distillation boundary

predicts a top product that is almost pure methanol for the investigated distillation process, the result from the model, which takes into account reaction kinetics, shows that the top product contains considerable amounts of formaldehyde and water, which is in agreement with the findings from technical distillations.

The formaldehyde + water + methanol system is just one example for many technically important systems, which are intrinsically chemically reactive. Other examples include aqueous solutions of weak electrolytes, such as amine solutions used to scrub carbon dioxide and other sour gases from gaseous streams, or the solutions containing bases used for chemical extraction of acids from aqueous streams. In many of these cases, the key to the development of predictive thermo-dynamic models is a quantitative model of the often complex chemical reactions in those mixtures. The necessary information often only can be obtained using spectroscopic methods [25],

In all these cases, it is essential to use thermodynamic consistent models. As an example, modeling results for the azeotrope in the formaldehyde + water system are discussed in more detail here using Fig. 4.14. Note that this azeotrope is in fact one of the technically most important examples for a reactive azeotrope. The left side of Fig. 4.14 shows results from a thermodynamic consistent model of the system formaldehyde + water [22]. At 140 °C the pressure is plotted as a function of the overall formaldehyde concentration xFA. There is a low boiling azeotrope at about xFA = 0.5 mol/mol. The right side of Fig. 4.14 shows the result that is obtained, when in that model arbitrarily K is set to unity, while the activity coefficients are still used in the phase equilibrium calculation. Hence, the resulting model is inconsistent. It is not important here that the vapor-liquid equilibrium calculated with that modified model is strongly shifted, which only underlines the influence of liquid-phase non-idealities in that system. The important message from the plot on the right side of Fig. 4.14 is that applying the inconsistent model leads to a completely erroneous shape of the dew and bubble line. The dew and the bubble line do not touch with a horizontal tangent as in the picture of the left hand side of Fig. 4.14 (which is a result of the Gibbs-Konovalev theorem applied to reactive solutions in equilibrium) but intersect twice. Formally, there are two concentrations at which the condition for the reactive azeotrope xFA = yFA is fulfilled and two other concentrations at which the dew and the bubble line pressures are maximal. Obviously, it is not even worth to try to interpret this and a use of such an inconsistent model can not be recommended.

The examples given in this and the previous sections underline the necessity for using thermodynamic consistent models of chemical equilibria and phase equilibria. It should be kept in mind that this also applies to reaction kinetic models, which always contain information on the chemical equilibrium. It should, however, be taken into account that there is also a price to pay for the advantages of thermodynamic consistency: the evaluation of phase equilibrium data and chemical equilibrium data can no longer be carried out separately and any change either in the chemical reaction model or in the phase equilibrium model will affect the other model too.

Was this article helpful?

## Post a comment