A direct methanol fuel cell (DMFC) is an electrochemical energy-conversion device that converts chemical energy of liquid methanol into electrical energy directly. Because of its unique advantages, such as higher energy densities, facile liquid fuel storage, and simpler system structures, the DMFC has been identified as one of the most promising power sources for portable and mobile applications. Although promising, the DMFC technology is facing some challenging technical issues. For given electrolyte and electrode materials, the performance and operating stability of a DMFC are determined by the mass transport of different species in the cell. A better understanding of mass transport behaviors in DMFC is thus essential. However, the opaque materials that form the constituent components of the fuel cell make it rather hard to experimentally reveal the real-world details of how different species are transported and distributed inside the cell. Mathematical modeling, therefore, becomes essential to shed light on the mass transport processes in the DMFC.
This thesis focuses on numerically investigating the mass transport processes in the liquid-feed DMFCs. First of all, an integrated mathematical model that can simulate the intricate mass transport processes occurred in different regions of the DMFC was developed. The model consists of three sub-models. The first one that describes the two-phase flow and mass transport in the anode and cathode porous regions was formulated based on the classical porous-medium multiphase flow theory. As compared with conventional models, the present model in the porous structures eliminates (a) the assumption of constant gas pressure in the unsaturated porous medium flow theory, (b) the definition of the liquid-gas mixture pressure in the multiphase mixture model, and (c) the assumption of thermodynamic equilibrium between phases in the multiphase mixture model. The second sub-model was developed based on the drift-flux and homogeneous theories to simulate, respectively, the flow in the anode and cathode channels. The third one is a microscopic agglomerate model that was applied to the effect of microstructure of the catalyst layer. This integrated model is solved numerically using a home-written computer code based on SIMPLE algorithm with finite-volume-method. The validations of the model against both the experimental and numerical data reported in the open literature show that the present model can predict the cell performance with fairly good accuracy. The model can then be applied to investigate the mass transport behaviors of reactants (methanol, water and oxygen) and products (carbon dioxide and water) in a DMFC.
The liquid-gas two-phase mass transport of methanol through the DMFC anode was numerically investigated. Predictions by different models with various considerations on methanol transport were compared. The results indicate that the present model that considers two-phase methanol transport with interface phase exchange can yield a more accurate prediction of cell performance than conventional DMFC models that ignore the interfacial mass exchange. Particularly, it is shown that the assumption of thermodynamic equilibrium between liquid methanol and vapor methanol invoked in conventional models will overestimate the mass transfer of methanol to the anode catalyst layer, thereby resulting in an inaccurate prediction of cell performance. Besides, the numerical investigations show that the mass transfer of methanol to the anode catalyst layer is coupled with the counter transport of gas CO
2 from the anode catalyst layer to the anode flow channel. The presence of gas CO
2 in the anode porous region will reduce the transfer path for liquid methanol, thus increasing the overall mass transfer resistance of methanol from the channel region to the catalyst layer. Also, it is revealed that the gas-void (gas CO
2) fraction in the anode porous region is influenced by the gas-void fraction in the anode flow channel which sensitively depends on the methanol flow rate. A higher methanol flow rate will result in relatively lower gas-void fraction in the channel, thereby increasing the mass transport of methanol from the channel region to the catalyst layer. Finally, the overall mass transfer coefficient of methanol from the channel to the catalyst layer is calculated, which is in well agreement with the experimentally-determined empirical correlation reported in the literature.
The water transport behaviors through the membrane electrode assembly (MEA) were then numerically investigated. The effects of the design variables of each constituent component of the MEA on each of the three water crossover mechanisms, namely electrosmotic drag, diffusion and back-flow, were examined. Emphasis is placed on studying the ways of how to reduce the rate of water crossover so that the anode water loss can be reduced and the cathode flooding can be alleviated. The numerical results shown that lowering the diffusion flux of water to the cathode and enhancing the back-convection flux of water to the anode are both feasible in suppressing the overall rate of water crossover to the cathode. It is revealed that the reduction in the diffusion flux of water to the cathode can mainly be achieved by lowering the liquid saturation level (or increasing the gas-void fraction) in the anode electrode, which depends on the optimal design of the anode porous layers. As for the enhancement of the water back-flow to the anode, it can be realized by building up a much higher hydraulic liquid pressure in the cathode electrode, which relies on the optimization of cathode porous layers.
In addition, the coupled oxygen and electron transport through the cathode gas diffusion layer (GDL) was modeled. The influences of both the inherent anisotropic GDL and the anisotropy caused by the GDL deformation were examined. The results indicate that the anisotropy of the GDL significantly influences the local distributions in cathode potential and current density. It is found that the GDL with deformation results in an increase in the concentration polarization due to the increased mass transfer resistance in the deformed GDL. However, the ohmic polarization is found to be smaller in the deformed GDL as the result of the decreased interfacial contact resistance and electrical resistance in the GDL. This result implies that an appropriate clamping force needs to be achieved so that the cathode performance can be maximized.
Moreover, the dynamic response of a DMFC to the change in the cell operation conditions was numerically investigated with a transient two-phase mass transport model. Various physicochemical processes that may affect the cell dynamics were examined. The results show that the dynamic cell voltage exhibits a significant overshooting in response to a sudden change in the cell current. This is due to the dynamic change in the methanol-crossover flux in response to the change of cell current results, which causes a strong cathode potential overshooting. In particular, it is revealed that the slow response in the methanol transport through the porous electrode and the membrane is one of the dominant factors that influence the cell dynamics, implying that the assumption of fast mass transport processes that employed by other transient DMFC models is unrealistic and will lead to inaccurate predictions.
Keywords: Direct methanol fuel cells (DMFCs), two-phase flow, mass transport, methanol crossover, water crossover, overpotential, liquid saturation, non-equilibrium, phase-change, dynamic behavior, model
Post a Comment