Novel polyelectrolyte membranes for fuel and flow batteries: Insights from simulations Soumyadipta Sengupta, Giorgos Kritikos, Konstantinos Karatasos, Arun Venkatnathan, Rakesh Pant, Pavel Komarov, and Alexey V. Lyulin
Citation: AIP Conference Proceedings 1981, 020004 (2018); doi: 10.1063/1.5045866 View online: https://doi.org/10.1063/1.5045866 View Table of Contents: http://aip.scitation.org/toc/apc/1981/1 Published by the American Institute of Physics
Articles you may be interested in Preface: 9th International Conference on “Times of Polymers and Composites” AIP Conference Proceedings 1981, 010001 (2018); 10.1063/1.5045862 Durability of crosslinked polyethylene in human-contact applications: Stabilisation challenges AIP Conference Proceedings 1981, 020002 (2018); 10.1063/1.5045864 Shape memory electrospun nonwovens based on crosslinked poly(ε-caprolactone) for multifunctional biological applications AIP Conference Proceedings 1981, 020006 (2018); 10.1063/1.5045868 Antimicrobial biodegradable coatings based on LAE for food packaging applications AIP Conference Proceedings 1981, 020010 (2018); 10.1063/1.5045872 Crystallization kinetics of blends of two poly(lactic acid) grades with diverse stereoregularity and molar mass AIP Conference Proceedings 1981, 020008 (2018); 10.1063/1.5045870 Frictional properties of cartilage loaded against cartilage by using a pin on disc tribometer AIP Conference Proceedings 1981, 020011 (2018); 10.1063/1.5045873
Novel Polyelectrolyte Membranes for Fuel and Flow Batteries: Insights from Simulations Soumyadipta Senguptaa, Giorgos Kritikosb, Konstantinos Karatasosb, Arun Venkatnathanc, Rakesh Pantc, Pavel Komarovd and Alexey V. Lyulina* a)
Theory of Polymers and Soft Matter Department of Applied Physics, Technische Universiteit Eindhoven, 5600 MB, Eindhoven, The Netherlands *Email:
[email protected] b) c)
d)
School of Chemical Engineering, Aristotle University of Thessaloniki, 54124 Thessaloniki, Greece
Department of Chemistry and Center for Energy Science, Indian Institute of Science Education and Research, Dr. Homi Bhabha Road, Pashan, Pune 411008, Maharashtra, India
Institute of Organoelement Compounds, Russian Academy of Sciences, Vavilova st. 28, Moscow 119991, Russia
Abstract. Recent experiments on polyelectrolyte membranes have clearly shown that at operating temperatures perfluoroimide acid (PFIA) has a higher electrical conductivity than widely used Nafion. In the present paper classical molecular-dynamics simulations were carried out to study the structural properties of both materials, and the proton and water transport in the corresponding membranes at T=300 K and T=353 K. In this temperature range, the temperature effects on the hydrated internal polyelectrolyte structure were found to be negligible. The PFIA has longer side chains across a wide range of hydration levels which would have promoted more trapping of water and hydronium ions in PFIA. Indeed, the average number of water molecules in the first hydration shell around the side-chain protogenic groups was found to be higher in PFIA than in Nafion. Our simulations showed the formation of large continuous water clusters and connected pore volumes in PFIA at high hydration levels which promotes conductivity. The diffusivity constants for hydronium ions and water increase with increasing hydration and increasing temperature. Unlike the experimental conductivities, the simulated data for PFIA were comparable to those of Nafion at high hydration levels. Note that the experimentally measured conductivity in PFIA is both due to vehicular transport of ions, which can be resolved using classical molecular dynamics, and structural Grotthuss diffusion, which cannot be resolved in our simulations. Interestingly, we observed a higher total number of water molecules in the first coordination shell around hydronium in PFIA than in Nafion at higher hydration levels. This should aid in more hydrogen bonding between hydronium and water in PFIA which, in turn, should help in structural diffusion. Finally, we discuss our preliminary results and some peculiarities of the proton transport in Nafion membranes filled with the graphene oxide nanoflakes. Keywords: molecular dynamics, flow battery, polyelectrolyte membrane, hydronium diffusion, nanocomposite. PACS: 82.35.Rs; 82.35.Np; 82.45.Np; 82.47.Nj;68.35.bm; 68.05.cf
INTRODUCTION Polyelectrolyte membrane (PEM) is a key component in flow batteries and polyelectrolyte membrane fuel cells (PEMFC). In organic and inorganic flow batteries, the polyelectrolyte membrane separates the catholyte and anolyte and allows the passage of protons1-3. Crossover of electrolytes in a battery results in higher charging and shorter discharging times4, and, ideally, should be prevented. In a fuel cell, the polyelectrolyte membrane separates the hydrogen and oxygen streams and also allows the passage of protons. The key characteristic - electrical conductivity of the membrane - defines the performance of these batteries5. Importantly, membranes need to retain conductivity at elevated temperatures of fuel cells due to the high amounts of heat generated.
9th International Conference on “Times of Polymers and Composites” AIP Conf. Proc. 1981, 020004-1–020004-4; https://doi.org/10.1063/1.5045866 Published by AIP Publishing. 978-0-7354-1697-0/$30.00
020004-1
Nafion is a widely used membrane for both flow batteries3 and PEMFCs6, has good conductivity and mechanical stability at room temperature, but it loses water at high temperatures. Due to this loss of water, the conductivity of Nafion degrades6,7. Alternative membranes, which have high conductivity at both low and high temperatures, are urgently needed. Multiple acid side chain (MASC) PEMs, which have multiple protogenic groups, like ortho bis acid8, meta bis acid9 and perfluoroimide acid (PFIA)8,10 have been developed to serve this purpose. The existing experiments suggest that the average current through a PFIA membrane is significantly higher than in some Nafion membranes11,12. The clear message is that the effects of the multiple side chain acidic groups and of a longer side chain in PFIA on the transport of ions need to be studied in more detail. Graphene oxide (GO) is a promising filler for improving the properties of PEMFCs. As the GO particle contains epoxy and hydroxyl groups, it is an amphiphilic filler that promotes more water retention. The latter is important for improving the conductivity of protons at high temperatures. Additional advantages of GO particles are large surface areas, high mechanical and chemical stability. In the present manuscript, PFIA and Nafion membranes (also modified by graphene oxide, GO) were simulated across a range of hydration levels, and for different temperatures, using classical molecular-dynamics (MD) simulations. The membrane density was calculated for different hydration levels to benchmark the present simulations with the experimental values, and good agreement was observed. The size of the separate water phase was studied using cluster distribution and pore size distribution. The connected volumes for water in the membranes have been analyzed to get information about the percolation of water clusters at a particular hydration level. The transport dynamics was simulated using mean-square displacements (MSD) of hydronium and water molecules; the corresponding diffusion coefficients were computed for both PFIA and Nafion membranes. A comparison between Nafion and PFIA conductivities was done.
MODELS AND SIMULATION DETAILS
a)
b)
c)
FIGURE 1. a) PFIA chain (n=6, m=10) b) Nafion chain (n=7, m=10); the red ovals highlight the protogenic side-chain groups. c) Graphene oxide (GO) 8 nm x 8 nm nanoflake, with highlighted hydroxyl and epoxy groups..
The PFIA and Nafion chains, FIGURE 1, were constructed using Materials Studio13. The structure of PFIA monomer is shown in FIGURE 1a). A common variety of PFIA has an equivalent weight, EW (i.e., the weight of the PFIA molecule divided by the number of acidic/protogenic groups) of 62514 and n=6. The number of monomers in a single PFIA and Nafion molecule is taken as 10 for this study. The PFIA chain has two different protogenic, sulfonate and the sulfonyl imide, groups. A previous simulation9 study has shown that these protogenic groups lose their hydrogen atoms at different hydration levels Ȝ, defined as the number of water molecules per side chain of the polyelectrolyte14. Five different 3D periodic boxes were constructed corresponding to the five different values of Ȝ using the Amorphous Cell module of Materials Studio13. Each simulation box had 14 chains of PFIA, and the corresponding number of hydronium ions and waters. The simulation (cubic) boxes were periodic in all dimensions, and had a length from ~52 Å to ~ 61 Å for each of the three dimensions for Ȝ in the range from 0 to 20. The maximum number of atoms in a simulation box was 19068 for Ȝ=20.
020004-2
The pcff (Polymer Consistent Force Field)15 and LAMMPS software were used for simulating the PFIA and Nafion matrices, water and hydronium ions in the all not-filled membranes. In the case of Nafion filled with GO, FIGURE 1c), we have used GROMACS software and the OPLSAA Force Field for both Nafion and GO. The GO sheet having dimensions of 8x8 nm2 was modelled with a 5:1 carbon to oxygen atom ratio and a 3:2 hydroxyl to epoxy group ratio.
MAIN RESULTS AND DISCUSSION
a)
b)
c)
FIGURE 2. Snapshots of Nafion box at the end RISURGXFWLRQUXQVIRUD Ȝ and E Ȝ at T=300 K. Green spheres represent hydronium and water molecules. c) Snapshot for Nafion + GO simulation box at the beginning RIWKHHTXLOLEUDWLRQȜ
The FIGURE 2a,b) shows the typical Nafion membrane snapshots at different hydration levels at 300 K. The water and hydronium atoms can be seen to form larger clusters as the hydration level increases. Due to the hydrophilic nature of the side chains in both Nafion and PFIA, they are (preferably) in contact with the water phase and are aligned away from the polymer backbone. Figure 2c) shows the snapshot of Nafion and one GO flake (in the presence of water which is clustering initially) at the end of the equilibration. -5
6
Nafion Sim1(300 K)
2
/s)
Nafion Sim2(300 K) Nafion Sim3(300 K) PFIA(300 K)
Diffusivity constant (cm
Diffusivity constant (cm
0.05
Nafion Sim2(300 K)
5 4
Nafion Sim3(353 K) PFIA(353 K)
1
0.5
0 5
-6
Nafion Sim1(300 K)
0.04
10
a)
15
20
Nafion Sim3(300 K) PFIA(300 K)
Nafion Sim2(353 K)
Nafion Sim2(353 K)
Nafion Sim3(353 K)
3
Nafion Sim1(300 K) Nafion Sim2(300 K)
Nafion Sim3(300 K) PFIA(300 K)
Nafion Sim2(353 K)
1.5
10
Conductivity (S/cm)
10
2
2
/s)
2.5
PFIA(353 K)
2 1 0 5
10
b)
15
20
`
0.03
Nafion Sim3(353 K) PFIA(353 K)
0.02
0.01
0 5
10
15
20
c)
FIGURE 3. PFIA (solid lines) and Nafion diffusion coefficients for a) water and b) hydronium. c) PFIA (solid lines) and Nafion conductivity values. Nafion Sim1 are the simulated values of Nafion using pcff force field. Nafion Sim 2 [16] and Nafion Sim 3 [17] are from previously published Nafion simulation studies.
The mean-square displacements have been simulated using the Einstein relation in the diffusive regime, as averages of all the water molecule atoms and, separately, for all the hydronium ions. The translational diffusion coefficients for water molecules and hydronium ions were computed from the corresponding slopes. The FIGURE 3a,b) shows the simulated vehicular diffusion coefficients in both membranes. Previously published16,17 simulated vehicular diffusion coefficients in Nafion are shown as well. At all temperatures for every Ȝ, the vehicular diffusion coefficients for hydronium ions are almost an order of magnitude smaller than those for water because of the strong attraction between sulfonate groups and hydronium ions. The vehicular diffusion coefficient for water molecules and hydronium ions are very low at
020004-3
Ȝ=5 because of the presence of negligible water-accessible volume which, in turn, restricts the water diffusion. The vehicular diffusion coefficients for water molecules and hydronium ions in PFIA show a significant increase for Ȝ=15 as compared to those for Ȝ=10, at both T=300 K and T=353 K. This increase can be explained due to the emergence of a large continuous water cluster. The vehicular diffusion coefficients in PFIA are comparable to those in Nafion. The FIGURE 3c) shows the simulated vehicular proton conductivity of PFIA and Nafion calculated by using the Nernst-Einstein equation. The PFIA vehicular proton conductivities are observed to be higher than those of Nafion which agrees with the experimental trends; this is connected to the enhanced concentration of hydronium ions in PFIA for Ȝ>10, due to the dissociation of the extra protogenic sulfonyl imide groups. The simulated vehicular proton conductivity values for PFIA at T=353 K are smaller than the experimental data at the same temperature17. The Grotthuss mechanism, which is not taken into account in classical MD, aids in diffusion of protons inside the water clusters, in addition to the vehicular diffusion of protons. Therefore, the difference between simulated and experimental proton conductivity values could be due to this limitations of classical MD simulations. We also discuss the very preliminary results of the molecular dynamics modelling of the Nafion/GO nanocomposites DW GLIIHUHQW K\GUDWLRQ OHYHOV Ȝ 10 and 15). We have successfully characterized the morphology and the concomitant effects on transport inside the hydrated nanostructure for new PEM materials; our simulations, a.o., depict the influence of the Nafion chains ordering (due to the presence of the nanosheet) on the conductivity and on the mechanical properties of the membranes.
ACKNOWLEDGMENTS This work was done as a part of the FOM-SHELL 15CSER13 project and was carried out on the Dutch national e-infrastructure with the support of SURF Cooperative. The project is part of the research programme of the Center of Computational Energy Research. Rakesh Pant thanks IISER Pune for graduate fellowship. Arun Venkatnathan thanks DST Nanomission Thematic Unit (SR/NM/TP-13/2016(G)). G. Kritikos acknowledges the support from the HPC-Europa3 Transnational Access programme.
REFERENCES 1. 2. 3. 4. 5. 6. 7. 8. 9. 10. 11. 12. 13. 14. 15. 16. 17. 18.
S. P. Badwal, S. S. Giddey, C. Munnings, A. I. Bhatt and A. F. Hollenkamp, Frontiers in Chemistry, 2014, 2. P. Alotto, M. Guarnieri and F. Moro, Renewable and Sustainable Energy Reviews, 2014, 29, 325-335. B. Huskinson, M. P. Marshak, C. Suh, S. Er, M. R. Gerhardt, C. J. Galvin, X. Chen, A. Aspuru-Guzik, R. G. Gordon and M. J. Aziz, Nature, 2014, 505, 195-198. S. Won, K. Oh and H. Ju, Electrochim. Acta, 2015, 177, 310-320. L. Napoli, J. Franco, H. Fasoli and A. Sanguinetti, Int. J. Hydrogen Energy, 2014, 39, 8656-8660. S. Bhadra, N. H. Kim, J. S. Choi, K. Y. Rhee and J. H. Lee, J. Power Sources, 2010, 195, 2470-2477. M. A. Ilhan and E. Spohr, J. Electroanalyt. Chem., 2011, 660, 347-351. M. S. Schaberg, J. E. Abulu, G. M. Haugen, M. A. Emery, S. J. O'Conner, P. N. Xiong and S. Hamrock, ECS Transactions, 2010, 33, 627-633. J. K. Clark II and S. J. Paddison, Electrochim. Acta, 2013, 101, 279-292. S. Sengupta, R. Pant, P. Komarov, A. Venkatnathan, A. V. Lyulin, Int. J. Hydrogen Energy, 2017, 42, 27254-268. Q. Duan, H. Wang and J. Benziger, J. Memb. Sci., 2012, 392, 88-94. N. J. Economou, A. M. Barnes, A. J. Wheat, M. S. Schaberg, S. J. Hamrock and S. K. Buratto, J. Phys. Chem B, 2015, 119, 14280-14287. Accelrys Inc., San Diego, CA, 2013. L. Puskar, E. Ritter, U. Schade, M. Yandrasits, S. Hamrock, M. Schaberg and E. Aziz, Phys. Chem. Chem. Phys, 2017, 19, 626-635. H. Sun, S. J. Mumby, J. R. Maple and A. T. Hagler, J. Am. Chem. Soc., 1994, 116, 2978-2987. S.S. Jang, V. Molinero, .Tahir, W.A.Goddard III, J. Phys. Chem. B 2004, 108, 3149–57. A. Venkatnathan, R. Devanathan, M. Dupuis, J Phys Chem B 2007, 111, 7234–44. L. Puskar, E. Ritter, U. Schade et al., Phys. Chem. Chem. Phys. 2017, 19, 626–35.
020004-4
