(629ae) Esterification of Iso-Amyl Alcohol with Benzoic Acid: A Kinetic Study
AIChE Annual Meeting
2011
2011 Annual Meeting
Catalysis and Reaction Engineering Division
Poster Session of CRE Division
Wednesday, October 19, 2011 - 6:00pm to 8:00pm
As a result of global ethanol production by fermentation, the by-production of fusel oil, principally composed by amyl-alcohol, has increased proportionately. The implementation of a process to transform it into higher added value products (i.e. esters) presents high potential. The iso-amyl benzoate, as a derived iso-amyl alcohol ester, is attractive because of its fruity balsamic odor.
The esterification of benzoic acid with iso-amyl alcohol using p-toluensulfonic acid and Lewattit K2431 as catalyst was evaluated and the kinetic expressions were determined. The catalyst reactivity tests were performed varying temperature (95-126°C), alcohol:acid molar ratio (4:1-6:1) and catalyst concentration (0,5-1.5 wt%) at 560 mmHg. However, in the heterogeneous reaction experiments the catalyst concentration used was higher (1,5-2 wt %) and the agitation was established in 500 rpm in order to limiting the external mass transfer problems.
The experimental test number was discriminate using a response surface design of experiments. In accordance with the chemical equilibrium limitations of the reactive system, this effect was considered and included in the kinetic model.
Since the thermodynamics data to the system is so limited, calculations using group contribution UNIFAC method were carried-out and compared with reported experimental values (i.e. azeotropic data, pure substance properties, etc.).
The homogenous catalyst reached the equilibrium conversion (97%) at 8 h; faster than heterogeneous catalyst with a time of reaction of 24 h
The kinetic model for homogeneous and the heterogeneous reactions were expressed considering a typical expression for elemental reaction of equilibrium. The estimation of intrinsic parameters of specific constant of reaction rate was fitted using the NelderMead optimization method.
Comparison between estimated and experimental data shows in the case for thermodynamics an absolute error no longer than 3% and for the kinetic models a media error function no longer than 9%.