High Rates of Oxygen Reduction over a Vapor Phase–Polymerized PEDOT Electrode

See allHide authors and affiliations

Science  01 Aug 2008:
Vol. 321, Issue 5889, pp. 671-674
DOI: 10.1126/science.1159267


The air electrode, which reduces oxygen (O2), is a critical component in energy generation and storage applications such as fuel cells and metal/air batteries. The highest current densities are achieved with platinum (Pt), but in addition to its cost and scarcity, Pt particles in composite electrodes tend to be inactivated by contact with carbon monoxide (CO) or by agglomeration. We describe an air electrode based on a porous material coated with poly(3,4-ethylenedioxythiophene) (PEDOT), which acts as an O2 reduction catalyst. Continuous operation for 1500 hours was demonstrated without material degradation or deterioration in performance. O2 conversion rates were comparable with those of Pt-catalyzed electrodes of the same geometry, and the electrode was not sensitive to CO. Operation was demonstrated as an air electrode and as a dissolved O2 electrode in aqueous solution.

Both fuel-cell technology for power generation and metal-air batteries for energy storage require an efficient electrode for O2 reduction. Such air electrodes are usually a Pt catalyst embedded in a porous carbon electrode. Despite having a high current density suitable for high-power applications such as vehicle drive systems (14), a number of issues with these electrodes may ultimately limit the use and lifetime of the fuel cell or the storage battery, despite recent improvements (5, 6). For example, the cost of the Pt alone in a polymer membrane fuel cell for a small 100-kW passenger vehicle is substantially greater (at March 2008 prices) than the current cost of an entire 100-kW gasoline engine (7). Several technical issues also arise with the use of Pt catalysts. The Pt particles present in the composite electrode are not fixed in place, and a well-known drift phenomenon (8) by which the particles diffuse and agglomerate over time ultimately diminishes the performance of the fuel cell. Further, Pt is very sensitive to deactivation in the presence of CO, either in the air supply or as a by-product from the use of methanol in the direct methanol fuel cell (913).

Other metallic electrode materials, such as cobalt and Ru/Pt alloys (14, 15), have been explored to overcome some of these problems, but in all cases one or more of the issues remain. In particular, the sensitivity to CO is a particularly difficult problem to overcome. In this work, we have developed a Pt-free air electrode based on a nano-porous, intrinsically conductive polymer (ICP) multilayer structure that offers performance similar to that of Pt-catalyzed electrodes under parallel testing. Because the material is homogeneous, the drift issue is avoided and, being nonmetallic, the catalyst is not sensitive to CO poisoning.

The use of ICPs for catalytic electrodes was investigated early in the history of conducting polymer research and applications. However, success was limited by low conductivity and efficiency, and the instability of the ICP in the environment required for the catalysis (1618). By incorporation of traditional catalytic centers such as Pt into ICPs, a range of catalytic electrodes have been reported (14, 19, 20), but these materials all suffer from many of the same problems as the Pt-C electrode. Recently, the development of chemical polymerization [in particular, a process known as vapor phase polymerization (VPP)] and “designed” ICP derivatives has produced materials with high conductivity, improved ordering and stability, and controllable porosity at the nanoscale (2124). These properties improve the potential of these materials for electrocatalytic applications. Previous studies of PEDOT (25) have not been able to demonstrate catalysis of oxygen reduction; however, the improved thin-film properties obtained via VPP prompted us to reexamine its potential as an electrocatalytic material.

One of the key features of an air electrode is that it must establish a high–surface area boundary between the three active phases: air, the electrolyte, and the catalyst/conductor. To achieve this three-phase interface, we coated a PEDOT electro-active layer onto one side of a hydrophobic, porous membrane (Goretex). The procedure developed here involves plasma polymerization of a binding layer to the polytetrafluoroethylene (PTFE) membrane, followed by VPP of the 3,4-ethylenedioxythiophene monomer to form the PEDOT conducting polymer (see fig. S1 for further details).

A schematic of the cross-sectional structure of the electrode in Fig. 1A shows the intended three-phase interface in the circled region. The designed structure allows access of the air stream from one side of the electrode to a high–surface area, electrochemically active layer of the PEDOT, which is simultaneously in contact with the electrolyte. The Goretex membrane provides a good, although not entirely optimized, starting point because it is highly porous at the micrometer level and, being hydrophobic, does not allow penetration of the aqueous electrolyte into the pores of the membrane. Because the electrical conductivity of PEDOT is still not high enough to provide a low-resistance path to the external circuit, a more electronically conductive under-layer was used. Coating one face of this membrane with a ∼40-nm layer of gold provided the conductor layer without altering the structure of the membrane (Fig. 1B). In the next stage, a 400-nm PEDOT layer is created on one side by VPP (Fig. 1C). The structure of the underlying membrane is still visible after these deposition steps, which indicates that the three-phase boundary has been obtained over a substantial fraction of the membrane. An image of the cross-sections of the complete membrane assembly (Fig. 1D) shows the thin PEDOT layer on the surface of the structure.

Fig. 1.

(A) Schematic representation of the PEDOT/Goretex air electrode. (B to D) Scanning electron micrograph images: (B) The Goretex membrane coated with gold. Scale bar, 5 μm. (C) The PEDOT/Goretex structure. Scale bar, 5 μm. (D) Cross-section of the electrode with thickness measurements of the PEDOT layer. Scale bar, 20 μm.

From the thickness measurements, the mass of the PEDOTcan be determined to be 0.05 mg/cm2 for the optimum layer for this particular Goretex membrane; this optimal thickness will change with pore size and shape of the membrane.

The Goretex/PEDOT electrode was subjected to testing as an air electrode at various pH levels and potentials in a cell that allowed direct contact with air from one side and electrolyte from the other (see fig. S2 for detailed experimental setup). The air reduction current under standard conditions (Fig. 2) shows that the electrode performs well over a wide pH range. The PEDOT membrane provides substantial oxygen reduction current densities at all of the pH conditions studied with the potential of onset of the reduction currents shifting as a function of pH in the expected way. To demonstrate that the underlying gold layer was not actively involved in the catalytic process, we conducted a separate experiment (fig. S3) using a Goretex membrane coated only with gold; this electrode produced substantially lower currents.

Fig. 2.

Steady-state measurements (each point after 1 hour of continuous operation) of the conversion current versus potential at different pH values (black line: 400-nm PEDOT/Goretex; gray line: 45-nm Pt/Goretex). (A) pH1, (B) pH 7, and (C) pH 13 for oxygen reduction from air.

Continuous operation in air was achieved at –0.3 V versus saturated calomel electrode (SCE) for more than 1500 hours at pH = 1; 3 A·hour/cm2 of charge was passed during this test. Testing of the electrochemical characteristics of the electrode after 1500 hours showed no change as a result of this period of operation (fig. S4).

Electrodes were constructed with PEDOT thicknesses ranging from around 40 to 1300 nm. It appears that the 400-nm coating shown in Figs. 1 and 2 is nearly optimal for this membrane pore size. When thicker coatings are applied, the diffusion of species, either in the electrolyte, or of oxygen through the PEDOT, becomes limiting and leads to lower currents.

For comparison purposes, we created a Pt-catalyzed electrode by depositing a 45-nm Pt layer onto the Au layer. The magnitudes of the conversion currents delivered by the PEDOT electrode are comparable to those of Pt for the same geometrical (membrane) area (Fig. 2). At pH 1, the Pt seems to perform better than the PEDOT electrode by a factor of ∼2, whereas at pH 7 and pH 13, the conversion currents are similar. The active surface area is actually considerably higher in the Pt case, because of the pore-filling effect of the thicker PEDOT layer in the Goretex membrane. It is also interesting that, although the thicknesses are different for the Pt (45 nm) and PEDOT (400 nm) layers, the difference in their densities (21.1 g/cm3 for Pt and ∼1.2 g/cm3 for PEDOT) means that the mass loading of active material is actually lower in the PEDOT case by a factor of ∼2. The electrocatalytic performance of the PEDOT material (up to ∼0.2 A/mg) is also similar to that of other recently reported cobalt-based materials (14).

The highest room-temperature current density observed with the membranes described here is ∼6 mA/cm2. This value is sufficient for some metal/air batteries and a number of fuel-cell technologies including small direct methanol fuel cells, micro fuel cells, and the various biofuel cell concepts (9, 26). Higher–current density fuel-cell application of the PEDOT electrocatalyst concept would require extension of the three-phase interface into a thicker membrane structure.

The performance of the PEDOT and Pt-based assemblies is compared in Fig. 3A for different levels of CO contamination in the air supply; the PEDOT electrode is not affected, whereas the Pt electrode is poisoned very rapidly under identical conditions. The formation of carbonyl complexes of Pt at the surface that poisons its activity is unlikely with PEDOT. The effect of oxygen partial pressure in the gas supply (air = 20%) (Fig. 3B) demonstrates that the electrode is capable of even higher currents than are generated in air and that no limit related to processes within the PEDOT is being reached over the range of oxygen contents probed.

Fig. 3.

Response of the PEDOT air electrode (black line) to different gas supplies (–0.3 V versus SCE, 0.1 M phosphate buffer, pH 7). (A) Current versus time in air contaminated by 10% CO compared to a similar Pt-catalyzed electrode (gray line). (B) Current as a function of oxygen content in the gas supply.

In a similar series of tests, the sensitivity of the electrode to the presence of methanol in the electrolyte was examined. Methanol crossover from the anode to the cathode is a major issue in the direct methanol fuel cell. The oxidation of methanol is seen to be a competitive reaction to O2 reduction on the cathode side (9). In the present work, a 1% addition of methanol was found to decrease the steady-state current at –0.3 V versus SCE by 20%, at which point the current again reached a steady value. Removal of the methanol brought the current back to its original value, indicating that no permanent damage to the electrode had been caused.

Further insight into the mechanism of the processes taking place in these PEDOT electrode assemblies is provided by Fig. 4, which shows the electronic conductivity of the membrane as a function of the applied potential in an aqueous system. PEDOT in the absence of O2 adopts a variable state of oxidation as a function of potential between about – 0.5 to +0.5 V versus Ag/AgCl in aqueous solution (fig. S5). It is transformed from a low-conductivity material in its reduced state to a highly conductive material in its fully oxidized state. Operating the PEDOT air electrode at various potentials shows a conductivity profile with much higher conductivity at lower potentials (Fig. 4) compared to PEDOT in the absence of air (fig. S5), indicating that the PEDOT is reaching a steady-state oxidation level according to the applied potential, which is greater in the presence of air. The mechanism of the air reduction electrocatalysis likely involves a redox cycling process where the PEDOT, which naturally rests in an oxidized form, is momentarily reduced by the action of the electrochemical cell. An O2 molecule then absorbs onto the surface of the PEDOT and rapidly reoxidizes the PEDOT to its preferred oxidized state and is itself reduced in the process. The role of the counterion in this mechanism, if any, is still unclear.

Fig. 4.

Conductivity (σ) versus potential (E): Conductivity in PEDOT air electrode (0.1 M phosphate buffer, pH 7).

Given the similarity between the Pt and PEDOT responses in Fig. 2, it seems likely that the O2 reduction proceeds via the four-electron pathway as it does on Pt, because there is no sign of an additional process that might indicate a contribution from H2O2 formation. Further investigation of the type described by Halseid et al. (27) is under way to probe the selectivity with respect to the four-electron pathway. Recent reports (28) describe efficient iron-loaded graphite catalysts for oxygen reduction. Given the very low Fe loadings involved, we cannot exclude the possibility of a role for residual Fe, at levels below the limit of x-ray photoelectron spectroscopy detection, in the mechanism reported here. However, iron-based catalytic centers would normally be expected to show signs of poisoning in the presence of CO; the resistance to CO poisoning seen here (Fig. 3) thus suggests that iron centers do not play a notable role.

A laboratory Zn/air battery was also constructed based on this PEDOT air-electrode assembly and a 1 M KOH electrolyte. An open-circuit voltage of 1.44 V was measured, comparable with other examples of this cell (29, 30). Discharge characteristics (fig. S6) as a function of current density and over a 48-hour continuous test were superior to similar devices constructed with a Pt/Goretex air electrode.

The electrode described here provides only a partial solution to some of the problems with the use of Pt discussed in the introduction, because Pt is also used in the anode (fuel) electrode in the fuel cell. However, the fundamental mode of catalysis at work in the present materials may be able to be extended to other reactions, such as the hydrogen oxidation reaction, by careful choice of the ICP. ICPs can be successfully used as a substitute for Pt in dye-sensitized solar cells for the I/I 3 redox reaction (31). Thus, the development of the gas-ICP-electrolyte three-phase interface electrode reported here may provide a platform for a new generation of metal-free electrocatalysts.

Supporting Online Material

Materials and Methods

Figs. S1 to S6


References and Notes

View Abstract

Stay Connected to Science

Navigate This Article