(100c) How Active Is Too Active? a Catalyst Selection Study

When selecting a catalyst formulation as part of a new reactor build, the process can be designed around the catalyst such that nearly any commercially-viable catalyst can be used. For these systems, higher activity catalysts generally give rise to higher product rates or reduced capital investment. However, most of industry’s active processes are not newly built, but have been in service for many years (if not decades). For these systems, higher activity catalysts can also enable higher product rates, but typically only to the extent constrained by the original process design. Indeed, many if not most of today’s active systems are now multiply constrained, such that it is not always true that increasingly higher activity catalysts will yield better performance.

As one example, many fixed-bed, vapor-solid reactor systems are constrained by heat transfer. For these systems, a circulating coolant absorbs the heat of reaction thereby controlling the temperature rise and making possible higher product rates. In this context, there is generally a maximum achievable product rate corresponding to the maximum heat removal capability of the system. This maximum product rate is itself a function of the catalyst (in particular, the catalyst activity and selectivity).

In choosing from next-generation, high activity catalysts for a system constrained by heat transfer, the impact of varying activity and selectivity must be balanced, taking into account also the free variables for optimization and their bounds. Here, we utilize detailed reactor simulations to investigate the catalyst selection problem for one such industrial system. Our steady-state simulations incorporate mass and energy balances and reaction kinetics while predicting temperatures and concentrations that vary in two spatial dimensions. We begin by demonstrating how such simulations, when informed by laboratory experiments, can predict the heat transfer-constrained operation as observed in the plant. Then, we utilize the tuned reactor model to explore how process optimization accounting for a catalyst change (e.g., changes to the base reaction rates, activation energies, and partial reaction orders) leads to varying and sometimes counterintuitive changes in reactor performance. To finish, we develop simplified analytical models from first principles and compare their predictions to those generated using the full reactor model. These simplified models can be utilized as a coarse guide for catalyst selection in the case that more detailed and highly tuned reactor models are not available.