1269
Effect of Oxidation Level on  Methanol and Proton Transport Characteristics of Graphene Oxide (GO) Based Proton Exchange Membrane (PEM)

Tuesday, May 13, 2014: 09:00
Bonnet Creek Ballroom XII, Lobby Level (Hilton Orlando Bonnet Creek)
A. Paneri and S. Moghaddam (University of Florida)
Development of a PEM with insignificant methanol permeability at highly concentrated methanol supply is considered a major advancement in the direct methanol fuel cell (DMFC) technology. GO, a popular precursor for graphene synthesis,  is decorated with surface oxidative groups (-COOH, -OH, C=O, etc.), which render proton conductivity to GO under hydrated conditions [1]. Its proton conductivity, along with a highly impermeable graphene backbone makes GO a suitable material for a highly selective PEM. We prepared a GO membrane, which exhibited two orders of magnitude higher proton to methanol selectivity, as compared to Nafion [2]. The membrane didn’t suffer any open cell voltage drop, even at 5 and 10 M methanol solution.

GO is prepared by chemically exfoliating graphite flakes via oxidation. Different recipes have been reported in literature for oxidation of graphite, which result in GO with different oxidation levels. Oxidation level of GO is known to impact its physiochemical properties (flake size, surface defects), which should affect mass transport of species across it. Herein, we present effect of oxidation level of GO on its proton and methanol transport characteristics. At first, we studied the impact of mean flake size on these transport processes, by preparing GO using Hummer’s method, and varying its mean flake size via sonication, while keeping a constant oxidation level. A 10 µm thick GO laminate is prepared by vacuum filtration of its aqueous dispersion. Methanol permeability of the GO laminate is observed to be varying linearly with flake size (Fig.1). This is because, being similar in size to water, methanol molecules also follow a tortuous transport path around a GO flake [3]. Hence, changing the flake size changes the diffusion path length for methanol. On the contrary, proton conductivity changed rather insignificantly with flake size (Fig.1). This could be due to the presence of proton selective atomic formations on the GO surface, which allows protons to be transported through the GO flake (Fig.3). We conducted a TEM study on the GO flakes, which revealed presence of surface defects of different sizes. Based on the size of these defects, possible proton transport paths across the GO flakes is discussed in Fig.4.

GO with three different oxidation levels is prepared using process developed by Marcano et al. [4] by varying the weight ratio of graphite flakes to KMnO. GO 1 to 3 (refers to products with an increasing oxidation level) is prepared by using weight ratios (Graphite: KMnO4) of 1:2, 1:4 and 1:6 respectively. Mean flake size of GO is reduced with increase in the oxidation level; mean flake sizes for GO-1, GO-2 and GO-3 measuring 77, 56 and 34 µm. Reduction in flake size is due to unzipping of GO flakes under oxidative environments. Therefore, methanol permeability across the GO laminate changes with oxidation level (Fig.2). However, unlike in Fig.1, methanol permeability doesn’t change linearly with flake size; rather it varies polynomially (order 3). This is because, in addition to breaking down of flakes, surface defects on GO surface also enlarge under the oxidative environment. This further adds to the methanol permeability across the GO laminate. Proton conductivity also increases with increase in the oxidation level. As mean flake size doesn’t have a significant effect on proton conductivity of GO flakes, this increase is primarily due to increase in the number of surface oxidative groups, which are responsible for the in-plane ionic conductivity of the GO flakes.

[1]       Karim et al. J. Am. Chem. Soc. 135 (2013) 8097−8100.

[2]       Paneri et al. J. Power Sources (2013)(Under review).

[3]       Boukhvalov et al.Nano Lett. 13 (2013) 3930.

[4]       Marcano et al. ACS Nano 4 (2010) 4806.