(635b) Sensitivity of Chemical-Looping Combustion to Particle Reaction Kinetics

Chemical-looping combustion (CLC) allows fossil fuels to be burnt whilst producing CO₂ undiluted by N₂. It is based on the redox cycling of an oxygen carrier, typically a transition metal oxide. Thus, for a hydrocarbon fuel:

            (2n + m) MeO + C_nH_2m -> (2n + m) Me + mH₂O + nCO₂.                   (1)

            Me + 1/2O₂ -> MeO                                                                             (2)

To utilise the reactions, two fluidised bed reactors, viz. an air and a fuel reactor, are used. The oxygen carriers are fed into the fuel reactor in their oxidised form, where the lattice oxygen in the solid reacts with the fuel via reaction (1). The gaseous product, after removal of H₂O, will be pure CO₂, suitable for sequestration. The reduced oxygen carrier is re-oxidised with air in the air reactor, reaction (2). Overall, the fuel has been burnt in air, but inherent separation of CO₂ has been achieved, at a fraction of the energy penalty associated with other techniques.

A significant challenge in CLC systems is determining experimentally the precise reaction mechanism and intrinsic kinetics of the particles.  This is because the reactions are usually fast and are often dominated by mass transfer effects.  Once the particles are in the looping reactor system, the detailed kinetics may however become masked by the large number of particles present, their range of sizes and the distribution of times that the particles spend in the two reactors.  If this is the case, a simple particle reaction model might be sufficient when modelling the overall system. 

The present research explores how sensitive the CLC system is to different particle reaction mechanisms and kinetics, with the aim of determining the extent to which they can be simplified. In order to do this, a general system is modelled, consisting of two coupled, well-mixed reactors with a steady circulation of particles between them.  The time the particles spend in each reactor is governed by residence time distributions.  Monte Carlo simulations have been used to determine the mean conversion and rate of reaction of particles as they leave the reactors.  The approach is similar to that used by Kimura et al (1979).  The oxygen carrier particles themselves have been modelled in detail, in which the expressions for intrinsic kinetics have been coupled with a multi-component flux model and an enthalpy balance.  In order to determine the sensitivity of the results, these detailed models have been compared with simple approaches where the particles are described by simple reaction theories; ranging from homogeneous-reaction to shrinking-core kinetics. 

The results show that the behaviour of the CLC system is insensitive to the precise kinetics.  This is especially clear when particles spend at least a significant fraction of their total reaction time in each reactor. This is an important result, because it is helpful for the future design and operation of the CLC process. Furthermore, CLC is just one example of a reactor-regenerator system.  Other examples include fluidised catalytic cracking (FCC), processes for the removal of H₂S, SO₂ and other gases from gasification or combustion processes, chemical looping reforming and calcium looping.  Thus, the insight that the CLC process is insensitive to the precise particle behaviour is directly applicable to these systems as well.