engleski

Mathematical Modeling of a Heating System for a SnO_2 CVD Reactor and Computational Fluid Dynamics Simulations (3D-CFD)

We studied the heat transfer of a continuous system that is being designed for the deposition of tin oxide thin films on soda-lime glass substrates of large dimensions (400 mm x 400 mm x 2 mm). This system will operate with infra-red heating (parallel mounted ceramics IR lamps in one unit) and will substitute the present CVD system, which is used in the CenPRA display pilot plant to coat substrates up to 150 mm x 150 mm x 2mm. In this system the substrates are heated by thermal conduction and the deposition is done by an "in house" process called vapor decomposition process (VDP). The infra-red heating has advantages to be adopted as faster processing, higher uniformity of temperature at the glass surface, higher lifetime of the system, shorter maintaining time and easier cleaning of the furnace. In this poster we present the first calculations of the necessary time to heat moving substrate up to 400°C as this is the optimal temperature for the deposition of highly transparent and conductive layers of SnO_2 for displays applications. Assuming that the heat flow is an unique flux between two flat and parallel surfaces, the temperature as a time depended function can be calculated. By solving the equation using MatLab package, Simulink, we could get the logarithmic behavior of the temperature, and determine that we need 23 s to heat soda-lime glass substrates of 2 mm thickness at 400°C the heat source covers all the substrate area. Using FLUENT software tools we simulated the VDP process at atmospheric pressure, taking in mind that the system consists of two parallelepiped-like chambers: the furnace containing the IR lamps in which the glass is heated, and the chamber where the deposition occurs. For this simulation, in a first approach we adopted that the substrate temperature remains constant during the deposition. Using GAMBIT, we created a real geometry assembly to which we applied the mesh function by dividing the volume into infinitesimal cubes called "control volumes". This approach enables the resolution of the elementary differential equations for mass diffusivity, full multi-component thermal diffusion effects and reactions on the heated substrate. Defining the boundary conditions for inlet, outlet, reactant surface, surrounded walls and the thermal, chemical and mass flux properties, we could establish the final mesh which was exported to FLUENT. That was used for solving the huge amount of differential equations (each control volume has his own set of equations) by iteration principle. The simulations permitted to get a 3D perspective of the deposited layer along the substrate surface and also to calculate the VDP process parameters as temperature, pressure, reaction rate for each species (O_2, Cl and HCl) and gas speed along the chamber. Although we used very simple chamber geometry, the simulation showed to be helpful if we intend to optimize the uniformity of the deposited layer, throughput and the yield of the coating glasses of large area.

mathematical modeling; heating system for a SnO_2 CVD reactor; computational fluid dynamics simulation

nije evidentirano