Carbon corrosion measurements

In acid electrolytes, carbon is a poor electrocatalyst for oxygen evolution at potentials where carbon corrosion occurs. However, in alkaline electrolytes carbon is sufficiently electrocatalytically active for oxygen evolution to occur simultaneously with carbon corrosion at potentials corresponding to charge conditions for a bifunctional air electrode in metal/air batteries. In this situation, oxygen evolution is the dominant anodic reaction, thus complicating the measurement of carbon corrosion. Ross and co-workers [30] developed experimental techniques to overcome this difficulty. Their results with acetylene black in 30 wt% KOH showed that substantial amounts of CO in addition to C02 (carbonate species) and 02, are... 

In this chapter, we will review the fundamental models that we developed to predict cathode carbon-support corrosion induced by local H2 starvation and start-stop in a PEM fuel cell, and show how we used them to understand experiments and provide guidelines for developing strategies to mitigate carbon corrosion. We will discuss the kinetic model,12 coupled kinetic and transport model,14 and pseudo-capacitance model15 sequentially in the three sections that follow. Given the measured electrode kinetics for the electrochemical reactions appearing in Fig. 1, we will describe a model, compare the model results with available experimental data, and then present... 

Figure 3. Carbon corrosion rate versus carbon weight loss for both conventional and graphitized KB-supported Pt catalysts. The carbon corrosion rate (in units of A/g( ) is based on the measured CO2 concentration at the exit of a 50 cnr cell using a GC, assuming 4e /( (T molecule. The carbon weight loss is obtained by integrating the measured CO2 evolution rate over time. The cell is operating on neat H2/N2 (95 °C, 80% RIIjn, and 120 kPaa, s) with potential held at 1.2 Volts versus RHE.

Figure 4. Polarization curves of carbon corrosion and oxygen evolution reactions based on measured carbon corrosion kinetics for Pt/Vulcan and Pt/Graphitized-Vulcan and oxygen evolution kinetics for Pt/C catalysts. The upper horizontal dotted line denotes a current density equivalent to oxygen crossover through membrane from cathode to anode.

Figure 16. (a) Real time CO2 measurement at the exit of the cathode during H2/air-front start-stop events with a residence time of 1.5 s based on anode void volume (including flow-field and diffusion medium). More carbon corrosion occurs during start than during stop, (b) Ratio of the integrated carbon loss (proportional to the integral of CO2 concentration over time) for stop over start versus residence time. The ratio approaches unity as residence time increases. 

Relationships such as that in Eq. (12) offer convenient means of testing the validity of mixed potential models by comparing electrochemically determined parameters (in this case, a reaction order based on measured Tafel slopes) to values measured by other means. One such example would be the corrosion of U02 (nuclear fuel) in aerated neutral solutions containing added carbonate (6). In the presence of carbonate, corrosion product deposits are avoided, since the U02+ corrosion product is solubilized by complexation with the carbonate. Measured Tafel slopes yield a predicted reaction order of n0l = 0.67 with respect to 02 for the overall corrosion reaction ... 

Figure 19.4. Intergranular corrosion measured by change in eiectrical resistance of 18-8 stainless steels containing nitrogen or carbon immersed in 10% CUSO4+ 10% H2SO4. All specimens sensitized at stated temperatures for 217 h [19], (Reproduced with permission. Copyright 1945, The Electrochemical Society.)...

Most work on linear polarization probes has been done in chloride corrosion condition. Ho vever, the only methods of assessing carbonated concrete are destructive drilling or coring for carbonation depth measurement and trying to interpret half cell potentials which is difficult (Section 4.7.2). Linear polarization is therefore very useful in asse.ssing carbonated structures, particularly as half cell potentials are so difficult to interpret for carbonation induced corrosion. 

Local hydrogen starvation is usually observed in the cell before cell reversal occurs. By measurement of the CO2 concentration in the cathode exhaust gas, carbon corrosion can be detected. However, it is extremely difficult to distinguish between CO2 evolution within the cell and its amount present in air due to low concentrations. Alternatively, critical conditions can be detected by integration of a reference electrode (see Section 20.2.1). Since the reference electrode has to be positioned within the starved region, several reference electrodes have to be used per cell in order to make detection of the starved region probable. This can only be realized in a laboratory cell and is not practical for automotive stacks. 

In order to overcome the carbon corrosion problem, several strategies have been developed. Carbon materials with more graphite component showed improved corrosion resistance capability [59-61], Carbon nanotubes (CNT) have been found to be more corrosion-resistant than Vulcan carbon [62-66]. In 168 hours of measurement under a simulated PEMFC cathode environment, multiwalled carbon nanotube (MWCNT) showed a 30% lower corrosion current than Vulcan carbon... 

The issue of carbon corrosion has received considerable attention in recent years. There are several drivers for this (1) the cost drivers for commercialization require the use of high performance catalysts with less durable carbon catalyst supports, (2) the need for system simplification and low cost prevents additional control systems to be implemented to avoid the carbon corrosion conditions, and (3) the use of the fuel cells subjected to "real world" conditions as opposed to carefully controlled demonstration projects, with very dynamic duty cycles and many start-up/shutdown cycles. This increased attention has resulted in new or improved measurement techniques and several studies and reviews on the high cathode potential and associated carbon corrosion mechanism [39,40,48-51]. 

This chapter therefore mainly addresses two questions how can one precisely and reliably measure the ECSA in HT-PEMFC and how can carbon corrosion be mitigated on an operation and material-based level ... 

Roen et al. (2004) examined the effect of platinum on carbon dioxide emissions three synthesized in-house MEAs (carbon, 10 wt%and 39 wt%platinum supported on carbon catalysts) were potential cycled between 0 and 1 V (vs RHE) at 65 °C (100% RH, hydrogen/air or oxygen as reactants) and their carbon dioxide emissions were measured by mass spectrometry. The presence of Pt enhanced carbon corrosion rate since Pt catalyzes CO2 formation at low potentials (-0.55-0.65 V vs RHE) (Willsau and Heitbaum, 1984) and increases CO2 emission rates at 1V (vs SHE) (Roen et al., 2004). It was also reported that the carbon corrosion rate is enhanced as the range of potential cycling is increasing (higher anodic and lower cathodic potentials) due to the formation of defects (Stevens et al., 2005) on carbon support by chemical oxidation in low potentials and the presence of a harsh electro-oxidation envirorunent at high potentials (Maass et al., 2008). 

To test the degradation of fuel cell catalyst and assess the carbon support degradation effect on fuel cell performance, many diagnostic tools are available. These tools may test the morphology of the catalyst support directly or may evaluate the carbon corrosion indirectly through the fuel cell overall performance. Common parameters analyzed to evaluate the electrocatalyst degradation include measurement of the catalyst layer areas (cross-sectional and smface area), the ECSA, fuel cell current density, surface morphology, and elemental composition of material or effluent gas. 

Fig. 2 The effect of potential on carbon corrosion rate versus time over commercial conventional-carbon-supported MEAs measured at 80°C and 80% RH. j. The working electrode was fed with Nj (50cm min ), while the counter/reference electrode was purged with (200cm min )...

The gas-phase oxidation of carbon blacks by oxygen and/or water is strongly catalyzed by the presence of catalytically active metals, such as platinum (Rewick et al. 1974, Stevens and Dahn 2005), whereby several weight percent of platinum on carbon can increase the gas-phase oxidation rate by orders of magnitude. This, however, is not the case for the electrochemical oxidation of carbon blacks, where at potentials of 0.8 V and higher (vs. RHE) the carbon corrosion rate is within a factor of 2 between that for noncatalyzed and platinum-catalyzed carbon blacks (Roen et al. 2004, Passalacqua et al. 1992, Kinoshita 1988). Therefore, gas-phase oxidation tests to screen potential carbon-black supports is not a reliable method for predicting their stability in the electrochemical environment, so it is essential to measure the carbon corrosion rates directly in an electrochemical cell. 

Fig. 20 Predicted relative carbon corrosion rate compared with the measured relative voltage degradation rates at two applied current densities and a H /air front residence time of 1.3 s. The steutup/shutdown improvement factor is defined as the inverse of the d radation rate of a cxmventional carbon MEA over the degradation rate of a Vulcan MEA or a graphitized carbon MEA. The mcxlel predictions were based on HOR kinetics (Neyerlin et aL 2007), ORR kinetics (Neyerlin et aL 2006), COR kinetics (Yu et al. 2006a), and OER kinetics (Yu et al. 2006b). (Reproduced with permission of Yu et til. (2006a), The Electrochemiccd Scxaety)...

Fundamental model analyses incorporating the measured carbon corrosion kinetics were developed for conditions of start/stop or local starvation. The combination... 

---