I. Thermal decomposition of N-butane. II. Flow in entrance section of parallel plates

Abstract

Part One The thermal decomposition of n-butane was investigated in a flow reactor at a pressure of 1 atm, in a temperature range of 460° to 560°C, and at low conversion levels, i.e. 0.06 - 0.68% for the 460° runs, 0.5 - 2.3% for the 510° runs, and 3.5 to 8.2% for the 560°C runs. Temperature, velocity, and concentration profiles at the exit end of the reactor were measured to study the effects of energy, momentum, and mass transports on chemical reaction. It was found after analysis of data that the reactor could be treated as an isothermal reactor with plug flow under the prevailing operating conditions. Two rate expressions were determined for the reaction; one corresponding to a first-order and the other to a second-order rate. They are First-order rate = 3.34 x 10^(12) e -54,600/RT (C_4H_(10) lb/ft^3 sec Second-order rate = 2.55 x 10^(14) e -56,800/RT (C_4H_(10)^2 lb/ft^3 sec These two expressions equally well represent the experimental data. On the basis of the products formed and the rates observed, a Rice-type, free-radical mechanism was proposed for the thermal decomposition of n-butane. The mechanism, which is presented in the section on correlation of data, quantitatively describes the reaction. One major feature of the mechanism is the consideration of secondary reactions at very low conversions. Part Two Flow of an incompressible fluid at the entrance section of parallel plates under isothermal, laminar conditions was investigated by solving the two-dimensional Navier-Stokes equations numerically. The Navier-Stokes equations were transformed into finite-difference equations in terms of stream functions ψ and vorticities ω with a technique developed by de G. Allen. The finite-difference equations were then solved by an iterative procedure on digital computers. From the solution, point velocities and pressure gradients were computed. Two cases were studied, both with a Reynolds number of 300. Case I had a flat velocity distribution at the entrance to the plates. Case II assumed that potential-flow conditions existed only far upstream from the entrance. For both cases, large velocity and pressure gradients were found near the leading edges of the plates, although they were comparatively smaller in Case II. Also the velocity profiles for small distances from the entrance were found to be slightly concave in the central portion between the plates. Schlichting and others have solved the boundary layer equation for Case I. Their solutions agree well with the present work at large distances from the entrance but deviate considerably near the leading edges as the boundary-layer equation does not describe the behavior of fluid flow near singular points.