Analysis of heterogeneous catalytic reactions by nonlinear estimation

Abstract

AbstractIntegral conversion catalytic data have been studied by nonlinear estimation to determine whether the Langmuir‐Hinshelwood heterogeneous catalytic models are more valid than the simpler homogeneous noncatalytic forms, and whether it is possible to discriminate among models. The experimental system chosen for this study was the vapor phase dehydration of ethanol. Data for a similar hypothetical catalytic system were also generated and analyzed to highlight the effects arising in the ethanol dehydration case.Nonlinear estimation is shown to be a valuable adjunct in the study of reaction kinetics and mechanism. For the hypothetical system, the correct model was found to provide the best fit to the data. Linear regression, coupled with the usual criterion that models are acceptable if all parameter estimates are positive, did not provide this discrimination. For the experimental system of ethanol dehydration, as corroborated by the hypothetical system, such problems as degeneracy of the heterogeneous rate equations and comparatively large errors in measurement are shown to be obscuring factors in determining the best model. When these effects are involved, it is doubtful that the Langmuir‐Hinshelwood equations are more warranted than the simpler homogeneous forms.