Electrocatalysis and Kinetics of the Direct Alcohol Fuel Cells: DEMS and ac Voltammetry Studies

Abstract

The main challenge in the area of fuel cell research is to find a good catalyst that helps in complete oxidation of the fuel at the anode and reduction of the oxygen at the cathode at a low overpotential which consequently would give rise to the maximum cell efficiency. However, for the direct methanol fuel cell (DMFC) operating at low temperature, the main problem that arises at the anode is its poisoning (deactivation) due to the accumulation of the fuel adsorption product (COad) which can only be oxidized at high potentials (> 0.7 V). For low temperature direct ethanol fuel cells (DEFCs), the main problem that arises at the anode, beside its poisoning by ethanol adsorption products (CO_ad and CH_x,ad), is the incomplete ethanol oxidation due to the difficulty of (C-C) bond breaking. In the previous types of fuel cells, a sluggish ORR kinetics was observed at the cathode which results in a large voltage drop. Such behavior is due to strong inhibition of the cathodic ORR, resulting in high overpotentials and therefore, significant deterioration in the energy conversion efficiency of the cell. The slow kinetic behavior stems from the difficulty of (O=O) bond breaking.<br /> In order to model the conditions of continuous oxidation/reduction in a fuel cell, the continuous mass transfer to the electrode surface is necessary. Therefore, mass spectrometry and ac voltammetry measurements presented here were done using the thin layer flow through cell. This thesis aims at a determination of the rate constant of single reaction steps during the oxidation of CO, methanol and ethanol at different platinum surfaces. Towards that aim, I investigated the electrocatalytic oxidation and adsorption rate of methanol (chapter 3) and the electrocatalytic oxidation of ethanol (chapter 4) at different Pt surfaces, using DEMS. In chapter 5, the potential dependence of the bulk and adsorbed methanol oxidation reaction rate (presented by the apparent transfer coefficient, α') and the corresponding Tafel slope of the reaction have been determined under convection conditions using a potential modulation ac voltammetry technique. Finally, as an application of the method presented in chapter 5, my work in chapter 6 aims at the determination of the apparent transfer coefficient and Tafel slope of the ORR at Pt(Poly) electrode.<br /> The electrooxidation of methanol proceeds via the dual pathway mechanism. The first pathway (named ''indirect pathway'') involves the dehydrogenation of methanol to adsorbed CO followed by its oxidation to CO₂. The second pathway (named ''direct pathway'') involves the formation of dissolved intermediates as HCHO and HCOOH which are transported away from the electrode surface by convection. CO₂ current efficiencies and the degree of surface poisoning with CO_ad have been shown to be independent of the electrolyte flow rate; both confirm the parallel pathway mechanism.<br /> As shown above, since CO_ad is the main poison of the anode catalyst layer in the direct alcohol fuel cell, it is better to catalytically oxidize it at a low overpotential. In the present thesis, it has been shown that Ru electrodeposited at Pt is better catalyst than pure Pt. It promotes the oxidation of CO_ad at low potentials according to bifunctional and electronic mechanisms, at high potentials, however, Ru losses its co-catalytic activity. On such bimetallic surfaces, Ru is preferentially adsorbed at steps. Complete blocking of the Pt step sites with Ru shifts the oxidation to the direct pathway (non-CO-pathway) and thus results in low CO₂ current efficiency. It leads also to the inhibition of the methanol oxidation current due to the blocking of the most active Pt step sites necessary for methanol adsorption and oxidation.<br /> Methanol adsorption rates have been determined: at Pt(Poly), the adsorption rate increases with increasing methanol concentration and adsorption potentials. At Pt(331) and Pt(332) electrodes, methanol adsorption rate was doubled with double step density, higher with higher Ru coverage and increase by a factor of 10 per 0.1 V. Increasing step density however lead to a decrease in methanol adsorption rate from 2.2 MLs^-1 at Pt(100) to only 1.8 MLs^-1 at Pt(11,1,1) due to the geometric ensemble effect and the much smaller activity of (111) sites as compared to (100) sites.<br /> A detailed evaluation of the CO₂ and acetaldehyde current efficiencies during ethanol oxidation at Pt(Poly), Pt(11,1,1) and Pt(311) as well as the same single crystal surface modified with Sn has been investigated. Under flow through conditions, during the potentiostatic ethanol oxidation, the amount of CO₂ is negligible. There is no further oxidation of the soluble product at the surface and acetaldehyde is the main oxidation product (current efficiency close to 100 %). At a Pt(311) electrode, a small amount of CO₂ is observed due to the oxidation of the ethanol adsorption product and not due to the oxidation of bulk ethanol as proved by a separate potential step experiments. Acetic acid in addition to acetaldehyde (current efficiency of ca. 50 %) are the main oxidation products. The onset of ethanol oxidation at Sn modified Pt(311) electrode is shifted negatively by 0.2 V. This shift is not associated with CO₂ production; rather acetaldehyde and acetic acid are the main oxidation products.<br /> At the above surfaces, the experimentally determined acetaldehyde current efficiencies are too high if calibration is simply achieved by an electrolyte with a known concentration of the product due to incomplete mixing in the dual thin layer flow through cell. By performing other experiments with i-propanol, I determined a correction factor for that: In that case, the product (acetone) with a faradaic current efficiency of 100 % has a similar diffusion coefficient and volatility as acetaldehyde.<br /> The apparent transfer coefficient (α') and consequently the corresponding Tafel slopes were determined quasi continuously as a function of potential or time (i.e. in the CV or in the potentiostatic experiments) for the oxidation of pre-adsorbed CO, and methanol as well as bulk methanol at Pt(Poly) electrode under convection conditions. This method involves a sinusoidal modulation of the potential and simultaneous recording of the ac and the dc current. This method has the advantage that the transfer coefficient can be determined at a single potential; a wide range of potentials with a constant Tafel slope is therefore not necessary. In control experiments, using adsorbed CO, values previously determined using the H-cell were reproduced. This demonstrates that the method is applicable to the thin layer cell despite of the high electrolyte resistance which was subtracted by applying a simple ac mathematical ac voltage correction. Contrary to the case of the oxidation of adsorbed CO, where the transfer coefficient varies from about 1.5 at low potentials to 0.5 at high potentials due to a change of the rate determining step, the apparent transfer coefficient for the methanol adsorption product is around 0.5 (Tafel slope of 118 mV dec^-1) at all potentials, suggesting that at all potentials the first reaction step, the adsorption of OH, is the rate determining step and not in equilibrium.<br /> As an application of the above mentioned ac voltammetry method, the apparent transfer coefficient and the Tafel slope have been also determined for the ORR at Pt(Poly) electrode under convection conditions. In addition to the high electrolyte resistance, also a correction for the adsorption resistance was performed. Apparent cathodic transfer coefficient of ca. 0.5 (Tafel slope of ca. –120 mV dec^-1) were calculated in all cases independent of the frequency and the electrolyte flow rate. This suggests that, at Pt(Poly) the first electron transfer to oxygen is the rate determining step.<br /> For future work, the following experiments would be most interesting:<br /> 1. Determination of the apparent transfer coefficient for ORR for some other ORR catalysts, e.g., metal oxide based catalysts.<br /> 2. Electrocatalytic oxidation of methanol on Sn modified surfaces vicinal to (100) plane, e.g., Pt(11,1,1) and Pt(311) in order to examine what would be the effect of Sn on the current efficiency for CO₂ during methanol oxidation. <br /> 3. Electrooxidation of CO at Ag modified Pt(311) electrodes whereas Ag is much stable adatom at Pt electrode. In this respect, it would be interesting to check what would be the behavior of adsorbed CO at these surfaces, where CO molecules could adsorb and what would be the effect of Ag on the CO coverage at the surface.