Cathode structure and principle of DMFC

The oxygen reduction reaction (ORR) is a multi-electron process consisting of numerous elementary steps, involving both series and parallel pathways. It is generally accepted that oxygen reduction on Pt occurs via dissociative adsorption of O2 followed by protonation of the adsorbed species, with the former being the rate-determining step. There have been several models that attempt to describe these pathways [148-150]. Several models successfully interpret the same data due to their similarity. One such model, illustrated in Fig. 2-5, is the bridge model of the ORR on Pt in acid. Because of the bridging oxygen system, it is obvious that optimal Pt particle spacing is of critical importance.

Fig. 2-5 Bridge model of oxygen reduction on Pt (z represents the oxidation state).

The complicated pathway of the ORR results in slow electrochemical kinetics. One measure of the rate of an electrochemical reaction is its exchange current density, jo.

The jo of ORR on Pt is 105 less than jo of hydrogen oxidation at Pt. This huge difference accounts for the chief influence of the cathode activity on hydrogen/air cell fuel performance. Therefore, cathode activity enhancement has been a major focus for PEMFC electrode development. In order to increase cathode activity, one must increase catalyst utilization. This will not only increase performance but can also lead to a lowering of the Pt loading required. In order to activate a catalyst site electrochemically, pathways for electron, proton and gas transports must all be present.

The active area of Pt is typically measured by using cyclic voltammetry (CV) in acid electrolytes. Specifically, the area under the hydrogen adsorption/desorption peaks is determined as shown in Fig. 2-6. Regions of oxide formation (QA) and reduction (QC) as well as formation of hydrogen (HA) and its reduction (HC) are indicated. A larger area per mass of Pt indicates a larger active area. One very successful method to increase catalyst utilization is to employ carbon supported Pt catalyst. Typically, Pt particles (3-10 nm) are dispersed onto the electronically conducting carbon particles, about 30-50 nm. The ideal carbon support should possess high chemical stability, good electronic conductivity and high surface area for suitable pore size distribution.

The best type of carbon for fuel cell catalyst support is carbon black. There are several types of commercial carbon blacks under study to be used in fuel cells, and Vulcan XC72 is the most common. Although carbon is an excellent electronic conductor, it is a very poor proton conductor because carbon is hydrophobic. However, the carbon surface does consist of hydrophilic moieties. Carbon-oxygen complexes, such as phenol, carbonyl, carboxyl, quinone and lactone groups can all be found on the carbon surface. In general, exposing the carbon to an oxidizing agent forms these complexes.

Fig. 2-6 CV of a platinum electrode in 0.5M H2SO4(aq).

One serious issue with Pt/carbon catalysts is the sintering of Pt particles [151].

Sintering occurs when Pt particles become larger over their lifetime. This decreases the Pt surface area, and ultimately leads to a decline in performance throughout the operation lifetime. An ideal catalyst support material that is both electronically and ionically conductive is desired. Pickup et al. have studied such a material, a conducting polymer composite [152]. The composite consists of polyprrole and polystyrenesulfonate, which are electronically conductive and proton conductor, respectively. This material was tested as a replacement of carbon and reasonable performance was achieved. However, low Pt utilization and polymer stability are still issues [153]. Another method to increase catalyst utilization is to add a proton-conducting polymer (such as Nafion) into the catalyst layer. Pt catalyst near or directly in contact with the Nafion membrane is utilized most efficiently. However, utilization drops off deeper into the catalyst layer, largely due to the limited proton conductivity of the catalyst layer. Nafion solution can be applied onto preformed

electrodes or directly mixed with the catalyst during ink preparation. This increases the proton conductivity of the catalyst layer. In 1986, Raistrick was able to demonstrate that carbon-supported Pt catalyst mixed with Nafion could outperform conventional Pt black electrodes that had ten times the Pt loading [154]. This was a major breakthrough in fuel cell development in that it made the cost of Pt required much more feasible. To design an electrode with carbon-supported catalyst and Nafion, they must be mixed in proper proportions to form a stable three-phase boundary where the gas, ion conductor and electronically conducting phase with catalytical activity are all present. This requirement limits the amount of Nafion that can be added since the morphology, low gas permeability and poor electronic conductivity of Nafion disrupts this boundary and adversely affects cell performances.

Because of this and the high cost of Nafion, alternative methods to provide proton conductivity in the catalyst layer are of interest. Another approach to increase Pt utilization is to simply deposit Pt only in the areas of the electrode where it would be electroactive. This can be done by sputter deposition where layers as thin as 2 nm can be deposited. There have been many studies that use sputter deposition to localize Pt catalyst at the front surface of the electrode or even directly onto the membrane surface. Srinivsan et al. applied a 50 nm thick layer of Pt onto an uncatalyzed gas diffusion layer (GDL) by sputter deposition and achieved a 10-fold reduction in Pt loading (from 4 mg/cm2 to 0.4 mg/cm2) without performance loss [155]. Hirano et al.

later showed that electrodes prepared by sputter deposition with Pt loading of 0.1 mg/cm² could perform the same as those prepared by using standard materials (Pt/C) at Pt loading of 0.4 mg/cm² [156]. Cha and Lee further reduced the Pt loading to 0.04 mg/cm² by alternating sputter deposited Pt layers and painted Nafion/Carbon ink layers with successively lower amounts of Pt in each layer [157]. This leads to very efficient utilization of Pt. Sputter deposition is promising for fuel cells since a larger

percentage of Pt is electrochemically active. It also allows the fabrication of very thin active layers to decrease ohmic and mass transport overpotentials in catalyst layers.

Sputter deposition is a well established industrial technique in areas such as thin films and integrated circuits and it is anticipated that this technique could be readily applied to micro-fuel cell applications.