of the deposition occurs in a small volume (tube length of ~2 inch) close to the entrance of
the reaction zone.
To better understand the effect of precursor concentration in the solvent on the film
growth rate, both half and double of the original concentration of the precursor solution (listed in
Table 4) were employed for growth at a substrate temperature of 600˚C, with all the other
process conditions held constant. The film growth increases almost linearly with increasing
concentration (0.39, 0.75, and 1.42 μm hr respectively). The film deposited with lower
concentration of precursors does not show any obvious change in surface roughness and
chemical composition. However, the enhanced growth rate obtained for deposition with the
higher concentration is compromised by increased surface roughness and an Fe rich
stoichiometry. From a mass-transfer limited assumption, it is possible that the increased
concentrations of Ni and Fe species in the hydrodynamic boundary layer affect their relative
diffusion rates,[85] however, mass transfer limitations are not easily determined in this set-up
81