The goal of volume production of very small diameter silicon-carbide fibrils at a price that is not cost prohibitive is a real possibility using a novel microwave processing technique.

Reaction zone during microwave-assisted silicon-carbide fibril growth in a cylindrical aluminum-oxide boat. Glowing annulus is the catalyst layer reacting to microwave energy.

Very fine diameter silicon-carbide (SiC) fibrils have excellent high-temperature properties that make them attractive for use in a variety of high-temperature applications such as reinforcements in metals and ceramics (ORNL has verified the use of SiC fibrils as reinforcements for Al2O3, for example) and high-temperature filter media. Key properties of interest include an elastic modulus of 84 x 106 psi (579 GPa), a tensile strength of 2,300 ksi (15.8 GPa) and good oxidation, chemical and creep resistance at temperatures to 1600 C (2910 F).

The U.S. Department of Energy (DOE) wants to use SiC fibrils as reinforcements in fiber-reinforced silicon-carbide matrix composite heat-exchanger tubes, which would be fabricated using chemical vapor infiltration (CVI). Long fibrils can be spun and CVI coated for the high-temperature tubes. In addition to this application, fibrils are being considered to solve some of DOE's Fossil Energy Program material challenges including improving the creep strength of combustion-chamber refractory tiles, producing high-temperature filter media for combustion gases and improving the toughness of refractory metals.

Also, companies are looking at using SiC fibrils in commercial applications including reinforcing CVI silicon carbide for heat management in silicon-carbide computer circuit boards, replacing hazardous SiC whiskers with a nonrespirable product in metal-cutting tools and using SiC fibrils as high-temperature filter media in diesel exhaust, chemical processing, and fossil energy-plant emissions.

Fig 1 Schematic of silicon-carbide fibril microwave reactor

Fibril process development

Companies have pursued the manufacture of SiC fibrils (vapor-liquid-solid or VLS, whiskers) since 1965. Carborundum was the last to do so, and achieved a projected volume price of $2,000/lb. The major limitations of the current state-of-the-art fibril growth are the high temperatures required (1600 to 1700 C, or 2910 to 3090 F), the slow fibril growth rate (~0.17 mm/hr) and the large quantity of excess of expensive methyl trichlorosilane (MTS) gas, which is wasted [1]. The commercial process is complicated by the processing of large quantities of hydrogen gas at high temperatures and the generation of corrosive hydrochloric acid [2].

In 1999, ReMaxCo Technologies verified a novel microwave-based VLS silicon carbide-fibrils concept. Current work continues the proof-of-concept, microwave-based VLS process. In the process, the catalyst is heated to the experimental temperature (1200 to 1300 C, or 2190 to 2370 F) while a mixture of MTS and hydrogen are introduced into an aluminum-oxide (Al2O3) boat. The MTS is dissociated and the carbon and silicon components are dissolved into the catalyst. The catalyst saturates and precipitates silicon carbide onto the surface of the reaction boat. These experiments yielded fibril growth rates of 0.75 mm/hr, which is an improvement of approximately 4.4 times faster than the best graphite furnace runs [3]. It also was demonstrated that some volume of fibrils could be produced on the surface of the aluminum oxide boats.

Research was sponsored by the U.S. Department of Energy Fossil Energy Advanced Research Materials Program. This project continues the commercial process development to a pilot-scale commercial reactor, which will lead to sufficient quantities of fibrils to allow expanded work by Oak Ridge National Laboratory (Oak Ridge, Tenn.) and others on heat exchanger tube development.

A semicontinuous, microwave-heated, vacuum reactor was designed, fabricated and tested in these experiments (Fig. 1). The major obstacle that had to be overcome during the project was the performance of the reactor. The original design of the reactor focused the microwaves in such a manner that they missed the catalyst/fibrils growth zone. The microwaves did react with the insulation and the reactor was heated by coupling with the insulation. Modifications were made to the reactor to focus the microwaves on the catalyst.

SiC Fibrils were produced using both MTS and Starfire SP4000 (produced by Starfire Systems) as feed-gas precursors. Both precursors produced fibrils at temperatures of less than 1000 C (1830 F). The new Starfire SP4000 produced fibrils as low as 800 C (1470 F) without the use of hydrogen and without producing the hazardous hydrochloric acid.

Fig 2 TEM micrograph of silicon-carbide fibril growth using MTS gas; fibril size is 2 to 5 um; Fig 3 TEM micrograph of perfect single-crystal silicon-carbide fibril grown using MTS gas.

Technology approach

The boundary conditions for the experiments were determined by running a computer thermodynamic analysis on the raw materials system. Cylindrical (7.6 cm diameter x 7.6 cm long) high-density aluminum-oxide reaction boats are coated on the inner surface with a catalyst and placed into the reactor under a light vacuum. The microwave reactor is evacuated to approximately 30 torr and flushed with nitrogen gas at a pressure of 150 torr. After the flush, the furnace is backfilled with hydrogen gas to a pressure of 150 torr and maintained at less than 180 torr throughout the microwave fibril-growth run. Ferrous silicon and iron powder catalysts (and several mixtures thereof) were tested. Fibril catalyst-seed paint is prepared using metallurgical grade -325 mesh ferrous silicon mixed in a dispersant paint from YZP Corp. in a 1:1 ratio.

A series of reaction boats were run (one at a time) through the reactor. Each boat is preheated using resistance heaters to a temperature between 850 to 900 C (1560 to 1650 F), and then moved to the microwave heated section where each of two 2-kW microwave sources is stabilized at 1.8 kW. The catalyst is heated to a temperature of 1200 to 1300 C (2190 to 2370 F), while introducing a mixture of MTS and hydrogen into the catalyst-coated area of the boat. The MTS forms the carbon and silicon components, which dissolve into the catalyst to grow the fibrils.

MTS reaction gas is generated by bubbling hydrogen through liquid MTS in a steel container. Replacing the steel container with a transparent, heated glass bubbler allows the operator to view the hydrogen flow through MTS liquid and control the vapor pressure of the MTS gas. Hydrogen flow is passed through the MTS bubbler at a rate of 0.13 liters/min for a period of one to three hours.

Process optimization experiments were conducted after achieving consistent furnace operation to achieve fibril growth and define operating parameters. Ferrous silicon was replaced with iron particles, then a mixture of 50% ferrous silicon and 50% iron by weight. Fibrils produced using optimal operating parameters were analyzed at ORNL, and the results from these experiments were used to design a second-generation microwave reactor. Experiments also were conducted to identify a less hazardous raw material gas than MTS. MTS liquid was replaced with a SP4000, a polysilymethylene CVD silicon carbide precursor. The SP4000 can be reacted in nitrogen gas rather than the more dangerous hydrogen required by the MTS liquid and gas.

Fig 4 SEM micrograph of silicon-carbide fibrils using SP4000 gas in nitrogen. 2,000X; Fig 5 SEM micrograph of SP4000 fibril from growth ball.

Test results

The only fibril growth in the initial microwave field configuration occurred after being in the microwave growth chamber for approximately three hours. Measured microwave intensity in the fibril growth area was zero. Rebuilding the furnace to focus more of the microwave field in the fibril growth zone improved the fibril growth quality and time. Figures 2 and 3 shows fibrils (2 to 5 um in size) grown in this sequence. The fibril quality is good, but the fibril yield is very low.

Issues with the fibril reactor that need to be addressed, and that will be implemented in the next phase of this work include:

  • Improving microwave field uniformity
  • Using flat ceramic plates in place of cylindrical ceramic boats in which catalyst paint flakes off the top and sides
  • Implementing a more accurate mass flow controller and a manifold-mixer for the reaction gas distribution because the current MTS gas feed mechanism had very little mass flow control and an irregular feed pattern to the fibril growth zone


One of the problems with scaling the fibril development to a large-scale commercial process is the generation of significant quantities of hydrochloric acid in the off-gas stream. The acid destroys the vacuum system and the exhaust ducts. Silicon-carbide fibrils made using the SP4000 in nitrogen produces no acid in the offgas. An unexpected advantage is that fibrils grew at a temperature of 850 C (1560 F) compared with required a temperature of 1200 to 1300 C (2190 to 2370 F) for the MTS reaction. Fibrils 5 to 15 um in diameter grown in the SP4000 experiments are shown in Figs. 4 and 5. Melt growth balls were observed with the fibrils indicating that they were VLS.

Conclusions

Silicon-carbide fibrils can be produced at temperatures as low as 850 C (1560 F) compared with the 1700 C (3090 F) in previous graphite furnaces, and fibril growth rate has been increased by a factor of four over previous technologies. Microwaves play a major role in improving the fibril-growth process. The SP4000 silicon-carbide precursor provides a reaction without hazardous offgas products, with a projected future volume cost of $600/kg.

An added benefit to the development of this microwave process is the ability to also economically produce TiN, TiB2 and TiC whiskers.

It is feasible to scale up to a commercial process by overcoming the equipment engineering problems encountered on this project including obtaining a uniform microwave field, good control and uniform distribution of reactant gases, and the use of flat ceramic reaction boats. These improvements will get the commercial process closer to the project goal of a fibril price of $300/lb. With these improvements in place, previous technology can be improved to accomplish lower energy consumption, higher growth rates, reduced reactant gas waste, lower cost raw materials and consistent high-quality fibril product.

A new pilot-scale reactor will be designed and fabricated incorporating these improvements, which will be capable of producing 200 g of fibrils per day. The reactor will be tested and then operated to supply sample quantities to various researchers in the Fossil Energy Materials Program. A production reactor capable of producing 200 g of fibrils per day will be designed, and will be built if a commercial application evolves. A scale-up to production volume will cost approximately $3 million.

Acknowledgement

The author acknowledges the support of the U.S. DOE Fossil Energy through the Advanced Research Materials Program under the project direction of Dr. Roddie R. Judkins for funding this work. Gratitude is expressed to Oak Ridge National Laboratory's High Temperature Materials Laboratory for the electron microscopy work of Larry Allard and Larry Walker and to Microwave Materials Technology for microwave equipment engineering and fabrication.