Low-Temperature Water Dynamics Study in Fully Hydrated Nafion® Membrane

Nafion^® membrane has been widely used as a key component in high performance fuel cells due to its high proton conductivity and slow chemical aging characteristics^[1]. Although many studies have been published on this subject, the hydrated membrane performance at low temperatures is not fully understood^[2]. The corresponding ambiguities reported in the literature may be attributed to the state of water in very narrow channels within the membrane. In this work, in order to better understand the water dynamics in this system, we combine the interpretation of our DSC study results with the data obtained using the Ru(bpy)₃²⁺ luminescence probe.

In the DSC experiment performed using 10 °C/min heating and cooling rates, the enthalpy of melting of water was found to be 253 J/g, which can be rationalized as if 24 wt% of water remained unfrozen when the sample was cooled to -150 °C. The difference in enthalpies of the water freezing and the ice melting peaks indicates that water freezing starts at -38 °C; about 34 wt% of water is frozen at -65 °C and water continues to freeze as the temperature decreases to -150 °C. The thermogram also reveals that the frozen water in Nafion^® membrane melts at -22 °C, which is consistent with the previously established fact that the water freezing and melting temperatures are sample size dependent^[3].

To gain deeper understanding of this peculiar behavior of water, the Ru(bpy)₃²⁺ luminescence probe, which is sensitive to spatial confinement, was applied^[4]. The emission spectrum maximum was measured as a function of temperature; the uninterrupted hypsochromic shift down to -105 °C was observed. Two crossover points were noticed (-30 °C, -71 °C); they can be attributed to some phase transitions of the polymer matrix. The crossover at -105 °C can be attributed to the complete freezing of water since the Ru(bpy)₃²⁺ spectrum hypsochromic shift is negligible below that temperature. The NMR measurements reported in Reference 2 were performed in the temperature range from +25 °C to -70 °C, while our measurements were performed from +25 °C to -150 °C and that is the reason for the discrepancy in the water dynamics interpretation.

References:

[1] V. Rao et al, CRC Press, 2015, pp. 567–6143.

[2] Ren-Hao Cheng et al, PCCP, 2021, 23, 10899–10908

[3] A. E. Carte, Proc. Phys. Soc. B 1956, 69, 1028–1037

[4] B. H. Milosavljevic and J. K. Thomas; Macromolecules, 1984, 17, 2244–2248