Gas-liquid diffusion layer

Considerable changes are needed in the anodic part of the membrane-electrode assemblies in order to accommodate the first two of the above-mentioned points. Instead of the porous gas diffusion layer that in polymer electrolyte membrane fuel cells ensures a uniform distribution of hydrogen across the surface, a gas-liquid diffusion layer that contains a set of hydrophilic as well as a set of hydrophobic pores is needed here. Through the hydrophilic pores, this layer must secure the unobstructed access of the aqueous methanol solution to the reaction zone and its uniform distribution. Through the hydrophobic pores, this layer must secure the unobstructed elimination of carbon dioxide, as the gaseous reaction product, from the reaction zone. Analogous changes must be made in the catalytically active anode layer of the membrane-electrode assemblies, where the gas is actually formed, and must be removed toward the gas-liquid diffusion layer. 

By introducing carefully determined amounts of Naflon and polytetrafluoroethy-lene into the gas-liquid diffusion layer and into the catalytically active layer, the ratio needed between hydrophilic and hydrophobic micropores is achieved. 

The clathrate hydrate growth model presented by Englezos and Bishnoi is based on crystallization and mass transfer theories. It describes the growth of the hydrate as a three step process. The first step is the transport of the gas molecule into the liquid phase. The second step is the diffusion of the gas molecule through a stagnant liquid diffusion layer which surrounds the hydrate particle. The last step is the incorporation of the gas... 

The best possible mode of gas-liquid contact for a given process depends upon a combination of effects, including hydrodynamics, mass transfer, and chemical kinetics. In treating this combination, a dimensionless parameter P has been defined as the ratio of total volume of the liquid phase to volume of the liquid diffusion layer. Krishna and Sie reported general values of jS to be 10 0 for thin liquid films and liquid sprays, and 10 -10" for gas bubbles within a continuous liquid. The relative rates of mass transfer and chemical reaction show whether high values or low values of jS best utilize available reactor volume. 

The equilibrium partition state is not established instantaneously, not only because of its dependence on the disproportionation of HOI, but also because of the mechanisms of iodine transport in the liquid and gas phases. Under isothermal conditions, i. e. when both phases are at the same temperature, I2 molecules present in the water phase have to be transported by diffusion to the gas-liquid boundary layer from where they can pass over to the gas phase. When, however, the liquid phase shows a higher temperature than the atmosphere, diffusion transport will be supported by convection in the water phase. In the case of a boiling iodine solution the rate of I2 carry-over to the gas phase is greatly enhanced as a consequence of the vigorous convection within the solution. Usually, the time taken to reach the equilibrium state of iodine partitioning is not of significance for the situation inside... 

It should be noted that in a vapour phase the liquid layer on the surface of a sensitive element of the sensor (zinc oxide) must be sufficiently thin, so that it would not produce any influence on the diffusion flux of oxygen through this layer. Possible lack of the film continuity (the presence of voids) does not prevent determination of concentration of oxygen in the bulk of the cell by the vapour - gas method. In this case, one deals with a semi-dry method. On the contrary, the presence of a thick liquid layer causes considerable errors in measuring t, because of different distribution of oxygen in a system gas - liquid layer -semiconductor film (this distribution is close to that in the system semiconductor film - liquid), in addition to substantial slowing down of oxygen diffusion in such systems. 

First, we must consider a gas-liquid system separated by an interface. When the thermodynamic equilibrium concentration is not reached for a transferable solute A in the gas phase, a concentration gradient is established between the two phases, and this will create a mass transfer flow of A from the gas phase to the liquid phase. This is described by the two-film model proposed by W. G. Whitman, where interphase mass transfer is ensured by diffusion of the solute through two stagnant layers of thickness <5G and <5L on both sides of the interface (Fig. 45.1) [1—4]. 

The Hatta criterion compares the rates of the mass transfer (diffusion) process and that of the chemical reaction. In gas-liquid reactions, a further complication arises because the chemical reaction can lead to an increase of the rate of mass transfer. Intuition provides an explanation for this. Some of the reaction will proceed within the liquid boundary layer, and consequently some hydrogen will be consumed already within the boundary layer. As a result, the molar transfer rate JH with reaction will be higher than that without reaction. One can now feel the impact of the rate of reaction not only on the transfer rate but also, as a second-order effect, on the enhancement of the transfer rate. In the case of a slow reaction (see case 2 in Fig. 45.2), the enhancement is negligible. For a faster reaction, however, a large part of the conversion occurs in the boundary layer, and this results in an overall increase of mass transfer (cases 3 and 4 in Fig. 45.2). 

Heterogeneous uptake on surfaces has also been documented for various free radicals (DeMore et al., 1994). Table 3 shows values of the gas/surface reaction probabilities (y) of the species assumed to undergo loss to aerosol surface in the model. Only the species where a reaction probability has been measured at a reasonable boundary layer temperature (i.e. >273 K) and on a suitable surface for the marine boundary layer (NaCl(s) or liquid water) have been included. Unless stated otherwise, values for uptake onto NaCl(s), the most likely aerosol surface in the MBL (Gras and Ayers, 1983), have been used. Where reaction probabilities are unavailable mass accommodation coefficients (a) have been used instead. The experimental values of the reaction probability are expected to be smaller than or equal to the mass accommodation coefficients because a is just the probability that a molecule is taken up on the particle surface, while y takes into account the uptake, the gas phase diffusion and the reaction with other species in the particle (Ravishankara, 1997). 

Upon dissolving Al into liquid Ga, the alumina layer that instantly forms from exposure to air or water at the surface is either discontinuous or porous. In either case the surface of the Ga-Al liquid is not passivated. As a result, when water contacts the surface of the liquid, Al atoms at the surface split the water, liberating hydrogen and heat with the formation of alumina. Since the liquid is fluid, the alumina cannot form a bonded layer at the liquid surface that would passivate pure, solid Al. Instead, the alumina is swept away by convection or agitation as a suspension of alumina particles in the water. The surface of the liquid alloy is now depleted of Al. This depleted region at the surface is replenished via diffusion or convection of Al from the bulk to the surface where it continues to split water. This process continues until all of the Al in the liquid alloy is converted to alumina. To summarize, liquid Al-Ga alloys rich in Ga split water because the Al component is not passivated as it is in solid pure Al. 

In electrochemical systems, metal meshes have been widely used as the backing layers for catalyst layers (or electrodes) [26-29] and as separators [30]. In fuel cells where an aqueous electrolyte is employed, metal screens or sheets have been used as the diffusion layers with catalyst layers coated on them [31]. In direct liquid fuel cells, such as the direct methanol fuel cell (DMFC), there has been research with metal meshes as DLs in order to replace the typical CFPs and CCs because they are considered unsuitable for the transport and release of carbon dioxide gas from the anode side of the cell [32]. 

Besides silicon, other materials have also been used in micro fuel cells. Cha et al. [79] made micro-FF channels on SU8 sheets—a photosensitive polymer that is flexible, easy to fabricate, thin, and cheaper than silicon wafers. On top of fhe flow channels, for both the anode and cathode, a paste of carbon black and PTFE is deposited in order to form the actual diffusion layers of the fuel cell. Mifrovski, Elliott, and Nuzzo [80] used a gas-permeable elastomer, such as poly(dimethylsiloxane) (PDMS), as a diffusion layer (with platinum electrodes embedded in it) for liquid-electrolyte-based micro-PEM fuel cells. 

After testing a number of DLs with and without MPLs, Lin and Nguyen [108] postulated that the MPL seemed to push more liquid water back to the anode through the membrane. Basically, the small hydrophobic pores in the MPL result in low liquid water permeability and reduce the water transport from the CL toward the DL. Therefore, more liquid water accumulated in the CL is forced toward the anode (back diffusion). This reduces the amount of water removed through the cathode DL, decreases the number of blocked pores within the cathode diffusion layer, and improves the overall gas transport from the DL toward the active zones. 

To design the optimal diffusion layer for a specific fuel cell system, it is important to be able to measure and understand all the parameters and characteristics that have a direct influence on the performance of the diffusion layers. This section will discuss in detail some of the most important properties that affect the diffusion layers, such as thickness, hydrophobicity and hydrophilicity, porosity and permeability (for both gas and liquids), electrical and thermal conductivity, mechanical properties, durability, and flow... 

In conclusion, at an intermediate optimum thickness, a diffusion layer allows for (1) gas diffusion toward the CL, (2) liquid transport from the CL toward the flow field channels, (3) good contact with both the bipolar plate... 

One of the most common ways to characterize the hydrophobicity (or hydrophilicity) of a material is through measurement of the contact angle, which is the angle between the liquid-gas interface and the solid surface measured at the triple point at which all three phases interconnect. The two most popular techniques to measure contact angles for diffusion layers are the sessile drop method and the capillary rise method (or Wihelmy method) [9,192]. 

In this section, we will briefly discuss different testing techniques that are widely used to measure most of the important mass transport properties of fhe diffusion layers. It is important to note that these techniques can also be used with MPLs. The first subsection will explain methods that deal with properties that affect both gas and liquid mass transport, and the other two subsections will discuss only techniques that measure gas and liquid transport properties, respectively. 

As stated earlier, CEP and CC are the most common materials used in the PEM and direct liquid fuel cell due fo fheir nature, it is critical to understand how their porosity, pore size distribution, and capillary flow (and pressures) affecf fhe cell s overall performance. In addition to these properties, pressure drop measurements between the inlet and outlet streams of fuel cells are widely used as an indication of the liquid and gas transport within different diffusion layers. In fhis section, we will discuss the main methods used to measure and determine these properties that play such an important role in the improvement of bofh gas and liquid transport mechanisms. 

However, this porosity takes into account all the open pores—even those that are not connected between each other, which are useless in fuel cell operation. Therefore, the effective porosity, which counts only the interconnected pores, is more critical when determining the optimal diffusion layer in a fuel cell. This porosity can be determined by using volume filtration techniques. For example, a porous sample is immersed in a liquid that does not enter inside the pores (e.g., mercury at low pressures) and then the total volume of the material can be determined. Next, the specimen is put inside a container of known volume that contains an inert gas, and the changed pressure is recorded. After this, a second evacuated chamber of known volume is connected to the system, and the new pressure is recorded. With these pressures and the ideal gas law, the volume of open pores and thus the effective porosity can be determined [195]. 

Nguyen et al. [205] designed a volume displacement technique that was used to measure the capillary pressures for both hydrophobic and hydrophilic materials. One requirement for this method is that the sample material must have enough pore volume to be able to measure the respective displaced volume. Basically, while the sample is filled wifh water and then drained, the volume of water displaced is recorded. In order for the water to be drained from fhe material, it is vital to keep the liquid pressure higher than the gas pressure (i.e., pressure difference is key). Once the sample is saturated, the liquid pressure can be reduced slightly in order for the water to drain. From these tests, plots of capillary pressure versus water saturation corresponding to both imbibitions and drainages can be determined. A similar method was presented by Koido, Furusawa, and Moriyama [206], except they studied only the liquid water imbibition with different diffusion layers. 

K. Nishida, T. Murakami, S. Tsushima, and H. Shuichiro. Microscopic visualization of state and behavior of liquid water in a gas diffusion layer of PEFC. Electrochemistry 75 (2006) 149-151. 

S. Litster, D. Sinton, and N. Djilali. Ex situ visualization of liquid water transport in PEM fuel cell gas diffusion layers. Journal of Power Sources 154 (2006) 95-105. 

Liquid water arrives in the CCL via transport through the PEM or it is generated in the electrochemical reaction. Invariably, PEECs require a medium that is highly effective in transforming liquid water into water vapor otherwise, liquid water will clog pores and channels in gas diffusion layers and flow fields that are needed for the gaseous supply of reactants. 

Obviously, the CCL not only determines the rate of currenf conversion and the major portion of irreversible voltage losses in a PEFC, but also plays a key role for the water balance of the whole cell. Indeed, due to a benign porous structure with a large portion of pores in the nanometer range, the CCL emerges as favorite water exchanger for PEFCs. Once liquid wafer arrives in gas diffusion layers or flow fields, PEFCs are unable to handle if. 

As can be seen in the different boundary conditions, the main effects of having ribs are electronic conductivity and transport of oxygen and water, especially in the liquid phase. In terms of electronic conductivity, the diffusion media are mainly carbon, a material that is fairly conductive. However, for very hydro-phobic or porous gas-diffusion layers that have a small volume fraction of carbon, electronic conductivity can become important. Because the electrons leave the fuel cell through the ribs, hot spots can develop with large gradients in electron flux density next to the channel. " Furthermore, if the conductivity of the gas-diffusion layer becomes too small, a... 

Figure 19. Liquid saturation and current density of the cathode as a function of position for the case of dry air fed at 60 °C. (a) Liquid saturation in the gas-diffusion layer where the channel goes from x = 0 to 0.05 cm and the rib is the rest the total cathode overpotential is —0.5 V. (b) Current-density distributions for different channel/rib arrangements. (Reproduced with permission from ref 56. Copyright 2001 The Electrochemical Society, Inc.)...

Divisek et al. presented a similar two-phase, two-dimensional model of DMFC. Two-phase flow and capillary effects in backing layers were considered using a quantitatively different but qualitatively similar function of capillary pressure vs liquid saturation. In practice, this capillary pressure function must be experimentally obtained for realistic DMFC backing materials in a methanol solution. Note that methanol in the anode solution significantly alters the interfacial tension characteristics. In addition, Divisek et al. developed detailed, multistep reaction models for both ORR and methanol oxidation as well as used the Stefan—Maxwell formulation for gas diffusion. Murgia et al. described a one-dimensional, two-phase, multicomponent steady-state model based on phenomenological transport equations for the catalyst layer, diffusion layer, and polymer membrane for a liquid-feed DMFC. 

Lim, C. Wang, C. Y. Measurement of contact angles of liquid water in PEM fuel cell gas diffusion layer (GDL) by sessile drop and capillary rise methods. Penn State University Electrochemical Engine Center (ECEC) Technical Report no. 2001 03, Perm State University State College, PA, 2001. 

---