10.1149/1.2209349, copyright The Electrochemical Society A Microfabricated Nickel-Hydrogen Battery Using Thick Film Printing Techniques W. G. Tam a, and J. S. Wainright a a Department of Chemistry Engineering, Case Western Reserve University, Cleveland, Ohio 44106, USA To utilize the distinctive cycle life and safety characteristics of the nickel-hydrogen chemistry while eliminating conventional high pressure limitations, a microfabricated nickel-hydrogen battery using low-pressure metal hydride for hydrogen storage was proposed. Thick film printing techniques which are simple and low cost were used to fabricate this proposed battery. Inks were developed for the different battery components, including electrodes, current collectors and separator, and SEM images on these printed components showed the desired characteristics for each. Positive electrode cycling tests were performed on the printed positive electrodes while cyclic voltammetry was used to characterize the printed negative electrodes. Consistent charge and discharge performance was observed during positive electrode cycling. Full cells with printed positive and negative assemblies were assembled and tested for cycling. Nickel-Hydrogen Battery Overview Nickel-hydrogen (Ni-H 2 ) battery is a secondary alkaline battery combining battery and fuel cell technologies. Its positive electrode is a Ni(OH) 2 /NiOOH electrode, commonly used in NiCd and NiMH batteries (Equation [1]). The negative electrode is a hydrogen electrode on catalytic platinum, commonly used in hydrogen-oxygen fuel cells (Equation [2]). Concentrated potassium hydroxide ranging from 26 to 31 wt. % has been used as the battery electrolyte (1). The net reaction of the battery during charge and discharge is shown in Equation [3]. Ni(OH) +OH 2 NiOOH+H O+e [1] - charge - discharge 2 [2] - charge - HO+e OH+ 1 H 2 2 discharge 2 ( ) charge 2 discharge Ni OH NiOOH+ 1 H 2 [3] As the net reaction only involves hydrogen reduction of NiOOH to Ni(OH) 2 without a net change in electrolyte concentration or in the amount of water within the battery, this battery system is referred to a maintenance-free battery. During overcharge, oxygen is evolved at the positive electrode, which will be recombined electrochemically at the catalytic negative electrode. Thus, the battery can tolerate continuous overcharge without any adverse effects. Moreover, the battery offers excellent cycle life, allowing 40,000 cycles at 40% DOD for low earth-orbit (LEO) satellites (1). 2 1
Limitations of Conventional Nickel-Hydrogen Batteries Conventional Ni-H 2 battery relies on high pressure hydrogen (55-70 atm) for operation. The use of individual pressure vessels (IPV) to seal the battery leads to high material cost and complicated vessel sealing. In addition, many other mechanical parts are needed for the battery assembly, taking up additional volume and leading to low volumetric energy density. Motivation for Microfabrication As current successes in micro-electromechanical systems (MEMS) have been advancing, there is an urgent need for microscopic power sources (2). Availability of these power sources will permit a truly integrated system to be fabricated, greatly reducing fabrication complexity. In light of this demand, evaluating the possibility of microfabricating a Ni-H 2 battery using thick film printing is attractive. To eliminate the high pressure hydrogen constraint as in the conventional Ni-H 2 battery, we proposed the use of low-pressure metal hydride as on-board hydrogen storage. As the metal hydride only acts as a hydrogen reservoir, metal hydride corrosion, which is the main performance failure for the nickel-metal hydride battery, can be eliminated. Instead of using conventional fabrication techniques, we proposed the use of thick film printing to fabricate this battery. Thick film printing is inexpensive to set up and simple to use. Recently, this technique has been investigated for battery (3) and fuel cell (4) applications. The resulting battery should be small, light, and reliable, with long cycle life. Objective The objective of this work is to investigate the use of thick film printing in fabricating a completely standalone sealed Ni-H 2 battery using low-pressure metal hydride as hydrogen storage. The resulting battery should be rechargeable, showing consistent charging and discharging performance upon cycling, and operate at near ambient hydrogen pressure. The work shown below attempted to produce these goals, but the major focus was on investigating if thick film printing could be used to fabricate a Ni-H 2 battery with rechargeable performance. Design Requirements of Microfabricated Nickel-Hydrogen Batteries Figure 1 shows the fundamental Ni-H 2 battery design for thick film printing. Each battery component, except the porous substrate, will be thick film printed. To ensure the battery performs properly, each battery component must be carefully designed and formulated. 2
Battery Substrate Figure 1: Fundamental Design for Microfabricated Ni-H 2 As this battery relies on metal hydride as on-board hydrogen storage, the battery substrate should be porous enough to allow effective hydrogen diffusion into the negative electrode for its proper operation. It should also be porous enough to allow any oxygen evolved during overcharging to recombine with the negative electrode, so as to maintain water balance within the battery. As each battery component is thick film printed layer by layer onto the substrate, it should also be mechanically robust to handle any squeeze pressure and repeated baking during the printing process. The substrate selected for this work is a Teflon membrane from GE Osmonics Labstore. It is 50 μm thick with pore size of 0.22 μm, and is stable up to 260 C. Positive Electrode For the positive electrode to function properly, efficient charge transfer must occur both from the current collector and within the solid-state active material (5). Thus, electrical conductivity is critical for the electrode s performance. The positive electrode should also be porous to allow any oxygen evolved during overcharge to diffuse out. In this work, β-nickel hydroxide powder (Kansai Catalyst Co., Ltd) has been used as the positive active material. The 12-μm powder is 2.5 wt.% cobalt surface-coated and has a 4.5 wt.% zinc and 0.7 wt.% cobalt mixed in a solid solution within the nickel hydroxide. Polyacrylic acid potassium salt (PAAK) is used as a binder and 26 wt.% Potassium Hydroxide (KOH) solution is used as solvent and electrolyte. Extra fine nickel powder (Novamet, Inco Type 210, 0.5-1 μm) was also added to further improve electrical conductivity of the positive electrode. The resulting positive electrode ink has 15 vol.% of fine nickel and 39 vol.% of nickel hydroxide. Electrical conductivity of the resulting ink is 0.065 S/cm with a specific theoretical capacity of 100 mah/g. Positive Current Collector Good electrical conductivity is the primary requirement for the positive current collector. It should also be stable within the cycling voltage range of the battery (~1.2 V to 1.5 V). Preliminary cyclic voltammetry tests with Ercon gold (Ercon R-464 (DPM- 78)) and silver (Ercon E1400-136) inks showed both inks were unstable within the voltage range of the battery. Custom made nickel current collector ink showed positive results and was further tested in this work. 3
Cyclic voltammetry was used to further evaluate the stability of the custom made nickel ink. The ink was first printed onto a Teflon porous substrate with a 1-cm 2 active area as shown in Figure 2. 1 cm 2 Figure 2: Current Collector Cyclic Voltammetry Sample A commercial platinum electrode (Etek A-2 Silver-Plated Nickel Screen Electrode) was used as the counter and reference electrode (Cyclic voltammetry tests performed using two of these commercial electrodes showed very little overpotential, allowing this commercial electrode to be used as a quasi-reversible hydrogen reference electrode). An electrochemical cell was assembled by sandwiching the working and the counter/reference with a Celgard separator (3501) soaked with 26 wt. % KOH. The whole cell was clamped together to ensure close contact between the electrodes. The top and the bottom plates of the clamping assembly were perforated to allow gas flow around the cell during test. The cell assembly was then placed in a hydrogen purged chamber to maintain constant hydrogen pressure. Cyclic voltammetry was performed at about 1.4 psig. hydrogen pressure from open circuit voltage to 1.5 V vs. reference and back to open circuit voltage at 2 mv/s scan rate. Figure 3 shows the cyclic voltammetry results for the custom made Nickel ink. Nickel oxidation and reduction was observed at 1.40 V and 1.23 V, respectively. An unknown anodic peak was also observed at 1.33 V whose height decreased upon cycling. This may due to some impurities in the test sample. Oxygen evolution was not observed during positive voltage scans. Moreover, hydrogen oxidation on the nickel surface was not observed, suggesting that the nickel surface is always oxidized. Since the nickel redox couple lies in the range of cycling voltages for the positive electrode and shows stability during cyclic voltammetry, the nickel ink was selected to be the positive current collector ink. This ink consists of 50 vol. % filamentary nickel powder (Novamet, Inco Type 225, 2.2-2.8 μm) in PVDF slurry with DMF as solvent. The electrical conductivity was measured to be 111 S/cm. I (Amps/cm 2 ) 0.0012 0.0010 0.0008 0.0006 0.0004 0.0002 0-0.0002-0.0004-0.0006-0.0008-0.0010 0 0.5 1.0 1.5 E (Volts) Figure 3: Cyclic Voltammetry Results for Nickel Ink 4
Electrolyte Filled Separator Very fine alumina powder was used as the basis for the separator. Polyacrylic acid potassium salt (PAAK) was used as the binder and 26 wt.% Potassium Hydroxide (KOH) solution was used as the solvent. As KOH was used as the solvent, the resulting ink contains KOH within the alumina matrix, which also acts as the battery electrolyte. A.C. impedance spectroscopy showed an ionic conductivity of 0.00149 S/cm at 35% relative humidity. Considering a battery capacity of 0.3 mah/cm 2 with an electrolyte filled separator layer about 100 μm thick, if the battery is cycled at C/10 rate (i.e. 0.03 ma/cm 2 ), the IR drop across this layer was only 0.2 mv, which is negligible. Handprinting of the negative electrode ink on top of the cured separator showed that this ink was rigid enough to withstand the printing process. Negative Electrode Fast electrochemical kinetics for hydrogen oxidation and reduction are the main requirement for the negative electrode. The electrode should also be porous enough to allow efficient mass transfer of hydrogen into the electrode for hydrogen oxidation. Therefore, the electrode thickness and platinum loading of the negative electrode are critical for the negative electrode performance. Preliminary tests using a 40 wt. % platinum on Vulcan XC-72 (Etek, C1-40) showed relatively little reactivity towards hydrogen oxidation. A 100 % platinum black (Alfa Aesar, #12755) ink in PVDF/DMF slurry was therefore used in this work. This ink was printed to a platinum loading of 6.3 mg/cm 2. Negative Current Collector Good electrical conductivity and compatibility with the negative electrode ink are the main requirements for the negative current collector. The current collector should also be porous to allow efficient transport of hydrogen into the negative electrode for hydrogen oxidation. Previous studies using Ercon gold ink (R-464 (DPM-78)) as the current collector in microfabricated fuel cells (4) showed good porosity. However, this ink is not compatible with the negative electrode ink as the solvent in the negative electrode ink (DMF) readily dissolves the binder in the Ercon gold ink. Thus, Ercon gold ink was not further investigated as negative current collector in this work. To avoid incompatibility with the negative electrode ink, sputtered gold was considered for the negative current collector. To create porosity in the negative current collector, a grid pattern mask was used to allow hydrogen transport into the negative electrode. A 5000-Å layer of metallic gold was sputtered onto porous Teflon substrates as the negative current collector. Gold sputtering works was done in the Center for Micro and Nano Processing at CWRU using a Denton Vacuum Discovery 18 Sputtering System. Metal Hydride As metal hydride serves as on-board hydrogen storage in this battery, hydrogen absorption and desorption of the metal hydride is critical. LaAl 0.3 Ni 4.7 metal hydride material with an additional surface modification (6) is considered as hydrogen storage in 5
this battery. This material has been successfully used in previous work on microfabricated fuel cells (4). Details of this material preparation can be found elsewhere (7). Positive Current Collector Nickel-Hydrogen Battery Fabrication Four layers of the metallic nickel ink were printed onto each substrate for a total thickness of 18 μm, as measured with a non-contact laser profilometer from Optical Gauging Products (OGP model Cobra ). Figure 4 shows the SEM image of the printed positive current collector surface at 100X. Considerable roughness and porosity was observed on the current collector surface. Both characteristics were important for the positive assembly to function properly. The former allows the positive electrode to adhere well to the current collector while the latter allows efficient oxygen diffusion from the positive electrode during overcharge. The porosity was estimated to be 24 %. In addition, the nickel powder appeared to be well mixed with the PVDF binder, providing a continuous path for the electrons to pass through. Figure 4: SEM Image of the Positive Current Collector's Surface (100X) Positive Electrode The printed positive electrode with the current collector was around 200 μm thick. Figures 5 and 6 show the SEM images of the printed positive electrode s surface (350X) and cross-section (250X), respectively. Nickel hydroxide particles (spherical particles) were shown to be well distributed among the extra fine nickel metal powder (filamentary particles), and considerable porosity was observed throughout the surface and the crosssection. A razor blade was used to cut across the positive assembly with caution to prevent disrupting the cross-section structure. 6
Figure 5: SEM Image of the Positive Electrode's Surface (350X) Positive Electrode Nickel Current Collector Teflon porous substrate Figure 6: Cross-section of the Positive Electrode (250X) As the extra fine nickel was added to enhance electrical conductivity of the positive electrode, good dispersion of the extra fine nickel was critical. Good porosity throughout the whole positive assembly allows efficient oxygen diffusion from the positive electrode during overcharge. The SEM images confirm that the printed positive electrode exhibits the critical characteristics for it to function properly. Negative Electrode The average negative electrode thickness was about 77 μm, which is quite thick compared to those used in the microfabricated fuel cells (4). This was mainly due to a thicker stencil used in this work. The thicker electrode provided improved electrical conductivity necessary to provide good connectivity with the grid pattern of the negative current collector. A.C. impedance measurements showed that the platinum black ink had an electrical conductivity of 31.8 S/cm. Figure 7 shows the SEM image of the negative electrode s surface at 350X. Similar to the positive electrode, considerable porosity is observed and the electrode material appears well mixed and homogeneous. 7
Figure 7: SEM Image of the Negative Electrode s Surface (350X) Positive Electrode Cycling Half Cell Electrochemical Testing Positive electrode cycling tests were performed with the positive electrode being tested as the working electrode and the Etek platinum electrode (Etek A-2 Silver-Plated Nickel Screen Electrode) as the counter/reference electrode. An electrochemical cell was assembled by sandwiching the working and the counter/reference with a Celgard separator (3501) soaked with 26 wt. % KOH. The whole cell was clamped together as in the cyclic voltammetry experiments for the positive current collector ink studies. The whole cell was then placed inside a hydrogen purge chamber to maintain constant hydrogen pressure at about 1.4 psig. A four-wire arrangement was used for the cycling test to eliminate IR drop across the positive current collector s contacts. Electrode formation was performed at a C/24 rate until a stable discharge capacity was obtained. Electrode cycling was then performed at the same formation current densities, resulting in cycling rates ranging from C/9-C/11. A 1.48 V cutoff voltage was applied during charging to avoid excessive oxygen evolution. The discharge cut-off voltage was 1.22 V. A.C. impedance spectroscopy was also used to monitor the cell solution impedance after both charging and discharging for each cycle. Two essentially similar cells were tested (cell details shown in Table I). The theoretical capacity of the positive electrode was determined by multiplying the specific theoretical capacity of the positive paste (100 mah/g) by the amount of positive paste printed onto the positive current collector. Electrode thickness was measured using the non-contact laser profilometer. Formation charging and discharging curves for cell 087 are shown in Figure 8. Table I: Cell Details for Positive Electrode Cycling Cell ID Positive Electrode Theoretical Capacity (mah) Electrode Thickness (μm) Electrode Area (cm 2 ) 084 1.7 228 1.00 087 1.7 212 1.00 8
1.48 1.46 1.44 1.42 1.40 1.38 E (Volts) 1.36 1.34 1.32 1.30 1.28 1.26 1.24 1.22-1.0-0.5 0 0.5 1.0 1.5 Q (mahs/cm 2 ) Figure 8: Formation Charging and Discharging for Cell 087 A steady discharge capacity was observed after 5 th cycle at 0.55 mah. Further cycling was performed at C/11 as shown in Figure 9. Consistent charging and discharging behaviors were observed with discharge capacity increasing slightly upon each cycle. The test was terminated after the 20 th cycle due to time constraints; however, the results in Figure 13 show that the cell was still functioning properly and cycling could have been continued. The discharge capacity as of the 20 th cycle was 0.34 mah/cm 2. Similar behaviors were also observed for cell 084 for which the discharge capacity after the 20 th cycle was 0.29 mah/cm 2. 1.48 1.46 1.44 1.42 1.40 1.38 E (Volts) 1.36 1.34 1.32 1.30 1.28 1.26 1.24 1.22-0.5-0.4-0.3-0.2-0.1 0 0.1 0.2 0.3 0.4 0.5 0.6 0.7 Q (mahs/cm 2 ) Figure 9: Charging and Discharging Curves from 5 th to 20 th Cycles for Cell 087 Figure 10 shows the cell solution impedance upon cycling for cell 087. Solution impedance slightly increased upon cycling with slightly higher solution impedance observed after discharge for both cells. Although solution impedance was increasing upon cycling, sign of stabilization were observed after electrode formation cycles. 9
Soln resistance (ohm) 1.0 0.9 0.8 0.7 0.6 0.5 0.4 0.3 0.2 0.1 0.0 soln R after charge soln R after discharge Formation Cycles Stabilization 0 5 10 15 20 Cycle Number Figure 10: Cell Solution Impedance upon Cycling for Cell 087 Negative Cyclic Voltammetry Negative electrode cyclic voltammetry tests were performed with the negative electrode being tested as the working electrode and the Etek platinum electrode (Etek A-2 Silver-Plated Nickel Screen Electrode) as the counter/reference electrode. The same electrochemical setup used in positive electrode cycling was used. Hydrogen pressure was maintained at about 1.4 psig. During testing, voltage was scanned from -0.3 V to 0.3 V at 2 mv/s. Electrode activation was performed at 50 mv/s and 10 mv/s before actual testing. A.C. impedance spectroscopy was used to measure cell solution impedance. Figure 11 shows the cyclic voltammetry results with the platinum black ink. H 2 oxidation current density reaching 100 ma/cm 2 was observed at about 0.3 V, without any sign of mass transport limitation. The noise observed during the negative voltage regime might due to an unstable connection. Based on the results shown in Figure 11 and considering that the proposed battery will be operating at current densities from 10 μa/cm 2 to 1 ma/cm 2 ; the negative electrode overpotential will be well under 1 mv, which is negligible. Thus, the platinum black ink was shown to be a good negative electrode ink for this battery. 0.10 0.05 I (Amps/cm 2 ) 0-0.05-0.10-0.15-0.20-0.3-0.2-0.1 0 0.1 0.2 0.3 E (Volts) Figure 11: CV Results with Platinum Black Ink 10
Full Cell Electrochemical Testing A full cell was assembled by sandwiching the screen printed positive electrode on nickel current collector (working electrode) and the screen printed negative electrode with sputtered gold current collector (counter/reference electrode), with a Celgard separator (3501) soaked with 26 wt. % KOH. The whole cell was again clamped together and placed in a hydrogen purge chamber, similar to that used in the positive electrode cycling tests. A constant hydrogen pressure was maintained at about 1.4 psig., and a four-wire arrangement was again used to eliminate IR drop across the current collectors contacts. Formation and cycling procedures similar to those used in the positive cycling test was used in the full cell testing, with cell details shown in Table II. Table II: Cell Details for Full Cell Testing Positive Electrode Theoretical Capacity (mah) Positive Electrode Thickness (μm) Negative Electrode Thickness (μm) Electrode Area (cm 2 ) 1.7 257 77 1.00 Similar formation curves to those observed in half cell tests were obtained in full cell tests. Cell discharge capacity after the 5th cycle was 0.58 mah. Further cycling was performed at C/12. Figure 12 shows the cycling data for full cell testing. E (Volts) 1.48 1.46 1.44 1.42 1.40 1.38 1.36 1.34 1cm2nicharge093kohadditionc6_june0905a 1cm2nicharge093kohadditionc7_june0905a 1.32 1cm2nicharge093kohadditionc8_june0905a 1cm2nicharge093kohadditionc9_june0905a 1.30 1cm2nicharge093kohadditionc10_june0905 1cm2nicharge093kohadditiondc6_june0905 1.28 1cm2nicharge093kohadditionc11_june0905 1cm2nicharge093kohadditiondc7_june0905 1.26 1cm2nicharge093kohadditiondc8_june0905 1cm2nicharge093kohadditiondc9_june0905 1.24 1cm2nicharge093kohadditiondc10_june090 1cm2nicharge093kohadditiondc11_june090 1.22-0.5-0.4-0.3-0.2-0.1 0 0.1 0.2 0.3 0.4 0.5 0.6 Q (mahs/cm 2 ) Figure 12: Full Cell Cycling Data from 6 th to 12 th cycles Premature cutoff at 1.48 V Slightly increasing discharge capacity was observed from the 6th to 9th cycles; however, diminishing discharge capacity was observed after the 9th cycle due to premature charge cutoff at 1.48 V. In contrast to the half cell results shown in Figure 9, two sudden shifts in charging curves were also observed (double headed arrows) in Figure 12. Figure 13 shows the total cell impedance at 20 khz. Total cell impedance at 20 khz rather than solution impedance was used due to difficulties in extrapolating of the impedance data to the real axis. However, as the total cell impedance was obtained at high frequency (20 khz), mostly ohmic impedance would be measured. Thus, it was a reasonable representation of the cell solution impedance although not exactly the same. 11
Total Cell Resistance at 20kHz(ohm) 6.0 5.5 5.0 4.5 4.0 3.5 3.0 2.5 2.0 Rtot at 20 khz after charge Rtot at 20 khz after discharge 1.5 1.0 0 1 2 3 4 5 6 7 8 9 10 11 12 Cycle Number Figure 13: Total Cell Impedance at 20 khz upon Cycling Total cell impedance at 20 khz was observed to be increasing rapidly upon cycling, reaching about 4.5 ohm-cm 2 after charge and 5.5 ohm-cm 2 after discharge at the end of 11th cycle. No sign of stabilization was observed. Based on this impedance data, the ohmic overpotential across the cell at most should be about 0.25 mv (0.0495 ma*5 ohm = 0.25 mv). However, an overpotential increase of about 20 mv was observed in Figure 12, suggesting that the cell was impeded by more than ohmic impedances. Figure 14 shows the total cell impedance at 0.02 Hz. Similar to the total cell impedance at 20 khz, the total cell impedance at 0.02 Hz was observed to be increasing rapidly, reaching about 55 ohm-cm 2 at the end of 11th cycle, which doubled that at the end of 4th cycle. Since the impedance data at 0.02 Hz includes all impeding processes within the cell (ohmic, charge transfer, capacitive, mass transport), it shows that the cell became more and more resistive in general upon subsequent charging. The steepest increasing slope was observed from the 7th to 10th cycles where sudden shifts in charging curves were observed in Figure 12. Clearly the sudden shifts in charging curves are correlated with the rapidly increasing total cell impedance. Total Cell Resistance at 0.02 Hz (ohm) 60.0 50.0 40.0 30.0 20.0 10.0 Formation Cycles Rtot at 0.02 Hz after charge 0.0 0 1 2 3 4 5 6 7 8 9 10 11 12 Cycle Number Figure 14: Total Cell Impedance at 0.02 Hz upon Cycling However, one should note that the cycle % utilization (Discharge Capacity/Charge Capacity) was improving upon cycling as shown in Figure 15. Utilization improved from 22 % after the 1st cycle to 82.6 % at the end of 11th cycle. The data suggested that although the deliverable capacity was diminishing due to premature cutoff as shown in Figure 12, more of the charge capacity of the cell was able to be delivered upon cycling, especially after the formation cycles. This might be explained in terms of the better conducting paths achieved after formation between the nickel hydroxide particles, which 12
allowed for more efficient charging of the nickel hydroxide particles without going into oxygen evolution. 100 % Cycle Utilization 80 60 40 20 Formation Cycles Cycle % Utilization 0 0 1 2 3 4 5 6 7 8 9 10 11 Cycle Number Figure 15: % Cycle Utilization upon Cycling Conclusions Thick film printing inks/pastes were developed for the different battery components of a microfabricated nickel-hydrogen battery, including the positive electrode, the positive current collector, the electrolyte filled separator, and the negative electrode. Sputtered gold was used for the negative current collector to avoid solvent incompatibility with the negative electrode ink. SEM analysis performed on the printed positive electrodes, the printed positive/negative current collectors, and the printed negative electrode showed the desired characteristics for each component to perform properly. Half cell electrochemical tests were performed separately for the printed positive electrodes and the printed negative electrodes. Consistent charge and discharge performance was observed during positive electrode cycling. The platinum black ink was shown to be a good negative electrode ink with reasonable reactivity towards hydrogen oxidation and reduction. Full cells were assembled using the printed positive assembly and the printed negative assembly with electrolyte-soaked separators, and tested for cycling. Sudden upward shifts in charging curves and rapid increase in total cell impedance were observed during cycling. However, improvement in the cycle % utilization was also observed, showing promise for acceptable cycling behavior if better control of the total cell impedance can be achieved. Acknowledgements The authors would like to thank Dr. C.C. Liu and Ms. Laurie Dudik for their helps in the thick film printing works. The authors would also like to acknowledge NIH grants (NINDS R01-NS-41809) for financially supporting this work. 13
References 1. D. Linden, and T. B Reddy, Handbook of Batteries, McGraw-Hill, New York, (2001). 2. P.B. Koeneman, I.J. Busch-Vishniac, and K.L. Wood, J. Microelectromech. Sys.6, 355 (1997). 3. G.A. Ghiurcan, C.C. Liu, A. Webber, and F.H. Feddrix, J. Electrochem. Soc. 150, A922 (2003). 4. J.S. Wainright, R.F. Savinell, C.C. Liu, and M. Litt, Electrochim. Acta 48, 2869 (2003). 5. Zimmerman, A. H., in Symposium on Hydrogen and Metal Hydride Batteries, P.D. Bennett and T. Sakai, Editors, PV 94-27, p. 268, The Electrochemical Society Proceedings Series, Miami Beach, FL (1994). 6. Shan, Xi, J.S. Wainright, and J. Payer, manuscript in preparation. 7. Shan, Xi, PhD. Thesis, Case Western Reserve University, Cleveland, 2004. 14