Masters Theses

Date of Award

8-1995

Degree Type

Thesis

Degree Name

Master of Science

Major

Metallurgical Engineering

Major Professor

Peter K. Liaw

Committee Members

Charlie R. Brooks

Abstract

Continuous fiber reinforced ceramic composites (CFCCs) produced by chemical vapor infiltration, due to their high temperature strength and damage tolerant behavior, are a candidate material for a number of high temperature applications. However, a major obstacle in the commercialization of this material is posed by the materials' high cost and limited thickness. In an attempt to reduce this cost and enable fabrication of thicker components, Oak Ridge National Laboratory developed Forced Chemical Vapor Infiltration (FCVI). This process has been used to fabricate fiber reinforced silicon carbide ceramic composites measuring over 12 mm. thick in less than 24 hours, as compared to a time span of weeks or months required to process material, a few millimeters thick using the isothermal, isobaric chemical vapor infiltration process (ICVI).

In spite of the significant advantages FCVI provides over ICVI, composites produced by FCVI may still be too expensive to be commercially viable. Furnace operating costs, during preform infiltration, make up the largest fraction of the composite's cost. Therefore, reducing the time to infiltrate the preforms would have the greatest impact on reducing the final cost of the composite.

In the current one-step process a SiC matrix material is deposited onto fibrous preforms at atmospheric pressure. Hydrogen gas is used to carry the reactant, methlytrichlorosilane (CH3SiCI3), into the hot zone of the reactor where it decomposes into SiC and HCI. A thermal gradient is imposed onto the preform, with reactant gases entering the cool-side of the preform and effluent gases exiting the hot-side of the reactor. Material initially forms in the hot side of the preform. As the preform densifies its' thermal conductivity increases, thereby driving the hot zone toward the cool-side of the preform. The process terminates when the preform densifies to the point where reactant gas can no longer flow through the preform.

A two-step process was developed to more efficiently infiltrate preforms of layered plain weave Nicalon® cloth. In the first stage the thermal gradient was adjusted to uniformly fill the micro-voids between fibers in the bundles. In the second stage, where bundle overcoating would not result in trapped micro-porosity, reactant concentration was increased to more effectively infiltrate the macro-voids between layers and bundles. By processing the material in two stages, final densities and microstructures comparable to the one-stage material were achieved, in 70% of the traditional processing time.

Georgia Tech. Chemical Vapor Infiltration Model (GTCVI), a three dimensional finite volume program for modeling chemical vapor infiltration processes, was used to gain a phenomenological understanding of the process and as a guide in identifying the relative importance of each of the involved variables. The modeling, in combination with experimental work, was integral in developing an overall process optimization scheme.

When comparing the modeling results to the experimental data a general pattern began to occur - the changes in process variables which reduced processing times also yielded lower densities. Process optimization came to represent a compromise between high final densities and short processing time. High final densities were a result of complete bundle infiltration and uniform macro-void filling. Short processing times resulted from fast infiltration rates. With this in mind, an overall optimization strategy was proposed.

First, the largest reductions in processing time should come from raising the temperature and increasing the internal surface area of the preform. From a practical point of view this means increasing the hot-side temperature to just below the thermal damage limit of the fiber and selecting preforms with the greatest internal surface areas. Using these fixed conditions the other variables should be adjusted to achieve the fastest overall infiltration rate possible for a set final density.

In the first stage, where fiber bundles are being infiltrated CH3SiCl3 concentrations should be kept as high as possible, up to a point just below where bundle over-coating is observed. H2 flows should be set at a point high enough that Si formation is prevented and the SiC deposition rate is not suppressed. Cool-side temperature should then be adjusted until a uniform infiltration rate is achieved throughout the preform.

In the second phase, where bundles have already been overcoated, CH3SiCl3, concentrations should be increased up to a point just below were over-coating at the edge of the preform would occur. Once again the thermal gradient should be varied to achieve uniform infiltration rates.

Through the use of the process modeling and computer process control, this two-step approach could be extended to a process where conditions are continually changed in response to the preform changing material properties. Additionally, since the process model data correlated closely with the experimental results, the effort required to extend these advanced processing techniques to larger scale components should be greatly reduced.

Download
Files over 3MB may be slow to open. For best results, right-click and select "save as..."

Share

COinS