On Mixing of Gases in a Laminar Chemical-Vapor-Deposition Reactor

Abstract

Introduction. Investigation of the effi ciency and time of mixing of gases in fi lling laminar chemical-vapor-deposition (CVD) reactors is a topical problem facing engineers and researchers. This issue is particularly important in the context of the development of a highly effi cient technology for production of carbon and silicon nanofi bers on metallic catalysts in CVD reactors [1–4]. A similar problem exists in cleaning the CVD-reactor atmosphere on completion of the step of nanofi ber growth. To produce nanofi bers, one must control gaseous mixing in a region smaller than a micron in size near the reactor wall. Clearly, in this case the existing engineering estimates of the time of cleaning or fi lling of the reactor volume are rough and inaccurate. Nonetheless, mixing in a region as small as this near the substrate is determined by the processes of convective heat and mass transfer and diffusion in the volume of the entire fl ow reactor. The present work seeks to experimentally and theoretically investigate isothermal and nonisothermal mixing in fi lling the atmosphere of a CVD reactor with the laminar gas-mixture fl ow. In the model experiments discussed here, nitrogen is a background gas in the reactor, and oxygen is a diffusing gas. Experimental Setup and Measurement Procedure. A diagram of the experimental setup is shown in Fig. 1. The setup is a classical fl ow CVD reactor. Hydrocarbon gas diluted with high-purity nitrogen (99.9 vol. %) is used as the raw material for production of carbon nanomaterials. The fl ow rate of each gas is prescribed in an independent manner using automatic controllers 1 (Omega FMA-5420 and FMA-5426 fl ow controllers). The fl ow-rate relation determines the concentration of the hydrocarbon raw material in the working mixture. The two gases are delivered by the common tube to the reactor, i.e., a quartz tube 3 passing through a furnace with electroresistive heaters 4. The length of the reactor is L = 0.738 m and the internal radius is R = 0.056 m. Inside the tube, at a prescribed value of the temperature within 600–900 o C we have the pyrolysis of the hydrocarbon gas and the liberation of carbon on a catalytic substrate 6 depending on the selected operating mode with the resulting formation of carbon nanotubes, nanofi bers, and virtually spherical nanoparticles. The spent gas mixture fl ows out of the reactor freely. The created control system makes it possible to hold prescribed values of the operating parameters and to carry out measurements in an automatic mode. Nitrogen is pumped through the reactor with a constant fl ow rate throughout the operating period of the setup, whereas hydrocarbon is only delivered in addition to it during the operating mode. Thus, in the initial step of the operating mode, we have the process of mixing of the gases on the way to the reactor and inside it. Analogously, once the delivery of hydrocarbon in the reactor is discontinued, the process of displacement of this gas by pure nitrogen occurs. The procedure of measurements for quantitative determination of the processes in question was as follows. To the channel of an empty reactor, we connected a Testo-350XL digital gas analyzer making it possible to continuously measure the concentration of a number of components of a gas mixture: O2, CO, CO2, NO, and others. An analysis of the concentration of