Previous studies based on bifurcation theory have shown that the key features of an ideal OCM reactor are high thermal backmixing (i.e. high effective thermal conductivity) and low species backmixing (i.e. narrow residence time distributions). Narrow residence time distributions, i.e. plug flow behavior, is necessary to control and maximize the selectivity towards the intermediate products ethane and ethylene. High effective thermal conductivity creates the opportunity to exploit the bifurcation behavior and operate an OCM reactor autothermally, in this way utilizing the reaction heat in the best possible way. Both these characteristics can be obtained in the gas-solid vortex reactor (GSVR) that is studied in this work.
Gas-solid fluidized beds are generally known for their enhanced heat and mass transfer characteristics and are therefore widely used for both reactive and non-reactive processes in the chemical industry. The efficiency of heat and mass transfer in fluidized beds is determined by the relative velocity, the so-called slip velocity, between the gas and the solid phase. As a consequence of the balance between the drag force and the gravitational force, the slip velocity in a conventional gravitational fluidized bed is limited by the terminal free-fall velocity of the particles. Higher gas velocities give rise to the formation of bubbles and slugs, which negatively affect the heat and mass transfer efficiency as the extensive gas bypass decreases the gas-solid contact. If the gas flow rate is increased even more, particle entrainment is unavoidable. Higher gas throughput, more uniform fluidization, higher slip velocities, and hence better heat and mass transfer, can be achieved by using a centrifugal force instead of the gravitational force. A centrifugally fluidized bed thus emerges as an excellent candidate for process intensification. There are two possibilities to achieve a centrifugal fluidized bed: using a rotating fluidized bed (RFB), where the particles are set in motion by rotating the operating vessel itself, or using a gas-solid vortex reactor (GSVR), where the particles are introduced in a swirling flow field of azimuthally injected gas in a static vessel. In a RFB, the tangential and radial velocity components can be varied independently, as the rotational velocity of the vessel and the injected gas flow rate can be controlled in a separate way. The downside of using RFBs however is the use of mechanically moving parts which imply the risk for mechanical abrasion. In a GSVR the fluidizing gas is injected tangentially through multiple gas inlet slots in the cylindrical wall of the fluidization chamber. The swirling gas transfers tangential momentum to the particles, which in turn start rotating and experience an outward centrifugal force. A fluidized state is reached when the radially inward drag force exerted by the gas overcomes the apparent weight of the solids in the centrifugal field. In contrast to the RFB, it is impossible to control the particle velocity components in a GSVR independently. Nevertheless, the GSVR is preferred to the RFB as the absence of moving parts makes the unit more appropriate for scale-up. At this moment, the first reactive GSVR unit is being tested at the Laboratory for Chemical Technology (LCT). Because of its excellent heat transfer characteristics and narrow residence time distributions, this reactor emerges as an ideal technology for OCM.
The reactive GSVR setup at the LCT consists of a cylindrical unit positioned along a vertical axis with eight gas injection slots of 1 mm width, equally distributed over the circumferential wall and tangentially inclined at a 10° angle. A reactor diameter of 80 mm and a reactor length of 15 mm, are defined. The design of the bottom end wall and exhaust profiles was based on preliminary calculations and CFD simulations. The design of the reactor is such that the inlet slots are easy to replace, so that the number and the angle of the inlet slots can be altered for parametric studies etc. For a given gas flow rate, the thickness of the injection slots determines the magnitude of the gas injection velocity, while the injection angle determines its radial and tangential components. The GSVR is currently operated in semi-batch mode. Solid (catalyst) particles are already present in the inner chamber at the beginning of an experiment. Then the hot fluidizing gas is introduced until a pseudo-steady-state swirling gas-particle flow is established in the reactor. Both PIV and infrared measurements are available for validation of CFD results. Preliminary reactive OCM test experiments are being performed as well. The product composition is analyzed using gas chromatography (GC).
In this work, the open-source CFD package OpenFOAM is used to simulate an adiabatic gas-solid vortex reactor using an Euler-Euler approach. A detailed microkinetic model is used, which was developed at the Laboratory for Chemical Technology for different types of catalyst (Sn-Li/MgO, Mn/Na2WO4/SiO2, Sr/La2O3). The model contains 26 elementary reactions on the catalyst surface and 39 gas phase reactions. Cantera is used as mechanism interpreter and linked to a newly developed OpenFOAM solver. The simulation method is validated by comparison with experimental data. The effect of operating conditions and reactor geometry is evaluated. The CFD simulations show that the narrow residence time distributions and efficient heat management in the GSVR can give rise to C2 yields as high as 25 %.