The identification of side reactions and by-products in process synthesis

• Main Author: Wahyu, Haifa
• Published: University of Edinburgh 1999
• Subjects: 547
• Online Access: http://ethos.bl.uk/OrderDetails.do?uin=uk.bl.ethos.663335

This work presents a systematic approach to the identification of possible by-products in process synthesis. The objective is to discover all possible reactions that might take place in a reaction system, in addition to already known main reactions. These side reactions may generate species in minor quantities which can be difficult to separate and therefore be into the recycle system. These species can thus remain in the process and accumulate. A literature survey indicated that generic languages for chemical reactions have been developed recently by other workers. These have adopted the concepts of fundamental chemistry with the aim of identifying major chemical reactions. For the purpose of the present work, strict interpretation of chemistry was not considered necessary since the aim was to generate minor species which <I>might</I> be formed. The work covers aspects of molecular representation, reaction generation and thermodynamic evaluation for screening. Molecular properties are estimated using group contribution methods. Unlikely species are ruled out using both stoichiometric and thermodynamic tests. The approach was first tested by comparison with experimental work, on the partial wet oxidation of p-coumaric acid and alkylation of toluene with ethanol. This shows favourable results. Process case studies were then undertaken including vinyl chloride monomer and the preparation of R-134a and R-22. These processes have been used to study the effect of by-product build-up. The study involved separation simulation to investigate the separability of the components. This confirmed that many potential by-products would be difficult to separate and might tend to accumulate in recycles. The results show that the approach has a promise for identifying possible side reactions, by-products and the problems that they will cause in a continuous process. Thus in the long term the approach affects the possibility of reducing and eliminating pilot plant studies in moving from laboratories to full scale processes.