A Nanoporous Silicon Membrane Electrode Assembly for ... - Chang Lu

JOURNAL OF MICROELECTROMECHANICAL SYSTEMS, VOL. 15, NO. 3, JUNE 2006

671

A Nanoporous Silicon Membrane Electrode Assembly for On-Chip Micro Fuel Cell Applications Kuan-Lun Chu, Scott Gold, Vaidyanathan (Ravi) Subramanian, Chang Lu, Mark A. Shannon, and Richard I. Masel

Abstract—Silicon-based fuel cells are under active development for chip-scale electrical power supply. One of the greatest challenges in micro-fuel-cell research is the development of a suitable proton conducting membrane material that is compatible with standard silicon microfabrication technology. In this paper, the use of nanoporous silicon as a novel proton conducting membrane material in a microscale fuel cell membrane electrode assembly (MEA) is demonstrated. The devices were fabricated by first creating 100- m-thick silicon windows in a standard silicon wafer, anodizing to create pores in the windows, and then painting catalyst layers and insulators onto the porous structures. Using 5 M formic acid and 0.5 M sulfuric acid as the fuel, the fuel cell peak power density reached about 30 mW/cm2 at current density level of about 120 mA/cm2 . These results represent the successful integration of a new class of protonic conductor into a microfabricated silicon fuel cell. [1455] Index Terms—Micro fuel cells, porous silicon. Fig. 1. An illustration of the fabrication scheme used to create the porous silicon membranes.

I. INTRODUCTION

T

HERE has recently been considerable interest in developing silicon-based micro fuel cells for chip scale power [1]–[5]. Fuel cells have several advantages over traditional batteries, including rapid recharging and much higher energy densities. Among the many types of fuel cell systems, proton exchange membrane fuel cells (PEMFC) have been most widely examined as major future power sources for microscale devices. Significant efforts have focused on the development of CMOS compatible processes and materials to produce silicon-based PEMFCs that can be integrated with microelectronic and microelectromechanical systems (MEMS) devices. Development of a suitable proton conducting membrane material has proved to be one of the greatest challenges in the manufacture of micro fuel cells. Nafion or a similar polymer is most widely used as a proton conductor in silicon micro fuel cells and in PEMFCs in general. However, Nafion is not

Manuscript received November 4, 2004; revised September 10, 2005. This work was supported by the Defense Advanced Research Projects Agency under U.S. Air Force Grant F33615-01-C-2172. Subject Editor J. Judy. K.-L. Chu, V. Subramanian, and R. I. Masel are with the Department of Chemical and Biomolecular Engineering, University of Illinois at Urbana-Champaign, Urbana, IL 61801 USA (e-mail: [email protected]). S. Gold was with the Department of Chemical and Biomolecular Engineering, University of Illinois, Urbana, IL 61801 USA. He is now with the Chemical Engineering Program and Institute for Micromanufacturing, Louisiana Tech University, Ruston, LA 71270 USA. C. Lu was with the Department of Chemical and Biomolecular Engineering, University of Illinois, Urbana, IL 61801 USA. He is now with the Agricultural and Biological Engineering, Purdue University, West Lafayette, IN 47907-2093 USA. M. A. Shannon is with the Department of Mechanical and Industrial Engineering, University of Illinois, Urbana, IL 61801 USA. Digital Object Identifier 10.1109/JMEMS.2006.872223

Fig. 2. Nanoporous silicon formation setup for fabrication step in Fig. 1(f).

readily integrated with standard microfabrication techniques used in making micro fuel cells, other microchemical systems, or MEMS-based devices, as it cannot be easily patterned using photolithography and bonding it to silicon and other commonly used materials is extremely challenging under working fuel cell conditions due to its volumetric changes with changes in hydration level [6]. Countless efforts have been made to develop a next-generation protonic conducting membrane material to replace Nafion. Solid-state protonic conductors include materials such as solid acids, polymers, oxide ceramics, and intercalation compounds [7], [8]. One common approach has

1057-7157/$20.00 © 2006 IEEE

672


(a)

(b)

Fig. 3. SEM image of nanoporous silicon membrane (cross section view) at (a) low magnification and (b) high magnification.

Fig. 4.

SEM image of porous silicon (top face view).

been to use inorganic–organic hybrid materials, often including Nafion as one of the components. Examples of such composite proton conducting membrane materials reported in the literature include Nafion-silica [9]–[11], Nafion-borosiloxane [12], and silica-polyethylene oxide [13] composites, among others. More recently, nanoporous silicon has been shown to have strong potential as a proton conducting fuel cell membrane material having proton conductivity and fuel crossover flux comparable to Nafion [14], [15]. Nanoporous silicon, which is readily formed by anodic etching in hydrofluoric acid, is compatible with conventional Si microfabrication technologies, presenting the possibility of fabricating different components of a fuel cell or even a fuel cell and its supported devices in a monolithic fashion. Nanoporous silicon should also be stable at elevated temperatures, unlike many polymeric materials. The objective of this paper is to demonstrate the potential of nanoporous silicon as a protonic conductor in a silicon micro fuel cell. In the study reported in this paper, a membrane electrode assembly (MEA), a crucial component of almost all fuel cell systems, was fabricated using nanoporous silicon. The performance of this MEA was evaluated by measuring its current–voltage characteristics and power output.

Fig. 5. An illustration of the fuel cell as tested.

Fig. 6. SEM image of the nanoporous device.

TiO

layer on the anode side of the

II. EXPERIMENTAL METHODS A. Nanoporous Silicon Membrane Fabrication The procedure used to fabricate the nanoporous silicon membranes is illustrated in Fig. 1. Generally, the procedure was to use deep reactive ion etching (DRIE) to form 100 M membranes in a standard silicon wafer. The membranes were then anodized, to convert them to porous silicon. Then catalyst layers,

CHU et al.: NANOPOROUS SILICON MEMBRANE ELECTRODE ASSEMBLY

673

TiO

Fig. 7. Polarization and power density curves for two nanoporous silicon MEAs that were insulated with a a solution containing: : 5 M formic acid plus 0.5 M sulfuric acid; euro: 5 M formic acid with no sulfuric acid.

insulators, and current collectors were painted onto the wafers to create fuel cells. In detail, the fabrication started with prime grade p-type (boron-doped) double side polished 100 silicon wafers with 100 mm diameter and 0.01–0.02 -cm resistivity from Wafer World Inc. Native oxide on the wafer surface was removed by immersion in a buffered oxide etch (BOE) solution prior to any other processing [Fig. 1(a)]. First, an approximately 800-nm-thick silicon nitride film was deposited on both wafer sides by low-pressure chemical vapor deposition [Fig. 1(b)]. Photolithography was used to pattern four small circles on the front side of each die with a total area (all four cells) of 0.0625 cm . Each die has a square shape with 2.25 cm area. Reactive ion etching with freon plasma was then used to remove the silicon nitride from wafer back side and to expose the four circles on the front side, as shown in Fig. 1(c). Silicon membranes with thickness of 100 m were formed by DRIE [Fig. 1(d)]. For each die, the indented side etched down by DRIE is referred as the front side, and the opposite flat side as the backside, in the rest of this paper. A 50 nm chrome and gold layer was then sputtered on the backside of the wafer for electrical contact following removal of any native oxide with a BOE solution. AZ 4903 photoresist was spin-coated on top of the chrome layer to protect it from acid etching in the next

film. The measurements were done using fuel

step [Fig. 1(e)]. A nanoporous silicon membrane was formed by immersing the die in an electrolyte solution composed of ethanol and 49 wt.% hydrofluoric acid (1:1 by volume) stirred by magnet rotor and passing a constant current of 40 mA/cm through the four circular membranes, while silicon die was the anode and a platinum wire coil acted as the cathode in the electrolyte solution, as illustrated in Fig. 2. Nanopores were formed by electrochemical etching of silicon. The die with porous silicon membrane was illustrated in Fig. 1(f). The photoresist and chrome on the die were removed by piranha (aq): 98 wt.% solution (1:3 volume ratio of 30 wt.% ) and chrome etchant, respectively [Fig. 1(g)]. Scanning electron microscopic (SEM) images of the nanoporous silicon membrane surfaces and cross sections were obtained using a Hitachi S4700 field emission gun SEM at the Center for Microanalysis of Materials (CMM), University of Illinois at Urbana-Champaign, and are shown in Figs. 3 and 4. B. Construction of the Fuel Cell The final fuel cell is illustrated in Fig. 5. Generally, the fuel layer onto the membrane, cell is constructed by painting a adding anode and cathode catalyst layers onto the nanoporous sol. silicon membrane, and then insulating with a

674


TiO

Fig. 8. Polarization and power density curves for two different nanoporous silicon MEAs. Both were insulated with a identical procedures. The measurements were done using fuel a solution containing 5 M formic acid plus 0.5 M sulfuric acid.

In detail, the backside surfaces of four circular nanoporous nanoparsilicon membrane were painted by brush with ticle colloid to form an insulating film on top of nanoporous silicon. When being painted, temperature of the silicon die was maintained between 40 and 50 C. This insulating film is to prevent short-circuiting between anode and cathode electrodes formed on the two opposite sides of nonporous silicon memcolloids were prepared brane in the following steps. The by the hydrolysis of titanium isopropoxide (0.6 M) ( 99%) obtained from Aldrich Chemical Co. in acetic acid for 12–16 h. The sol thus prepared was autoclaved at 230 C for 12 h and allowed to cool to room temperature. An alternative to the above film on the method for insulating film was depositing backside surfaces of the four circular nanoporous silicon membranes (and also other area on the die) by PlasmaLab plasma-enhanced chemical vapor deposition system. The thickness of this film is around 6 nm. The catalysts for both anode and cathode were prepared by ink-painting method, which is commonly used for macroscale fuel cell systems. The catalyst ink for anode was prepared by mixing 10 mg of palladium black

film. Both were fabricated with

powder (Sigma-Aldrich), 40 mg of 5% Nafion solution (Solu, and then sontion Technology Inc.), and 100 mg Millipore ication for 1 min. This catalyst ink for the anode was painted by or thin brush on top of the insulating film, which is film on the back side of the porous silicon membrane. The ink for cathode was prepared by mixing 10 mg of platinum black powder (Alfa-Aesar), 40 mg of 5% Nafion solution, and 100 mg , and sonication for 1 min. The catalyst ink for Millipore the cathode was painted by brush on the front side surfaces of the porous silicon membranes. To make the anode catalyst layer surfaces hydrophilic so that fuel can be attracted to the catalyst thin film was to facilitate anode reaction, an additional painted by brush on top of the palladium film. The complete nanoporous silicon membrane electrode assembly is illustrated in Fig. 5. C. Micro Fuel Cell Performance Testing Nanoporous silicon membrane electrode assembly was tested by dropping about 50 of fuel solution on the anode side of


675

SiO

Fig. 9. Polarization and power density curves for two different nanoporous silicon MEAs. Both were insulated with a identical procedures. The measurements were done using fuel a solution containing 5 M formic acid plus 0.5 M sulfuric acid.

silicon die, which was held horizontally. The fuel solution was either 5 M formic acid or 5 M formic acid with 0.5 M sulfuric acid. The formic acid/sulfuric acid mixture is quickly wicked and catalyst layers but does not flow through into the the porous silicon. The testing setup is illustrated with the nanoporous silicon MEA structure in Fig. 5. In all tests, the cathode was air-breathing. III. RESULTS AND DISCUSSION SEM images of the nanoporous silicon membranes are shown in Figs. 3 and 4. Generally, the films contain branched, interconnected pores with an average diameter of less than 10 nm. thin film (Fig. 6) and the catalyst The SEM images of the layers show that they are also porous. The polarization curves (voltage–current density curves) for nanoporous silicon MEA with insulating film were taken using two kinds of fuel solutions, as illustrated in Fig. 7(a). One of the fuel solutions is 5 M formic acid and the other is 5 M formic acid plus 0.5 M sulfuric acid, in which formic acid is

film. Both were fabricated with

the fuel oxidized at the anode. Open cell potentials (the voltage output of MEA in the zero current density limit) shown by both curves in Fig. 7 were much lower than the theoretical electromotive force of formic acid fuel cell, which is 1.45 V, and also lower than Nafion membrane-based MEA, which is usually 1.0 V. This is attributed to higher fuel crossover through the membrane from the anode side to the cathode side in nanoporous silicon membrane than in Nafion membrane [14]. This is also the evidence that fuel solution penetrated into the MEA from the top film, the anode catalyst layer, insulating film, and into the pores of the nanoporous silicon membranes. Fuel solution inside the pores contributed to proton conductivity of the membrane. It can be seen from both Fig. 7(a) and (b) that nanoporous silicon MEA showed better performances when 0.5 M sulfuric acid was added in the fuel solution than when just using 5 M formic acid alone as fuel. This improvement in performances is attributed to increase in proton conductivity (contributed by sulfuric acid) of fuel solution inside the pores of nanoporous silicon membranes. Therefore in the following figures, all data were obtained by using fuel solution of 5 M formic acid plus 0.5 M sulfuric acid.

676


In order to explore the reproducability of the findings, Figs. 8 and 9 show polarization curves and power density curves for nanoporous silicon MEAs made in two different batches. It can be seen that there were differences in their performances though they were prepared with the same procedure. The MEA with higher open cell potential also has lower short-circuit current density (the current density output when anode and cathode are short-circuited). This is as expected if there were slight differences in the proton conductivity in the porous silicon the two different batches of porous silicon membranes. Physically, as the conductivity of the membrane decreases, the parasitic currents due to fuel transport through the membrane are reduced. That raises the open cell potential. Unfortunately, the proton conductivity is reduced. That reduces the short circuit current. At this point we are not completely clear why the batch-tobatch variations in the conductivity of the membranes exists. In our pervious paper [15], we noted that the anodization of the porous silicon stops when the chrome on the backside of the membrane dissolves. That can result in incomplete anodization of the membrane where some of the pores do not extend completely through the membrane.

[9] K. A. Mauritz, “Organic-inorganic hybrid materials: Perfluorinated ionomers as sol-gel polymerization templates for inorganic alkoxides,” Mater. Sci. Eng. C, vol. 6, pp. 121–133, 1998. [10] N. Miyake, J. S. Wainright, and R. F. Savinell, “Evaluation of a sol-gel derives Nafion/silica hybrid membrane for proton electrolyte membrane fuel cell applications I: Proton conductivity and water content,” J. Electrochem. Soc., vol. 148, pp. A898–A904, 2001. [11] , “Evaluation of a sol-gel derived nafion/silica hybrid membrane for polymer electrolyte membrane fuel cell applications II: Methanol uptake and methanol permeability,” J. Electrochem. Soc., vol. 148, pp. A905–A909, 2001. [12] H. Suzuki, Y. Yoshida, M. A. Mehta, M. Watanabe, and T. Fujinami, “Proton conducting borosiloxane solid electrolytes and their composites with nafion,” Fuel Cells, vol. 2, pp. 46–51, 2002. [13] I. Honma, S. Hirakawa, K. Yamada, and J. M. Bae, “Synthesis of organic/inorganic nanocomposites proton conducting membrane through sol-gel processes,” Solid State Ion., vol. 118, pp. 29–36, 1999. [14] S. Gold, K.-L. Chu, C. Lu, M. A. Shannon, and R. I. Masel, “Acid loaded porous silicon as a proton exchange membrane for micro-fuel cells,” J. Power Sources, vol. 135, pp. 198–203, 2004. [15] S. A. Gold, K. L. Chu, M. A. Shannon, and R. I. Masel, “Nanoporous silicon as a proton exchange membrane for micro-fuel cells,” presented at the 206th Meeting Electrochemical Society, Honolulu, HI, 2004. [16] M. P. Stewart and J. M. Buriak, “Chemical and biological applications of porous silicon Technology,” Adv. Mater., vol. 12, pp. 859–869, 2000.

IV. CONCLUSION In this paper, nanoporous silicon is demonstrated to be a novel proton conducting membrane material for membrane electrode assemblies in a microscale silicon fuel cells. With 5 M formic acid as the fuel and 0.5 M sulfuric acid added to the fuel to increase proton conductivity, the fuel cell peak power density reached about 30 mW/cm at current density of about 120 mA/cm .

Kuan-Lun Chu received the B.S. degree in physics from National Taiwan University, Taipei, Taiwan, in 1994, and the M. Eng. degree in electrical and computer engineering from Cornell University, Ithaca, NY, in 2001. He is currently pursuing the Ph.D. degree in electrical and computer engineering at the University of Illinois, Urbana-Champaign. His research area is microelectromechanical systems for micropower generation

ACKNOWLEDGMENT The CMM Department of Energy National User Center for Electron Beam Microcharacterization, University of Illinois at Urbana-Champaign, provided access to the SEM used in this paper.

REFERENCES [1] K. Shah, W. C. Shin, and R. S. Besser, “A PDMS micro proton exchange membrane fuel cell by conventional and nonconventional microfabrication techniques,” Sens. Actuators B, vol. 97, pp. 157–167, 2004. [2] G. Q. Lu, C. Y. Wang, T. J. Yen, and X. Zhang, “Development and characterization of a silicon-based micro direct methanol fuel cell,” Electrochimica Acta, vol. 49, pp. 821–828, 2004. [3] T. J. Yen, N. Fang, and X. Zhang, “A micro methanol fuel cell operating near room temperature,” Appl. Phys. Lett., vol. 83, pp. 4056–4058, 2003. [4] J. S. Wainright, R. F. Savinell, C. C. Liu, and M. Litt, “Microfabricated fuel cells,” Electrochimica Acta, vol. 48, pp. 2869–2877, 2003. [5] S. C. Kelley, G. A. Deluga, and W. H. Smyrl, “Miniature fuel cells fabricated on silicon substrates,” AIChE J., vol. 48, pp. 1071–1082, 2002. [6] R. F. Service, “Shrinking fuel cells promise power in your pocket,” Science, vol. 296, pp. 1222–1224, 2002. [7] T. Norby, “Solid-state protonic conductors: Principles, progress and prospects,” Solid State Ion., vol. 125, pp. 1–11, 1999. [8] G. Alberti and M. Casciola, “Solid state protonic conductors, present main applications and future prospects,” Solid State Ion., vol. 145, pp. 3–16, 2001.

Scott Gold received the B.S. degree in chemical engineering from the University of Kentucky, Lexington, the M.S. degree from the Georgia Institute of Technology, Atlanta, and the Ph.D. degree from Arizona State University, Tempe, in 2002. He followed with a postdoctoral research position at the Univesity of Illinois, Urbana-Champaign. He joined the faculty at Louisiana Tech University, Ruston, in 2004. He works in micromechincal systems and micropower generation.

Vaidyanathan (Ravi) Subramanian received the Ph.D. degree in chemical engineering from the University of Notre Dame, South Bend, IN, in 2004. He is currently a visiting research Assistant Professor with Prof. Richard Masel at the University of Illinois, Urbana-Champaign, in the Chemical and Biomolecular Engineering Department. His research area is synthesis and applications of semiconductor-metal nanocomposites to alternate energy systems, sensors, and environmental systems.


Chang Lu received the B.S. degree in chemistry from Peking University, Beijing, China, in 1998, and the M.S. and Ph.D. degrees in chemical engineering from the University of Illinois, Urbana-Champaign, in 2001 and 2002, respectively. He followed with a post doctorate at the Nanobiotechnology Center at Cornell University, Ithaca, NY. He is now an Assistant Professor at Purdue University, West Lafayette, IN. His research group focuses on the use ofmicro/nano scale devices and materials in the study of biological systems and for harnessing biological energy. More specifically, part of the effort is directed toward development of micro/nano devices for biological sample treatment and processing.

Mark A. Shannon received the B.S. and Ph.D. degrees in mechanical engineering from the University of California, Berkeley, in 1989 and 1993, respectively. He joined the faculty of the University of Illinois, Urbana-Champaign, in 1994. He is Willett Professor of Mechanical and Industrial Engineering, Electrical and Computer Engineering, Bioengineering, and Director of the Center of Advanced Materials for the Purification of Water with Systems. His research is in microfabrication, micro electrical mechanical systems, micro/nanofluidics, and BioMEMS. Recently, he and Richard Masel founded Cbana Laboratories, a developer of microfluidic products.

677

Richard I. Masel received the M.S. degree in chemical engineering in 1973 from Drexel University, Philadelphia, PA, in 1973. He joined the University of Illinois, Urbana-Champaign, in 1978. He is now Professor of Chemical and Biomolecular Engineering and Electrical and Computer Engineering. Masel has worked on kinetics and catalysis, and more recently fuel cells, microchemical systems, and MEMS. He helped found Tekion, a developer of fuel cells for portable electronics, in 2003. He served for three years as Chief Technical Officer of Tekion before stepping down to the position of Chief Technical Advisor. Recently, he and Mark Shannon founded Cbana Laboratories, a developer of microfluidic products.