Design of reactive distillation systems using a process systems engineering approach

DOMANCICH ALEJANDRO O.; BRIGNOLE NELIDA B.; HOCH, PATRICIA MÓNICA

Salt Lake City, EEUU

Congreso; AIChE Annual Meeting 2007; 2007

American Institution of Chemical Engineers

During the last years, process intensification received a great deal of attention because of thepotential cost savings. By process intensification it is meant the significant improvement achievedby a process, at a lower total cost. For example, the entire plant design is upgraded by reducingthe number of units involved in a conventional process to obtain the same products. Reactivedistillation columns constitute a clear example, because it is possible to use a single piece ofequipment in order to perform two tasks: reaction and separation of products, otherwiseperformed separately.However, reactive distillation is not suitable for every process where reaction and separationsteps occur. Operating conditions, such as pressure and temperature of the reactive andseparation processes and perhaps other requirements, must overlap in order to assure thefeasibility of the combined process. This limitation can be overcome by fixing adequate operatingconditions in the cases where this is possible.Reactive distillation has been presented as a multifunctional reactor, where the reactive andseparation tasks are combined into a single unit, thus reducing investment costs. Theadvantages of using this configuration have already been reported (Stankiewicz, 2003). Athorough review on the design methods for reactive distillation can be found in Almeida-Rivera etal. (2004). However, a systematic analysis of the performance of reactive distillation columnschanging structural variables has not been presented previously. The simplest case exhibitsreactive and non-reactive sections with no interrelationship between them, except at theequipment level. Even when the interrelations between sections remain unchanged comparedwith traditional technology; significant savings (2.5 times as big as the equipment expenditure)can be obtained because of the smaller and cheaper plant needed to perform the same tasks. Aclear example is the Urea2000plus™ technology, by Stamicarbon B.V., where a pool reactor isused to combine the urea reactor, the carbamate condenser and the inerts scrubber. However,further improvements can be obtained by really integrating the tasks, on the following items:• On the reaction: because there is an equilibrium displacement, since the products are beingwithdrawn• On the separation: because of the changes in the driving force for mass transfer due to thereaction.This case is formally presented as a reactive separation, also called a separative reaction. Bothtasks occur at the same level. Several disadvantages can be overcome by using these combinedschemes, because the selectivity and the reaction yield increase. This makes it possible to avoidthermodynamic constraints, such as azeotropes, thus allowing a considerable reduction not onlyin the investment costs but also in the operating costs (energy, water and solvents, otherwiserequired as entrainers). On the other hand, the mathematical model for this piece of equipment ismore complex, and the requirements for simulating the combined process are higher. Non-linearcoupling of reaction, phase equilibria, and mass and energy transport can give rise to manyundesired effects, for example the appearance of reactive azeotropes (called arheotropes) andthe multiplicity of steady states.The performance of reaction with separation in one piece of equipment offers distinct advantagesover the conventional, sequential approach (Podrebarac et al, 1998). As advantages of thisintegration, chemical equilibrium limitations can be overcome, higher selectivities can beachieved, the heat of reaction can be used for distillation, auxiliary solvents can be avoided, andazeotropic mixtures can be more easily separated than in conventionally distillation. This maylead to an enormous reduction of capital and investment costs and may be important forsustainable development due to a lower consumption of resources (Al-Arfaj and Luyben, 2000).Some industrial processes where reactive distillation is used are the sterification processes, suchas the synthesis of methyl acetate (Agreda et al, 1990); and the preparation of ethers, like MTBE(Jacobs and Krishna, 1993), TAME and ETBE, used as fuel additives. An explanation for theoccurrence of steady-state multiplicity was provided by Hauan et al (1995). A thorough review onthe modeling aspects of reactive distillation can be found in Taylor and Krishna (2000), wherespecial emphasis is put on the differences between equilibrium and non-equilibrium basedmodels with their advantages and drawbacks. Doherty and Malone (2001) gave valuablecommentaries on future trends and challenges, and a thorough review on the design methods forreactive distillation can be found in Almeida-Rivera et al. (2004).We present the optimal design of a reactive distillation column, based on a model presented byAlmeida-Rivera (2005) and further modified by Domancich et al (2006). The model comprises acombination of pure separation stages and a reactive distillation sector, where both reaction andseparation take place at the same time.For this study, we use a rigorous modeling of the column, with an equilibrium model for thereactive sector. The model of the column includes integer variables for defining the total numberof stages, the number of reactive stages and feed locations. The economic optimization of thereactive distillation process is performed using a MINLP algorithm implemented in GAMS(Brooke et al, 2004) and the cost of the resulting scheme is compared to the cost of theconventional process considering that reaction and separation take place in different equipment,showing significant savings.