Modeling can provide crucial insight into the magnitude of the carbon dioxide poisoning effect and possibilities for performance improvement. As an initial investigation, we have created a 1‑D model of carbonate transport through the thickness of the membrane electrode assembly (MEA) based on dilute solution theory. The model considers the kinetics of CO2 hydration and dehydration by both the carbonic acid and alkaline reactions2,
CO2 + H2O ↔ H2CO3 | [1a] |
H2CO3 + OH- ↔ H2O + HCO3- | [1b] |
CO2 + OH- ↔ HCO3- | [2] |
Additionally, the acid-base equilibrium between carbonate, bicarbonate, and hydroxide,
HCO3- + OH- ↔ CO32- + H2O | [3] |
is considered. The model is used to predict the pH profile through the MEA and quantify the effects of temperatures, flow rates, and CO2concentration on HEMFC performance.
Figure 1 shows the pH profile in the anode, membrane, and cathode for the case of O2 with 400 ppm CO2 as the oxidant. At open circuit, the pH is lower at the cathode than at the anode due to the difference in CO2 partial pressure, although the gradient is small. Once current is applied, the ionic potential gradient drives bicarbonate and carbonate towards the anode, and the pH gradient reverses. At moderate and high current densities, nearly all CO2is removed from the cathode gas stream and electrochemically pumped into the anode gas stream.
The electrochemical pumping of carbonate and bicarbonate towards the anode lowers the anode pH until the rate of CO2 dehydration balances the flux of CO2 through the membrane. Carbonate and bicarbonate carry both CO2 and charge simultaneously. When the CO2flux through the membrane is small relative to the current density, the current is carried either by hydroxide ions or by counterbalanced fluxes of carbonate and bicarbonate. A large pH gradient is required to drive the counterbalancing diffusion of bicarbonate from anode to cathode. At high pH, charge transport shifts from the carbonate-bicarbonate shuttle to hydroxide ions, and the pH gradient flattens. In Figure 1, these dynamics are illustrated by the steep pH gradient from ca. 10-12, and the much flatter gradient above a pH of 12.
Figure 2 shows the effect of 400 ppm carbon dioxide on MEA performance for varying reactant flowrates. Increasing the cathode flowrate increases the amount of CO2 electrochemically pumped into the anode gas stream. Reducing the anode flowrate provides less dilution for the pumped CO2, increasing the anode partial pressure of CO2. Both of these factors force the anode pH lower to accelerate CO2 dehydration, leading to larger pH gradients and reduced performance. Even at the high operating temperature of 95 °C, the introduction of 400 ppm CO2into the cathode gas stream can cause roughly 0.2 V of performance loss.
References
1. J. R. Varcoe et al., Energy Environ. Sci., 7, 3135–3191 (2014).
2. X. Wang, W. Conway, R. Burns, N. McCann, and M. Maeder, J. Phys. Chem. A, 114, 1734–1740 (2009).
3. B. P. Setzler, Z. Zhuang, J. A. Wittkopf, and Y. Yan, Nat. Nanotechnol., 11, 1020–1025 (2016).
Figure Captions
Figure 1: pH profile through the thickness of the MEA over a range of current densities. Conditions: temperature: 95 °C; anode gas: 100 kPa H2 at 80 sccm/cm2; cathode gas: 100 kPa O2 with 400 ppm CO2 at 200 sccm/cm2; ionomer bulk conductivity: 12.5 mS/cm; electrode ionomer volume fraction: 0.15; electrode ionomer tortuosity: 3. Water and gas transport are excluded from the model. Other conditions match ref. 3.
Figure 2: Simulated polarization curves for HEMFCs with varying CO2 concentration and flow rates. Key lists CO2 concentration in cathode feed and (anode:cathode) flow rates expressed as fixed flow rates in sccm/cm2 or stoichiometric flow rates (‘x’). Anode and cathode fixed flow rates of 80 sccm/cm2 and 200 sccm/cm2 correspond to stoichiometries of 11.5 and 12.0, respectively, at 1000 mA/cm2. Other conditions listed in Figure 1 caption.