Articles techniques et présentations

Increasing energy demands in society have led to intensive fossil fuel usage, resulting in increasing carbon dioxide (CO2) emissions, one of the most significant greenhouse gases and major contributors to global warming. Considerable efforts have been made towards CO2 capture, storage, usage and conversion (CCUS) processes. One such technology of increasing interest is non-thermal plasma (NTP) CO2 reduction, due to its effectiveness at ambient temperature and pressure, unlike thermal plasma technologies. Understanding the effectiveness of CO2 conversion to carbon monoxide (CO) and oxygen (O2) in an NTP, as accurately as possible, is key. In this work, a non-thermal plasma with a dielectric barrier discharge (DBD) for CO2 conversion with argon (Ar) as the diluent gas is studied with COMSOL Multiphysics®. This study is coupled with previous contributions tackling thermodynamic and fluid-dynamic behaviors in the reactor (of pure compounds/mixtures), as well as experimental results.

The NTP is generated by applying a potential difference between two electrodes in the reactor, filled with a CO2/Ar mixture. When electrons are accelerated by the electric field generated, they collide towards gas molecules, resulting in reactions and collisions (excitation, ionization, and dissociation). The dissociation collisions create radicals, forming new compounds (CO and O2). The plasma physics study of this reactor implements the COMSOL® solver and the Plasma Module, solving a series of equations: Drift diffusion, Heavy species transport, Bulk gas flow transport, etc.

Firstly, the reactor is drawn as a 1D geometry (coordinates): dielectric walls (left/right) and plasma in the middle. Secondly, the dielectric material is specified with relative permittivity of quartz. Thirdly, input variables are provided: applied frequency, voltage, reactor dimensions, plasma temperature, etc. In addition, only the Ar cross-section (imported: Phelps database) is required to define the plasma reactions. Additional reactions (in plasma Domain) and specifications of species (e.g., initial number density of Ar+ equal to initial number of e-) are introduced. Boundary conditions are set as surface reactions, walls, and charge accumulation on both dielectric points. The ground and terminal are specified for each dielectric wall respectively (anode/cathode). Finally, a mesh with sufficient element size is defined, prior to computing the Time Dependent study. A comparable model can be found with Argon 1D Dielectric Barrier Discharge example (COMSOL Application Library).

The Excited Ar Mass Fraction shows the generation of plasma after the first start-up cycle. The discharge reaches a periodic state solution only after two RF cycles. The Electric Potential and Field indicate a homogeneous behavior across the discharge gap. The Electron Density shows a large initial spike, at the start of the plasma, then reduced to a continuous flow. The obtained results present a correct description of the plasma behavior involving Ar (ground state, metastable and ions), allowing further reactor optimization to maximize CO conversion with limited computational burden. Future work of this plasma model could aim at: introducing user-default thermodynamic characterization of involved compounds, modelling a 2D geometry, inserting copper-based catalysts, etc. all possible with the COMSOL Multiphysics® software.