MEMS fuel cell system integrated with a methanol reformer for a portable power source

MEMS fuel cell system integrated with a methanol reformer for a portable power source

Sensors and Actuators A 154 (2009) 204–211 Contents lists available at ScienceDirect Sensors and Actuators A: Physical journal homepage: www.elsevie...

2MB Sizes 0 Downloads 45 Views

Sensors and Actuators A 154 (2009) 204–211

Contents lists available at ScienceDirect

Sensors and Actuators A: Physical journal homepage: www.elsevier.com/locate/sna

MEMS fuel cell system integrated with a methanol reformer for a portable power source Taegyu Kim ∗ , Sejin Kwon Department of Aerospace Engineering, Korea Advanced Institute of Science and Technology, 335 Gwahangno, Yuseong-gu, Daejeon, 305-701, Republic of Korea

a r t i c l e

i n f o

Article history: Received 26 March 2008 Received in revised form 2 July 2008 Accepted 11 July 2008 Available online 25 July 2008 Keywords: Micro fuel cell Methanol reformer Micro power source Hydrogen MEMS

a b s t r a c t MEMS fuel cell system was designed and fabricated for a portable power source in the present study. The system consists of a methanol–steam reformer, catalytic combustor, preferential oxidation (PROX) reactor, and polymer electrolyte membrane fuel cell (PEMFC). Methanol reformer is an essential part for hydrogen supply, in which a pre-heater, vaporizing/reforming channels and catalytic combustor were integrated. All components were fabricated using MEMS fabrication technologies combined with catalyst loading processes. Performance of the MEMS fuel cell system was measured with the optimal conditions of the reformer and PROX. Power density was 195 mW/cm2 when the potential was 0.64 V. The performance was low compared to the result for pure hydrogen because the feed at the fuel cell included undesired CO, CO2 , and N2 . MEMS fuel cell system with a weight specific energy density of 225 W h/kg was accomplished as an alternative micro power source. © 2008 Elsevier B.V. All rights reserved.

1. Introduction New microsystems, such as micro aerial vehicles (MAV), microbots, nanosatellites, and soldier power systems are being introduced and developed in recent years. Contrary to ordinary electronic devices, these microsystems perform mechanical work. In addition, they must be mobile and require the extended operation. As their functions are getting complex and advanced, their energy consumption is also increasing exponentially. However, present portable devices extract power from existing batteries. No power sources powerful enough to activate the mobile microsystems are available yet. Therefore, an alternative power source is essential for the successful development of new microsystems [1]. Various concepts for micro power generation have been introduced such as micro engine, micro gas turbine, thermoelectric generator combined with combustor, and micro fuel cell. Among these, a fuel cell has drawn attention as a primary candidate for the portable power source. The fuel cell is a device that converts directly chemical energy to electric energy. Therefore, it is efficient and able to store much energy because the energy density of chemical fuel is higher than that of existing batteries.

∗ Corresponding author. E-mail address: diafi[email protected] (T. Kim). 0924-4247/$ – see front matter © 2008 Elsevier B.V. All rights reserved. doi:10.1016/j.sna.2008.07.010

Direct methanol fuel cell (DMFC) has been widely investigated as a possible candidate for micro power generation. It uses liquid methanol with much higher density than gaseous hydrogen and suits therefore applications in microsystems. In addition to the phase of fuel, DMFC is simple to make and easy to refuel. However, an inherent problem of DMFC caused by fuel crossover severely limits the output power, despite many advantages. Polymer electrolyte membrane fuel cell (PEMFC) is known to produce much higher output power than DMFC and has no fuel crossover. However, there is a major obstacle in the successful development of PEMFC. That is the storage of gaseous hydrogen [2]. Though possible to use hydrogen in either compressed gas or liquid form, it gives significant hazards due to its explosive nature. Metal hydride suffers from high weight per unit hydrogen storage and low response for a sudden increase in hydrogen demand. Storage in the form of liquid fuel such as methanol has significantly higher energy density compared to the suggested technologies. Methanol can be reformed to generate hydrogen gas when needed, as expressed in Eq. (1). CH3 OH + H2 O → 3H2 + CO2

(1)

Though a fuel cell combined with reformer is more attractive, it is complex and bulky compared to the DMFC due to the fuel reformer. Therefore, the miniaturization of a fuel reformer has been a major research activity for the successful development of PEMFC system in recent years [11]. MEMS technology is a useful tool to reduce the size of a reformer and fuel cell [3]. It allows the miniaturization of conventional reactors while keeping its throughput and yield

T. Kim, S. Kwon / Sensors and Actuators A 154 (2009) 204–211

205

Fig. 1. Schematic of the fuel cell system that consists of methanol reformer, combustor, PROX reactor, and PEM fuel cell.

because of its large surface area-to-volume ratio, which provides the increased rate of heat and mass transport and short response time [4]. Entire fuel cell power system consists of a fuel reformer and fuel cell as presented in Fig. 1. The fuel reformer module is classified into four units: fuel vaporizer/pre-heater, steam reformer, combustor/heat-exchanger, and preferential oxidation (PROX) reactor. First, methanol is fed with water and is heated by the vaporizer. The methanol is reformed by the reforming catalyst to generate hydrogen in the steam reformer. To supply heat to the steam reformer, part of hydrogen that is un-utilized in fuel cell anode can be fed to the combustor, expressed in Eq. (2), which generates sufficient amount of heat to sustain the steam reforming of methanol. H2 + 0.5O2 → H2 O

(2)

Typically, the reformate gas includes undesirable byproducts such as carbon monoxide, carbon dioxide, and methane. PEMFC can be severely poisoned by extremely small amount of carbon monoxide. Therefore, carbon monoxide should be reduced to below 10 ppm by PROX as expressed in Eq. (3). CO + 0.5O2 → CO2

(3)

This paper presents the design, fabrication and evaluation of a MEMS fuel cell system that consists of a methanol–steam reformer, catalytic combustor, PROX reactor, and PEM fuel cell [5].

2. Fabrication 2.1. Structure design Fig. 2 depicts the construction of the MEMS methanol reformer. Mixture of methanol and water enters the steam reformer at the top, and reforming products, such as hydrogen and carbon dioxide, leave the reactor. Mixture of hydrogen and air flows into the catalytic combustor at the bottom with counter flow stream against the reforming stream. The heat generated from catalytic combustion is transferred to the steam reformer. Structure of the MEMS methanol reformer was made of four wafers; two for top and bottom, one for a steam reformer layer, and the reminder for a catalytic combustor layer. Silicon wafer was selected as a substrate of the steam reformer due to high thermal conductivity that enhances the heat-exchange between the reformer and combustor. The covers and combustor chamber were made of the glass to reduce the heat loss to the surrounding. Substrate of the reformer cover is Pyrex® glass for the anodic bonding with the silicon wafer, and the photosensitive glass (FORTURN® ) was selected as a substrate of the

Fig. 2. Construction of a MEMS methanol reformer that was made of four wafers; two for top and bottom, one for steam reformer layer, and the reminder for catalytic combustor layer (PSG: photosensitive glass).

206

T. Kim, S. Kwon / Sensors and Actuators A 154 (2009) 204–211

Fig. 3. a–j Fabrication procedure of the MEMS methanol reformer. Silicon-based reformer layer was assembled with a glass combustor layer, after catalysts were loaded.

combustor with its cover because the etching with a high aspect ratio was possible with a high tolerance [6].

2.2. Methanol reformer Overall fabrication procedure of the MEMS methanol reformer is presented in Fig. 3. First, a photoresist (PR) (PR1-4000A, Futurrex, Inc.) was deposited on the silicon wafer by spin coating. By a typical lithography process, PR pattern was obtained to use as a mask of the silicon deep reactive ion etching (DRIE) process. Microchannels were engraved on the silicon wafer using a silicon DRIE process (Fig. 3(b)). The remaining PR was removed in acetone (Fig. 3(c)). Inlet and outlet were fabricated on the Pyrex® glass wafer using a sandblast (Fig. 3(d)). The Pyrex® glass wafer was joined to the silicon wafer using anodic bonding (Fig. 3(e)).

Commercially available Cu/ZnO/Al2 O3 was selected as a reforming catalyst due to its proven reactivity and selectivity [7]. One of the important challenges is to load this catalyst into microchannels. The Cu/ZnO/Al2 O3 catalyst was coated on the silicon substrate using a slurry injection method [7]. Catalysts were ground with ball-mill machine to obtain the grain size smaller than 1 ␮m. Bentonite was used as an inorganic binder to enhance the adhesion between the catalyst and silicon substrate. The ground catalyst powders and binders were dispersed into distilled water to prepare the catalyst slurry. The slurry was stirred vigorously for sufficient dispersion. Before the injection of the prepared slurry, Al2 O3 sol was washcoated on the surface of microchannels as an adhesion layer. The prepared slurry was injected in the channel structure in the reformer by two syringes (Fig. 3(f)). To avoid the clogging of catalyst particles in the microchannel, one of syringes was used for a suction of the air inside the microchannel, and the

T. Kim, S. Kwon / Sensors and Actuators A 154 (2009) 204–211

207

Fig. 4. Fabricated results of the MEMS methanol reformer. The individual wafers, microchannel, complete device, and test holder are shown.

other was used to inject the catalyst slurry into the microchannel.

glass combustor layer), microchannel, complete MEMS methanol reformer, and test holder.

2.3. Catalytic combustor

2.4. Prox reactor

Combustor chamber was fabricated by the photosensitive glass process that consists of Cr mask pattering, UV light exposure, heat treatment, and HF etching [6]. As a catalyst of the catalytic combustor, Pt was selected due to its high activity. Carbon nanotube (CNT) was used as a supporting material of Pt catalyst [8]. The CNT support has large surface area to enhance the catalytic activity, and has low process temperature. In order to prepare the Pt/CNT catalyst, multi-walled carbon nanotube (MWNT) was used (CVD MWNT95, Iljin Nanotech Co. Ltd.). First, MWNT powders were mixed with a CNT vehicle that is composed of additive inorganic materials, organic binder materials and nitro cellulose to enhance the dispersion and adhesion. Solution of H2 PtCl6 ·xH2 O (Aldrich, 99.9+%) in de-ionized water was compounded into the prepared MWNT. The MWNT was dispersed in the ethanol solution with sonification for 1 h to prevent MWNT tangling. Finally, this ethanol solution with MWNT impregnate with Pt precursor was poured into the combustor chamber, and reduced at 230 ◦ C in the hydrogen flow after dried for 12 h (Fig. 3(i)). The two fabricated parts of the reformer layer and combustor layer were united with high temperature epoxy bond (Fig. 3(j)). Cu0 in the catalysts of the fabricated reformer was in an oxidized state. In order to recover the activity of the catalysts, the fabricated device was reduced in an environment of 4% H2 in N2 that is flowing at a rate of 10 sccm at 250 ◦ C for 3 h. Fig. 4 shows photographs of the fabricated wafers (two glass covers, silicon reformer layer, and

Pt/Ru was selected as a catalyst of PROX. Microchannel was fabricated on the photosensitive glasses. Al2 O3 was coated as a washcoat layer on microchannels using sol–gel method and the catalyst was loaded by wet impregnation [9]. Al2 O3 sol was prepared by the same procedure of Yoldas process. Aluminum isopropoxide was hydrolyzed in de-ionized (DI) water with vigorous stirring for 1 h at 80 ◦ C. The sol was peptized by adding nitric acid (HNO3 ) with adjusting the pH. Polyvinyl alcohol (PVA) solution was prepared by dissolving the PAV in DI water at 75 ◦ C. The presence of PVA can reduce crack formation of the washcoat layer at the drying time. The Yoldas sol and the PVA solution was mixed with adding the ␥Al2 O3 powder with the surface area of 100 m2 /g. The mixture slurry was ball-milled for 72 h, and the glass substrate was then dipped into the ␥-Al2 O3 slurry, dried for 2 h at 120 ◦ C after blowing off the excess slurry. This procedure was repeated until the weight of the washcoat ␥-Al2 O3 is 15% of the total weight of the substrate. The washcoated microchannel was then calcined at 350 ◦ C for 4 h. Mixture of a 0.5 M aqueous solution of H2 PtCl6 and a 0.5 M aqueous solution of RuCl3 was prepared. The substrate was immersed in the mixture. Moisture was removed by drying the catalyst loaded substrate in a convection oven at 70 ◦ C for 12 h. Calcination procedure followed in a furnace at 350 ◦ C for 3 h. After completion of coating, the catalyst surface was activated by reduction in a steady flowing hydrogen environment at 350 ◦ C for 5 h. Fabricated microreactor is shown in Fig. 5.

208

T. Kim, S. Kwon / Sensors and Actuators A 154 (2009) 204–211

Fig. 5. Microreactor for PROX reaction. The substrate was photosensitive glass and the Pt/Ru catalyst was coated.

2.5. Micro PEM fuel cell

Fig. 7. Schematic of experimental apparatus to measure the performance of the fuel cell system. Reformer, PROX, and fuel cell were tested for the separate performance and the integrated test was carried out.

3. Performance evaluations

ature of each reactor is recorded by thermocouples. Product gas of the reformer was cooled and the condensable portion was removed in a cold trap. Non-condensable product gas was analyzed by a gas chromatography (HP6890, Agilent Co.). The flow rate of dry gas was measured by a flow meter. Column in the gas chromatography was Porapok-Q (80/100 mesh, 1/8 in., 6 ft) that can separate H2 , CO, CO2 , CH4 and others. The gas composition was detected by a thermal conductivity detector (TCD) with Ar as a reference gas. Product gas of catalytic combustion process of hydrogen was analyzed, after moisture was removed in a cold trap. Inlet flow rate of methanol was 0.1 ml/h and the steam to carbon ratio (S/C) was 1.1. Reformer temperature was varied between 200 ◦ C and 300 ◦ C by controlling hydrogen flow rate in the combustor. Performance of PROX was measured at the various temperature and O2 /CO ratio. Micro fuel cell was tested with pure hydrogen to compare to the result with the reformate gas.

3.1. Experimental setup

3.2. Performance of MEMS methanol reformer

Fig. 7 shows the experimental layout for the performance measurement of the fuel cell system. Syringe pump supplied a mixture of methanol and water to the reformer at a controlled rate. Temper-

CFD analysis was carried out to predict the accurate performance in methanol conversion and hydrogen production rate [10]. Fig. 8 shows the results of CFD analysis of the reforming channel in

Micro fuel cell was fabricated for an integrated test with the micro reformer. Membrane electrode assembly (MEA) was prepared by coating 0.3 mg/cm2 Pt-Ru/C as an anode catalyst and 0.3 mg/cm2 Pt/C as a cathode catalyst on a Nafion-112 membrane. Pt-Ru/C was selected as an anode catalyst because Pt/C is poisoned by CO in the reformate gas even if removed via PROX reaction. Carbon paper (TGP-H-090, 260 ␮m) was used as a gas diffusion layer (GDL). Flow channels were fabricated by etching the photosensitive glass wafer and the current collectors, Ag/Ti layer, were sputtered. Overall fabrication process and the assembled micro fuel cell are presented in Fig. 6. The fuel cell assembly was compressed between two aluminum plates for hydrogen gas tightness.

Fig. 6. Fabrication process and assembled micro fuel cell. Pt-Ru/C and Pt/C were coated on a membrane. Carbon paper and Ag/Ti electrode were assembled with the flow channel.

T. Kim, S. Kwon / Sensors and Actuators A 154 (2009) 204–211

209

Fig. 8. Results of CFD analysis of the reforming channel in the MEMS methanol reformer at 250 ◦ C with 0.1 ml/h feed rate. Pressure drop (Pa), velocity contour (m/s), temperature distribution (K), and mole fraction of hydrogen are shown.

the MEMS methanol reformer. Pressure drop through the channel was negligible so that the pressure in the reformer is regarded as ambient pressure. Comparing the reforming channel proposed in the present study with a serpentine channel that is a conventional structure in reactor applications, the pressure drop of the proposed channel is lower than that of serpentine shape. In addition, the velocity distribution of proposed structure is more uniform than that of serpentine type. As shown in the velocity contour, in the entire reforming zone, the flow had uniform velocity that provides constant resident time of reactants on the catalyst. Maintenance of the optimal resident time in the channel makes the micro reformer display the maximum performance. At temperature distribution in the reformer layer, temperature field was uniform within 7 ◦ C due to the high thermal conductivity of the silicon wafer. It can be seen that methanol was perfectly converted to hydrogen at the outlet. Performance of the micro reformer was measured at various test conditions. Fig. 9 shows the methanol conversion as a function of the reformer temperature at 0.1 ml/h feed rate of methanol. In this case, the electric heater was used to provide the heat to the reformer instead of the combustor. The methanol conversion increased with the reformer temperature when the temperature was below 250 ◦ C. For the reformer temperature higher than 250 ◦ C, the conversion was maintained with its value at 250 ◦ C. Optimal methanol conversion was 93.1% at 250 ◦ C with 0.1 ml/h feed rate of methanol. The production rate of hydrogen was approximately 2.8 ml/min that is sufficient amount to generate 0.24 W electric power for typical PEM fuel cell. Fig. 10 shows the CO content in the reformate gas as a function of the reformer temperature at 0.1 ml/h feed rate. The CO content increased with the reformer temperature. For the reformer temperature higher than 280 ◦ C, the CO was generated exponentially. It is difficult to remove the high CO content by using PROX reactor.

Considering the trade-off of methanol conversion and CO concentration, the optimal temperature was 250 ◦ C. 3.3. Methanol reformer operation by a catalytic combustor Catalytic combustor was used as a heat source of the reformer. Methanol conversion and CO concentration of the methanol reformer as a function of the reaction temperature are shown in Fig. 11. Methanol feed rate was 0.1 ml/h with S/C = 1.1 and hydrogen feed rate was varied as the reaction temperature. The

Fig. 9. Methanol conversion as a function of the reformer temperature at 0.1 ml/h feed rate. In this case, the electric heater was used to provide the heat to the reformer instead of the combustor.

210

T. Kim, S. Kwon / Sensors and Actuators A 154 (2009) 204–211

Fig. 10. Content of carbon monoxide in the reformate gas as a function of the reformer temperature at 0.1 ml/h feed rate.

combustor generated the sufficient amount of heat to sustain the methanol–steam reforming. The results of the performance at each case of using an electric heater and catalytic combustor were nearly identical, which means that combustor provided the uniform heat flux to the reformer layer because of the high thermal conductivity of the silicon wafer and excellent heat-exchange performance of the microchannels. 3.4. CO removal through PROX reactor Conversion of carbon monoxide in PROX reactor as a function of the reaction temperature with varying O2 /CO is shown in Fig. 12. Mixed gas including 69.91% H2 , 3.06% CO, 2.03% CH4 , and 25% CO2 was used for PROX tests. Optimal temperature was 200 ◦ C and the CO conversion increased with O2 /CO ratio. When O2 /CO ratio was 4 at 200 ◦ C, the CO conversion was 85.15% with 1 wt% Pt/Ru/␥-Al2 O3 catalyst, thus the remaining 0.45% CO was still the amount to be able to deactivate the anode catalyst of PEMFC. Loading amount of Pt/Ru/␥-Al2 O3 catalyst was increased up to 5 wt%, which showed 100% CO conversion when O2 /CO ratio was 4 at 200 ◦ C.

Fig. 11. Methanol conversion and CO concentration as a function of the reaction temperature (methanol feed rate = 0.1 ml/h, S/C = 1.1, hydrogen feed rate was varied as the reaction temperature, ˚ = 1.0).

Fig. 12. Conversion of carbon monoxide as a function of the reaction temperature with varying O2 /CO.

3.5. Integrated tests of a micro fuel cell with a methanol reformer Performance curve of MEMS fuel cell system with a pure hydrogen and reformate gas is shown in Fig. 13. When a methanol feed rate was 2 ml/h, the flow rate of reformate gas was 71 ml/min which included 74% hydrogen concentration, thus a hydrogen flow was approximately 50 ml/min after passed through the PROX reactor. Pure hydrogen test used a gas flow rate of 50 ml/min. Power density was 195 mW/cm2 when the potential was 0.64 V. The performance was low compared with the result for pure hydrogen because the feed at the fuel cell that included undesired CO, CO2 , and N2 . The fabricated fuel cell system including a reformer, combustor, PROX reactor, and PEM fuel cell had a mass of 18.4 g and a volume of 10 cc. For the duration of 20 h, a methanol storage of 31.4 g (40 cc) was required and water feed requirement was 19.5 g (19.5 cc). Therefore, the mass and volume of the total system are 69.3 g and 69.5 cc, respectively, thus the system would have a weight specific energy density of 225 W h/kg and a volume specific energy density of 224.5 W h/kg. MEMS fuel cell system in the present study had an energy density two times higher than the

Fig. 13. Performance curve of MEMS fuel cell system with the reformate gas generated by a methanol reformer. The test with pure hydrogen was carried out for a performance comparison.

T. Kim, S. Kwon / Sensors and Actuators A 154 (2009) 204–211

state-of-art rechargeable batteries and it shows the possibility that a micro fuel cell system can be an ideal alternative solution for a portable power source in the future.

211

Design, fabrication and evaluation of the MEMS fuel cell system were described. Methanol reformer, catalytic combustor, PROX reactor, and fuel cell were fabricated using MEMS fabrication combined with the catalyst processes. Methanol reformer had a performance of methanol conversion higher than 90% in the range of 0.1–2 ml/h feed rate. Catalytic combustor generated the sufficient amount of heat to sustain the methanol–steam reforming reaction. PROX reactor was able to eliminate carbon monoxide from the reformate gas at the optimal temperature of 200 ◦ C. Micro PEM fuel cell was fabricated and the integrated tests with the methanol reformer were carried out. The system generated the power of 195 mW/cm2 at the potential of 0.64 V and the weight specific energy density was 225 W h/kg.

[3] K. Yoshida, S. Tanaka, H. Hiraki, M. Esashi, A micro fuel reformer integrated with a combustor and a microchannel evaporator, J. Micromech. Microeng. 16 (2006) S191–S197. [4] N.-T. Nguyen, S.H. Chan, Micromachined polymer electrolyte membrane and direct methanol fuel cells—a review, J. Micromech. Microeng. 16 (2006) R1– R12. [5] T. Kim, S. Kwon, MEMS fuel cell system for portable power source: integration of methanol reformer, PROX, and fuel cell, in: Proceedings of the MEMS 2008, Tucson, USA, January 13–17, 2008, pp. 980–983. [6] T. Kim, S. Kwon, Design, fabrication and testing of a catalytic microreactor for hydrogen production, J. Micromech. Microeng. 16 (2006) 1752– 1760. [7] T. Kim, S. Kwon, Preparation of Cu/ZnO for fabrication of a micro methanol reformer, Chem. Eng. J. 123 (3) (2006) 93–102. [8] D.-E. Park, T. Kim, S. Kwon, C.-K. Kim, E. Yoon, Micromachined methanol steam reforming system as a hydrogen supplier for portable proton exchange membrane fuel cells, Sens. Actuators A 135 (2007) 58–66. [9] S. Srinivas, A. Dhingra, I. Hong, G. Erdogan, A scalable silicon microreactor for preferential CO oxidation: performance comparison with a tubular packed-bed microreactor, Appl. Catal. A: Gen. 274 (2004) 285–293. [10] T. Kim, S. Kwon, A MEMS methanol reformer heated by catalytic combustor, in: Proceedings of the 18th International Symposium on Transport Phenomena, Daejeon, Korea, August 27–30, 2007, p. 258. [11] T. Kim, J.S. Hwang, S. Kwon, A MEMS methanol reformer heated by decomposition of hydrogen peroxide, Lab Chip 7 (7) (2007) 836–847.

Acknowledgement

Biographies

This work was supported by the Korea Science and Engineering Foundation (KOSEF) grant funded by the Korea government (MOST) (No. R0A-2007-000-20065-0)

Taegyu Kim received the BS degree from Korea Aerospace University (KAU) in 1997, the MS and PhD degree in Aerospace Engineering from KAIST in 2005 and 2008, respectively. He is now a postdoctoral in Rocket laboratory from KAIST. His present research area includes micro power system, micro fuel cell, hydrogen generator, and power source for unmanned systems.

4. Conclusion

References [1] J.D. Holladay, J.S. Wainright, E.O. Jones, S.R. Gano, Power generation using a mesoscale fuel cell integrated with a microscale fuel processor, J. Power Sources 130 (2004) 111–118. [2] A.V. Pattekar, M.V. Kothare, A microreactor for hydrogen production in micro fuel cell applications, J. Microelectromech. Syst. 13 (1) (2004) 7–18.

Sejin Kwon received the BS degree from Seoul National University in 1982, the MS degree in Aerospace Engineering from KAIST in 1984, and PhD in Aerospace Engineering from University of Michigan, Ann Arbor. In 1997, he joined the Department of Aerospace Engineering at KAIST, where he is now an Associate Professor. His current research area includes micro catalytic reactor, micro fuel cell, and micro propulsion devices. He is a member of AIAA.