Glimpses of MEMS. G. K. Ananthasuresh. Mechanical Engineering Indian Institute of Science Bangalore, INDIA

Transcription

1 Glimpses of MEMS Presented to ME237 class in Jan.-Apr. term of 2005 G. K. Ananthasuresh Mechanical Engineering Indian Institute of Science Bangalore, INDIA

2 2of 80 Any sufficiently advanced technology is indistinguishable from magic. Arthur C. Clark

3 What s in a name? 3of 80 Micro-Electro-Mechanical Systems (MEMS) Widely used in Americas. MicroSystems Technology (MST) Popular in Europe. Micromachines Used in Japan. Microscience Some people prefer to call it this way as they begin to explore scientific aspects of MEMS. NEMS When a feature becomes less than a micron and is respectably measured in nanometers.

4 What? How? Why? 4of 80 What are they? How small are they? How are they useful? How do they work? What are they made of? How are they made? How do you design them? What are the modeling and design issues and challenges for us?

5 Why MEMS? 5of 80 M C M $ Because C M $ MEMS Device and Application 1996 Sales (Millions of $$) 2003 Sales (Millions of $$) Inertial measurement: Accelerometers and gyros Microfluidics: Ink jet printers, mass flow, lab chip Optics: Optical switches, displays Pressure measurement: Automotive, medical, industrial RF Devices: Cell phones, devices for radar none Other Devices: Microrelays, sensors, disk heads Source: Systems Planning Corporation, 1999

6 Where is MEMS field today? 6of 80 What does this mean? 1. MEMS field is somewhat mature because people are asking routine questions about how to design them. 2. There must be something different about designing them. 3. There is a need to learn the basics of the field and be familiar with the jargon of the field. And so we begin

7 Standards for MEMS are emerging 7of 80 3 ASTM test methods in the round-robin testing stage. (from NIST National Institute of Standards and Technology)

8 What are they? 8of 80 MEMS are systems that integrate sensing actuation computation control communication power They are smaller more functional faster less power-consuming and cheaper!

9 How small are they? 9of 80 Nanotechnology Microsystems Meso Macrosystems 10 nm 1 um 100 um 10 mm 1 m 0 A 1 nm 0.1 um 10 um 1 mm 100 mm 10 m Atoms Molecules DNA Nanostructures Virus Smallest microelectronic features Bacteria Biological cells Dust particles Dia. of human hair MEMS Optical fibers Packaged ICs Packaged MEMS Lab-on-a-chip Plain old machines Humans Animals Plamts Planes, trains, and automobiles Micro-machining Nano-machining Macro-machining Precision machining

10 Why is it exciting? 10 of 80 Money matters aside Wavelength of visible light 400nm 700nm Micro-technology brings engineering to the size scale of the workshops of the biological world. The motions of micro-mechanical devices overlap with the wavelength of the visible light and thus allowing us to play with light in interesting and useful ways.

11 Why are they small? 11 of 80 Micro size is almost incidental. They are small because of the technologies used to make them. And it is economical to make them small when made in large volumes just like microelectronics. Of course, there are some MEMS devices that would not work if they are any bigger.

12 A bit of history 12 of 80 There is plenty of room at the bottom - A 1959 lecture by Richard Feynman Pioneered by Professor James Angell at Stanford University, researchers at Westinghouse in late 1960 s into 1970 s Infinitesimal Machinery - A 1983 lecture by Richard Feynman Formal identity ( MEMS ) to the field came in late 1980 s

13 Ink-jet printer head MEMS devices in 70 s 13 of 80 Bassous E., Taub H.H., Kuhn L. Ink jet printing nozzle arrays etched in silicon Appl. Phys. Lett. 31, 135 (1977) Micro mirrors for steering light Petersen K.E. "Micromechanical light modulator array fabricated on silicon" Appl. Phys. Lett. 31, (1977) Petersen K.E. Silicon torsional scanning mirror IBM J. Res. Dev. 24, (1980) Accelerometer Roylance L.M., Angell J.B. A batch fabricated silicon accelerometer IEEE Trans. on Electron Devices 26, (1979) Optical fiber connector Schroeder C.M. "Accurate silicon spacer chips for an optical fiber cable connector" Bell. Syst. Tech. J. 57, (1977) Microfluidic device Terry S.C., Jerman J.H., Angell J.B. A gas chromatograph air analyzer fabricated on a silicon wafer IEEE Trans on Electron Devices 26, (1979)

14 MEMS devices in late 70 s and early 80 s 14 of 80 Pressure sensors Clark S.K., Wise K.D. Pressure sensitivity in anisotropically etched thin diaphragm pressure sensors IEEE Trans. on Electron Devices TED-26, (1979) Ko W.-H., Hynecek J., Boettcher S.F. Development of a miniature pressure transducer for biomedical applications IEEE Trans. on Electron Devices T-ED26, (1979) Other types of sensors Kimura K. Microheater and microbolometer using microbridge of SiO2 film on silicon Elect. Lett. 17, (1981) Najafi K., Wise K.D., Mochizuki T. A high-yield IC-compatible multichannel recording array IEEE Trans on Electron Devices 32, (1985) Stemme G. A monolithic gas flow sensor with polyimide as thermal insulator IEEE Trans. on Electron Devices TED-33, (1986) Optical switching and multiplexing Gustafsson K., Hök B. Fiberoptic switching and multiplexing with a micromechanical scanning mirror Proc. 4th Int. Conf. on Solid-State Sensors and Actuators, Tokyo, June 3-5, P 212 (1987)

15 What is common to all of them? 15 of 80 A beam or a diaphragm A bulk-micromachined silicon, glass, etc. Electrical and electronic components for sensing a signal Micro-electro-mechanical systems (MEMS)

16 A 1979 micro-accelorometer 16 of 80 Top view Bulk micro machined; Piezoresistor -based sensing Side view Roylance L.M., Angell J.B. A batch fabricated silicon accelerometer IEEE Trans. on Electron Devices 26, (1979)

17 Widespread attention came with a micromotor 17 of 80 The excitement began only after a rotary motor, revolute (pin) joints, and prismatic (sliding) joints were demonstrated. At U. C. Berkeley, MIT, and Bell Labs The reason for the excitement was batchfabrication of assembled micro-mechanisms without assembly. Crucial development: sacrificial layer process using polysilicon as the structural layer.

18 Electrostatic micromotor 18 of 80 Sacrificial layer process to make a revolute joint MUMPs process (MCNC) Ravi Jain, undergraduate at Penn.

19 Early references on micromotors 19 of 80 Gears M. Mehregany, K.J. Gabriel, and W.S.N. Trimmer, "Micro Gears and Turbines Etched from Silicon," Sensors and Actuators, vol. 12, pp , Nov./Dec Revolute joints and linkages L.S. Fan, Y.C. Tai, R.S. Muller, "Integrated Movable Micromechanical Structures for Sensors and Actuators," IEEE Trans. on Electron Devices, Vol. ED-35, No. 6, pp , June M. Mehregany, K.J. Gabriel, and W.S.N. Trimmer, "Fabrication of Integrated Polysilicon Mechanisms," IEEE Trans. Electron Devices, vol. ED-35, no. 6, pp , June Micro rotary motors Y.C. Tai and R.S. Muller, "IC-processed Electrostatic Synchronous Motor," Sensors and Actuators, Vol. 20, No. 1&2, pp , Nov. 15, M. Mehregany, S.F. Bart, L.S. Tavrow, J.H. Lang, S.D. Senturia, and M.F. Schlecht, "A Study of Three Microfabricated Variable-Capacitance Motors," Sensors and Actuators, vol. A21 A23, pp , 1990.

20 What (more) are they? 20 of 80 Early on Solid state transducers And later Integrated systems MEMS sensors actuators are batch fabricated are economical have more functionality involve physical, chemical, biochemical phenomena at small scales act upon macro scale too Take leverage of the enormously successful VLSI technology

21 How are they useful? 21 of 80 Commercial successes Pressure sensors (Motorola and several others) Accelerometers (Analog Devices, Delphi, Motorola) Ink-jet printer heads (HP) Projection display with micro mirror array (TI) Portable clinical analyzers (Abbott) etc. Movable solids and fluids at microscale made possible lots and lots of sensors and actuators.

22 Portable Clinical Analyzer (PCA) 22 of 80 The i-stat Portable Clinical Analyzer used in conjunction with disposable cartridges for determination of a variety of parameters in whole blood. stores up to 50 patient records permits on-screen viewing of test results transmission of records to a data management system using infrared signals. all with couple of drops of blood under a minute at the point of care

23 Opto A remote gas sensor (a MOEMS device) 23 of 80 Source Sample Cell Reference Cell Filter Detector Modulation Source Traditional Correlation Spectroscopy Approach Polychromator: Honeywell-MIT-Sandia project Source Sample Cell Diffractive Element Detector Holographic Correlation Spectrometer θ d Modulation Source Filter Broadband Light In Polychromatic Spectrum Out Deflectable Micromirror Grating Elements C B A D C B A Si Wafer (Source for figures: Honeywell and S. D. Senturia) Individually Addressable Array of Driver Electrodes C

24 One more way to play with light 24 of 80 (a QualComm acquisition) Interference-modulation by electrostatic actuation of vertically moving membranes.

25 Truly digital sound using MEMS 25 of 80 Analog Advantages of DRS: Large dynamic range is not necessary. Digital Nonlinearity can be controlled distortion is minimized. Fault tolerance. Intensity control. Speaklets are combined to produce the sound effect. With low-pass filters, the sound is smoothened. Diamond et al., 2003

26 More applications 26 of 80 Inertial measurement devices Accelerometers, gyroscopes Mass data storage Opto-mechanical devices Projection displays, photonics, optics-on-a-chip Flow control Bio-chemical sensors/actuators lab-on-a-chip, drug delivery, bio-sensors Communication hardware Mechanical filters, RF-switches and relays, wireless MEMS Chemical microreactors ( plant-on-a-chip ) Power MEMS Micro engines, generators

27 Consumer electronic products 27 of 80 Accelerometers: In laptops, PDAs, camcorders, CD, DVD, and MP3 players Microphones, filters, switches, etc.: Cell phones Revenues in 2001: $124.3 M Projected revenues in 2006: $613.5 M (Marlene Bourne)

28 Toys! 28 of 80 Teaching a new dog some new tricks SONY AIBO dogs reportedly use microaccelerometers. Source: SmallTimes 9/10,2003

29 MEMS affecting the macro world 29 of 80 UCLA MEMS creating large effects an example

30 MEMS to power the macro world? 30 of 80 MIT Microengine (Source: Epstein, 2003) Power on a chip Demo engine with H 2 fuel Turbine-compressor test Detail of the DRIEetched blades

31 Outline 31 of 80 What are they? How small are they? How are they useful How do they work? Pressure sensor V Capacitive sensing Piezoresistive sensing

32 How do they work? 32 of 80 Accelerometer Side view V Top view

33 Many ways to do one thing. 33 of 80 Convective accelerometer No moving parts! Dao et al., 1996; Leung et al., 1997.

34 How do they work? 34 of 80 Texas Instrument s digital light processor InFocus digital projectors

35 How do they work? 35 of 80 TI s digital light processor (DLP) torsional beam Tiltable mirror torsional beam actuating electrode 1 actuating electrode 2 (Source:

36 What lies beneath TI s digital light processor (DLP) and deformable mirror display (DMD) 36 of 80 Anatomy of DLP/DMD Ant s leg on the DMD array

37 How do they work? 37 of 80 Lucent/Bell Labs two-axis micro-mirrors optical cross-connects

38 Lucent s optical cross-connect 38 of 80 Optical fibre bundle Micro-mirror array Micro-mirror array Optical fibre bundle

39 How do they work? 39 of 80 Ink-jet printer head paper orifice ejected ink droplet Weight: ng resistive heater drive air bubble Electronics are integrated to trigger the drive bubble

40 How do they work? 40 of 80 A mechanical relay Dielectric Signal output V Signal input

41 How do they work? 41 of 80 A normally closed fluidic valve Trapped fluid Glass Silicon Glass Flow (Redwood Microsystems)

42 Two commercial micro-valves 42 of 80 Redwood Microsystems s thermo-pneumatic normally closed valve A normally open valve;

43 How do they work? 43 of 80 A diaphragm pump Diaphragm V Passive inlet valve Passive outlet valve

44 Micro-cage with cantilevers 44 of 80 C. J. Kim, UCLA Bi-metal cantilevers curled due to residual stress. Opened with actuating the bottom membrane

45 Outline 45 of 80 What are they? How small are they? How are they useful How do they work? What are they made of? How are they made?

46 What are they made of? 46 of 80 Phase 1: Old materials and old processes Silicon, its oxide, nitride, and some metals IC-chip processing technology Lithography Thin film deposition (e.g., chemical vapor deposition CVD) Etching Doping Phase 2: Old materials and new processes Silicon, its oxide, nitride, glass, polysilicon, and some metals IC-chip processing techniques enhanced as micromachining techniques Sacrificial layer process Dissolved wafer process Wafer bonding LIGA Hexil Deep reactive ion etching Etc.

47 What are they made of (contd.) 47 of 80 Phase 3: New materials and old processes Polymers More metals Ceramics Silicon carbide Piezoelectric films Ferroelectric films Shape-memory materials, etc. e.g., PDMS George Whitesides at Harvard Phase 4: New materials and new processes Processes unconventional to the microelectronic field Processes that re-define the size of MEMS micro to meso or nano Deposition and etching for the new materials

48 How are they made? 48 of 80 Surface micromachining Deposition of thin films (mainly polysilicon) Etching using masks Layered construction Bulk-micromachining Carving features into bulk wafers by etching Wafer bonding Patterning individual wafers Wafer-to-wafer bonding LIGA, HEXIL, and other HARM processes DRIE Others: laser, micro EDM, etc.

49 Micromachining is not precision machining! 49 of 80 Precision machining Relative tolerance (feature to part size) is better than For micromachining, it is 10-2 to Roughly what we have for building houses. With micromachining, You can make it small, but not precisely.

50 Surface micromachining 50 of 80 Deposit or grow silicon dioxide Silicon wafer Pattern the oxide using a mask Deposit polysilicon Pattern polysilicon Sacrifice oxide layer by dissolving The sacrificial layer process to make released structures (Berkeley)

51 Isotropic etching Types of etching 51 of 80 With agitation Without agitation Anisotropic etching (111) plane (111) (100) silicon (110) silicon Slanted surfaces

52 Bulk micromachining 52 of 80 Etch using a mask Silicon wafer Boron doping using a mask Flip and bond to a glass Glass Dissolve undoped silicon Boron doped dissolved wafer process (Michigan)

53 Wafer bonding 53 of 80 Etch a cavity in a wafer Bond another wafer Thin down / polish and etch Released cantilever using MIT s wafer bonding process

54 Making an electrostatic micromotor using surface micromachining 54 of 80 Side view Top view After sacrificing oxide layers Rotor Stator poles Cronos MUMPs (formerly MCNC MUMPs)

55 Making a micromotor 55 of 80 Deposit poly0 Etch poly0 Deposit oxide1 Dimples in oxide1 Etch oxide1 Deposit poly1

56 Making a micromotor (contd.) 56 of 80 Etch poly1 Deposit oxide2 Cross-section up to this point Cronos MUMPs (formerly MCNC MUMPs)

57 Making the micromotor (contd.) 57 of 80 Etch oxide2 Deposit poly2 Etch poly2 Deposit and etch metal Cronos MUMPs (formerly MCNC MUMPs) Cross-section before sacrificing oxide layers

58 Finished micromotor 58 of 80

59 Micromotor after release 59 of 80 Cronos MUMPs (formerly MCNC MUMPs)

60 How much should we know about u-fab? 60 of 80 Being able to draw the process flow diagrams from a description. Visualizing a process from a crosssection.

61 Visualize device from a verbal description of the process 61 of 80 Being able to draw the process flow diagrams from a description. Shallow pits were etched into n-type substrates, and p-type deflection electrodes were diffused in the above pits, followed by fusion bonding of a second wafer above the first. The top wafer was then ground and polished down to a thickness of 6 um. A passivation layer was then formed on the top wafer and sensing piezoesistors were formed using ion implantation, after which contact holes for metallization to connect to the diffused deflection electrodes were etched. Bond pads and interconnect metallization were then deposited and patterned, followed by etching of the diaphragm from the back of the wafer. Finally, two slots were etched next to the beam to release it over the buried cavity. (Petersen et al., 1991) See

62 Process flow 62 of 80 Wire bond

63 Verbalize the process steps from a device cross section 63 of 80 verbalizing a process from a cross-section. How was this made? See

64 MEMS Foundries 64 of 80 You don t have to make them if you don t want to or can t. A successful MEMS foundry MUMPs (multi-user MEMS processes) MCNC (RTP) Cronos MUMPs Uniphase MEMSCAP A recent survey shows that there are about 360 commerical MEMS fabrication facilities with an average staff of % in USA, 38% in EU, 12 % in Japan, and 9% in Asia. MEMS is growing at a CAGR of 25%. Source: Yole Development, 2003.

65 Making elements of mechanisms 65 of 80 A surface micromachined hinge (Kris Pister, Berkeley) Substrate hinge

66 Floating hinges 66 of 80 Mask layout

67 Electrostatic comb drive 67 of 80 Misaligned comb capacitors align creating actuation. Folded-beam suspension Moving combs Shuttle mass anchor Fixed combs

68 Sandia s micro mechanical lock 68 of 80 Pin in a maze

69 Close-up of Sandia s micro lock 69 of 80

70 Merry go-around for mites 70 of 80 See Sandia s web site for animation

71 Sandia s SUMMiT process revolute joint 71 of 80 Pin Rotor Substrate

72 Deep etching 72 of 80 Deep RIE (reactive ion etching) to get vertical sidewalls over large depths of several hundred microns. E.g., SCREAM (Cornell) bulk-micromachining

73 SCREAM process Noel MacDonald, Cornell university 73 of 80 Deposit and pattern mask oxide Etch bottom sidewall oxide Deep RIE silicon etch Second deep RIE silicon etch Deposit sidewall oxide Isotropic silicon etch

74 Packaging! 74 of 80 Packaging access to and protection from the external macro world Packaging is a big problem with MEMS. Sometimes, it may be better not integrate sensor/actuator and electronics. Packaging serves Signal redistribution Mechanical support Power distribution Thermal management Fluidic fittings Etc. Some techniques Ball and wire bonding Flip-chip Sandia s process Research continues

75 Chip/package level integration of M and E of MEMS 75 of 80 Sandia s chip-level integrative process Mechanical Electronic Motorola s package-level integration Divide and rule! Unity is strength

76 Outline 76 of 80 What are they? How small are they? How are they useful How do they work? What are they made of? How are they made? How do you design them? What are the modeling and design issues and challenges for us?

77 Why do MEMS devices often pose with US coins? 77 of 80 In God We Trust

78 Modeling and design of MEMS What is different? 78 of 80 Integration of sensor, actuator, mechanism, processor, power, and communication makes system level tasks challenging -- common representation for multiple energy domains Device level too has multiple energy domains -- macromodels Component level -- coupled energy domain equations Mask level -- geometric modeling

79 Modeling and design of MEMS 79 of 80 System Representing as block diagrams of multi-domain subsystems Device Reduced order macro models of the components Component (physical) Multiple, coupled energy behavioral simulations Artwork of masks and process Defining mask geometry for the process steps Each level involves design There is analysis (forward) problem and synthesis (inverse) problem.

80 Books Further reading Principles of microfabrication Marc Madou Micromachined transducers: A source book Greg Kovacs Microsystem Design Steve Senturia MEMS: Advanced materials and fabrication methods National Research Council (NRC) committee report, 1997 An Introduction to Microelectromechanical Systems Engineering N. Maluf Microsensors J. W. Gardner Sensor Technology and Devices L. Ristic Transducers, Sensors, and Detectors R. G. Seippel Microactuators: Electrical, Magnetic, Thermal, Optical, Mechanical, Chemical, and Smart Structures M. Tabib-Azar Nano- and Microelectromechanical Systems: Fundamentals of Nano- and Microengineering S. E. Lyshevski 80 of 80 The world wide web