Vacuum Deposited Non-Precious Metal Catalysts for PEM Fuel Cells

By: Dr. David G. O’Neill, Dr. Radoslav Atanasoski, Alison K. Schmoeckel, Alison K. Schmoeckel, George D. Vernstrom, Dennis P. O’Brien, Dr. Manish Jain, Dr. Thomas E. Wood, Material Matters Volume 1 Article 3

Dr. David G. O’Neill, Dr. Radoslav Atanasoski (pictured) 
Alison K. Schmoeckel, Alison K. Schmoeckel, George D. Vernstrom, Dennis P. O’Brien, Dr. Manish Jain, Dr. Thomas E. Wood, The 3M Company

Current PEM fuel cell technology uses platinum as a catalyst. A large body of work seeking to find a Pt replacement involves Oxygen Reduction Reaction (ORR) catalysts produced using nitrogen coordinated transition metal compounds, such as Fe porphyrins or Fe phthalocyanine.1,2,3 Traditionally, these catalysts are made by pyrolyzing transition metal compounds dried onto carbon support particles at temperatures up to 950 ºC in reducing and/or nitrogen-containing atmosphere. Unfortunately, ORR activity of these catalysts has not been enough and attempts to increase the oxygen activity, for example by increasing the Fe content, have not been successful.4 Analytical work by Dodelet’s group, including TOFSIMS, XPS and other techniques, led to their proposal that a pyridinic structure is required for good oxygen activity.5,6


Figure 1. Structure proposed by Dodelet’s group6 for the species required for ORR catalytic activity.

A new, alternative process to make non-precious ORR metal catalysts for the oxygen reduction reaction is to use vacuum deposition techniques to combine the three elements believed to be required (a transition metal, carbon, and nitrogen).7,8 Processes including pulsed arc plasma deposition, sputter deposition, evaporation and combinations thereof have been used to produce a variety of materials whose catalytic activity was determined by assembling and measuring the oxygen generated electrical current in 50 cm2 fuel cells. The pulsed arc process is well suited to deposit carbon/Fe based catalysts because it can produce a carbon/iron plasma by erosion of a high-purity graphite cathode containing Fe wires.9 The plan is to vary deposition processes and conditions over a wide range while checking their catalytic activity in a 50 cm2 test fuel cell fixture.10,11 However, solely relying on fuel cell testing does not provide clues to process changes that would be most beneficial, but an analytical technique has not been identified that uniquely qualifies catalytic behavior. Nevertheless, an ability to measure a material characteristic that is known to relate to ORR catalysis would significantly improve process R&D.

Therefore, we began using local, element-specific, analytical probes that have the ability to differentiate between Fe-based catalyst materials that have had different process histories but similar compositions. Vacuum deposited C-Nx:Fe catalysts were analyzed using Extended X-ray Absoption Fine Structure (EXAFS) analyses of the Fe k-edge and resonant ultraviolet photoelectron spectroscopy (UPS) to measure the valence electronic structure around Fe atoms. Both techniques show significant differences between samples that received and those that did not, receive thermal treatment.

back to top

Results

Carbon-nitrogen-iron coatings were deposited onto either a carbon non-woven cloth or Si wafers, depending on the analytical technique to be used, and processed either with or without a thermal treatment. Surface composition measurements by XPS showed that the nitrogen and iron content could be varied between 1–5 at. % for nitrogen and between 0.1 to 7 at. % for iron. This range is much larger than can be obtained with molecular pyrolysis methods (2–3 at. % nitrogen and 0.05 at. % Fe). Analysis of the Fe core level indicates that the iron in these coatings is neither metallic nor an oxide; however some carbide phase cannot be ruled out. High resolution XPS shows that the N 1S core level in vapor deposited C-Nx:Fe coatings has the same doublet structure with a component at 398.5 eV, identified as pyridinic nitrogen, that has been reported for pyrolyzed Fe-derived catalyst materials.6 The fact that the nitrogen exhibits the same XPS fingerprint as pyrolyzed Fe catalysts combined with the fact that the nitrogen and iron content is much higher than in pyrolyzed Fe catalysts suggested that the catalytic currents would be higher. However, our electrochemical fuel cell testing showed the opposite to be true. Although a direct comparison is difficult, after correcting for large differences in surface area, the ORR catalytic activity of vapor deposited coatings is lower than that of pyrolyzed Fe catalysts. Although a thermal treatment above 650 °C, either during or after deposition, significantly increased catalytic activity the level was still below that of pyrolyzed molecular compounds.

The similarity of vacuum deposited coatings based on XPS spectra and the difference in their electrochemical behavior suggests that more detailed materials characterization is needed. Detailed analyses of the N 1S core-level shows that the intensity and binding energy of the 398.5 eV component, the component normally associated with pyridinic nitrogen, changes with iron content. For example, in C-Nx:Fe(7 at. %) the intensity is 75% of the total N 1S intensity, whereas in a nitrogenated carbon (i.e. without iron) the intensity is only 45% of the total. Indeed, the pyridinic component intensity steadily increases as the Fe content increases, as shown by the graph in Figure 2. In addition, the pyridinic components binding energy increases with increasing iron content. The shift to deeper binding energy is similar to findings by Dodelet’s group for pyrolyzed Fe-derived catalysts, although in those materials the shift is much less.6 The fact that a small amount of iron has such an effect on the nitrogen core level binding energy suggests that Fe and N atoms are near each other instead of being uniformly dispersed.


Figure 2. Relative intensity of the N 1S peak at 398.5 eV associated with pyridinic nitrogen as a function of Fe content.

Figure 3. The k3-weighted Fourier transform magnitudes of the Fe K-edge of a vacuum deposited catalyst before and after the thermal post-treatment. The transforms are not phase-corrected.

Figure 3 shows EXAFS Fourier transforms for vacuum deposited catalysts. These transforms are not phase-corrected and therefore they have a small shift in the x-axis from real values. Examination of Figure 3 shows that the unheated material has a single main peak while the heated sample has two well-resolved peaks. This indicates that the as-deposited material is fairly disordered and that heating increases atomic order (narrower peaks) with the introduction of a second coordination shell around Fe atoms. Photoelectron spectroscopy is another technique that has the potential to identify unique aspects of C-Nx:Fe materials because it directly probes valence electrons that are involved in bonding, and many believe catalysis. Photoemission is not, in general, element specific, however, element specific information can be obtained by utilizing resonant photoemission. In resonant photoemission, large intensity variations occur with small changes in the exciting photon energy and this can be used to detect where Fe d-states are located in the valence densityof- states.12 The role of Fe and it’s importance for catalytic behavior, many believe, is key to understanding Fe-based catalysts. We have successfully used photoemission to study C-Nx:Fe materials including the use of resonant techniques to detect valence Fe d-states. These studies, to be published, show that metallic Fe d-states exist near the Fermi level in as-deposited coatings as well as hybridized d-states between 4 to 10 eV below the Fermi level. Photoemission studies combined with ab initio density functional theory simulations, using VASP,12,13 has the potential to provide new insights into bonding in ORR Fe-based catalysts.

back to top

Conclusions

Vacuum deposited carbon-nitrogen-iron coatings deposited by pulsed cathodic arc plasma deposition were characterized and the effects of a high temperature thermal treatment were evaluated. Characterization by XPS shows that the iron is not metallic iron and that the N 1S core level matches that found in pyrolyzed Fe-based catalysts. High resolution XPS studies of the N 1S core level vs. Fe content shows that Fe atoms cause the N 1S core level component at about 398.5 eV, also associated with pyridinic nitrogen, to increase in intensity and to shift to deeper binding energy. Clearly, Fe atoms must be near nitrogen atoms in order to change N 1S binding energies.

Subjecting the vapor deposited catalysts to a thermal treatment, either during or after deposition, significantly increases the catalytic activity of these materials. However, comparison of these materials, with and without a thermal treatment, using high resolution XPS does not show significant differences. EXAFS studies of the Fe k-edge show that atomic rearrangements occur after high temperature treatments, thermal treatments that also increase ORR catalytic currents. The thermal treatment increases atomic order and introduces a second coordination shell around Fe atoms. Resonant UV photoemission techniques show that Fe d-electrons are located in two energy regions of the electronic density of states. There are metallic Fe d-states located near the Fermi level and hybridized Fe d-states, most likely with ligand p-states, located between 4 to 12 eV below the Fermi level. More work is needed with different materials before more conclusive statements can be made.

Element specific techniques such as EXAFS and resonant photoemission show promise in being able to detect changes around given atoms, e.g. iron, in a ternary compound such as the C-Nx:Fe materials that exhibit catalytic activity for the oxygen-reduction reaction. Changes that are associated with better catalytic activity. Although the vapor deposited materials made so far have not exhibited significant ORR catalytic activity, future developments and improved understanding of these materials could allow vapor deposition to be a viable process for economical catalyst production.

back to top

Acknowledgments

This research was supported in part by the Department of Energy, Cooperative Agreement No. DE-FC36-03GO13106. DOE support does not constitute an endorsement by DOE of the views expressed in this article. The authors would like to thank our colleagues at 3M, particularly Dr. Shih-Hung Chou, Dr. Allen Siedle, and Terry Pechacek; Drs. Xiao-Qing Yang and Won-Sub Soon at Brookhaven National Laboratory; Dr. Cliff Olson at the Synchrotron Radiation Center; and Professor David Wieliczka at University of Missouri - KC. The authors would also like to thank Professor Jeff Dahn and his research group, Dalhousie University, for useful conversations, and Professor Jean-Pol Dodelet and his group, INRS-Energie, Materiaux et Telcommunications, for the samples provided for this study.

back to top

References

  1. Terminology Convention: “pyrolyzed, Fe-derived” will refer to the catalyst syntheses using Fe containing molecules, such as Fe porphoryns, Fe phthalocyanines, Fe-acetate, etc.
  2. N4-Macrocyclic Metal Complexes: Electrocatalysis, Electrophotochemistry & Biomimetic Electroanalysis. Edited by J. Zagal, F. Bedioui, and J. P. Dodelet, Springer 2006.
  3. H. A. Gasteiger, S. S. Kocha, B. Sompalli, F. T. Wagner, Applied Cat. B: Environ., 2005, 9, 56.
  4. J.-P. Dodelet in N4-Macrocyclic Metal Complexes: Electrocatalysis, Electrophotochemistry & Biomimetic Electroanalysis. Edited by J. Zagal, F. Bedioui, and J. P. Dodelet, Springer 2006.
  5. F. Jaouen, M. Lefevre, J.-P. Dodelet, and M. Cai, J. Phys. Chem. B, 2006, 5553, 110.
  6. F. Jaouen, S. Marcott, J.-P. Dodelet and G. Lindbergh, J. Phys. Chem. B, 2003, 1376, 107.
  7. R.T. Atanasoski, Novel Approach to Non-Precious Metal Catalysts, DOE Hydrogen Program FY 2004 Progress Report (V.C.5).
  8. G. Lalande, D. Guay, J. P. Dodelet, S. A. Majetich, M.E.McHenry, Chem. Mater., 1997, 784, 9. E. Bradley Easton, Arman Bonakdarpour, and J.R. Dahn, Electrochem. and Solid- State Lett., 2006, A463, 9.
  9. E.I. Tochitsky, O.V. Selifanov, V.V. Akulich, I.A. Kapustin and A.V. Stanishevskii, Surface and Coatings Technology, 1991 522, 47.
  10. R.T. Atanasoski, Novel Approach to Non-Precious Metal Catalysts, DOE Hydrogen Program FY 2005 Progress Report (V.C.5).
  11. R.T. Atanasoski, Novel Approach to Non-Precious Metal Catalysts, DOE Hydrogen Program FY 2006 Progress Report.
  12. G. Kresse and J. Hafner, Phys. Rev. B, 47, 558 (1993); G. Kresse and J. Furthmüller, Phys. Rev. B, 1996, 11169, 54. G. Kresse and J. Furthmüller, Comput. Mater. Sci., 1996, 15, 6. G. Kresse and D. Joubert, Phys. Rev. B, 1999, 1758, 59.
  13. M. Jain, S.-H. Chou, A. Siedle, J. Phys. Chem. B, 2006, 4179, 110.

back to top

Related Links