6974
Langmuir 2004, 20, 6974-6976
Microfluidic Devices for Energy Conversion: Planar Integration and Performance of a Passive, Fully Immersed H2-O2 Fuel Cell Svetlana M. Mitrovski, Lindsay C. C. Elliott, and Ralph G. Nuzzo* Frederick Seitz Materials Research Laboratory and Department of Chemistry, University of Illinois at Urbana-Champaign, 600 South Mathews, Urbana, Illinois 61801 Received June 25, 2004 We describe the fabrication and performance of a passive, microfluidics-based H2-O2 microfluidic fuel cell using thin film Pt electrodes embedded in a poly(dimethylsiloxane) (PDMS) device. The electrode array is fully immersed in a liquid electrolyte confined inside the microchannel network, which serves also as a thin gas-permeable membrane through which the reactants are fed to the electrodes. The cell operates at room temperature with a maximum power density of around 700 µW/cm2, while its performance, as recorded by monitoring the corresponding polarization curves and the power density plots, is affected by the pH of the electrolyte, its concentration, the surface area of the Pt electrodes, and the thickness of the PDMS membrane. The best results were obtained in basic solutions using electrochemically roughened Pt electrodes, the roughness factor, Rf, of which was around 90 relative to a smooth Pt film. In addition, the operating lifetime of the fuel cell was found to be longer for the one using higher surface area electrodes.
We describe in this paper the fabrication and performance of a microfluidic H2-O2 fuel cell which is comprised of electrodes fully immersed in a liquid electrolyte to which the fuels are supplied solely through a gas-permeable membrane. The fuel cell is of a passive design which does not require the use of pumps to supply fuel to or circulate electrolyte at the electrodes. The device consists of two 1.23 mm2 platinum electrodes placed 1 cm apart and embedded in gas-exchange-membrane sealed poly(dimethylsiloxane) (PDMS) microfluidic channels. The electrode arrays are fabricated as 0.1 µm thin films deposited on quartz, with the remainder of the device comprised of PDMS. The microchannel network defines the internal boundaries within which the liquid electrolyte is physically confined while serving also as a thin, gas-permeable membrane through which the reactants are supplied to the electrodes. The three layers comprising the device, their design dimensions, and their order of assembly are shown schematically in Figure 1 along with a photograph of the system tested in this study. The details of the fuel cell fabrication, the characteristic dimensions of the device, and the methods of measurement are given as Supporting Information. In the present fuel cell design, the cathodic reduction of oxygen and the anodic oxidation of hydrogen take place at platinum electrodes embedded in the microfluidic channels. The currents sustained by the cell arise solely due to transport through the PDMS gas-exchange membrane that seals the device. Figure 2 shows a series of representative polarization curves recorded by taking the current and voltage readings of the cell for imposed resistances in the range 1-900 kΩ. Several electrolytes were used in the study of the device performance including 0.1 M H2SO4, 0.1 M NaOH, and 1.0 M NaOH. The data in Figure 2 show several important features relevant to the performance of the device. First, the polarization curves and the power density plots demonstrate that the H2-O2 fuel cell can be operated at current densities of around 1 mA/cm2 even in the absence of an * Corresponding author. E-mail:
[email protected].
Figure 1. (a) Schematic of the microfluidic fuel cell design. (b) Photograph of the device. (c) Top view of the three PDMS layers (bottom, middle, and top) comprising the fuel cell and a crosssectional view along the dashed line shown along the bottom and middle layers.
anode/cathode separator membrane by taking advantage of the selective feed of gases toward the electrodes of interest. Using a bipotentiostat, we have been able to measure the amount of gas crossover by monitoring the hydrogen oxidation or oxygen reduction currents at both electrodes, while the gas was supplied solely to one electrode. Results indicate that ∼30% of the gas fed to the electrode of interest was being lost by crossover to the adjacent electrode in the course of 5-10 min. Decreasing the gas partial pressure of either reagent above the membrane resulted in a partial suppression of the reactant crossover but also led to an overall decrease of the power output due to lower gas-exchange rates across the PDMS membrane. The current densities measured in these fuel cells are similar to the mass-transfer-limited current
10.1021/la048417w CCC: $27.50 © 2004 American Chemical Society Published on Web 07/24/2004
Letters
Langmuir, Vol. 20, No. 17, 2004 6975
Figure 3. Variation of the voltage (red), current density (blue), and power density (green) of the cell based on smooth (clear symbols) and roughened (full symbols) electrodes with time, recorded in 1.0 M NaOH.
Figure 2. (a) Polarization curves for the H2-O2 fuel cell recorded at smooth electrodes covered by a thick PDMS membrane in 0.1 M H2SO4 (red), 0.1 M NaOH (blue), and 1.0 M NaOH (purple) and by a thin membrane in 1.0 M NaOH (pink) and at roughened fuel cell electrodes in 1.0 M NaOH covered by a thick (green) and a thin (gray) membrane. (b) Power density vs current density plots derived from the data presented in part a.
densities in microfluidic half-cells with a smooth working electrode covered by an identical PDMS membrane (3.5 mA/cm2 for oxygen reduction and 1.2 mA/cm2 for hydrogen oxidation). Most interestingly, these latter values exceed those measured at an identical bare electrode immersed in bulk electrolyte in a conventional electrochemical cell (0.7 mA/cm2 for oxygen reduction and 0.2 mA/cm2 for hydrogen oxidation, see the Supporting Information). Earlier work on microfluidic fuel cells that do not incorporate an anode/cathode separator membrane utilized the laminar flow of two separate fuel streams along electrodes deposited inside the microchannels.1 The performance of immersed-electrode systems is generally limited by the low solubility of the reactant gas in the liquid stream and requires pumps to sustain dynamic flow. For example, a direct formic acid fuel cell based on laminar flow and using dissolved oxygen as an oxidant was recently reported to produce a maximum current density of 0.8 mA/cm2 and a maximum power density of 170 µW/cm2.2 Only in the case when liquid fuels such as potassium permanganate or vanadium VII/VV compounds were used, the maximum current densities of the laminar flow fuel cell were markedly higher (8 mA/cm2, as reported by Kenis et al.,2 and 35 mA/cm2 at 1.1 V, as reported by Whitesides et al.,1 respectively). As shown in Figure 2, the integration of electrodes within microfluidic channels covered by a gas-permeable membrane yields passive devices (ones not requiring pumps to sustain the reagent flows) that exceed the performance of laminar flow systems that utilize gaseous fuels. Perhaps more important, though, our data (1) Ferrigno, R.; Stroock, A. D.; Clark, T. D.; Mayer, M.; Whitesides, G. M. J. Am. Chem. Soc. 2002, 124, 12930-12931. (2) Choban, E. R.; Markoski, L. J.; Wieckowski, A.; Kenis, P. J. A. J. Power Sources 2004, 128, 54-60.
reveal that the PDMS gas-exchange membrane does not limit the sustainable currents of the device over its full range of operation but, rather, enables current densities that exceed the current diffusion limit reported for similar dynamic systems (1.4 mA/cm2 vs 0.8 mA/cm2).2 We attribute these properties to the fact that the performance of this system is not limited by the low bulk solubility of the gases in the electrolyte but by that of the PDMS membrane, which is higher by ∼5-6 times for both gases.3 These impacts on the sustainable current densities, then, are dynamical, given their clear origins in mass-transfer effects. We note that the use of a basic supporting electrolyte (0.1 M NaOH vs 0.1 M H2SO4) results in both a higher open circuit potential (900 mV vs 640 mV) and larger current densities. These differences appear to correlate with expectations for a device limited by the kinetics of the cathode.4 Still, some uncertainties remain. For example, the performances of the devices are sensitive to many aspects of their design rules, dependences that are not fully optimized or understood at the present time. Systems constructed using electrochemically roughened platinum anodes and cathodes with a roughness factor, Rf, of around 90 gave higher limiting current densities and power outputs by a factor of 2 (an increment significantly smaller than the geometric scaling of the electrodes). The latter electrodes are exceptionally porous, however, and depletion effects are expected to greatly complicate their geometric sealing. Using a thinner PDMS membrane (≈30 µm vs ≈0.75 mm) also results in higher limiting current densities. The cell design used in this study purposely omits an integrated anode/cathode separator. The stability and lifetime of the device is limited therefore by the rate at which crossover depolarizes the cell. Figure 3 shows that, when the cell operated continuously at maximum power for 3 h, it lost almost 75% of its output in the first 1.5 h after which the observed currents stabilized. The performance stability is markedly improved for cells based on high surface area electrodes, a result expected for a system better able to deplete the reagent feed reaching the electrodes. At this point, one can envision explicit modifications to the system that would limit the diffusive (or convective)5 mixing of the gases within the electrolyte as well as interdiffusion within the PDMS membrane layer. First, an ionomer separator membrane (e.g., Nafion) between (3) Merkel, T. C.; Bondar, V. I.; Nagai, K.; Freeman, B. D.; Pinnau, I. J. Polym. Sci., Part B: Polym. Phys. 2000, 38, 415-434. (4) Carrette, L.; Friedrich, K. A.; Stimming, U. Fuel Cells 2001, 1, 5-39. (5) Morris, R. B.; Fischer, K. F.; White, H. S. J. Phys. Chem. 1988, 92, 5306-5313.
6976
Langmuir, Vol. 20, No. 17, 2004
the two compartments inside the channels would serve to inhibit crossover. This would enable the two electrodes to be more closely spaced and further compensate for the high Ohmic losses observed within the microfluidic channels. In addition, a system incorporating a fully immersed Nafion membrane sealed by a hydrophobic polymer such as PDMS would eliminate the instability problems associated with the dehydration of the ionomer at higher temperatures, a problem commonly observed in PEM (proton exchange membrane) fuel cell systems.4 The overall output power, however, might best be improved by using elastomer membranes of higher reactant solubilities. Our future reports will explore such modifications along with the use of admetal promoters to enable selective oxidations of other fuel sources (e.g., methanol and formic acid). The results of such studies, ones that give perfor-
Letters
mances that exceed all current microfluidic devices,2 will be reported shortly. Acknowledgment. This work was supported by the Department of Energy (DEFG02-91ER45439) and used the central facilities of the Frederick Seitz Materials Research Laboratory. The authors thank William R. Childs for fabricating the thin PDMS membranes. Supporting Information Available: A discussion and figures on the fabrication of the electrodes and microfluidic devices, the electrochemical measurements in the fuel cells, and the electrochemical measurements of the oxygen reduction and hydrogen oxidation reactions. This material is available free of charge via the Internet at http://pubs.acs.org. LA048417W
