Surface modification of Nafion membranes by ion implantation to reduce methanol crossover in direct methanol fuel cells

Abstract

The surface of Nafion membranes was modified to reduce methanol crossover by ion implantation with proton ions. The methanol crossover of ion-implanted Nafion membranes was reduced without influencing the other properties including the membrane properties. This simple process could offer advantages for direct methanol fuel cells at high methanol concentrations.

---

Sulfonated fluorocarbon polymers such as Nafion have been widely used as proton exchange membranes for direct membrane fuel cell (DMFC) applications due to their high proton conductivity and good thermal, chemical, and mechanical properties.^1–3 However, Nafion is a poor barrier to methanol crossover, which limits its practical applications in DMFC. The methanol crossover to the cathode not only reduces fuel efficiency, but also increases the over-potential of the cathode, thus resulting in lower cell performance.^4,5 Therefore, various surface modification methods, including metal alloy deposition, electron beam irradiation, radiation-grafting, microwave excitation with hexane/hydrogen gases, plasma polymerization with ethylene/ammonia, and plasma etching followed by palladium sputtering, have been developed to suppress the methanol crossover of Nafion membranes.^6–10 However, most of these proposed methods have exhibited a dramatic reduction in methanol crossover together with detrimental changes in other crucial factors, such as proton conductivity, ion exchange capacity, and water uptake.^7,11 Hence, there is still an urgent need to develop an innovative surface modification technique for the significant diminishment of methanol crossover without deterioration of the other factors.

Ion implantation is one of the useful techniques to modify the surface of Nafion membranes. This technique offers several benefits, including surface-specific modification without influencing other bulk properties, room-temperature solid-state processability, and excellent processing controllability and reliability.^12,13 By taking these advantages, the physical surface modification of Nafion membranes by heavy ion implantation has been reported to increase the surface roughness without any change in the chemical structure and shown enhanced performance of polymer electrolyte membrane fuel cells (PEMFC).^14,15 However, to the best of our knowledge, it has not been reported to reduce methanol crossover of Nafion membranes by light ion implantation-induced surface modification without affecting important membrane properties for application in DMFC.

In this study, the surface modification of Nafion membranes using ion implantation with light ions was to reduce methanol crossover in DMFC applications. To this end, the commercially-available Nafion membranes were implanted with accelerated H⁺ ions to reduce the methanol crossover of the Nafion without a significant change in the surface morphology. The surface-modified Nafion membranes were investigated by FT-IR spectroscopy, X-ray photoelectron spectroscopy, and contact angle measurement. The resulting implanted membranes were characterized with the parameters for DMFC applications.

The well-known effects of ion implantation on the surface of a polymer are the chemical bond scission of a polymer chain and the generation of functional groups such as hydroxyl, carboxyl and carbonyl groups. Because the energy of the H⁺ irradiation is 150 keV, the chemical bonds of the Nafion surface can be broken easily by the irradiation, which have quite lower bonding energies than the energy of the H⁺ irradiation (Fig. 1a).^16,17

Fig. 1 (a) The chemical structure of the Nafion and the possible sites of partial segment resulting from ion implantation. FT-IR spectra of original Nafion (b) and ion-implanted Nafion (c) at fluences of 1 × 10¹⁴, 1 × 10¹⁵ and 1 × 10¹⁶ ions cm⁻².

The chemical changes on the Nafion surface were investigated by ATR FT-IR spectroscopy and the results are shown in Fig. 1b. The characteristic peaks of the original Nafion were observed at 1199 and 1139 cm⁻¹ (symmetric CF₂ stretching), 1055 cm⁻¹ (SO₃⁻ symmetric stretching), and 974 cm⁻¹ (C–O–C stretching in the side chain).^18,19 All of the peaks can be assigned to the terminal SO₃H group and two C–O–C linkages of the hydrophilic side chain and the C–F stretching vibrations of the PTFE backbone. After ion implantation, the intensity of the peaks at 1139, 1055 and 974 cm⁻¹ was decreased due to the fact that the hydrophilic side chains (Fig. 1c) were detached from the Nafion backbone by ion implantation.²⁰

However, the broad peaks corresponding to the carbonyl (C=O) at around 1750 cm⁻¹ were generated at a fluence of 1 × 10¹⁵ ion cm⁻² due to oxidation caused by the reaction of the generated radicals during ion implantation with oxygen in the air. At a fluence of 1 × 10¹⁶ ion cm⁻², the peak intensity was decreased due to carbonization.²¹

To investigate the effect of ion implantation on the Nafion surface more accurately, XPS analysis was performed and the results are presented in Fig. 2. The C 1s spectra of the original and implanted Nafion membranes at fluences of 1 × 10¹⁴, 1 × 10¹⁵ and 1 × 10¹⁶ ions cm⁻² are shown in Fig. 2a. The CF₃, CF₂, –OCFSO₂ and C–C peaks of the original Nafion appeared at 293.4, 291.7, 289.5, and 285 eV, respectively.²² After ion implantation, the intensity of the peaks for the CF₃, CF₂ and –OCFSO₂ groups was decreased due to defluorination of the PTFE backbones and scission of the hydrophilic side chains caused by ion implantation. From Fig. 2c, it was re-confirmed that the atomic concentration of the fluorine and sulfur decreased by increasing the fluence. On the other hand, the generation of new peaks such as C–O and (C=O)–O was seen in Fig. 2a and c. These changes could be induced by oxidation caused by the ion implantation.²³

Fig. 2 C 1s (a) and O 1s (b) XPS spectra of the original (1) and ion-implanted Nafion membranes at fluences of 1 × 10¹⁴ (2), 1 × 10¹⁵ (3), and 1 × 10¹⁶ (4) ions cm⁻². Atomic concentration of the ion-implanted Nafion as a function of the fluence (c).

The O 1s spectra of the implanted Nafion are also shown in Fig. 2b. Two kinds of O 1s peaks were observed in the original Nafion, one from the sulfonic acid (around 534.3 eV) and the other from the ether linkage (around 536.9 eV). As the ion fluence increased, the intensity of the peak at 534.3 eV corresponding to the –SO₃⁻ group was decreased, but the peak at 532.5 eV corresponding to the –O–O– group was increased. This change could be ascribed to the formation of peroxide groups, as indicated by the increased peak intensity at 532.7 eV. The scission energies of the chemical bonds in the Nafion structure are 4.86 eV for C–F, 3.64 eV for S–O, 3.6 eV for C–C, 3.58 eV for C–O, and 2.72 eV for C–S bond.²⁴ Therefore, these chemical bonds in the Nafion can be easily broken to beget radicals when it is irradiated with proton ions. The oxygen containing hydrophilic functional groups such as peroxide, COOH, OH, and CO were formed on the Nafion surface when exposed to the ambient air. In addition, in view of the relative amount of the sulfur with respect to other atoms given in Fig. 2c and the O 1s XPS peak of the sulfonic acid (at around 534.3 eV) in Fig. 2b, the sulfur content decreased as the fluence was increased. This result can be explained by the chemical bond breakage of the ether linkage (–COC–) and the carbon–sulfur linkage (C–S) due to its low scission energy confirmed by ATR-FTIR and XPS. Therefore, the intensity of the sulfur and oxygen peaks for the sulfonic acid group decreased significantly by increasing the fluence.

Although ion implantation removed sulfonic acid groups on the surface, hydrophilic functional groups such as COOH, OO, CO, and OH were generated on the implanted surface. The newly generated hydrophilic functional groups helped to maintain the hydrophilicity of the implanted Nafion surface, but the contact angle was increased compared to the pristine Nafion surface (Fig. S1†). This can be attributed to the fact that the reduced sulfonic acid group concentration on the surface increased the contact angle.

The surface roughness of the ion-implanted Nafion was observed with AFM. Fig. 3a–d show the 3D AFM images of the untreated Nafion surfaces having an RMS roughness of 3.06 nm, while the Nafion membranes implanted at fluences of 1 × 10¹⁴, 1 × 10¹⁵ and 1 × 10¹⁶ ions cm⁻² have a roughness of 3.26, 2.61 and 14.82 nm, respectively. In case of the ion-implanted Nafion at 1 × 10¹⁴ and 1 × 10¹⁵ ions cm⁻², there does appear to be a slight smoothing effect due to the ion implantation.²⁵ However this change was very small, being only just above the error of measurement. In the case of the ion-implanted Nafion at 1 × 10¹⁶ ions cm⁻², the RMS roughness was abruptly increased due to the carbonization and roughening of the Nafion surface by high ion fluence.

Fig. 3 AFM images of the original (a) and ion-implanted Nafion membranes at fluences of (b) 1 × 10¹⁴, (c) 1 × 10¹⁵, and (d) 1 × 10¹⁶ ions cm⁻²; (e), (f), (g) and (h) are the dried samples for (a), (b), (c) and (d) after immersion in a 10.0 M methanol solution for 24 h, respectively.

After immersion of the samples in a 10.0 M methanol solution, the surface morphology was dramatically changed as shown in Fig. 3e–h. All of the immersed samples showed a higher RMS roughness than the non-immersed samples due to the formation of hydrophilic clusters by immersing them in a methanol solution. However, the RMS roughness of the immersed samples was decreased by increasing the fluence. This could be attributed to the fact that the size and number of the formed the hydrophilic clusters were decreased by the scission of hydrophilic side chains in the Nafion backbone.

To investigate the ion implantation effect on the surface and bulk properties of the membranes, the membrane properties related to transport were measured such as the water uptake,²⁶ ion exchange capacity,²⁷ methanol permeability,²⁸ and proton conductivity.²⁹ Fig. 4a shows the water uptake as a function of the fluence. The water uptake of the original Nafion membrane was 22.57%, and the water uptakes of the ion-implanted membranes were also approximately 22%. This result implies that the surface of the membranes was not significantly changed by ion implantation, and the water uptake was unchanged. As shown in Fig. 4b and c, the ion exchange capacity and proton conductivity were not significantly affected by ion implantation.³⁰ These results are attributed to the fact that the ion penetration depth through the membrane was extremely thin and well-matched with the simulation data shown in Fig. S3.† Thus, the ion implantation does not have any detectable effect on the bulk properties related to the DMFC performance.

Fig. 4 Water uptake (a), ion exchange capacity (b), proton conductivity (c), and (d) methanol permeability of original and ion-implanted Nafion membranes.

Fig. 4d shows the methanol permeability of the ion-implanted Nafion membranes. With increasing fluence, the ion-implanted Nafion showed an excellent barrier property for methanol. The methanol permeability of the membrane irradiated at a fluence of 1 × 10¹⁶ ions cm⁻² was 58% lower than that of the original Nafion membrane, which may be one of the highest results in the literature, taking into consideration no significant changes in other key factors for membrane properties, such as proton conductivity, ion exchange capacity, and water uptake.^7,11,20 This phenomenon could be explained as follows: (i) the hydrophilic groups generated on the Nafion surface by ion implantation had a lower affinity to methanol than the original sulfonic acid groups; and (ii) the methanol crossover is suppressed by the reduced size of hydrophilic clusters on the Nafion surface.³¹

In conclusion, a simple method to reduce the methanol crossover of Nafion membranes was demonstrated by using ion implantation. Ion implantation caused the scission of hydrophilic side chains of the Nafion and the generation of other hydrophilic functional groups by oxidation in the irradiated regions. Although the implantation depth of proton ions was negligibly thin compared to the bulk membrane thickness, it rendered large resistance to methanol by the decreased affinity to methanol and the reduced size of the hydrophilic clusters on the Nafion surface. The Nafion irradiated at a fluence of 1 × 10¹⁶ ion cm⁻² showed the lowest methanol permeability. The proton conductivity of the ion-implanted Nafion membranes was not significantly affected by ion implantation due to the fact that the ion penetration depth through the membrane was extremely thin and thus ion implantation does not have any detectable effect on the bulk properties. This simple process could be applied in practical DMFC applications of Nafion membranes at high methanol concentrations.

Acknowledgements

This work was supported by the Radiation Technology R&D program through the National Research Foundation of Korea (NRF) grant funded by the Korea government (MSIP, No. 2012M2A2A6013198 and No. 2015M2A2A6A01043822) and the research fund of Chungnam National University.

Notes and references

1. R. Souzy and B. Ameduri, Prog. Polym. Sci., 2005, 30, 644–687.
2. V. Neburchilov, J. Martin, H. Wang and J. Zhang, J. Power Sources, 2007, 169, 221–238.
3. B. Smitha, S. Sridhar and A. A. Khan, J. Membr. Sci., 2005, 259, 10–26.
4. A. Heinzel and V. M. Barragán, J. Power Sources, 1999, 84, 70–74.
5. W. C. Choi, J. D. Kim and S. I. Woo, J. Power Sources, 2001, 96, 411–414.
6. Z. Q. Ma, P. Cheng and T. S. Zhao, J. Membr. Sci., 2003, 215, 327–336.
7. L. J. Hobson, H. Ozu, M. Yamaguchi, M. Muramatsu and S. Hayase, J. Mater. Chem., 2002, 12, 1650–1656.
8. K. Scott, W. M. Taama and P. Argyropoulos, J. Membr. Sci., 2000, 171, 119–130.
9. M. Walker, K. M. Baumgärtner, J. Feichtinger, M. Kaiser, E. Räuchle and J. Kerres, Surf. Coat. Technol., 1999, 116–119, 996–1000.
10. R. Zeng, Z. Pang and H. Zhu, J. Electroanal. Chem., 2000, 490, 102–106.
11. B. Bae, D. Kim, H.-J. Kim, T.-H. Lim, I.-H. Oh and H. Y. Ha, J. Phys. Chem. B, 2006, 110, 4240–4246.
12. F. Z. Cui and Z. S. Luo, Surf. Coat. Technol., 1999, 112, 278–285.
13. R. L. Clough, Nucl. Instrum. Methods Phys. Res., Sect. B, 2001, 185, 8–33.
14. M. Prasanna, E. A. Cho, H. J. Kim, T. H. Lim, I. H. Oh and S.-A. Hong, J. Power Sources, 2006, 160, 90–96.
15. S. A. Cho, E. A. Cho, I. H. Oh, H. J. Kim, H. Y. Ha, S. A. Hong and J. B. Ju, J. Power Sources, 2006, 155, 286–290.
16. Y. Momose, Y. Tamura, M. Ogino, S. Okazaki and M. Hirayama, J. Vac. Sci. Technol., A, 1992, 10, 229–238.
17. K. L. Tan, L. L. Woon, H. K. Wong, E. T. Kang and K. G. Neoh, Macromolecules, 1993, 26, 2832–2836.
18. Z. Liang, W. Chen, J. Liu, S. Wang, Z. Zhou, W. Li, G. Sun and Q. Xin, J. Membr. Sci., 2004, 233, 39–44.
19. D. M. Hillman, S. H. Stephens, S. D. Poynton, S. Murphy, A. L. Ong and J. R. Varcoe, J. Mater. Chem. A, 2013, 1, 1018–1021.
20. L. J. Hobson, H. Ozu, M. Yamaguchi and S. Hayase, J. Electrochem. Soc., 2001, 148, A1185–A1190.
21. E. A. Hoffmann, Z. A. Fekete, L. S. Korugic-Karasz, F. E. Karasz and E. Wilusz, J. Polym. Sci., Part A: Polym. Chem., 2004, 42, 551–556.
22. C. Chen, G. Levitin, D. W. Hess and T. F. Fuller, J. Power Sources, 2007, 169, 288–295.
23. M. A. Golub, E. S. Lopata and L. S. Finney, Langmuir, 1994, 10, 3629–3634.
24. G. Beamson and D. Briggs, J. Chem. Educ., 1993, 70, A25.
25. D. Ramdutt, C. Charles, J. Hudspeth, B. Ladewig, T. Gengenbach, R. Boswell, A. Dicks and P. Brault, J. Power Sources, 2007, 165, 41–48.
26. C. Zhao, X. Li, Z. Wang, Z. Dou, S. Zhong and H. Na, J. Membr. Sci., 2006, 280, 643–650.
27. J.-S. Lee, C.-H. Jung, S.-Y. Jo, J.-H. Choi, I.-T. Hwang, Y.-C. Nho, Y.-M. Lee and J.-S. Lee, J. Polym. Sci., Part A: Polym. Chem., 2010, 48, 2725–2731.
28. H. Lin, C. Zhao, W. Ma, K. Shao, H. Li, Y. Zhang and H. Na, J. Power Sources, 2010, 195, 762–768.
29. C. Li, J. Liu, R. Guan, P. Zhang and Q. Zhang, J. Membr. Sci., 2007, 287, 180–186.
30. M. A. Navarra, F. Croce and B. Scrosati, J. Mater. Chem., 2007, 17, 3210–3215.
31. J. Won, S. W. Choi, Y. S. Kang, H. Y. Ha, I.-H. Oh, H. S. Kim, K. T. Kim and W. H. Jo, J. Membr. Sci., 2003, 214, 245–257.

---

Footnote

† Electronic supplementary information (ESI) available: General experimental methods, contact angle, thermal analysis and penetration depth of 150 keV proton ions. See DOI: 10.1039/c6ra12756h

---