eee eee REPORT DOCUMENTATION PAGE prom ceeds TEE sins San wpa mieng 7 STEN © Soe & Geen, Love Rame aay be ine wimg Gavaton, sac a aaa, ares SEL PINE SON TLE PIAS Teale Te Sieetn a "tioman” See ements moeaeg ssi erle 9 sow eet ian Sot Sa. weno vi Be sate he Ba hinges aco Papo Paceien Pee OPbeSTOO, Wangan Be zeke SE TEN a 1. AGENCY USE ONLY (Leave Blank) | 2 REPORT DATE, “REPORT TYPE AND DATES COVERED rd FINAL 01 May 95 ~ 31 Jul 00 _TILEAND SUBTE FUNDING NUMBERS Micro Gas Turbine Generators Grant DAAHOS-85-1-0083, &_AUTHORS A Epstein. K Breuer, J. Lang, M. Schr. Sentra, M. Spearng,C. Tan | Watz 7, PERFORMING ORGANIZATION NAME(S) AND ADDRESS(ES) 3 PERFORMING ORGANIZATION REPORT NUMBER Massachusetts Institute of Technology TT Massachusetts Ave, 31-264 Cambridge, MA 02199 '& SPONSORING / MONITORING AGENCY NAME(S) AND ADDRESS(ES) 10 SPONSORING / MONITORING AGENCY REPORT NUMBER US. Army Research Ofce ARO 33888.2-CH-MUR P.O. Box 12211 Research Thangle Park, NC 27708-2211 7_ SUPPLEMENTARY NOTES & OSTOTCRSTHIBUTION STATEMENT A 2 DISTRIBUTION CODE Approved for Public Release Distribution Unlimited TS ABSTRACT (Maximum 200 words) LMIThas developed the technology for mico-gas turbine generators. These are milimeter-o centimeter-size heat engines fabreated with semiconductor nusty meremacnining tecinigues (MEMS), utmatly capable of producing 10-100 W of power ness than a cubic centimeter. ‘ppicatons incude compact power soufoes ofleing energy and power densities an order of magneude beter than current battery technology propulsion for small a vehicles; anda variety of miroblowers, compressors, and heat pumps. The work was divide into 8 microscale dscpinary frees (1) engine systems design, (2) tubomachinery fuld dynamics, (3) comeuston, () structures, (5) bearings, (6 elecromechanics, (7) siicon fabrication technology, and (8) micoftrcaion of hightemperature mateals and structures. Advances in he discoinary technologies enabled the sign and constuaton of preotpencpe “demo engine”. This 20 mem square by 4 men thick smple cycle gas turbine s designed to produce about 11 grams of fust or 17 wats of shaft power. The design turbine inet emperature 1600 K and the rotational speed is 1 2M pm, Atthe ‘oncuson ofthis MUR the ft engines had Been bul ang were just beginning lestng. A companion miccurbogenerator a few months behind the gas tabine, 74 SUBJECT TERMS 75. NUMBER OF PAGES MEMS, compact poner, mcroturine, micromoor,microcombustion 2 76. PRICE CODE 7. SEGURITY CLASSIFICATION | 18. SECURMTY CLASSIFICATION | 19, SECURITY 2), LIMITATION OF ABSTRAGT OF REPORT (OF THIS PAGE ‘CLASSIFICATION OF ue UNCLASSIFIED UNCLASSIFIED ABSTRACT UNCLASSIFIED NSN 7500-07-280500 ‘Siandard Form 208 Rev, 255) Presorbes by ANSI Std, 239-1208-102 pric QuLLtiy migrzo=D 20010116 13Gas Turbine Laboratory Department of Aeronautics and Astronautics Massachusetts Institute of Technology Cambridge, MA 02139 Final Technical Progress Report ‘on Grant #1DAAH04-95-1-0093 entitled MICRO GAS TURBINE GENERATORS prepared for US Army Research Office P.O. Box 12211 Research Triangle Park, NC 27709 ATIN: Dr. Richard Paur CO-INVESTIGATORS: _ Dr... K.S. Breuer Prof. A.H. Epstein (Principal Investigator) Prof. JH. Lang Prof. M.A. Schmidt Prof. S.D. Senturia Prof, S.M. Spearing Dr. C.S. Tan Prof. LA. Waitz PERIOD COVERED: April 1, 1996 - July 31, 2000 DISTRIBUTION STATEMENT A Approved for Public Release December 2000 Distribution Unlimited1.0 REPORT OUTLINE ‘This is the final technical progress report on ARO Grant DAAH04-95-1-0093, a five- year Multidiscipline University Research Initiative (MURI) program. Because the program has generated lengthly annual technical reports and a large number of technical publications and graduate theses (which are available upon request) this final technical report is relatively brief, It consists of four sections in addition to this one: (2) A short summary and list of accomplishments, (3) a list of participants, (4) a list of publications and theses, and (5) a technical paper which gives a more detailed overview of the technology. 2.0 SUMMARY AND ACCOMPLISHMENTS Under this MURI support, MIT has developed the technology for micro-gas turbine generators. These are millimeter- to centimeter-size heat engines fabricated with semiconductor industry micromachining techniques. As such, they are micro electrical and mechanical systems (MEMS) devices. They also represent the first application of a new technical field conceived under this program, Power MEMS. Calculations show that these microdevices may ultimately be capable of producing 10-100 W of power or 10-50 grams of thrust in less than a cubic centimeter. Applications include compact power sources offering, energy and power densities an order of magnitude better than current battery technology; propulsion for small air vehicles; and a variety of microblowers, compressors, and heat pumps. Much of this technology is also applicable to MEMS microrocket engines (which are the enabling technology for very small launch vehicles and missiles), the technology of which can be considered derivative of this program. The promise of Power MEMS is sufficient that DARPA and the Japanese government have started programs in this area and that Power MEMS international workshops and meetings are now held This technical effort has been an admixture of research and engineering design since the project goals are device-oriented. The work is nominally divided into 8 principal microscale disciplinary areas: (1) engine systems design, (2) turbomachinery fluid dynamics, (3) combustion, (4) structures, (5) bearings, (6) electromechanics, (7) silicon fabrication technology, (8) microfabrication of high temperature materials and structures. The centerpiece of this effort has been the design and construction of the world’s first MEMS micro-gas turbine engine. Realization of such a device has necessitated significant advances in many disciplinary technologies. Some specific technical achievements include the establishment of the enabling basic technology for and demonstration of: The first MEMS micro-gas bearing technology, with speeds above 14M rpm demonstrated. ‘© Micro-high-speed turbomachinery, transonic centrifugal turbines and compressors.‘© The first high power density MEMS H, and hydrocarbon fuel microcombustor technology; 400 watts of thermal energy in 200 mm’ has been demonstrated. ‘* The first cooled silicon microturbine airfoils, which have operated in a gas temperatures above the melting point of silicon, 1800 K. © Micromotor performance 100x higher than previously achieved. * Advances in the SOA of high aspect ratio silicon etching by a factor of 3-5. © The first multi-wafer (5-6) precision-aligned (1-2 micron) silicon structures. © Packaging for very high temperature silicon chips, including high pressure fluid interconnections. These advances in disciplinary technology enabled the design and construction of a so- called “demo engine”, as a proof-of-principle. This 20 mm square by 4 mm thick simple cycle gas turbine is designed to produce about 11 grams of thrust or 17 watts of shaft power ‘The design turbine inlet temperature is 1600 K and the rotational speed is 1.2M rpm. At the conclusion of this MURI, the first engines had been built and were just beginning testing. A ‘companion microturbogenerator is a few months behind the gas turbine. Demo Gas Turbine Engine This success of the technological advances made in this MURI prompted DARPA to match MURI funding. Discussions with the Army suggest that further 6.1 and 6.2 support may be available both to continue the basic research and to begin the technology transition process. MIT has initiated discussions with several potential industrial partners interested in developing the technology further.3.0 CONTRIBUTING TECHNICAL PERSONNEL Name Faculty: Prof. Kenneth Breuer Prof. John Brisson Prof, Alan H. Epstein Prof. Jeffrey H. Lang Prof. Martin A. Schmidt Prof, Stephen D. Senturia Prof. Mark S. Spearing Prof. Jan A. Waitz Technical Staff: Dr. G.K. Ananthasuresh (Post Doc) Dr. A. Ayon (Post Doc) Dr. Christopher Cadou (Post Doc) Dr. Fredric F. Ehrich (Senior Lecturer) Eric Esteve (Visiting Eng.) Dr. Anthony Forte Dr. Gautam Gauba (Post Doc) Dr. Reza Ghodssi Dr. Yifang Gong (Post Doc) Dr. Paul Holke (Post Doc) Dr. Eugene W. Huang (LL Tech. Staff) Dr. Stuart A. Jacobson (Engineer) Dr. Ravi Khanna (Research Eng.) Dr. Carol Livermore (Post Doc) Steven Lukachko (Research Eng.) Dr. Paul Maki (LL Tech Staff) Dr. James Paduano (Principal Eng.) Larry Retherford, Jr. (LL Tech. Staff) Dr. Choon S. Tan (Principal Eng.) Dr. Steven Umans Dr. Richard Walker (C.S. Draper Labs) Paul Warren Dr. Wenjing Ye (Post Doc) Patrick Yip (LL Tech Staff) Dr. Xin Zhang (Post Doc) Primary Discipline Fluids, Instrumentation ‘Thermal Systems, Heat Transfer Engine Design, Fluids Electromechanics Fab, Processes HWFab, Processes & Materials Structures, Materials ‘Combustion hFab Modeling Fab, Processes Fluids, Combustion Rotor Dynamics, Design Fluids, Engines wEabrication Combustion uEabrication Turbomachinery pEabrication Structures Fluids Fabrication Electromechanics Combustion Fabrication Controls Packaging ‘Turbomachinery Electromechanics Gas Bearings Electronics Structures ‘MAV Avionics hBabricationGraduate Students: Dye-Zone Chen Fabrication, Instrumentation Kuo-Shen Chen Structures Dongwon Choi Structures & Materials Luc Frechette Turbomachinery Systems Tod Harrison Structures, Packaging Kashif Khan Turbomachinery Jin-Wook Lee Combustion Chuang-Chia Lin Fabrication Chunmei Liu Turbomachinery Kevin Lohner Structures & Materials Adam London Packaging Amit Mehra Turbomachinery Bruno Miller Structures Jose Miranda Electric Bearings Hyug-Soo Moon Structures & Materials Steve Nagle Electric Machinery DJ. Or Fluid Bearings Baudoin Philippon Turbomachinery Ed Piekos Fluids Modelling John Protz Engine Systems Nicholas Savoulidis Bearings Gregory Shirley Turbomachinery Chris Spadaccini Combustion Shaun Sullivan Fluids, Heat Transfer David Tang Instrumentation Sheng-Yang Tzeng Combustion Douglas Walters Structures Chee Wei Wong Bearings4.0 LIST OF PUBLICATIONS AND THESES 1. Trinh, T., "Large-Scale Test of Position Detection for Microengine Rotor,” Technical Report, May 1996, 2. Esteve, , "Secondary Flow System Modeling,” Technical Report, 1996. 3. Waitz, LA., Gauba, G. and Tzeng, Y.-S., "Combustors for Micro-Gas Turbine Engines,” Proc. of the International Mechanical Engineering Congress and Exposition, November 1996. 4. Spearing, S. M., Chen, K.S., “Micro-Gas Turbine Engine Materials and Structures”, presented at 21" Annual Cocoa Beach Conference and Exposition on Composite, Advanced Ceramics, Materials and Structures, January 1997. 5. Epstein, A. H., and Senturia, $. D., “Macro Power from Micro Machinery”, Science, Vol. 276, May 1997, p. 1211. 6. Epstein, Senturia, Anathasuresh, Ayon, Breuer, Chen, Ehrich, Esteve, Gauba, Ghodssi, Groshenry, Jacobson, Lang, Lin, Mehra, Mur Miranda, Nagle, Orr, Piekos, Schmidt, Shirley, Spearing, Tan, Tzeng, Waitz, "Power MEMS and Microengines,” IEEE Conference on Solid State Sensors and Actuators, Chicago, IL, June 1997 7, Epstein, Senturia, Al-Midani, Anathasuresh, Ayon, Breuer, Chen, Ehrich, Esteve, Frechette, Gauba, Ghodssi, Groshenry, Jacobson, Kerrebrock, Lang, Lin, London, Lopata, Mehra, Mur Miranda, Nagle, Orr, Piekos, Schmidt, Shirley, Spearing, Tan, Tzeng, Waitz, “Micro-Heat Engines, Gas Turbines, and Rocket Engines", ALAA 97-1773, 28th AIAA Fluid Dynamics Conference, 4th AIAA Shear Flow Control Conference, Snowmass Village, CO, June 29-July 2, 1997. 8. Piekos, E.S., Orr, D.J., Jacobson, S.A., Ehrich, F.F. and Breuer, K.S., "Design and Analysis of Microfabricated High Speed Gas Journal Bearings," AIAA Paper 97-1966, 28th AIAA Fluid Dynamics Conference, Snowmass Village, CO, June 29-July 2, 1997. 9. Waitz, LA., Gautam, G., Tzeng, Y.-S., "Combustors for Micro-Gas Turbine Engines,” ASME Journal of Fluids Engineering, Vol. 120, March 1998. 10. Mehra, A., and Waitz, I. A., “Development of a Hydrogen Combustor for a Microfabricated Gas Turbine Engine”, Solid-State Sensor and Actuator Workshop, Hilton Head Island, SC, June 1998. LL. Jacobson, S. A., “Aerothermal Challenges in the Design of a Microfabricated Gas Turbine Engine”, ALAA 98-2545, 29th AIAA Fluid Dynamics Conference, Albuquerque, NM, June 1998. 12. Ay6n, A.A., Lin, C.C., Braff, R., Bayt, R., Sawin, HLH. and Schmidt, M., “Etching Characteristics and Profile Control in a Time Multiplexed Inductively Coupled Plasma Eicher,” 1998 Solid State Sensors and Actuator Workshop, Hilton Head, SC, June 1998. 13, Huang , E., “Thermal Design Trade Studies for A Silicon Turbojet Engine”, (Initial Report), October 1998 14. Piekos, E.S. & Breuer, K.S. "Pseudospectral Orbit Simulation of Non-Ideal Gas- Lubricated Journal Bearings for Microfabricated Turbomachines," Paper No. 98-Trib-48, presented at the Tribology Division of The Americal Society of Mechanical Engineers for presentation at the Joint ASME/STLE Tribology Conference, Toronto, Canada, October 1998. Also, to appear in Journal of Tribology.15. Ayén, A.A., Ishihara, K., Braff, R., Sawin, HLH. and Schmidt, M., “Application of the Footing Effect in the Microfabrication of Self-Aligned, Free-Standing Structures,” 45th Intemational AVS Symposium, Baltimore, MD, November 1998. 16. Mehra, A., Jacobson, S. A., Tan, C. S., and Epstein, A. H., “Aerodynamic Design Considerations for the Turbomachinery of a Micro Gas Turbine Engine”, presented at the 25" National and 1" International Conference on Fluid Mechanics and Power, New Delhi, India, December 1998. 17. Chen, K-S, Ayon, A. A., Lohner, K. A., Kepets, M. A., Melconian, T. K., and Spearing, S. M,, “Dependence of Silicon Fracture Strength and Surface Morphology on Deep Reactive Ton Etching Parameters”, presented at the MRS fall Meeting, Boston, MA, December 1998. 18, Ay6n, A.A., Ishihara, K., Braff, R., Sawin, HH. and Schmidt, M., “Deep Reactive Ion Etching of Silicon,” Invited Presentation at Materials Research Society Fall Meeting, Boston, MA, November 30-December 4, 1998. 19. Mirza, A.R. and Ay6n, A.A., “Silicon Wafer Bonding: Key to MEMS High-Volume “Manufacturing,” SENSORS, Vol. 15, No. 12, December 1998, pp. 24-33. 20. Chen, K-S, Ayon, A., and Spearing, S. M., “Silicon Strength Testing for Mesoscale Structural Applications”, MRS Proceedings, Vol. 518, 1998, pp. 123-130. 21. Lin, C.C., Ghodssi, R., Ayon, A.A., Chen, D.Z., Jacobson, S., Breuer, K.S., Epstein, A.H. & Schmidt, M.A. "Fabrication and Characterization of a Micro Turbine/Bearing Rig" presented at MEMS '99, January 1999, Orlando, FL. 22. Mirza, A.R. and Ay6n, A.A., “Silicon Wafer Bonding: The Key Enabling Technology for MEMS High-Volume Manufacturing,” Future Fab International, Issue 6, January 1999, pp. 51-56. 23, Ay6n, A.A., Braff, R., Lin, C.C., Sawin, H.H. and Schmidt, M., “Characterization of a Time Multiplexed Inductively Coupled Plasma Etcher,” Journal of the Electrochemical Society, Vol. 146, Number 1, January 1999, pp. 339-349. 24. Ayon, A.A., Ishihara, K,. Braff, R.A., Sawin, H.H., Schmidt, M.A., “Microfabrication and Testing of Suspended Structures Compatible with Silicon-on-Insulator Technology”, submitted to the Journal of Vacuum Science and Technology, February:1999. 25. Mirza, A. R., Ayon, A. A., "Advanced Silicon Wafer Bonding," Micromachine Devices, Vol. 4, No. 2, February 1999. 26. Ayén, A.A., Epstein, A.H., Frechette, L., Nagle, S. and Schmidt, M. A., “Tailoring and Controlling Etch Directionality in a Deep Reactive Ion Etching Tool,” submitted to Transducers'99, Sendai, Japan, June 1999. 27, Mebra, A., Waitz, LA., Schmidt, M. A., "Combustion Tests in the Static Structure of @ 6- Wafer Micro Gas Turbine Engine," 1999 Solid State Sensor and Actuator Workshop, June 2- 4, 1999. 28. Mirza, A.R. and Ay6n, A.A., “Silicon Wafer Bonding for MEMS Manufacturing,” Solid State Technology, Vol. 42, No. 9, pp. 73-78, August 1999. 29, Mehra, A., Ayon, A. A., Waitz, I. A., and Schmidt, M. A., “Microfabrication of High Temperature Silicon Devices Using Wafer Bonding and Deep Reactive Ion Etching”,IEEE/ASME Journal of Microelectromechanical Systems, Vol. 8, No. 2, June 1999, pp. 152-160, 30. Ayén, A. A., Chen, D.-Z., Braff, R. A., Khanna, R., Sawin, H. H., Schmidt, M. A., "A novel Integrated Process Using Fluorocarbon Films Deposited with a Deep Reactive Ion Etching (DRIE) Tool," Fall Meeting of the Materials Research Society, Boston, MA , November 29 - December 3, 1999. 31. Chen, K-S, Ayon, A., and Spearing, S. M., “Controlling and Testing the Fracture Strength of Silicon at the Mesoscale”, to be published in the Journal of the American Ceramic Sociery, 1999. 32. Ay6n, A.A., Braff, R.A., Bayt, R., Sawin, HLH., Schmidt, M.A., “Influence of Coil Power in the Etching Characteristics in a High Density Plasma Etcher,” Journal of the Electrochemical Society, Vol. 146, No. 7, 1999. 33. Epstein, A.H., Jacobson, S.A., Protz, J.M., Frechette, L.G., "Shirtbutton-Sized Gas ‘Turbines: ‘The Engineering Challenges of Micro High Speed Rotating Machinery," Plenary Lecture, 8* International Symposium on Transport Phenomena and Dynamics of Rotating Machinery (ISROMAC-8), Honolulu, HI, March 2000. 34, Epstein, A.H., "The Inevitability of Small," Aerospace America, March 2000, pp. 30-37. 35, Ayon A.A., Zhang X., and Khanna R., "Ultra Deep Anisotropic Silicon Trenches Using, Deep Reactive Ion Etching (DRIE)," Hilton Head Solid-State Sensor & Actuator Workshop, Hilton Head Island, SC, June 4-9, 2000, pp. 339-342. 36. Epstein, A.H., Jacobson, $.A., Protz, Livermore, C., Lang, J., Schmidt, M.A., "Shirtbutton- Sized, Micromachined, Gas Turbine Generators," presented at 39 Power Sources Conference, Cherry Hill, NJ, June 2000. 37. Ayon A.A., Protz J.M., Khanna R., Zhang X., and Epstein A.H., "Application of Deep Silicon Etching and Wafer Bonding in the MicroManufacturing of Turbochargers and Micro- Air-Vehicles," the 47" International Symposium of the American Vacuum Society, Boston, MA, October 2-6, 2000. 38. Zhang X., Ghodssi R., Chen K-S, Ayon A.A., and Spearing S.M., "Residual Stress Characterization of Thick PECVD TEOS Film for Power MEMS Applications," Hilton Head Solid-State Sensor & Actuator Workshop, Hilton Head Island, SC, June 4-9, 2000, pp. 316- 319. 39, Frechette, L.G, Jacobson, S.A., Breuer, K.S., Ehrich, F-F., Ghodssi, R., Khanna, R., Wong, C.W,, Zhang, X., Schmidt, M.A., and Epstein, A.H., "Demonstration of a Mictofabricated High-Speed Turbine Supported on Gas Bearings," Hilton Head Solid-State Sensor & Actuator Workshop, Hilton Head Island, SC, June 4-9, 2000, pp. 43-47. 40. Orr, D.J, and Jacobson, S.A., "High Order Galerkin Models for Gas Bearings,” submitted to the Proceedings of the ASME/STLE Tribology Conference, paper ASME/2000-TRIB-131, Seattle, WA, October 2000.41. Mehra A., Zhang X., Ayon A.A., Waitz I.A., and Schmidt M.A., “A Through-Wafer Electrical Interconnect for Multi-Level MEMS Sevices,” Journal of Vaccum Science and Technology B, Vol. 18, No. 5, pp. 2583-2589, September/October 2000. 42, Mehra A., Zhang X., Ayon A.A., Waitz I.A., Schmidt M.A., and Spadaccini C.M., “A 6- ‘Wafer Combuston System for a Silicon Micro Gas Turbine Engine,” to appear in Journal of MicroBlectroMechanicalSystems, December 2000. 43. Ayon, A.A., “Time Multiplexed Deep Etching,” Sensors, Vol. 16, No. 9, September 2000, pp. 64-73. 44. Ayon, A.A., Bayt, R.L., Breuer, K.S., “Deep Reactive Ion Etching: A Promising Technology for Micro and Nanosatellites,” submitted to Journal of Smart Materials and Structures: Special Issue on MEMS in Space, June 2000. 45. Ghodssi R., Frechette L.G., Nagle S.F., Zhang X., Ayon A.A., Senturia S.D., and Schmidt M.A., "Thick Buried Oxide in Silicon (TBOS): An Integrated Fabrication Technology for Multi-Stack Wafer-Bonded MEMS Processes," Proceedings of the 1999 International Conference on Solid-State Sensors and Actuators, Sendai, Japan, June 7-10, 1999, pp. 1456- 1459. 46. Ghodssi R., Zhang X., Chen K-S, Spearing S.M., and Schmidt M.A., "Residual Stress Characterization of Thick PECVD Oxide Film for MEMS Application," the 46" International Symposium of the American Vacuum Society, Seattle, WA, October 25-29, 1999. 47. Chen K-S, Zhang X., and Ghodssi R., "Residual Stress and Failure Modeling of Thick PECVD Oxide Films for MEMS Application,” Proceeding of the 1" joint China/Taiwan Symposium on Microsystem Technology, Tainan, Taiwan, May 2000, pp. 264-269. 48. Chen K-S, Zhang X., and Spearing S.M., "Processing of Thick Dielectric Films for Power MEMS: Stress and Fracture," Materials Science of Microelectromechanical System (MEMS) Devices III, Materials Research Society Symposium, Boston, MA, November 27 - December 1, 2000. 49, Khanna R., Zhang X., Protz J.M., and Ayon A.A., "Microfabrication Protocols for Multi- Stack Projects Involving Deep Reactive Ion Etching and Wafer-Level Bonding," Sensors, accepted, will be published in March issue of 2001. 50. Ayon A.A., Zhang X., and Khanna R., "Anisotropic Silicon Trenches 300 jum to 500 um Deep Employing Time Multiplexed Deep Etching (TMDE)," Sensors and Actuators, will be published in the special issue for the Hilton Head Solid-State Sensor & Actuator Workshop. 51, Zhang X., Ghodssi R., Chen K-8, Ayon A.A., and Spearing S.M., "Stress and Fracture in ‘Thick Tetraethylorthosilicate (TEOS) Films," Sensors and Actuators, will be published in the special issue for the Hilton Head Solid-State Sensor & Actuator Workshop. 52, Savoulides, N., Breuer, K.S., Jacobson, S., Ehrich, F-F., “Low-Order Models for Very Short Hybrid Gas Bearings,” ASME Paper 2000-TRIB-12, presented at the STLE/ASME Tribology Conference, Seattle, WA, October 2000; also to appear in J. of Tribology.4.1: Graduate Theses 1. Groshenry, C., "Preliminary Design Study of a Micro-Gas Turbine Engine," M.S. Thesis, MIT Department of Aeronautics and Astronautics, September 1995, 2. Mehra, A., “Computational Investigation and Design of Low Reynolds Number Micro- Turbomachinery,” M.S. Thesis, MIT Department of Aeronautics and Astronautics, June 1997, 3. Mur Miranda, J.O., “Feasibility of Electrostatic Bearings for Micro Turbo Machinery,” MEng. Thesis, MIT Department of Electrical Engineering and Computer Science, December 1997. 4, Tzeng, Y-S, “An Investigation of Microcombustion Thermal Phenomena,” M.S. Thesis, MIT Department of Aeronautics and Astronautics, June 1997. 5. Shirley, G., "An Experimental Investigation of a Low Reynolds Number, High Mach Number Centrifugal Compressor,” M.S. Thesis, MIT Department of Aeronautics and Astronautics, September 1998. 6. Chen, K-S, “Materials Characterization and Structural Design of Ceramic Micro Turbomachinery,” Ph.D. Thesis, MIT Department of Mechanical Engineering, February 1999, 7. Lin, C.C., "Development of a Microfabricated Turbine-Driven Air Bearing Rig,” Ph.D. Thesis, MIT Department of Mechanical Engineering, June 1999. 8, Lohner, K., "Microfabricated Refractory Ceramic Structures for Micro Turbomachinery," MS. Thesis, MIT Department of Aeronautics and Astronautics, June 1999. 9. Chen, D-Z, "Design and Calibration of an Infrared Position Sensor,” M.S. Thesis, MIT Department of Mechanical Engineering, June 1999. 10. Walters, D., "Creep Characterization of Single Crystal Silicon in Support of the MIT Microengine Project," M.S. Thesis, MIT Department of Mechanical Engineering, June 1999. 11. Mehra, A., "Development of a High Power Density Combustion System for a Silicon Micro Gas Turbine Engine," Ph.D. Thesis, MIT Department of Aeronautics and Astronautics, February 2000. 12. Orr, D.J., "Macro-Scale Investigation of High Speed Gas Bearings for MEMS Devices," Ph.D. Thesis, MIT Department of Aeronautics and Astronautics, February 2000. 13, Piekos, E., "Numerical Simulation of Gas-Lubricated Journal Bearings for Microfabricated Machines,” Ph.D. Thesis, MIT Department of Aeronautics and Astronautics, February 2000. 14, Savoulides, N., "Low Order Models for Hybrid Gas Bearings," M.S. Thesis, MIT Department of Aeronautics and Astronautics, February 2000. 15. Liu, C., “Dynamical System Modeling of a Micro Gas Turbine Engine,” MS Thesis, MIT Department of Aeronautics and Astronautics, June 2000. 16. Lee, Jin-Wook, “Numerical Simulation of a Hydrogen Microcombustor,” M.S. Thesis,MIT Department of Aeronautics and Astronautics, May 2000. 17. Nagle, S.F., "Analysis, Design and Fabrication Of An Electric Induction Micromotor for a Micro Gas-Turbine Generator", Ph.D. Thesis, MIT Department of Electrical Engineering, October 2000. 18, Frechette, L.G., “Development of a Microfabricated Silicon Motor-Driven Compression System,” Ph.D. Thesis, MIT Department of Aeronautics and Astronautics, September 2000. 105.0 OVERVIEW TECHNICAL PAPER SHIRTBUTTON-SIZED GAS TURBINES: THE ENGINEERING CHALLENGES OF MICRO HIGH SPEED ROTATING MACHINERY Alan H, Epstein, Stuart A. Jacobson, Jon M. Protz, Luc G. Frechette Gas Turbine Laboratory Massachusetts Institute of Technology Cambridge, MA 02139, USA Fax 617-258-6093, epstein@mitedu KEYWORDS: MEMS, microturbine, microcombustion, ‘microbearings ABSTRACT MIT is developing micro-electro-mechanical systems (MEMS)-based gas turbine engines, turbogenerators, and rocket engines. Fabricated in large numbers in parallel using semicon ‘ductor manufacturing techniques, these engineson-a-chip are based on micro-high speed rotating machinery with power den- sites approaching those oftheir more familiar, fllsize breth- ren. The micro-gas turbine is a em diameter by 3 mum tick Si (or SiC heat engine designed to produce about 10 W of electric power oF 0.1 Nof thrust while consuming about 15 grams of Hi. Later versions may produce up to 100 W using hydtocar- bom fuels. This paper gives an overview ofthe project and dis- cusses the challenges faced in the design and manufacture of| high speed microrotatng machinery. Fluid, structural, bearing, and rotor dynamics design isues are reviewed. INTRODUCTION ‘High speed rotating machinery comes in many sizes. Ia recent years much emphasis has been placed on the large end of| the business 10 m diameter hydroelectric turbines, 300 ton ground-based gas turbine generators, 3m diameter aircraft en ines. These machines are engineered to produce hundreds of ‘megawatts of power. The focus of this paper isthe opposite end ofthe rotating machine size scale, devices a few millimeters in diameter and weighing agramortwo. These machines ae about ‘one thousandth the linear scale oftheir largest brethren and thus, since power level scales with fuid mas flow rate and flow rate seales with intake are, they should produce about one millionth the power level, afew tens of wats. "The imerest in rotating machinery of this size ange is fu- led by both a technology push and a user pull. The technology pushis the development of micromachining capability based on semiconductor manufacturing techniques. This enables the fab- rieation of complex small parts and assemblies ~ devices with ‘dimensions in the 1-10,000 mievon size range with micron and ‘even submicron precision. Such parts a produced using pho- tolithography defined features and many can be made simults- neously, holding out the promise of low production cost. Such assemblies are known as micro-lecrical-mechanical systems (MEMS) and have been the subject of thousands of publications ‘over the last decade. Early work in MEMS focused on sensors and many devices based on this technology are in large scale production (such as pressure sensors and airbag accelerometers for automobiles). More recently, work has been done on actus- tors of various sors. Fluid handling is receiving attention as wel, for example MEMS valves are commercially availabe. “The ser pallis predominately one of electric power. There is proliferation of small, portable electronics ~ computes, digi- tal assistants, cell phones, GPS receivers, ete. — which require compact energy supplies. The demand is for energy supplies whose energy and power density exceed that ofthe best batter- ies available today. Also, the continuing advance in microelee- tronics permits the shrinking of electronic subsystems of mobile Gevices such as robots and air vehicles. These small, and in ‘sone case very small, systems reqite increasing compact power ‘and propulsion For compact power production, hydrocarbon fuels bumed fn air have 20-30 times the energy density of the best current lithium chemisty-based batteries. Thermal eyles and high speed rotating machinery offer high power density compared to other power production schemes and MEMS technology is advancing rapidly. Recognizing these tends, a group at MIT began re- searchin the mid 1990's on a MEMS-based “micro-gas turbine _generator” capable of producing tes of watts of electrical power‘fom a cubic comimetersied package (Epstein and Sentra, 1997; Epstein ea, 1997) Since that ime related efforts have ‘ean stared a mieromotordeven air compressor an a bi- propellant, guid rocket motor wbich uilze much ofthe same Technology asthe ens turbine. The attractiveness ofthese de- ioe is predicated to large parton thee high power density The power density isa function ofthe design ofthe rotting machinery. This paper reports on his work in progress, with emphasis on the rotating machinery development ‘SYSTEM DESIGN CONSIDERATIONS ‘Atlenth seals of afew millimeters, thermodynamic ean- sideration ace no different than for much arger device, Thus, highpower density forasingle Brayton esclerequzeshigh com- bust exittemperstures(1400-1800K) and pressureratis above ‘and preferably above 4. This an be see nthe thermodyoa- ies cyte calelation llseated in Figare 1 which shows that several es of wats canbe expected from a machine with a rim? ntake area. The power density of rotating machinery, both flydand elect, sale with the square of the peripheral sped, dos the tess in the rotor. Thus, high power density implies highly reseed rotating structures. Peripheral speeds of 300- (600 mae needed to ahieve pressure ratios inthe 2:15: per stagerange, assuming centrifugal turbomachinery. Thisimplies ‘tor esnifugal stresses a the order of hundreds of megaPascl ‘These periheral speeds in rotors afer milimetersin diameter require roatioal rats oa the order of 1-3 millon rpm. Ths, tow fiction bearings are needed. Als, high speed rotating ma- bier generally requieshigh precision ranufatring to main- fain tghtclearane aod good balance. For millimeter sized ma- chins to have the same factional precision as metersized de ‘ies, the geomeric precision equiemens are onthe order ofa 'A Brayton cycle is ao the only choice for power produc tion from a MEMS-based thermodynamic cyele. Rankine, ‘Sting, nd Ou cycles can all be considered candidates. The ‘vantages and disodvantages of each ifr. The principal ad- vantages ofthe Brayton cycles simplicity (ony cae moving ar, a rotor, i needed), highest power density (due tothe high throughiow Mach aumber and his high mas flow per nits), andthe avalabiliyofcompressed ar for cooling and oer wes, ‘The primary disadvantage is that a mislmum component eff ciency (onthe exder of 40-50%) must be met for the eyle 0 be “4:1 Pressure Ratio Speci Fost Consumption (as) ‘hat Pomar Output wat) pr rn et Ae ‘igare 1: Simple eyele gas turbine performance with H fuel ‘Table 1: Micro vs: Macro Material Properties Tiles! [Taam | Mace] Miso ‘Ser | Alloys” | er) | Siow iss Geames Canisgasien | 30 | <0 | co | 1000 Teele) om Tmmaasiess [27x103 12x 103] 20%103 | 09x10" Ue, so) siees | -26 ss | -0 epi oPsKens Mactenp ce) | -1000 | 300 | 100 | 00 Tinalogtacor | Gxep) [Gens | coxdaton) | Gren) self-sustaining; nly them can net power be produced. AS thisis A fistofits-kind effort that challenges the capabilities of se ‘ral iciplines, especially miroforicstion, simplicity isa very Assirable vue, ‘The Brayton cycle seems mos tractive in this egar ‘Mechanics Sealing “Thermdynamic considerations do not change as machines ‘become smaller, but mechanics considerations do, Structural mechanics, fluid mechanics, atdelectomechanicsallchangein ‘amanaerimportant to machine performance and design asength ‘eal ie decreased by a factor of order 1000. Forstructral mechanics, its the change in material prop- ceties wth length scale thats most important. A telativel smal ‘set of mates re acessibe to cuzent microfabrication tech ology. Te most commonly usd by far is Silica (Si, while Silicon Carbide (SiC) and Gall Arsinide (GaAs) are used to date mainly i niche applications. For many of the mets in- portant to high sped rotting machinery, Si and SiC are sape- flor fo moet commonly used mals suchas tel, tania, and ‘ike base superalloys (Spearng and Chen, 1997). This can ‘beseenin Table which compares materials tems of proper- ties impotnt for centrifugal stress, thermal ses, vibrations, nd hot szength Figure 2) (Mba, 2000), Materials such as Si fd SiC are nt use in conventional sized rotating machinery because they are rite. Ther usable strength is dominated by flaws introduced in manufacturing and flaw population gene~ ally scales with part volume However, these mazras are aval able inthe form regied for semiconductor manufacture (hin wafers) with an esentialy perfect, single crystal structure. AS Jhchtheyhavehighusble strength, values after micromachining shove 4 GBa have been reported (Chen ef aly 1998), several timeshigherthan that of rotating machinery metalic loys, This Higher strength can beused either to realize higher ation speeds (and this higher power densities) at constant geometry, oft ‘simplify the geomexy (and dus the manufataring) at constant peripheral speed. To dae, we have adopted the later approach for expediency, An addtional mitral cosideration i that thermal shock increases with length scale. Thus, materia which ‘ave very high temperature capabilites but are not considered igh tempersare stoctral ceramics (such as alumina or sap hire) du to their susceptibility to Ubermal shock, ae viable a the milmeter and below legth scales. Since these have not: SS See | E04 , Nn cmse | 8 | coed i i 7 ‘ a boot . : Be g bol : & se os aaa 2 aa 2 Tae Teper 09 “renrertre 4) “rence 0) [Figure 2: Material properties relevant o rotating machinery. ‘ben considered as MEMS materials in the past tee is cur. rently litle suitable manufactring technology aval for these materials (Sparing and Cha, 1997), laid Mechanics Sealing hid mechanics are ko scale dependent (Jaobs00, 1998) (One aspect stat Viscous fees are more important at sal, sale, Pressure ratios of 2:1-4:1 per stage imply turbomachin- fry ip Mach numbers that ar in the high subsonic or supe tonic range. Aso chords a the order ofa milimeter imply ‘hata device wih room temperature inflow, such sa compres sor, wil operate a Reyuoids ures he tens of thousands ‘Wit higher gas temperanres, turbines of similar size wll oper. ate at Reynolds numberof few thousand. These ae small "ales compared to the 10810 range of lige sale turboms Chinery and viscous loses will be concomitzatly ager. But ‘scous losses make up only about tid ofthe ttl id Hoss imabighspeed wrbomachine (3D, tpleakage, and shock wave losses account forthe rest) o thatthe decreas in machie ef ciency with size isnt so dramatic. “The increased viscous forces also mean that Mud dag in small gaps and on rotating disks wil be relatively higher. Un Tess gar ow pastages are stller thas oaemtoa, the Maid Iehavior can be represented as cominsum flow so that Knudsen umber considerations are not important Heat wanfer is another aspect of fid mechanics in which ricro devices operate in a diferent design space than large scale mechines, ‘The fid temperatures and velocities ae the ame but the viscous forces are larger so thatthe heat transfer ‘oeficints ar higher, by a factor of about 3. Notonyis thre ‘more beat ansfer oo from the structure bt thermal conduc: tivity within the strotureis higher dueto the short eng scale "Thus, temperature gradients within the structure are reduced, ‘Tis ie bepful in reducing thermal stress but makes thermal ‘solation challenging. Fabcation Considerations ‘Compared to manufactarng technologies familie at arge seal, euret mirofabrication techaoiogy i quite constrained ine geometries that canbe produce. The primary fabrication tools etching of photlthographicaly-efined planar geom- tries, The resultant shapes are min prismatic or "extruded" (yon etal 198). Concepaly 3D shapes canbe consticted of muiplepecsion-aliged 2-D layers. To this end, i wafers can te stacked and dffusion-ooded with bond streagh ap- poaching that ofthe native material (Mirza and Ayon, 1959), ‘Butlayeringis expensive with current echnology and 10 son sidered a arg numberof precision aligned layers fora mio device (wafer can cure be aligned about! micron). Thus $5 rotting machine gometes are dificult o realize so that planar geometries sre prefered. While 3D shapes are dificult, in-plane 2-D geometric complexity i essentially ein manu: facture since photolithography and etching proces an entize wafer at ope tne (wafes range fom 100 to 300 mm in dam ‘ter and may contain doaeas of hundreds of devices). These are ‘ouch diferent manufacturing consrains than common to the Terge-sale world soit is ot surprising the optimal machine de sgn may also be diferent. “Two-dimensional is ot ripping constant onthe de sign of igh sped rotating machinery. Figure 3is an image of ‘ma rotor diameter, radial inflow turbine designed to produce 60 wats of mechanical power at tip sped of 500 ms (Lin et ‘al, 1998). The aso eight is 200 micros. The clinical ftructre inthe cete a thrust pd for an axial si bearing. "The circumferential gap between th ote and stator las i 8 1S micron wide air journal bearing required to suppor teal loads, Te tiling edge of th rotor blades are 25 ncroos thick (aniform to witha 0 microns) andthe blade roots have 10 mi fron ras fillets for sess rele. This Si structure was pro- ced by deep reactive fon etching (DRIE). While te ais ‘sppear planar inthe gue, they ae acmally slighly tapered Figure 3: 4mm rotor diameter radial inflow tarbine.Tusine Tuttne edhust — Canterne Name "Baer See, from bub to tp. Curent technology can yet a taper unifor- sity of sbout 30-50: wih ether a postive or negative slope. ‘The constraints on airfoil heights ae th ech rate about 3 mi- rons per minute) and centrifugal bending stress at the blade Shae” ‘have bee produced ‘The effort described herein has been focussed on smiccomachinery which are prodaced with semiconductor fabri- cation technology (MEMS). Other manfacturingtecniques ‘may be feasible as well especialy the device sie grows into the cemimeter range. The MEMS approach was chosen here ‘because itis intisialybigh recision and pralie preuctin, ‘offering the promise of very low eosin large quay produe- Yio, Tatil estimates suggest that the cost per unit power might timely approach that of large gas bine engines. GAS TURBINE ENGINE ‘Considerations such ss those discuss above led in 1996 ‘othe proliminary or “baseline” gas turbine engine design lus teuedin Figure 4. The cm diameter engine isa single spot mangement with acenifglcompresor nd radia inion tar bine, separated bya hollow shaft or ‘wastouseahigh turbine inlet temperature to schieve acceptable work per units low, recognizing that componeateficlences ‘would be relatively low an parasite losses high. With a4 mm foto danee, the unit was sized to pomp 0.15 gram/sec of air ‘and produce 10-20 watts of power at 2.4 li sine is consucted from 8 ‘Theturbine wafer was assumed tobe SiC. This design served as baseline forthe esearch in composenttecbuologies described (One primary goal ofthe project isto show tat « MEMS: based gas turbines indeed posible, by demonstating benchiop ‘operation of such 8 device. This implies thay, fora fst demon- ratio, it would be expedient to trae engine performance for simplicg, expecially fabricaion simplicity. By 1998, te req- site ecologies were judge suficienly advanced to begin bulcing such an engine withthe exception of fbr fine cyce (.e. produce net power) withthe vantage of simplicity and the great disadvantage of lowering {he presare ratio ofthe DoW oon-adiabatic compresoe from 41 to 21 with «concomitant decrease in eyee power ouput end efcincy, This expedient arrangemeat is refered to sth Hp demo engine. Ii a gas genertoturbojt designed wit the objetivo demonstrating he concep of MEMS gas bine. Tedoee not contin electrical machinery. "The Hy demo engine design is shown in Figure 5. The compressor and trbine rotor diameters re 8 mm ard 6 mm respectively oce the turbine does nt extract power to drive generator its size and thus its cooing lad cond be reduce). ‘The compressor discharge ar wraps around the outside ofthe ‘combustorto cool he combastor walls capturing the wast het tnd soincreasng the combustor ficiency and reducing the ex. teal pctageemperate,The ors suport noun Goring ‘COMPONENT TECHNOLOGIES ‘Giver the overview ofthe system design eguirements ou lined above, the following sections discuss technical consider ation of the component echaologies. For each component, he ‘overtiding design objectiveis to devise a geometry whic yields the performance required by the cycle while being consistent ‘with eater realizable microfabrcai Figure S: Hy demo engine with sion, cooled turbine.rons to hundreds of micros, instrumentation eannot be pu Chased and thea instal, rather it must be fabricated into the ‘evice fom the tart. While technically posible, this approach ‘an eaily double the complexity of the microfabrication, and these devices are alresdy 0 he edge ofthe state-of shear. To txpeite the development proces therefore, whenever possible ‘development was dove in supersale igs, rigs lage enough for ‘convertonl instrumentation, Bearings "a mast be ‘nal operation ti loud i simply the weight of the rotor times the actcerations imposed (g's for aieraf engines). Ifa small, ‘evi is droppedonabard loo rom twomeers, several thw sand gs are impulsively applied. An aiiona reqirement for portable equipments that the support be independent of device Sresation, The bearings and any asscited equipment must ‘so be compatible with the miro device's environment, high temperate inthe ease ofthe ga tarbne engine. Previous (MEMS rotating machines have bees mainly micromotors tm ing at signfcanty lower speeds than of interest ere and so rating fr Limited Se desire for implemented wit ther rmagncticor ec elds providing the oor supportforce, Mag- neti bearings hve two disadvantages for this pplication. Fst, ‘magnetic mails arena comple with mest microfabrication technologies, limiting device fbrcation options. Second, C te poinsconideratias limit the temperatures at which mag ‘etc designs can operat. Slace these temperature ae below those encountered inthe miro-ga turbine, cooling would be require, For these reason, effort was frst concentrated onde ‘gos employing electe elds. These designs examined did ‘ot appear promising in hate forces produced were marginal Figure 6: Gas bearing geometry and nomendature. The ‘ap, erelly exaggerated in this figure. compared tthe bearing loads expected (Miranda, 1997). Also, sinoe electromagnetic Dearings are unstable, feedback stabliza- tions needed, adding to system complet. ppor ther load on tin layers of pressur- i sr pessre supplied from an external sore, the bering is known as ydrstatc. Whe ai pressure is de- ‘ive from the mation of the rou, then the design is hydady- ‘namic, Bybid implementations combining aspects of both ap- proaches are alto posibe-Since the micromachines in qustion Include aireompressore, both desigas ae applicable. tbe ap- ‘proach can reaiy support the lads of machine i his size Tange and canbe used on high temperature devices. All elie being te same, therelativelead-bearng capability ofan et bear ing improves as siz decreases since the surface areato-volume rato Cand thus the inerial os) scales inversely with size. Ro- {or and bearing éynacs scaling i more complex, however “The simplest journal baring is acylnaical rotor within a _lose-fiting creas joual Figure 6). This geometry was ‘opted ist asthe estet to mirofabrcate. Other, more com plex variations might include wave bearings and fil bearings "The relevant physical parameses determining the bearing be- havior are the lengt:-¢0diameter ratio (LD); the gap between the rotor and journal rtioed tothe rotor radius (e/R), and rnondimensional forms ofthe peripheral Mach number of the otor a measure of compressible Reynolds number, snd the mass ofthe rotor Fora bearing supported on a hydrody- namic film, the lad bearing capability scales inversely with (AR) which ends o dominate the design coesiderations. Load ‘earng alo sales with LD (Pies eta, 1997). “The design space available for the mirortating mach ‘exyis constrained ty manufacturing copay. Weave chosen to fabricate the otr and ouralstrvture tthe same ime to facilitate low ost, volume manafctring. The most important constraint i te eching of verical side walls. By pushing the imitation of published etching techaology, we have been able to achive taper ais of abou 30-1-S0:1 on narow etched ver- tical chanel for chanel depts of 30-500 microns as show igure 7 (Lin etal, 1999). This capably defines the bear 3g length while the (oper rato delimit the Baring gap, c. TO ‘Figure 7: Narrow trenches can be cched to serve as “journal bearings.Ae Exhaust Ac” Instrumentation Tee Pont ‘a Figure Sa: Exploded view of ve layers comprising the turbine bearing rig. rinimize gaadus the bearing shoud be onthe lagest dams ‘eer sible, the periphery of the rear. The penalty forthe high diameters relatively high rea and surface speed (has bear ‘ng drag) and low 1D (erefore reduced stability). Inthe radial turbine showa i Figure 3, the journal bearing is 300 microns long with an L/D of 0.075, fof 0.0, an peripheral Mach number of 1. Ths relatively shot, wideapped, high speed ‘bearings wel outside the range of analytical and experimental resis reported inthe gas bearing erature. ‘Stblity san important consideration for al highspeed rotating machines. When centered, hyrodyoamic bearings are ‘nstable, especially at low rotational speed. Camsmony, such ‘bearings re sailed by the application of ausiier ona force ‘whic pushes the rotor foward he journal wall, as measured by the ecentricty, the minimum approach dance of te rotor the wal faction ofthe average gap (0= centered, 1 = wall ste), Atcoavestonal scale, therotor weights often the source ofthis side force. At micro scale (1) the rotor weights nepi> gible, and 2 insensitivity to entation is desirable, so we Rave ‘opted a scheme which ues dicen gas pressure o force the rotor eccentric. Extensive numerical modeling of these ‘mictobeaing flows have shown that the rotor wil be stable at eccensicies above 08:09 (Pickos and Brent, 1998), Forthe geometry ofthe trbine ia Figure 3, the rotor mst thes operate between -2:microa from he jourtl val (Picks ral, 1997). “This implies that deviations from icaariy ofthe journal and oor mast be small compared t0 Irion. ‘Totes these dos, two geometrically similar turbine bear “Treustbearing Tippy pont Forwars ensue tnyat bearing Figure 8b: Fivelayer miroturbine bearing rig with ‘4mm da rotor ing tet igs have been built and tested sng the sume bearing ‘omer, one a micro sile with a4 mm diameter rotor and the ‘thera micro scale unit 26 times lager The macro version was tetensively instrumented for pressure and retr mation meas tment (Or, 1955). The micromrbine bearing test rg shown in Fire 8, consist of five sack lyes, each fabricated from a single Si wafer (Line al, 1999), The ceter wafer ste radia ‘now turbine of Figure 3, with 4,200 micron dames, 300 mi- con thik roto. Thetrbne roto isa pall sded disk with Plates cantilevered ffom one sige, While such a simple designs viable in slicon above SOD rs, the cena stresses are 100 igh for met without tapering of the disk (othe macro version slime 400m). Te wafers on eter side contain the st. ‘bearings nd plumbing forte side prestrizatin noded to p= atethe oto eccenically. The ouside wafers contain the ins, thant and vot oles. Int est device tots bearings ae byétsti, pressurized by eteral ai andthe jurmal bearing canoperis neither hydrodynamic or hydrostatic mode. Figure 9 {dna taken fom an optical speed sesor during hyo beer- ing operation. ‘Turbomachinery Fuld Mechanics Tr many ways the fluid mechanics of micotrbomachinery ae similar to that of lage seale machines, for example, high Froquency (Ha) Figure 9: Speed data from microturbine at 1.2M rpm.yo go® vat Normalized Total Pressuce Loss 34a low compressor and turbine as estimated witha 22D CFD code Gacobson, 198). "The design chalegesinsoducedby the low Reynosa ters ar exacted by geometric restcions imposed by eur teat micrfubcation technology. In pariculr, constant pas- ‘ageheighsaproblem in these highspeed designs. High work ‘nthe Mui means lage density changes. In conventional cea- ‘ofa turomachiney, Mud ent chaoge examina. cong Ur epee) aces tne) he Inightaite los pat: However cureat icoabicatice ec logy ot woes ping Fes igi Eom ED (which ad oss to the rot) ‘spproaces are being pursued in paale! ‘ere. Note thay although (eich is ot permited by mi tree eae over ld desig chlenge mig ean example, a the 2 mm diameter {nletn acompressorimpelle,3-D CFD simulations show that right angle tun cost $3 in compressor efficiency and 15% in ‘ass low compared io a smooth tun (Mehra et a, 198). Ea- sneering approaches to this probleminciue lowering the Mach umber athe ras (by increasing he flow area), smocthing the tums with steps or angles (which ads significant fabriation complexity), and adding externally. prodoeed contoured pars ven the tars ar tthe inlet o ote to he chip. ‘While extensive 2-D and 3-D numerical simulations have ‘been used tohepin the design and analysis ofthe micromachines, in ll high speed trbomachinery development, test data is needed. AS an aid to detailed flow measurement, 2 75 times linear sale model of «400 ms tip speed, nominally 2:1 pres sire rato, 4 nm diameter eompressor was bait and operated at reduced inlet pressure to match the design Reynolds number of ‘out 20,000 (Shire, 1998). A comparison of CED simalation to dataisshown in Figure 12. The simelation correctly pedis Combustion a Se eee them can be. Combustion requires the mixing of fel and air fol = emypskion sedienre Kore sane D eeivenny (ae)22. 20. Pressure Ratio 2 Square tne, Experiment 12] —2— Smooth no, Experiment wibes, 20 09 Denes, 30 OFD a 08 Corrected Mass Flow (traction of design) oe roa ‘Figure 12: Data nd simulations ona 400 stip speed ‘compressor at Re= 2,000. ime can sale with, ‘evice sie Dut the chemical reaction tims do not (peally Iixing soc for more thn 0% of omar sien tine) “Tus, thecombustr volume sa rat acon ofa miroeasie tan large egin, bya factor of abou 40 forthe devs dx odate Anche difference beweei IgE aHT ao scale ‘machines i increased surface areao-wolume ratio at mall Sat offs mar area for exalt "he design dts re dependent on the fel chosen “The disadvantage to this approach sa suscepbily to ast back from the eomaastor into the premix zose, which must be void. To expedite the demonstration ofa micro-gas turbine feaginc, hydrogen was chosen 2s the initial fuel because of ts ‘vide ammabilty its and fast reaction time is ste same sppreach taken by von Ohain when developing the first jet e- ‘ine in Genmany inthe 1930). Specifically, tyrogen wil aa _tegeivalence rats as low a 03 which yields adiabatic eon bustiontemperatres blow 1500 K. Hydrocarbon fucls must ‘eoperte lotro stochionetic and therefore at higher te- ‘peratures, above 2000 K. The reduced het load from the low temperature combustion, combined with the high thermal coe: ‘uci of scon, means tht silicon (which mets 1600 K) {sa viblesouctral material foe aH combustor (Waite al, 1995; Metra ea, 1999). ‘Silicon combustor: have been built wich duplicate the ge- ‘omety ofthe engines in Figures 4 and , with tb tosaing pas replaced with stationary sil vanes (Metra and Wat, 1998) spectively Fig ues designs tae vantage of micrfarication's Combustion ‘Chamber Fame Hotcers Turtine Newaae Figure 13: Mlerocombustor which comprises the tatie structure ofthe demo engine of Figure S. bility to produce similar geometiofeanres simultaneously. For exarpl, the larger combustor has 90 fue njetion pots, sch 120 microns in dlamete, to promote uniform fue-r mixing. “The smaller combustor operating a power lve of 200 wats is stow niga along with aCFD simulation of the oil ‘These ters demonstrated that conductivlj-cooled silicon tr- bin vanes cn servive 1800 K gas temperate forbs with lide degradation (Figue 15). Measurements have shown that these mcrocombustors can schieveeciacies in the mid 90% range (Metra eal, 19998), Data various equivalence ratios (@) are shown in Figure 16. (The gaps inthe Hines are due 10 ‘hemmocoaples failing at high temperamre,) Ignition bas not proven a problem: simple hot wire ignitor has prove sf ‘leat, even at room pressure (Meta, 2000), ‘Hydrocarbons are mere dificult o burn. Initial tests show that the existing combustors can bam ethylene an propane, but a reduced efficiencies since the residence times are too short, Figure M4: 200 watt mlcrocombustor.ae \ ae buat ‘rer tee he ha forcompete combustion. Other approaches being pursed here incide stoichiometric 2one-dltion schome similar to con ‘ventional a urine combustors and catalytic combustion. lecrial Machinery Microeletical machinery i required for power genera tion and asa prime mover fra starter or various pumps and compressors. There i an extensive iterature on microeecc motors, wich will a be reviewed bere, bu itl work on ge- ‘ators, The requirements forthe devies of interest ere fer fiom previous work i that te power denses needed are Teast two orders of magnitude above those repoted in the itera- tureta date, Alo, th thermal envirmeatisgeerlly harsher. Inegrating the elec machi within the device gn ubine, compressor, ee) offers the advantage of mechanical simplicity in than additional barags or structures ae required over that needed forth uid machine. Alo, theres supply of cooing ‘ad avalable As in the case of elec bearings, both elecostatic and magnetic machine designs canbe considered and fist ode, both approaches ean yield about equivalent power densities. Since the magnetic mochies ae material property mie st high temperature and because of the challeages of rmirofebresing magnetic materials, elecwostatic (commoaly refered to cleo) designs wee fr examined. Power dew 2009 at gas temperature (K)_ cee Cr a Air Mass low (glee) Figure 16: Mierocombustor performance with hydrogen fuel as function of equivalence ratio (). sity seaes with cece field strength (torque), requeney, and Fotaional speed. The micromachinry of interest here operates st peripheral velocies 1-2 orders of magnitude higher than pre- ‘ioasly reported micrometers, ano yields concomitantly more power, Elecvic machines may be configured in many ways. Here a induction design was chosen (Bat and Lang, 1985), “The operation ofan electric induction machine ean be un derstood with reteence to Figure 17 (Nagle and Lang, 1955) "The machine coasts of two components, a rotor and a stator. “The rotor is comprised of 2520 um thick good insulator ov cred with afew microns of «poor condocce (200 MO sheet resstviy). Testator consists of set of conductive radial ele trodes sported by a goed insulsfor. A moving electric elds. Jmposed onthe stor electrodes wih the ad of externa lee: tronics, Te stator fed then imposes a charge disibution on the rotor. Depending on the relative phase becween the motion of rotor eharges (et by the rotor mechanical speed) and that of the stator fld the machine can operate a a motor, generator, forbrake, Torque increases with eletic fc strength and fe (quency. The maximum electc field swength hat an ar gap ean antain without breakdown i funtion ofthe gop dimension Thai itis maximum ata few microns so that micromacines can potentially realize bigher power density tha age machines TSI en as eee Fioreaion Figure 17b:Eletric induction conceptual layout of motor ‘at expanded vertical seale.Phase teronnect “ue tar (er ean ole paces) Stator eleeroce Figure 18: High power density electric stator structure. ofthe same design. Frequency is contained by external elee- ‘woncs design and by faccation constants on he stator elec ‘uote geometry. Weare exlrig the design space centred about 100-300 vols and 1-3 Mila, Thi is consistent wih a machine power of about 10 watte with a4 um diameter and 23 ym sit fap. A si-phase,131-poe (786 electrodes) stator for such a ‘machin is showa in Figure 18 (Ghods ea, 1959), OTHER APPLICATIONS ‘While the fecus of this papers ben microfabricated mi- cro-gu turbine engines, ther MEMS devices which us the same fundamental high speed rating machinery technology and de- sign approach ar also underdevelopment. One such machine Js a miroir turbine generator. This is basically the wibine- betrngrig of Figures 3and& witha generator on the sie ofthe ‘isk oppoite the turbine blades. Ik was designed a a testbed for component development but could serve as an electric gea- erator runing from asuppy of compressed ga (With let tem parures upto about 900K). Another variants a mtor-iven air compressorflower (Figure 19). In ths eas, the turbine blading is replaced by compressor blading and the electric ma chine on the opposite side of the rotor rans a8 @ molar. 15 designed to pressurize smal fel els nd appt scientific struments Iecould also serve asthe compressor in aeigert- tio os for the cooling of electrons, sates, or people Ph Sark Figure 19: Mlcromotor-compressr test rg, igure 20: Mierofabreated 15 Nt thrust, bipropeliant, ‘cooled rocket engine. pn 2. Peay Sarre cooled thrust chamber promising. A complete engine system woul incade integrated ‘urbopumgs and conto, (CLOSING REMARKS 1 now appears tha microfabricated high speed rotating machinery is Feasible. Such micromachies are the enabling tecinology for micro-hest engines such a gs turbines, Rank: ine cycle, and rockat engines. These devices will id ahost of plication a energy conversion and power prodsctio, eool- 5 and heat pumps, snd propulsion for bo micro and macro ‘eels, Furthermore, the components - pumps, compressors, turbines, motors, generators ~ ae themselves useful fora host of fli handling, eansduer, and energy sytem applications. ‘One lest leave to dat stat conventional engineering wisdom doesnot necessurily apply to micromachines. Dilfer ‘at physical epmes demand diferet design solutions. Ditfer fat manufacturing constraints les to difereat coafgurations. [Not necessarily inferior, just ferent. Te fist devices mate are cre snd have low performance compared to their more ‘amir large sized been, jas as pas turbines did when they were fst invoduced. It sems likely, however, that the eco ‘omic pote of high power density micronsehineywillspur the investment in research and development nosed to greatly evolve the perfomance ofthe devices. ACKNOWLEDGEMENTS “Tae work summarize her the atellecsl accomplish rmentof. team of eurentand former researchers at MIT includ. ing:G. Ananthasresh, A. yoo, K. Breuer J Bisson, C.Cdon, DA Chen, K-S Chen, A. Deux, M.Drela,F Etch, A. Epstein, E.Exteve, L.Frecete, G. Gaba, R. Gods, ¥. Goag, C. Grosheny, T Hanison, E. Huang, $. Jacobson, K. Khan, R Khanna. Lang, C. Lis, C. Lie, . Liverere, K. Laine, A London, S.Lukachk, A. Mebra,B, Mill, J Miranda, . Nagle,. Orr, B, Philipgon, E.Pickos, J. Prot, L.Rethefor, N. Sevoulds, M. Schmit, S. Sentra, G. Shi, M. Sparing. Sullivan, Takacs, C. Tan, D. Tang, -Y Tezog, S. Umas, Walker, D. Wats, P Waren, 1. Wait, C. Wong, X. Yang, W Ye,X. Zhang “The works supported by the US Army Research Office, Drs, R Paut and T. Doligalski, technical managers and by DARPA, Des. Nowak, D- Fils and 8. Wilson, program man ees. ‘REFERENCES, ‘Aydo, AA, Lin, CC, Braff, R, Bay, R., Sawin, HEL Schmid, M, 1998, “Etching Characters and Profile Con trol in a Time Mulplxed Inductively Coupled Plasma Eicher” presented at 1998 Solid Sate Sensors and Actuator Workshop, Hilton Head, SC. Bart, SF and Lang, UHL, 1989, “An Analysis of ecroquasistat Induction Micromosurs Sensors and Actua tors, Nol 20, pp. 97-106. Chen, KS, Ayoa, A.A. Speatag, S. M, 1998, “Silicon ‘Seagth Testing foe Mesoscale Sctral Applications", MRS Proceedings, Vo S18,pp- 123-130. ‘pein, AH, Sentra S.D., 1997, "Macro Power ftom -Mioro Machinery” Science, Vl. 276, p 1211. ‘Epstein, A. et al, 1997, "Mero Heat Egines, Gas Tur vines, an Rocket Engines", AIAA 97-1773, presented at 28th [AIAA Fluid Dynamics Conference, 4th AIAA Shea: Flow Coa- ‘tol Conference, Jane 29-aly 2, Snowmass Vilage, CO. Ghodes,R, Freche LG, Nagle, SF, Zhang, X. Avon, AA. Seaturia, SD. Schmit, M.A., 1998, “Thick Buried Ox ide in Sitoon (TROS) An Integrated Favicaton Techaology for Mult-Stack Wafer-Bonded MEMS Processes” preseted at ‘Transducers 99, Sead, opus Taeebson, S.A. 1998, "Acrothermal Challenges inthe De sign ofa Microabicated Gas Turbine Engine”, IAA 98-2545, presented at 29 AIAA Fld Dysanis Conference, Albaguer que, Lin, CC, Gods R, Avon, A.A. Chen, DZ, Jacobson, S. Breuer, KS, Epsteia, AH, Schmid M.A, 1999, "Fabics- tion and Chareterization of a Micro Trbne/Bearng Rig" pre ‘ened at MEMS 99, Orlando, FL. ‘Metra, A, 2000, "Developmeat of a High Power Density ‘Combustion System for Silicon Micro Gas Turbine Engine.” PhD. Thesis, MIT Department of Aeronautics and Astonat- ‘Meira, A, and Wait, I-A, 1998, “Development of Hy- drogen Combustor fora Microfaricate Gas Turbine Engine presented t Solid-State Seasor and Actuator Werkshop ion Head Ilan, SC. ‘Metra, A. Jaobson, S.A, Tan C.S, Epstein, AH. 1998, "Aeroyaanic Design Considerations forthe Turbomachinery of Micro Gas Turbine Engine”, preseated at the 25° National ‘ad I"Totemational Conference en Pid Mechanics and Powes, [Now Det, Inia. ‘Mea, A. Ayon, A.A, Wait, LA, Schmidt, M.A. 19993, -Micrfuriaton of High Temperature Silicon Devices Using Wafer Bonding and Deep Reacive Ton Etching” Journal of Microelecrmchanlal Systems, Mich. Metra, A, Whit, LA. Schmit, M.A, 1990, “Combs tion Tests in the Static Structure of a6-Wafer Micro Gas Tur bine Engine, presented a Transducers 99, Senda epan, Mirza, AR, Aya, A.A, 1989, “Silicon Wafer Bonding “The Key Enabling Technology for MEMS High-Volume Mant factring?” Faure Fab Inernationa, Issue 6, pp. 51-56. "Mr Mitend, 3.0, 1997, "Feasibility of Electrostatic Bea ings for Micro Turbo Machinery” M.Eng, Tess, MIT Depa ‘ment of Bestel Engineering and Compute Science. ‘Nagle, SE. and Lang LF, 199, "A Miero-Scale Electric. Induction Machine for 8 Mico Gas Turbine Generate” pre ‘seated the Elecstatc Society of America 27h Anal Con ference, June 2-25, Boston, MA. (Orr. D., 1998, "Maero-Scale lvestigaton of High Speed Gas Bearings for MEMS Devices" PuD. Thesis, MIT Depat- sent of Acronatcs and Astonauts, Piekos, ES, Breuer, KS. 1998, “Pseudospectal Orit Simulation of Non-Ideal Gat Lubricated Joural Bearings for Microfabricsted Tuebomachines" Paper No. 98-rib, pe seated atthe ToiatASMBISTLE Tribology Conference, Toronto, (Canada. Alo, to appear in Journal of Tribology. iekos,£.S, Or DJ, Jacobson, S.A. Each, FE, Breve, KS, 1997, ‘Design and Analysis of Mcrofaricatd High Speed Gas Tourn Bscings” AIAA Paper 97-1966, presented at 25th [AIAA Fluid Dyoamis Conference, June 29-July 2, Snowmass Village, CO. ‘Shi G. 1998,"An Experimental Investigation of Low ‘Reynolds Number, High Mach Number Centrifugal Cmpres ‘sor? MS. Thess, MIT Department of Aeroautes and Astro Speaing, S. M, Chon, KS. 1997, "Miro-Gas Turbine ‘Eng Materials ad Statuses, presented at 21* Annual Co- ‘oa Beech Conference and Exposition on Composite, Advanced (Ceramics, Materials and Seucures ‘Wait, LA, Gatba, G, Tzeng, YS, 1998, “Combustors for Mico-Gas Turbine Engibes” ASME Jourat of Fuld Eng- neering, ol. 120