Properties of Fuel Cell Membranes

 

Chetan Mahajan


cmahajan@mail.che.utexas.edu

Office: CPE 3.472

Phone: 512-471-4966

 

In the pursuit of economical options for sustainable clean energy alternatives, fuel cells are emerging as a promising candidate (1). In this context, polymer electrolyte membrane fuel cells (PEMFCs) have attracted considerable interest by possessing amongst the highest power densities. Much of the research in PEMFCs have explored the use of a material termed NAFION, which produces high performance, but also suffers from certain disadvantages such as a limit on the operating temperature, water and methanol crossover, and high costs of the membrane (2-5). Consequently, there is active research in alternative polymeric proton exchange membranes and of particular interest to this work is one such alternative to NAFION membrane, viz., Sulfonated Poly(ether ether ketone) (SPEEK) and the membranes based upon SPEEK blending with other polymers. In pure form, SPEEK membranes have been demonstrated to offer potential for achieving a combination of desirable features such as chemical and thermal stability, mechanical strength under fuel cell conditions, reduced water, methanol crossover and low cost (3, 6-8). On the other hand, some limitations of SPEEK include reduced proton conductivity and swelling effects at temperatures as low as 80 C (the corresponding swelling temperature for NAFION is around 140 C), which results in lower morphological stability (3, 6-9). Recent studies have demonstrated that SPEEK blended with polysulfone bearing base like benzimidazole or its derivative improves stability and leads to better DMFC performance than pure SPEEK or NAFION at significantly reduced costs compared to NAFION (10-12).

Computer simulations provide an attractive means to study structural and dynamical features of polymer membranes and shed light on the experimental observations. However, there has been much less simulation work reported in the context of solvated SPEEK membranes in pure form (13). No simulation work has been reported in the blended form. Using atomistic simulations, we undertake a fundamental study on the characteristics of solvated SPEEK membranes in pure as well as in a form blended with polysulfone(Psf) bearing base like benzimidazole(BIm) used in the context of direct methanol fuel cell applications (DMFC). The analysis may, however, be extended to hydrogen fuel cell. Specifically, we are interested in studying the interplay between basic chemistry, structure, and observed dynamical properties (water, methanol, proton diffusivities) of the novel candidate membranes. This knowledge can be further utilized for designing better candidate membranes.

Methodology

All atom classical simulations have been performed using Accelrys Materials Studio (14) (for generating initial configurations) and LAMMPS (15) (Large-scale Atomic/Molecular Massively Parallel Simulator) for molecular mechanics and molecular dynamics simulations with self-written C, bash and awk programming tool-kit. The simulations have been performed using Ranger supercomputing system from Texas Advanced Computing Center. (16)

Results and Discussion

Dynamical properties: The diffusivity values of water and hydronium ion (which constitutes main component of vehicular proton conduction, thus referred to as vehicular proton henceforth) with change in water wt % are plotted and compared with NAFION studies (both simulation and experimental) (19, 23-26) in Fig.1 (a).  Water diffusivity is lower in SPEEK at lower water contents such as 10 wt % and becomes comparable to NAFION at water contents of 20 wt % and higher. Vehicular proton diffusivity is order of magnitude less than that in NAFION. Both the diffusivity trends agree well with the earlier experimental studies. Following results rationalize these trends using insights into microstructures of the two systems.

Phase segregation and water clusters: As shown in the Fig. 1 (b), (3) general experimental speculation is that solvated polyelectrolyte membranes like NAFION (left) and SPEEK (right) phase segregate into polymeric backbone containing hydrophobic phase (shown as green) and solvent rich aqueous phase (blue), the latter forming channels for water, proton and methanol (parasitic) conduction. Fig. 1(c) shows snapshot of our system at 40 wt % hydration, which clearly shows phase segregation with water forming large continuous cluster at such high water content. At this stage, it may be speculated that difference in water cluster characteristics may be influencing the difference in diffusivity trends. Fig. 1(d)-top shows how clusters are bigger, on an average, in NAFION (27) compared to SPEEK at any water wt %.  On the other hand, fig. 1(d)-Bottom depicts that water percolating fraction in SPEEK is lower at lower water wt % such as 10, but it becomes comparable to that in NAFION (23) at water wt % of 20 and higher. These two observations regarding water clusters explain the differences and the trends in the diffusivity in both the systems. This also explains observation of lower methanol crossover in SPEEK than NAFION, a particularly attractive feature of SPEEK for DMFC application.

Localization of vehicular proton and physical origins of water clusters: The above analysis of water clusters does not however explain peculiarly low vehicular proton diffusivity in SPEEK compared to NAFION. In addition, it would also be worthwhile to rationalize the physical reasons underlying differences in the nature of water clusters in NAFION and SPEEK membranes. We address these two issues as follows. Fig.2 (a) shows how coordination number of positively charged hydronium ion near negatively charged sulfonate anion decreases more steeply with increase in water wt % in NAFION (28) than in SPEEK, thus amounting to localization of vehicular proton in SPEEK than NAFION. This explains lower proton diffusivity of SPEEK further. In addition, this reveals much more basic nature of sulfonate anion and therefore more weakly acidic and less hydrophilic nature of its conjugate sulfonic acid in SPEEK than in NAFION. Water clusters form due to phase segregation and thus the nature of phase segregation can affect the water cluster characteristics. The nature of phase segregation can be expected to depend on sulfonic acid hydrophilicity, hydrophobicity of polymer backbone and flexibility of backbone. Our simulations also confirm less hydrophobic and less flexible nature of SPEEK backbone compared to NAFION (Figures not shown here –available upon request).  These three observations, for the first time using simulations, prove that phase segregation is less pronounced in SPEEK compared to NAFION, which gives rise to observed water cluster and dynamic properties.

Overcoming low proton conductivity in SPEEK: As discussed, solvated blends of SPEEK with polysulfone tethered with benzimidazole were simulated. As shown in Fig.2 (b), the vehicular proton conduction of blends is far higher than in SPEEK and in fact, is comparable to NAFION (26) (Note the temperature difference). This significant observation partially supports the conclusion of earlier experimental study. As can be seen clearly in the plot of coordination number of vehicular proton around sulfur in Fig. 2 (c)., the coordination numbers in blends decrease at a much faster rate relative to SPEEK. Moreover, coordination numbers in the projection shell or the outermost second shell are higher or substantially higher than in the first shell for blends, which is reversed in the case of pure SPEEK. Thus, it is appropriate to conclude that hydronium ions get delocalized faster in blends than in pure SPEEK with addition of water. This can be rationalized from the fact that some sulfonic groups are ionically bonded to protonated basic sites of polysulfone backbone.

Future Task:

In our work so far, purely classical framework based extensive all-atom simulations of fuel cell membranes based on SPEEK and its blends with PSf tethered with bases have been completed. The main future task would be to incorporate proton hopping details in our current classical simulation framework, using quantum mechanical principles.

Publications

1. Mahajan, C.V.; Ganesan, V. “Atomistic Simulations of Structure of Solvated Sulfonated
Poly(ether ether ketone) Membranes and their Comparisons to NAFION: I. Hydrophilic
and Hydrophobic Domains” J. Phys. Chem. B, submitted.

2. Mahajan, C.V.; Ganesan, V. “Atomistic Simulations of Structure of Solvated Sulfonated
Poly(ether ether ketone) Membranes and their Comparisons to NAFION: II. Structure
and Transport Properties of Water, Hydronium Ions and Methanol” J. Phys. Chem.B, submitted.

References

  1. Song, C. Catalysis Today 2002, 77, 17.
  2. Ren, X.; Wilson, M.S.; Gottesfeld, S. J. Electrochem. Soc. 1996, 143, 1, L12.
  3. Kreuer,K.D. J.Mem.Sci. 2001, 185, 29.
  4. Heinzel, A.;Steele, B.C.H. Nature 2001 414, 345.
  5. Deluca, N.W.; Elabd, Y.A. J. Polym. Sci. Part B: Polym. Phys. 2006 44, 2201.
  6. Li, l; Zhang, J.; Wang Y. J.Mem.Sci. 2003, 226, 159.
  7. Yang, B.; Manthiram, A. Electrochem. Solid State Lett. 2003, 6(11), A229.
  8. Yang, B.; Manthiram, A. J. Power Sources 2006, 153, 29.
  9. Kreuer,K.D. Solid State Ionics 1997, 97, 1.
  10. Fu, Y.; Manthiram, A.; Guiver M.D. Electrochem. Commun. 2006, 8, 1386.
  11. Fu, Y.; Manthiram, A.; Guiver M.D. Electrochem. Commun. 2007, 9, 905.
  12. Fu, Y.; Manthiram, A.; Guiver M.D. Electrochem. Solid State Lett. 2007, 10 (4), B70.
  13. Brunello, G.; Lee, S.G.; Jang, S.S.; Qi, Y. J. Renew. Sust. Ener. 2009, 1, 033101.
  14. MS Modeling V3.2, Copyright 2005, Accelrys Software Inc.
  15. Plimpton, S. J. J. Comput. Phys. 1995, 117, 1, http://lammps.sandia.gov/.
  16. www.tacc.utexas.edu.
  17. Vegt, N.F.A.V.; Briels, W.J.; Wessling, M.; Strathmann, H. J. Chem. Phys. 1996, 105, 19, 8849.
  18. Mayo, S. L.; Olafson, B. D.; Goddard, W. A. J. Phys. Chem.1990, 94, 8897.
  19. Jang, S.S.; Molinero, V.; Cagin, T.; Goddard, W.A. J. Phys. Chem. B 2004, 108, 3149.
  20. Levitt, M.; Hirshberg, M.; Sharon, R.; Laidig, K.E.; Daggett, V.  J. Phys. Chem. B 1997, 101(25), 5051.
  21. Honma, T.; Liew, C.C.; Inomata, H.; Arai, K. J. Phys. Chem. A 2003, 107(19), 3960.
  22. Jorgensen, W.L.; Maxwell, D.S.; Tirado-Rives, J. Am. Chem. Soc. 1996, 118, 11225.
  23. Blake, N.P.; Petersen, M.K.; Voth, G.A.; Metiu H. J. Phys. Chem. B 2005, 109, 24244.
  24. Devanathan, R.; Venkatnathan, A.; Dupuis, M. J. Phys. Chem. B 2007, 111(28), 13006.
  25. Perrin, J.C.; Lyonnard, S.; Volino, F. J. Phys. Chem. C 2007, 111, 3393.
  26. Cui, S.; Liu, J.; Selvan, M. E.; Keffer, D.J.; Edwards, B.J.; Steele, W.V. J. Phys. Chem. B 2007, 111, 2208.
  27. Urata, S.; Irisawa, J.; Shinoda, A.T.; Tsuzuki, S.; Mikami, M. J. Phys. Chem. B 2005, 109, 4269.
  28. Devanathan, R.; Venkatnathan, A.; Dupuis, M. J. Phys. Chem. B 2007, 111(28), 8069.