The Kinetic Investigation of the Deposition of Alumina and Aluminosilicates from Mixtures of SiCl 4 , AlCl 3 , CO 2 , an

  • PDF / 765,960 Bytes
  • 6 Pages / 391.5 x 607.5 pts Page_size
  • 57 Downloads / 204 Views

DOWNLOAD

REPORT


141 Mat. Res. Soc. Symp. Proc. Vol. 555 ©1999 Materials Research Society

microbalance for continuous monitoring of the weight, and thin refractory wires placed along the centerline of the reactor. EXPERIMENTAL Chemical vapor deposition experiments are carried out in a vertical hot-wall reactor coupled with a sensitive microbalance (1 ug sensitivity), used for continuous monitoring of the weight of the deposit. The reactor provides 23 cm of heating zone. A detailed description of the CVD system for the case in which it is employed for SiC deposition from CH 3SiC13 and H2 mixtures is provided by Papasouliotis and Sotirchos [51, and the reader is advised to consult that publication for more details. A carrier-based mass flow control system is used for the supply of silicon tetrachloride in the reactor, whereas aluminum trichloride is formed in situ in a packed-bed

reactor (chlorinator), loaded with high purity aluminum granules. A hydrogen-hydrogen chloride mixture is supplied in the chlorinator, and AIC13 is produced through the reaction: 6HC1 + 2A1 -- 2AIC13 + 3H2

(1)

The water vapor needed for the hydrolysis of the metal chlorides is formed via the water gas-shift reaction. Small SiC or Si substrates (1.-1.4 cm length, 0.8 cm width, and 0.2 mm thickness) are used in the experiments, and they are hung from the sample arm of the microbalance at a distance of 4 cm from the top of the heating zone, with the deposition surface parallel to the flow of the reactive mixture, which enters the chemical reactor from the top. Experiments are also carried out on very thin molybdenum wires in order to obtain information on the profiles of deposition rate, deposit morphology, and deposit composition along the reactor. RESULTS AND DISCUSSION Figure 1 presents typical results on the variation of the deposition rate of the single oxides (A120 3 and SiO 2) and the codeposition process with the temperature in Arrhenius coordinates. For all three cases, it is seen that the rate increases with increasing temperature. When deposition is carried out at a temperature of 1000°C or higher, the deposition of silica proceeds at significant rates, but the decrease of the temperature from 1000 to 950 0C is followed by a dramatic reduction in the deposition rate. (The smallest deposition rate shown in the figure, of the order of 10-6 mg/cmý-min, lies within the limitations of our microbalance for small surface area (nonporous) substrates.) Further decrease of temperature led to insignificant weight changes even when the deposition was allowed to occur for relatively long time. The apparent activation energy, the slope of the InRd vs. l/T curve, decreases with increasing temperature for the case of silica deposition. Linear regression over the entire temperature range in which data are shown in Figure 1 gave an activation energy value of 71.5 kcal/mol, while a much lower value of 28.5 kcal/mol is obtained when low temperatures (