| OCR Text |
Show 1.7.6 The results of the experiments conducted are plotted in Figs. 7-9, which also show the analytical results discussed in the following section. Table 1 is a summary of the average velocity, the residence time, and the diffusion coefficients for ammonia in argon and helium. The latter were calculated using the equations listed by Hirschfelder, Curtiss and Bird [73. IV. DATA ANALYSIS. The data obtained were analyzed using the two simple models described in [13. The fit that could be achieved using the simplest or 'Seery-Zabielski-Fenimore' model for the data obtained with argon as the diluent is shown in Fig. 7. It is seen that this fit is quite good. The fitting procedure followed is as described in [13, with the representative low and high temperatures being chosen as 850 and 1250 K, respectively. Fig. 7 is based on K3[OH3 = 3.41 x 105 exp(-i.ii x 10H/T) s"1 k1k3/k2 = 2.2 x 108 cm3/(gmole s), where the notation is that of [13. The latter value is about 40 times smaller than that obtained in C13 using the data of [2-43. It is believed that the difference arises because in the present work the reduction process at low temperatures was influenced by surface reactions (see below). This causes a shift of the temperature at which maximum reduction occurs. In turn, this results in a shift of the apparent value of kjk3/k2. Using the value k3 = 3.2 x 1 0 1 2 exp(-1067/T) reported by Silver and Kolb [83 and Fuji et al. [93, numerical results for [ O H ] ^ were estimated. They are given by [OH3Kins9.6 x 103 exp(-l.ll x 10H/T) ppm. At low temperatures, these are much larger than the values computed for [OH3eq, which are given by COH3e q = 7.88 x 10* exp(-i.95 x 10H/T) ppm. The difference consists of a factor of about 25 at 850 K, and of about 3 at 1100 K. The reductions at low temperature obtained with helium as the diluent (Fig. 9) are comparable to those obtained with argon as the diluent (Fig. 7). In order to investigate whether the unexpectedly large reaction rates at low temperatures with both argon and helium as the diluent may be due to wall reactions, use was made of approximate analytical results for first order reactions available in the literature C10-14, and references cited therein3. For present purposes, the most useful formulation of these results is that presented by Brown [133, who described a numerical method yielding the quantities of interest. The model used by Brown and others is based on the following assumptions: steady Poiseuille flow, constant pressure and temperature, trace reactant subject to first order chemical kinetics. With these assumptions, the solution for the concentration of the reactant is given |
