High Throughput Search for New Fuel Catalysts
PEM fuel cells require electrocatalysts to operate efficiently. At the anode, a catalyst is needed to oxidize the fuel, and at the cathode a catalyst is needed to reduce oxygen. Typically, Pt is used for the reduction of oxygen, and a Pt-Ru alloy is used for the oxidation of hydrogen. Our reliance on Pt-based catalysts poses several challenges for the widespread use of fuel cells. Platinum's relatively high price and small global supply limit the quantity of Pt that we can afford to use in fuel cells. Pt is also easily poisoned by trace impurities in hydrogen fuel (carbon monoxide and sulfur compounds), which leads to increased costs in fuel production. In addition, there is increasing interest in the use of liquid fuels (methanol, ethanol etc.) in fuel cells, and neither Pt nor Pt-Ru are effective catalyst for the complete oxidation of such fuels.
These issues can only be overcome through the development of new catalyst materials. Ideally, a non-precious metal catalyst will be identified, but the problems described above can also be remedied by the development of Pt-based catalysts that exhibit both higher reaction rates and improved tolerance to poisons. We have identified approximately 30 elements on the periodic table that could be combined to make a new fuel cell catalyst. Thus, there are approximately 30 Pt-based binary and 900 Pt-based ternary chemical systems that should be explored. To search all of these chemical systems using traditional chemistry techniques would require the synthesis of hundreds of thousands of samples. To more efficiently perform this search, we are employing high-throughput fabrication and electrochemical testing techniques.
By first exploring the vast number of Pt-based catalysts, we will better understand which material
properties lead to improved catalysis. This information will help guide our search of the millions of possible non-Pt and non-precious metal catalysts.
Combinatorial Materials Synthesis
To explore a new ternary chemical system, for example Pt-Ru-Au, we synthesize a composition spread thin film, which contains a broad range of the possible PtxRuyAu1-x-y compositions. This is achieved by the co-deposition of the 3 elements using magnetron sputtering in a custom built vacuum chamber. Fig. 1 shows a photograph of the deposition system and a diagram of the assembly of 4 sputter guns. For our Pt-Ru-Au example, Pt, Ru and Au sources would be placed in the three outer sputter guns.
A 76mm-diameter Si substrate is placed above the deposition sources to provide a carefully designed geometric relationship between the substrate and each deposition source. When the 3 deposition sources are operated simultaneously, a thin film is grown on the substrate, and this thin film has a unique composition at each location on the substrate. In some cases the gradients in composition across the substrate yield visible features, such as those on the thin film shown in Fig. 2. The deposition system includes many other features, including substrate heating, which can be used to incite the formation of different crystallographic phases in the composition spread thin film. Using this deposition technique, we quickly produce a library of potential catalysts on a single substrate. Details of our deposition techniques can be found in  and .
For high-throughput electrochemical evaluation, we employ an electrochemical fluorescence technique which was developed for catalyst studies by Reddington et al . The Si substrate with composition spread thin film is incorporated in the testing cell shown in Fig. 3. This technique is used for evaluating the catalytic activity towards the oxidation of a number of fuels as well as the reduction of oxygen. Here we will use methanol oxidation and a Pt-Ru-Au composition spread as an example.
The entire Pt-Ru-Au thin film is exposed to a testing solution containing the methanol fuel as well as quinine, a fluorescent indicator that begins fluorescing (under UV illumination) as the pH is decreased below ~4. To test for catalytic activity, the potential of the entire composition spread is slowly increased. Increasing the catalyst potential will increase the methanol oxidation rate, but the absolute oxidation rate at a given potential is a measure of the effectiveness of the catalyst. If a certain composition in the thin film is a good catalyst, it will promote the reaction CH3OH+H20 -> CO2+6H++6e-, which will result in the release of the H+ into solution. This corresponds to a local decrease in the pH and the incitement of fluorescence. The detection of fluorescing regions over the substrate thus leads to the identification of active catalysts.
For the Pt-Ru-Au example, Fig. 4 illustrates an experiment that identifies the region of the thin film composition spread that contains the best catalysts. The fluorescence images indicate that the active region is closer to the Ru and Pt sources than to the Au source. To quantify this relationship, the fluorescence image can be mapped onto a ternary composition diagram, as depicted in Fig. 5. Since each substrate position corresponds to a point in the composition diagram, this plot shows the range of compositions contained on the single wafer and the composition range of the active catalysts. After the identification of the active regions, the substrate can be used for subsequent electrochemical testing to better understand the catalyst. Many of our electrochemical testing methods are described in  and .
Materials Characterization and Trends in Catalytic Activity
The identification of a catalytically active region in a composition spread is an important aspect of our high-throughput research, but the significant advancement along our path to new catalysts comes if we can understand why that particular region contains good catalysts and the surrounding regions contain bad catalysts. As discussed above, we can study the relationship between composition and catalytic activity. In addition, by further characterization of our thin films, we can study a large number of material properties that may correlate with catalytic activity. One of the most important material properties is the crystalline phase behavior of the chemical system. Previous catalyst studies in the Abruña and DiSalvo labs at Cornell University demonstrated that the formation of ordered intermetallic phases can lead to great improvements in catalytic activity .
To determine which crystallographic phases are formed at each composition, we perform x-ray diffraction experiments at the Cornell High Energy Synchrotron Source (www.chess.cornell.edu). The x-ray diffraction experiment is shown in Fig. 6 and is described in . The substrate is placed on a stage so that an x-ray beam can be rastered over the thin film composition spread. For an example Pt-Ta binary composition spread, the measured x-ray diffraction patterns are shown in Fig. 7. The diffraction patterns indicated that over the composition range, 4 different crystallographic phases are formed. In the electrochemical fluorescence measurements, the half-wave potential for the oxidation of methanol is measured. By comparing the plot of this potential with the crystallographic phase regions, we discover that the formation of an orthorhombic ordered intermetallic phase leads to significant improvements in catalytic activity (see  for more details). We can use this result to design new catalyst systems that will be explored on our quest to discovery the catalysts that will enable the widespread deployment of fuel cells.
 M. Prochaska, J. Jin, D. Rochefort, L. Zhuang, F. DiSalvo, H. Abruña, and R. van Dover, Rev. Sci. Instrum. 77 (2006).
 J. M. Gregoire, R. B. van Dover, J. Jin, F. J. DiSalvo, H. D. Abruña, Rev. Sci. Instrum., 78, 072212 (2007).
 E. Reddington, A. Sapienza, B. Gurau, R. Viswanathan, S. Sarangapani, E. S. Smotkin, T. E. Mallouk, Science, 280, 1735 (1998).
 J. M. Gregoire, M. Kostylev, M. E. Tague, P. F. Mutolo, R. B. van Dover, F. J. DiSalvo, and H. D. Abruña, J. Electrochem. Soc. 156, B160 (2009).
 E. Casado-Rivera, D. Volpe, L. Alden, C. Lind, C. Downie, T. Vazquez-Alvarez, A. Angelo, F. DiSalvo, and H. Abruña, J. Am. Chem. Soc. 126, 4043 (2004).
 J. M. Gregoire, D. Dale, A. Kazimirov, F. J. DiSalvo, and R. B. van Dover, Rev. Sci. Instrum. in press (2010).
 J. M. Gregoire, M. E. Tague, S. Cahen, S. Khan, H. D. Abruña, F. J. DiSalvo, and R. B. van Dover, Chem. Mat. in press (2010).