Microsystems, MEMS Applications, Manufacturing Methods for MEMS PDF

See discussions, stats, and author profiles for this publication at: https://www.researchgate.net/publication/323069589 Microsystems, MEMS-applications, manufacturing methods for MEMS Presentation · February 2018 CITATIONS...

See discussions, stats, and author profiles for this publication at: https://www.researchgate.net/publication/323069589 Microsystems, MEMS-applications, manufacturing methods for MEMS Presentation · February 2018 CITATIONS READS 0 11,912 1 author: Kari Vierinen Helsinki Metropolia University of Applied Sciences 67 PUBLICATIONS 714 CITATIONS SEE PROFILE All content following this page was uploaded by Kari Vierinen on 09 February 2018. The user has requested enhancement of the downloaded file. Metropolia University of Applied Sciences Microsystems (3.0 op) 3/2015 Lecturer: Kari Vierinen Contents (Version 6/2015) 1.0 Introduction 1.1. Some of the advantages of MEMS devices are 1.2. Typical MEMS/MST-applications 1.3. Additional specific list of MST/MEMS-applications 2.0 What are the MST/MEMS materials? 2.1. Silicon (Si) 2.2. Silicon Dioxide SiO2 2.3. Silicon Nitride Si3Ni4 2.4. Metals 2.5. Polymers 2.6. Glass and Quartz 2.7. SiC and Diamond 3.0 MEMS/MST Applications 3.1. Capacitive Accelerometer 3.2. Pressure Sensors 3.3. Digital Micromirror Device™ (DMD™) 3.4. Microvalves 3.5. Micropumps 3.6. Microfluidics, Inkjet nozzles, Microfluidic Devices, Flow sensors 3.7. Gas and chemical sensors 3.8. Gyro and Angular rate Sensor 4.0 Thin-Film Deposition 4.1. Vacuum evaporation chamber 4.2. DC Sputtering system 4.3. CVD Reaction 4.4. Plasma-Assisted CVD Reaction 4.5. Epitaxial Growth 4.6. Thermal oxidation of Silicon 5.0 Etching 5.1. In isotropic Wet etching 5.2. Dry etching 5.2.1. Ion-beam etching (IBE) and glow-discharge sputtering  5.2.2. Chemical removal; Plasma etching. 5.2.3. Reactive Ion Etching RIE 5.2.4. Deep Reactive Ion Etching (DRIE), a special case of RIE. 6.0 Anodic bonding; Silicon Fusion bonding 7.0 Wire bonding; Wedge, Ball 8.0 Packaging; Hermetic sealing 9.0 References 1.0 Introduction Microsystems are the 2nd revolution of the Silicon based technologies. Silicon is very special material because of its unique electrical, mechanical, optical and thermal properties. Different physical and mechanical functional properties can be integrated into the Silicon based microchip. Electromechanical and optical operations are mostly used in different microsystems and MEMS-applications. Microsystem technology consists of microcircuits, micromechanics, micro-optics, microsensors and microfluidistics. Fabrication of microsystems are based on surface structures or so called 3D bulk technologies. Small micrometer size optical gas analyzers, micropumps and integrated microsensors and high frequency resonators and filters are examples of already existing applications. Main application fields are in Medical Instruments, Bioinstrumentation, Automotive , Consumer Electronics, Heavy Duty Vechiles, Avionics, Game applications, Mobile Phones, Tabloids,... Links for Introduction of Microsystems: MEMS-applications Youtube. Macro-Micro-Nano Scales. Goal for this course is to get familiar with Microsystems and to be ready to use Microsystem technology products for different applications at work and in product development projects. Electrical output signal coupling of micromechanics and other effects are studied. Mechanical energy interference with Electrical energy is going to be studied in detailed way in this course (Aplac, COMSOL Multiphysics, VHDL-Ansys,... What are Microsystems? Applications, markets and new technologies and solutions? Materials in micromechanics and special properties of MEMS-materials? - Year 1959 Richard Feynman: “There’s Plenty of Room at the Bottom“. - Year 1989 Micro Tele Operated Robotics Workshop (MTORW) Salt Lake City USA. - One of the first applications was so called micro-oscillator of the resonance modes of the mechanical system. Mechanical oscillations were coupled to produce electrical output signals. Exact timing and gating were one of the the main applications. - MEMS = Micro Electro Mechanical Systems, USA 1989 --->. - MST = MicroSystem Technologies is the name used in European countries. - “To learn the basics of the MEMS and going up the MEMS-learning curve happens in the traditional way: learning by doing“ - At year 1997 about 80 companies in USA and markets 2. 109 $; Year 2003 about 500 companies and markets 8. 109 $; Year 2005 about 2000 companies in USA and year 2010 about 4000 companies. An estimate for year 2012 was more than 6000 companies in the market area of world. - Data storage is one interesting application and one A4 page of text and pictures will take about 5.10 -12 m2 of space on Silicon surface. In the space of one A4 Silicon surface we could store more than 5000 pages of text. This mechanical storage is very interesting, because it is more or less permanent way to store information for very long time. - The micro mechanical device embedded with electronics/electrical system fabricated through a mix of integrated circuit manufacturing and micro-machining process where material is shaped by etching away micro layers is called Micro Electro Mechanical Systems (MEMS). The intelligent electronic system part is integrated in the same way than ASIC-device fabrication. The mostly used material for MEMS is Silicon. - Micro-Electro-Mechanical Systems might consists of mechanical elements, sensors, actuators, and electrical and electronics devices (ASIC= Application Specific Integrated Circuit) on a common Silicon or Ceramic substrate or separated in close distance from each other. - The sensors in MEMS gather information from the environment through measuring mechanical, thermal, biological, chemical, optical, and magnetic phenomena. The electronics then process the information derived from the sensors and through some decision making capability direct the actuators to respond by moving, positioning, regulating, pumping, and filtering, thereby controlling the environment for some desired outcome or purpose. The advantages of semiconductor IC manufacturing such as low cost mass production, reliability are also integral to MEMS devices. The size of MEMS sub-components is in the range of 1 to 100 micrometers and the size of MEMS device itself measure in the range of 20 micrometers to a millimeters. 1.1. Some of the advantages of MEMS devices are: 1. Very small size, mass, and volume 2. Very low power consumption 3. Low cost 4. Easy to integrate into systems or modify 5. Small thermal constant and heat capacity 6. Can be highly resistant to vibration, shock and radiation 7. Batch fabricated in large arrays 8. Improved thermal expansion tolerance 9. Parallelism The size of the MEMS-sensors is decreasing, MEMS-sensors are going to be very important part of ICT-systems, the amount of MEMS-sensors is increasing very rapidly, MEMS-sensors are going to be one of the most important elements of measurement systems, MEMS-sensors integrate to be part of the measurement systems and they are more intelligent and self supporting, Cloud- services and IoT-applications,... Figure 1. Number of Sensors as a function of time. 1.2. Typical MEMS/MST-applications: There are plenty of applications for MEMS. As a breakthrough technology, MEMS is building synergy between previously unrelated fields such as biology and microelectronics, many new MEMS and Nanotechnology applications will emerge, expanding beyond that which is currently identified or known. MEMS technology finds applications in the general domains: Automotive domain: 1. Airbag Systems, Controlling automotive movement changes 2. Vehicle Security and Alarm Systems 3. Intertial Brake Lights 4. Headlight Leveling 5. Rollover Detection 6. Automatic Door Locks 7. Active Suspension Consumer domain: 1. Appliances 2. Sports Training Devices 3. Computer Peripherals 4. Car and Personal Navigation Devices 5. Active Subwoofers Industrial domain: 1. Earthquake Detection and Gas Shutoff 2. Machine Health 3. Shock and Tilt Sensing Civil Engineering Services and Military: 1. Tanks, heavy duty Automobiles  2. Airplanes, Helicopter control 3. Equipment for Police forces, Firemen and Soldiers 1.3. Additional specific list of MST/MEMS-applications: - Sensors for airbag levitation - Micropumps, microfluidic products, Medical applications,... - Pressure sensors, Tires, Home appliances,... - Inertia Sensors, 3D accelerometers, 3D Gyros, navigation, autopilots,... - Inkjet printers, Biochips,... - Micromirrors, multimedia tools, optical routers,... - High density permanent storage of information - Micro analysis devices, micro mixers, Biosensors,... - Measurement of angle with respect to gravity,... - Optical network and optical network components, optical sensors,... - Accelerometers, Rollover, Coriolis,... - Fast switches, Optical,... - Multicomponent MEMS coupling, Sensor Fusion,... - Microrobotics, microautomation,... - Medical MEMS devices, sensors, medicin fluid controlling, mixing. - Bioinformatics - Applications in sport and recreationand health products - Games, kinestetics, Robotic applications for Games?,... - Human and Animal Sensing - Wireless applications - Context Sensing - Gas sensors - Sensor Fusion A. Biotechnology: 1. Polymerase Chain Reaction (PCR) microsystems for DNA amplification and identification. 2. Micromachined Scanning Tunneling Microscopes (STMs). 3. Biochips for detection of hazardous chemical and biological agents. 4. Microsystems for high-throughput drug screening and selection. 5. Bio-MEMS in medical and health related technologies from Lab-On-Chip to biosensor & chemical sensor. B. The commercial applications include: 1. Inkjet printers, which use piezo-electrics or thermal bubble ejection to deposit ink on paper. 2. Accelerometers in modern cars for a large number of purposes including airbag deployment in collisions. 3. Accelerometers in consumer electronics devices such as game controllers, personal media players/cellular phones and a number of Digital Cameras. 4. In PCs to park the hard disk head when free-fall is detected, to prevent damage and data loss. 5. MEMS gyroscopes used in modern cars and other applications to detect yaw; e.g. to deploy a roll over bar or trigger dynamic stability control. 6. Silicon pressure sensors e.g. car tire pressure sensors, and disposable blood pressure sensors.  7. Displays e.g. the DMD chip in a projector based on DLP technology has on its surface several hundred thousand micromirrors. 8. Optical switching technology, which is, used for switching technology and alignment for data communications. 9. Interferometric modulator display (IMOD) applications in consumer electronics (primarily displays for mobile devices). 10. Improved performance from inductors and capacitors due the advent of the RF-MEMS technology C. MEMS devices: Few examples of real MEMS products are: 1.Adaptive Optics for Ophthalmic Applications (Ophthalmology is the branch of medical applications that deals with the anatomy, physiology and diseases of the eye ). 2. Optical Cross Connects 3. Air Bag Accelerometers 4. Pressure Sensors 5. Mirror Arrays for Televisions, Dataprojectors and Displays 6. High Performance Steerable Micromirrors (DMD) 7. RF MEMS Devices 8. Disposable Medical Devices 9. High Force, High Displacement Electrostatic Actuators 10. MEMS Devices for Secure Communications D. MEMS devices used in Space exploration field include: 1.Accelerometers and Gyroscopes for inertial navigation 2. Pressure sensors 3. RF switches and tunable filters for communication 4. Tunable mirror arrays for adaptive optics 5. Micro-power sources and turbines 6. Propulsion and attitude control 7. Bio-reactors and Bio-sensors, Microfluidics 8. Thermal control 9. Atomic clocks 2.0 What are the MST/MEMS materials? 2.1. Silicon (Si) is the most important material, Glass (Pyrex), Ceramics (LiNb,...), Polymers, Semiconductors (GaAs, InGaAs,...), Metals ( Ti, Al, Cu, W, Ni, Au, Cr, Ag,...), SiO2, Si3N4, SiC, C AlN, Al2O3, Quartz, Graphene,... Silicon (bulk) has excellent mechanical and electrical properties, high heat capacity, special optical properties, not solvable for normal fluids, not poisonous material,... High purity Silicon is available for Medical and other applications. Silicon is in the form of single crystalline structure or also so called polycrystalline Silicon (Poly Silicon) structure is available for surface MEMS- applications. Amorphous Silicon is not in crystalline form and amorphous Silicon in some respects is like solids and in other respects is like liquids. Liquid crystalline structures are of growing interest because of display and other applications. Silicon 2-sided wafers are available in diameters of D = 100 mm, D = 150 mm, D = 200 mm, D = 300 mm (year 2003) and D = 450 mm (year 2012). Silicon material properties are an excellent compromise between performance and stability. High electrically resistive Silicon material can be made of extremely high degrees of purity Silicon. Also with dopants Boron, Phosphorus and Arsenic conductivity of Silicon can be adjusted an order of 8 orders of magnitude. Optically strong absorption of visible light makes Silicon very good material for photo detectors and solar cells. However Silicon is transparent for infrared wavelengths longer than 1100 nm and that property is used in IR-MEMS and -Microsystems. Silicon is hard and strong as steel with Young’s Modulus of 190 GPa for crystal orientation and 140 GPa for crystal orientation. Silicon in crystalline form is very brittle and the breakage pattern is different from steel. After so called yield point there is fracture of Silicon material with practically no plasticity. The yield strength for Silicon is about 7 GPa and for steel between 2 GPa and 4 GPa. For crystalline Silicon the so called fracture strain is very large about 4 %. Chemical etching rates are much faster in Miller index direction than in direction and this property is used in chemical etching processes. In Polycrystalline Silicon mechanical properties are very similar than in single crystalline Silicon. However there are mechanical tensions between different kind of crystalline structure areas inside Poly Silicon material and usually thermal heating or treatment is needed at about > 900 degrees of Celsius. Silicon is excellent thermal material because of its high heat capacity and good heat conductivity. For thermic MEMS applications Silicon is good choice. However Silicon is not good material for active optics, because of thermal molecular excitations. There are no excitations in Silicon releasing high frequency light emission. Mobility of free charge carriers in Silicon is limited. Lasers and LED- sources can’t be made from Silicon. Light detecting is possible and photodiodes can be fabricated from Silicon. Silicon is transparent for IR-radiation with wavelengths higher than 1100 nm. We could say that Silicon behaves like window glass for visible light if wavelengths are higher than 1100 nm. This is very good property for optical sensor applications in IR-radiation wavelengths. For short wavelengths < 400 nm reflection is > 60 % in smoothed Silicon surface. Mechanical properties of Silicon are very stable until temperature + 500 oC and at higher temperatures > + 500 oC Silicon is softening. In Poly Silicon structural changes are starting to occur at temperatures of > + 250 oC. Silicon is resistant or not solvable for many fluids and chemical compounds in normal temperature. Gases, biological liquids, brake liquids in mechanical systems, salt vapor, water, Freon,... In Medical MEMS-application this is very important aspect. Silicon is also not very sensitive to normal environmental variable changes. Figure 1. Properties of the most important MEMS materials. 2.2. Silicon Dioxide SiO2 SiO2 is electrically insulator and very stable compound. Silicon dioxide is excellent electrical and also thermal insulator. Oxide layer for crystalline Silicon can be fabricated in high temperature T > + 800 o C. Dry or wet oxidation processes are used for SiO2 layers. Silicon dioxide is thermally grown by oxidizing on Silicon at high temperatures above + 800 oC. Silica glass (SiO2) is excellent dielectric material for guiding and channeling light (n = 1,48 for λ = 1500 nm). Optical planar waveguide structures, dielectric mirrors and optical fibers are examples. Thermal treatment is usually needed to release mechanical tensions in SiO2 structures in fabrication processes or after fabrication. Figure 2. Properties of Silicon crystalline as a function of temperature. Figure 3. TEM (Tunneling Electron Microscope) image of boundary between SiO2 and crystalline Silicon. 2.3. Silicon Nitride Si3Ni4 Thin film applications and insulating layers are often made of Silicon nitride. Si3Ni4 is an excellent insulating barrier layer for ion diffusion. In medical MEMS applications Na+ and K+ ions are often problematic and they have to be kept outside MEMS sensing elements. Si3Ni4 is very strong and hard material and Young’s modulus is 320 GPa - 385 GPa depending fabrication method. There are no problems with internal mechanical stress in Silicon nitride layers. Si3Ni4 is very good masking material for chemical etching with alkaline chemicals (KOH). Si3Ni4 is widely used in optical MEMS applications and refraction index n is about 2.0 - 2.5. Si3Ni4 layers are fabricated wit CVD (Chemical Vapor Deposition) methods with low or high pressure. 2.4. Metals Thin metallic layers are fabricated by vacuum evaporation, sputtering or using CVD (Chemical Vapor Deposition). Electrical coating (electroplating) is also used for Au, Ni, NixFey,... Aluminum is one of the mostly used metal in coating and sputtering method for coating is usually used. Vacuum evaporation is also used for Aluminum, however controlling of the layer thicknesses for Aluminum might be in some cases problematic. Corrosion is one problem in Aluminum layers and temperature has to be kept below + 300 oC. For higher temperature applications Au, Ti, Cr and W can be used. Aluminum layers are sensitive to mechanical stress and there might be changes in metallic layer structure after several stress cycles. However Al is excellent optical material, because of good reflectivity of visible light and also IR-radiation. Au is good reflector for IR-radiation but for wavelengths λ < 500 nm there is practically no reflection. Platinum and Palladium are very stable and not very reactive materials with different chemicals. Electrochemically they are very stable. Usually we need electrical contacts in MEMS applications and Au, Pt (and Ir) are used in wire bonding contacts. Very thin adhesion layers are necessary and Cr, Ti and Ti-W are usually used, because of very good adhesion properties. Metallic multilayers consist of an adhesion layer and an intermediate Ni- or Pt-layer is usually used to solder with silver-tin or tin- lead alloys. On top of these adhesion contacts and intermediate layers also Au is used. Figure 3. Electrical properties of thin metallic layers and typical applications in MEMS. 2.5. Polymers Mechanical yield strength is low, flexible structures, insulating protective layers, photoresist masking, conductive plastics, piezoelectric polymers,... are some of the properties of polymers. They might  be deposited with varying thicknesses from nanometers to micrometers. In spin-coating thicknesses of photoresist are 1 µm - 10 µm. Ultraviolet (UV) light can be used to hardening of the polymer coating. Temperature limits are very low for polymers < 200 oC. Some polymers have unique and very sensitive absorption and adsorption for some gases. They are used for different gas sensors and for example to measure relative humidity (HumiCap Vaisala). In Deep Reactive Ion Etching (DRIE) Polymers are used for decreasing etching in sideways direction inside vertical cavities. 2.6. Glass and Quartz Glass and Silicon are used in many MEMS applications as companion materials. Glass and Silicon are bonded each other in many different ways. Pyrex glass and Silicon surface can be bonded together using temperature, contact pressure and electric field. This process is called Anodic bonding and Pyrex glass and crystalline Silicon surface are attaching each others by molecular forces. Strong and hermetically solid bonding can be reached. Glass has higher temperature expansion coefficient than Silicon and thermal interfacial stresses might be created in temperature changes. In special Pyrex glass mixture the thermal expansion coefficient is very close to the value of crystalline Silicon. Quartz is piezoelectric material and can be used for sensing applications. Mechanical stress will created opposite electric charges on surfaces. Measured potential difference or voltage is related to mechanical stress. Also external electric field variation in piezoelectric quartz will create mechanical stress variation. 2.7. SiC and Diamond Silicon carbide SiC is very hard and strong material and Young’s modulus is 450 GPa and yield strength 21 GPa. SiC has very good thermal properties. Thermal conductivity is about 4 times higher than with Silicon and specific heat capacity is about same. Melting temperature is very high about 2830 oC. SiC is piezoresistive material and its conductivity depends on mechanical stress. SiC can be grown or deposited on crystalline Silicon surface or epitaxially grown on crystalline Silicon surface. Polycrystalline SiC layers using CVD-process has been used for hard protective coating. Diamond has high hardness, resistance to harsh environments and excellent thermal properties. In accelerometers and some other applications so called DLC (Diamond Like Carbon) layers are used to protect against unwanted electrical contacts. 3.0 MEMS/MST Applications 3.1. Capacitive Accelerometer Please check the files: “Acceleration Sensors.pdf “, “Capacitive-accele-lecture1.pdf” and “Microaccelero Basics.pdf” in Folder “Accelerometer” Links to Muratamems accelerometers: Murata; Murata Digital Accelerometer Youtube 3.2. Pressure Sensors Please check the file: “Pressure Sensors.pdf” in Folder “Pressure Sensor” MEMS Pressure Sensor Operation Youtube The MEMS/MST- pressure sensors are based on bulk micromachining manufacturing technology. Piezoresistive sense elements are used to detect stress in a thin Silicon diaphragm in response to a pressure load. Capacitive methods are also used to sense the displacement of a thin Silicon diaphragm. The structure of a piezoresistive pressure sensor consists of four resistive elements in a Wheatstone bridge configuration. The stress is measured within a thin, crystalline Silicon membrane. The stress in membrane is, to first order approximation, linearly proportional to the applied pressure differential. The membrane deflection in typical measurement is less than one micrometer. The thickness and geometrical dimensions of the membrane affect the sensitivity, and consequently, the pressure range of the sensor. Devices rated for low pressure usually incorporate complex membrane structures. Figure 4. Pressure sensor with diffused piezoresistive sense elements. Piezoresistive sense elements R1 , R2, R3 and R4 are located at the points of highest stress. Two resistive sense elements R1 and R4 have primary axes parallel to the membrane edge and stress is decreasing the resistance. The other two resistive elements R2 and R3 have primary axes in perpendicular direction with respect to edge and resistance is increasing with the pressure load. The fabrication process of a typical pressure sensor relies mostly on steps standard to the integrated circuit processes. An n-type epitaxial layer of Silicon is grown on a p-type wafer. A thin, preferably stress-free, insulating layer is deposited or grown on the front side of the wafer, and a protective Silicon Nitride film is deposited on the back side. The piezoresistive sense elements are formed by locally doping the Silicon p-type using the masked implantation of Boron, followed by a high- temperature diffusion cycle. Etching of the insulator on the front side provides contact openings to the underlying piezoresistors. A metal layer, typically Aluminum, is then sputter-deposited and patterned in the shape of electrical conductors and bond pads. A square opening is patterned and etched in the Silicon Nitride layer on the back side. The process forms a membrane with precise thickness defined by the epitaxial layer. Anodic bonding in vacuum of a Pyrex glass wafer on the back side produces an absolute pressure sensor, which measures the pressure on the front side in reference to the cavity pressure (low pressure). For differential- or gauge-type pressure sensors, precisely drilled holes in the glass wafer provide vent ports. Figure 5. Fabrication steps for a piezoresistive gauge, or differential, bulk micromachined pressure sensor. The Silicon-fusion bonding has proven to be very useful to the design of bulk micromachined pressure sensors (Figure 6). Silicon-fusion bonding allows the forming of the membrane, after the etching of a reference cavity with inward sloping walls. Consequently, extremely small pressure sensors are feasible. The Silicon-fusion bonded sensor is fabricated with a first step etching the cavity in a bottom handle wafer. Silicon-fusion bonding of a second top wafer encapsulates and seals the cavity. Electrochemical etching or standard polishing thins down the top bonded wafer to form a membrane of appropriate thickness. The last and remaining process steps define the piezoresistive sense elements as well as the metal interconnects, and are very similar to those used in the fabrication of standard, bulk micromachined pressure sensors (Figure 5). Calibration and correction of error sources are necessary for the manufacture of precision pressure sensors. A specification on accuracy of better than 1.0 % over temperature range of - 40 oC to - 125 o C is typical of many automotive, medical, and industrial applications. Figure 6. A small size Silicon-fusion absolute-pressure sensor. Silicon-on-insulator (SOI) technology becomes very useful at elevated temperatures because the thin Silicon sense elements exist over a layer of Silicon dioxide, thus eliminating all p-n diode junctions. Adjacent Silicon sense elements are isolated from each other by shallow, moat-like trenches. The dielectric isolation below the sense elements completely eliminates the leakage current through the substrate as long as the applied voltages are below the breakdown voltage of the insulating oxide layer. Silicon-fusion bonding plays an important role in the making of the Silicon-on-insulator (SOI) substrates. A heavily doped, thin p-type layer is formed on the surface of one wafer, and an oxide layer is thermally grown on another wafer. Silicon fusion bonding brings the two subtrates together such that the p-type layer is in direct contact with the oxide layer. Etching thins down the stack and stops on the heavily doped p-type Silicon. A front-side lithography step followed by a Silicon etch patterns the piezoresistive sense elements. Gold (Au) metallization is sputtered or evaporated, and then lithographically patterned to form electrical interconnects and bond pads. The final step forms a thin membrane by etching a cavity from the back side. Double-sided lithography is critical to align the cavity outline on the back side with the piezoresistors on the front side. The front side does not need to be protected during etch of the cavity if EDP (Ethylenediamine Pyrocathecol) is used instead of Potassium Hydroxide; EDP is highly selective to heavily doped p-type Silicon, Silicon Dioxide and Gold (Au). Figure 7. Manufacturing of a SOI high-temperature pressure sensor.  3.3. Digital Micromirror Device™ (DMD™) Figure 8. DMD- element There are remarkable things about the Digital Micromirror Device (DMD) from Texas Instruments. The first is that it works at all: the second is that it works so reliably. The Digital Micromirror Device is at the heart of Digital Light Processing technology. It allows images to be projected and displayed which are brighter, sharper and more realistic than has previously been possible with alternative technologies. Applications : DLP- projector, DLP/HD- display monitors, DLP-Cinema display technologies, Optical telecommunication routers, 3D- displays, Optical sensing tools,... Figure 9. DMD- matrix. The DMD is an array of microscopically small, square mirrors - some half a million or more in a space no larger than a finger nail - each of which can be turned on and off thousands of times per second. Each mirror corresponds to a single pixel in the projected or displayed image. The development of new materials, new manufacturing processes and new technologies have enabled us to engineer mechanical products whose performance and reliability have shown to be unique. Individual mirrors are able to vibrate in the range of ± 20 o angles many thousand times in one second. Are the MEMS-mirrors mechanically strong enough for long operation cycles? Testing cycles have been repeated 10 9 - 1012 times and estimated lifetime for DMD’s is over 200 years. The micromirrors are tilting back and forth with very high frequencies and are they going to break under the strain of constant twisting? To demonstrate this concept, sets of devices have been tested through 1 trillion cycles in weather chambers, and they operated well in excess of the requirement for a ‘normal’ commercial lifetime. No broken DMD- MEMS devices were observed. Figure 10. DMD- elements before mirror deposition. Figure 11. Single DMD-element and its layers. The DMD: changing the rules How many of the mechanical devices with which we’re familiar (even today) have been the subject of a development program lasting close to 20 years - like the DMD? How many of them have been precision-engineered to tolerances of a micron or less, as the DMD is? (The average human hair is between sixty and one hundred microns wide). How many of them have been manufactured using the same manufacturing techniques, that are used to make the world’s most powerful microprocessors and computer memories, as is the case with the DMD? The DMD-questions So what do we think we know? What are the questions we ask ourselves about the DMD? Q1: We wonder whether the hinge - which allows the micromirror to tilt back and forth - will eventually break under the strain of constant twisting? Q2: We ask ourselves whether the micromirrors are likely to stick in one position, causing defects in the projected image? Q3: We ask whether such microscopically small components aren’t so fragile that they’ll break if handled roughly? Q4: We wonder if the enormous speed at which the micromirrors must switch will, over time, cause them to become damaged and inoperative? Q5: And we ask ourselves about the effect of dust: after all, if a human hair could cover five micromirrors, wouldn’t a piece of dust in the DMD surely cause one or more mirrors to become inoperative? The answers Each of these questions seems entirely reasonable, based as it is on our preconceptions. What follows is an explanation of what’s different about the DMD that helps us to understand why the answers to the above questions aren’t what we might expect. Question 1: won’t the hinges bend or break? Each micromirror is hinged, allowing it to rotate on its diagonal axis. Given that each mirror will be switched through twenty degrees thousands of times per second, it seems that it must, sooner or later, break. In fact, “hinge fatigue” has never been a problem for the Digital Micromirror Device. The mirror hinge is manufactured using ‘thinfilm’ technology. Thin films have distinctly different properties from the general macroscopic concept of bending metal. A thin-film material is said to be more ‘compliant’; in other words, it has less stiffness. Stiffness is the property of a material that causes the material to resist bending. The more the material resists, the greater the likelihood of its breaking. Consider, for example, a bar of aluminum which is repeatedly bent against itself: it will ultimately snap. Few of us have not at some time bent and re-bent a soft drink can until the metal has torn. But how many of us have taken a sheet of aluminum foil and attempted to cause it to break by bending it back and forth? The important relationship is between the size of the crystals - or ‘grains’ - which comprise the material of the item and the size of the item in question. When a material breaks, it is because of dislocations caused to the crystal structure within it. These dislocations migrate to, and accumulate at, joints between the grains. This has the effect of concentrating mechanical stress until the yield point of the material is exceeded - at which point, breakage occurs. In something as microscopically small as a DMD hinge, there is, in effect, no internal crystal structure - all crystals are at the surface of the material. What this means is that the stresses caused by crystal dislocations are relieved immediately on the hinge surface - before the hinge’s crystalline structure can be damaged. To demonstrate this concept, sets of devices were tested through 1 trillion cycles, well in excess of the requirement for a ‘normal’ commercial lifetime. No broken hinges were observed. In this case, the DMD ‘changes the rules’ because of the microscopic physical size of its mechanical parts: the laws that apply to the everyday objects in our lives work differently at this level of miniaturization. Figure 12. DMD layer structure. Question 2: are the mirrors likely to stick? When the micromirrors are switched between their ‘on’ and ‘off’ positions, they are held in place by electrostatic forces. It is true that during early DMD prototype development, some mirrors tended to stick to the underlying surface due to large (in sub-micron technology terms) adhesive forces. This, in turn, caused the mirrors to fail to switch. What would be the cause of such an adhesive force? There are two phenomena at work. The first phenomenon is the relatively straightforward one in which capillary water condensation will cause the landing tip and the landing surface to become ‘stuck’. To understand better how this happens, consider the case of two glass plates stuck together with water-an extreme example of the adhesive force of capillary condensation. By wetting the common interface of the two plates, a partial vacuum is produced at this interface due to the surface tension of the water on the glass. As a result, enormous forces are required to pull the plates apart. Figure 13. DMD-mirror angle as a function of voltage. Pull-in voltage has higher value than pull-out voltage The second phenomenon, which is probably familiar to those with an advanced semiconductor physics studies, is the phenomenon known as ‘van der Waals forces’. Van der Waals forces are short range forces ( strong force at close distances 1/r6 ) which cause materials to become attracted at the molecular level. In order to prevent the mirrors from sticking, three actions are taken in the design and manufacture of the DMD. First, the effect of van der Waals forces is significantly mitigated by the application, during the fabrication process, of a thin anti-stick layer which lowers the surface energy of the contacting parts. Second, the mirror was redesigned in 1995 to improve the ability of the mirror to overcome the remaining forces. The redesign added miniature springs to the mirror landing tips, as can be seen in Figure. These springs store energy upon landing and push the mirror away from the surface upon release. And third, to minimize the effect of capillary condensation of water, the DMD is sealed in a dry environment ( Argon gas with low pressure in cavity ) using special hermetic packaging designed to ensure that it stays dry throughout its lifetime. Thus the forces which would otherwise cause the DMD to stick were eliminated - and to improve the DMD’s operating margin, its ability to “bounce back” was dramatically improved. Question 3: will the mirrors fall off if the DMD is treated roughly? Most people find it hard - if not impossible - to imagine a device in which over half a million individual moving parts are arrayed in an area measuring less than 1.5 cm2. What they can imagine, however, is how fragile such a device must be. Surely it must be true that only the smallest shock or vibration will be enough to cause severe dislocation of the micromirrors when subjected to, for example, the robust treatment a projector will typically receive during its working life? Again, our thinking is conditioned by the phenomena we observe during everyday life. However, the objects with which we are familiar, and which exhibit such characteristics, have very different profiles in terms of their modes of vibration. Why are ‘modes of vibration’ important? Simply because it is the vibration that causes fracture and breakage to occur. A simple illustration of this is the classic example of a wine glass shattered by an opera singer’s voice. The sound waves - at the appropriate frequency - cause the glass to vibrate, and it is the vibration which causes the glass to break. The DMD superstructure has modes of vibration with frequencies which are at least two orders of magnitude above the frequency of vibration generated during normal handling and operation (the lowest frequency mode of the device is about 100 kHz: all of the other resonant modes are measured in MHz). As such, there is virtually no vibration coupling from the environment to the DMD array: the DMD has a much higher vibration frequency than can be generated with conventional shock and vibration sources. This is the theory: the fact is that laboratory testing has shown that the mirrors do not fall off as a result of shock or vibration. Many are surprised, when dropping a DMD on the floor and it does not cause thousands of tiny mirrors to start rattling around behind the transparent enclosure. Figure 14. Fabrication process of DMD-mirror. Only main fabrication processes are shown. Question 4: as the mirrors move so rapidly, surely they must eventually become damaged? To achieve the high quality image for which Digital Light Processing is renowned, the individual micromirrors - each of which corresponds to a single pixel in the projected image - must be switched on and off thousands of times per second. As such, each mirror is subjected to (relatively) enormous forces. To understand why the speed at which the mirrors switch does not cause them to fail, it must be understood that the size of the mirrors and the air gap are fantastically small (the weight of an individual micromirror is measured in millionths of a gram!) and that the mirrors rotate through only twenty degrees. In fact, in 5 microseconds, the mirror’s tip moves through just 2 micrometers: as such, the average velocity of the micromirror is just 40 cm/second. By comparison, the terminal velocity of an oak leaf falling from a tree is of the order of 100 cm/second - which makes the landing of the DMD mirror a relatively gentle event! Knowing this, it’s easy to understand why the speed with which the mirrors move will not cause them to break. Figure 15. Time response of DMD- mirror deflection in different pressure values inside the mirror cavity. Gas damping is really one of the most important things in mirror operations. Question 5: surely, tiny particles of dust could cause mirrors to fail? Yes, they could. So small are the dimensions of the DMD that even the smallest particle of contaminant could cause one or more micromirrors to become non-functional. The single most significant potential problem in fabricating a DMD is the existence of particles in the manufacturing environment (just as it is in the manufacture of microprocessors and other sophisticated electronic devices). The DMD’s are manufactured in exactly the same way, using exactly the same facilities, as the world’s most sophisticated semiconductor components. As such, the DMD manufacturing process - which occurs in a world-class wafer fabrication clean room facility - can take advantage of the huge amount that has been learned about contaminant particle elimination. In the manufacture of any sophisticated semiconductor device, initial yields are expected to be low. Process improvements are identified and implemented, and, over time, yields increase substantially. It is this “yield curve” which gives rise to the significant reductions in price that we have become accustomed to during the lifetime of, for example, a memory chip or a microprocessor. The DMD is, of course, hermetically sealed so that once the manufacturing process is complete, no contaminants can enter the enclosure. By early 1996, exhaustive tests (120,000 operating hours, which is approximately equivalent to 120 years in the life of a typical portable business projector) on 150 DMDs revealed only nine - 6% - which exhibited an increase in the number of nonfunctional pixels/mirrors. Detailed examination of each particle which had caused a failure allowed its source to be identified and eliminated. This effort has seen particulate contamination improved by over 80%. The very small possibility that a microscopic particle of dust was sealed into the DMD package during manufacture is now the only potential factor which might affect the operation of the DMD during its lifetime - all others having been progressively eliminated during its development. The Digital Micromirror Device has now reached a level of reliability which is the envy of competing technologies. Figure 16. Total efficiency of light power. The DMD: confounding our preconceptions So why does the reliability of the DMD appear to defy and contradict everything which we hold to be true about mechanical devices? Intuitively, it should not work: the fact is, though, that laboratory testing has thus far shown that it is capable of 100% reliable operation for periods of time measured in decades (and, in fact, testing has not yet revealed the point at which failure might be expected under normal operating conditions). Testing has put each micromirror through hundreds of billions of cycles - of being switched ‘on’ and ‘off ’ - yet there has been no sign of sticking or of fracture. Moreover, the DMD has been put through comprehensive shock and vibration tests - tests which exert far more significant forces on it than could be expected in normal commercial operation - and, despite its apparent fragility, micromirrors haven’t broken off or become misaligned as a result. There are three key reasons for the DMD’s ability to confound our preconceptions. The first relates to the microscopic size of the mechanical parts. Many of our preconceptions are based on our experience with objects that are readily observed by the naked eye. The microscopic size of the DMDs mechanical parts is therefore a part of the explanation for the way in which the DMD apparently contradicts those truths which we hold to be self- evident. The second reason is that the DMD is one of the first commercially produced mechanical technologies based upon conventional semiconductor techniques. As such, DMD technology derives significant benefit from building on the past experience and the current knowledge base of the semiconductor industry. Finally, while Digital Light Processing products featuring the Digital Micromirror Device only came to market during 1996, the DMD itself had been under development for twenty years in Texas Instrument and other laboratories. Much of that development effort was spent in progressively refining the design, architecture and manufacture of the DMD such that its reliability to commercial standards could be assured. The Digital Micromirror Device is an astonishing phenomenon. That it works, and works reliably, is almost beyond belief. It challenges our preconceptions about what should and should not be achievable in much the same way that seeing moving pictures did a hundred years ago, that seeing television pictures did fifty years ago and that seeing television pictures of a man walking on the moon did twenty-five years ago. The DMD defies and redefines - our understanding of what is possible. Figure 17. Three positions of the DMD mirror. Figure 18. Projection system of DMD. Materials used on DMD’s are Aluminum in mirrors, Gold in some special mirrors, metal-oxide semiconductor substrate for electrodes, Silicon dioxide for sacrificial layer and Silicon nitride for electrical insulation. Torsional bar hinge: a plate held by two bars of rectangular cross-section. When there is torque on the plate, the bars twist and thus rotate the plate. Address electrodes: used to actuate the mirrors. The information about the image is digitized and actuation of the mirrors to reproduce the image is done by a processor on the chip. Based on this, the electrodes attain voltage states corresponding to the ON or OFF position of the mirrors. CMOS: micromirrors are fabricated using a CMOS-compatible process. It is actually produced by bulk micromachining of layers of SiO2, a photoresist, and aluminum deposited over a CMOS substrate. Familiarity with and available expertise in CMOS fabrication, coupled with the ease of integrating electronics also on the same substrate, have made this processing technology popular. S-RAM: a memory cell fabricated on the CMOS substrate. It provides bias voltage to the electrodes, thereby determining the ON or OFF position. DMDTM-chips: made by Texas Instruments. 3.4. Microvalves Microvalves (FluistorTM) are used in industrial systems that require precision control of gas flow for manufacturing process, or in biomedical applications such as in controlling the blood flow in artery. Electronically programmable gas stoves, currently under development, require low-cost, electronically controlled gas valves. Miniature valves are important for the control of fluid-flow functions in portable biochemical analysis systems. A growing market for microvalves is in the pharmaceutical industry, where these valves are used as principal component in microfluidic systems for precision analysis and separation for constituents. Microvalves can be used for accurate drug delivering systems, controlling gas flow in gas chromatography systems and in controlling operations electrohydraulic braking systems (EHB). Microvalves operate on the principles of microactuation. Electrical ring shaped resistor on flexible Silicon diaphragm is closing the passage of the flow when heated by electric current. Removal of the heat from the diaphragm opens the valve. Diaphragm is 2 - 3 mm in diameter and 10 µm thick. The heating ring is made of Aluminum and is 5 µm thick and the valve capacity is 300 cm3/min at fluid pressure up to 690 kPa and electric power 1,5 W is needed. Figure 19. Microvalve operated by thin film heating resistor. Thermal actuation principle are also used for microvalves where heating liquid are used. When the temperature of the liquid rises, its pressure increase, thus exerting a force on a thin diaphragm wall, and flexing it outward. In this design flow rate is controlled from normally open valve to a fully closed state. The bending of the Silicon diaphragm regulates the amount of valve opening. Bending of the diaphragm is activated by heat supplied to a special liquid in the sealed compartment above the diaphragm. The heat source in this case is the electric resistance foil attached at the top of the device. Figure 20. Microvalve with heating liquid. Heating liquid FluorinertTM has boiling temperatures of (56 - 250) oC and very large temperature coefficient of expansion 1.30. 10-3 1/oC. This liquid is very good insulator and have a high dielectric coefficient value. Typical switching times are 0,50 s and corresponding electric power 500 mW. Figure 21. Microvalve by Redwood Microsystems. 3.5. Micropumps A primary application for micropumps is in the automated handling of fluids for chemical analysis systems and drug-delivery systems. Micropumps can be constructed by using the electrostatic actuation of a flexible diaphragm. The deformable Silicon diaphragm forms one electrode of a capacitor. It can be actuated and deformed toward the top electrode by applying a voltage across the electrodes. The upward motion of the diaphragm increases the volume of the pumping chamber and hence reduces the pressure in the chamber. This reduction of pressure causes the inlet check valve to open to allow inflow of fluid. The subsequent cutoff of the applied voltage to the electrode prompts the diaphragm to return to its initial position, which causes a reduction of the volume in the pumping chamber. This reduction of volume increases the pressure of the entrapped fluid in the chamber. The outlet check valve opens when the entrapped fluid pressure reaches a designed value, and fluid is released. The gap between the diaphragm and the electrode is 4 µm. The size of the micropump has a square shape with diaphragm 4 x 4 mm2 and 25 µm thick. The actuation frequency is between 1 - 100 Hz. At frequency 25 Hz, a pumping rate of 70 µl/min is achieved. Another type of micropump is called piezopump, where the wave motion of tube wall is created by piezocrystal. The wave motion in tube walls exerts forces on the contained fluid for the required motion. Figure 22. Micropump operated by electric force. The basic structure of the micropump consists of a stack of four wafers. The bottom two wafers define two check valves at the inlet and outlet. The top two wafers form the electrostatic actuation unit. The application of a voltage between the top two wafers actuates the pump diaphragm, thus expanding the volume of the pump inner chamber. When the applied AC-voltage goes through its null point, the diaphragm relaxes and pushes the drawn liquid out through the outlet-check-valve. This design is able to pump fluid in either a forward or reverse direction at frequencies higher than natural frequency of the flaps. Figure 23. Micropump operated by electrostatic actuation. 3.6. Microfluidics, Inkjet nozzles, Microfluidic Devices, Flow sensors Please check the file: “Microfluidics.pdf” in subfolder “Microfluidics”. 3.7. Gas and chemical sensors Gas sensing is concerned with surface and interface interactions between the molecules on the surface and the gas molecules to be detected. Many reactions are possible on the surface of sensors, and they can be accepted as the gas-sensing transduction schemes. However, the dominant reaction is a reversible gas-adsorption mechanism that occurs on the sensor’s surface. The adsorbed gas atoms inject conducting free electrons into or extract electrons from the semiconducting material, depending on whether they are reducing or oxidizing, respectively. The resulting change in electrical conductivity is directly related to the amount of the analyzed gas present in the sensed environment, thus resulting in a quantitative determination of the concentration of the gas present in the environment. The principle is that a suitable catalyst, when heated to an appropriate temperature, either promotes or reduces the oxidation of the combustible gases. The additional heat released by the oxidation reaction can be detected. The fundamental sensing mechanism of a gas sensor relies on a change in the electrical conductivity due to the interaction between the surface complexes such as O-, O2-, H+, and OH- reactive chemical species and the gas molecules to be detected. For process control and laboratory analytics, large and expensive gas analyzers are used. Micromachining simplifies this through miniaturization and reduces the cost. In the so called Pellister sensor, which is basically a heater resistor embedded in a sintered ceramic pellet, on which a catalytic metal Platinum is deposited. Other methods, such as nondispersive IR methods using pyroelectric IR sensors and solid electrolyte gas- sensing mechanisms, can also be used for sensing. Applications are in environmental monitoring, exhaust sensing in automobiles, air conditioning in airplanes, houses and sensor networks, breath analyzers, odor sensing in food-control applications,... Materials used in gas sensor thin films are metal oxides such as SnO2 and TiO2. Conductivity: a material property that quantifies the material’s ability to conduct electric current when a electric potential difference is applied. It depends on the number of free electrons available. Adsorption: the process of collection and adhere of ions, atoms or molecules on a surface. This is different from absorption, a much more familiar term. In absorption, the species enter into the bulk material, that is, the volume; in adsorption, they stay put on the surface. Desorption: the reverse of adsorption; species (ions, atoms or molecules) are given out by the surface. Combustion: a technical term for burning: a heat-generating chemical reaction between a fuel (combustible substance) and an oxidizing agent. It can also result in light (flame ). Principle of Operation: An active area suspended by four beams off the fixed part of the wafer is the main element of this gas sensor. The adsorption or reaction of a gas on this active surface of the semiconducting material induces a change in the density of the conducting electrons in the polycrystalline sensor element. The change in the conductivity is detected using electronic circuitry. The chemical reaction can be described in four steps: 1. Pre adsorption of oxygen on semiconducting material surface; 2. Adsorption of a specific gas; 3. Reaction between oxygen and adsorbed gas; 4. Desorption of reacted gas on the surface. In the above process of delivering electrons between the gas and the semiconductor actually represents the sensitivity of the gas sensor. While reacting with the gas, the conductivity of the semiconductor gas sensor decreases when the adsorbed oxygen molecules play the role of the acceptor, whereas the conductivity increases when the adsorbed oxygen molecules play the role of the donor. Metal electrodes deposited on the top of the formed membrane that contains the active area make the measurement of the resistance of the gas-sensitive layer possible. Generally, the electrodes are located underneath the sensing film. Usually, the electrode materials are Gold and Platinum and, in some cases also Aluminum or Tungsten. Figure 24. Metallic electrodes and resistance measurement of gas sensitive layer. 3.8. Gyro and Angular rate Sensor A micromachined vibrating gyroscope is an interesting dynamic device. It is shown schematically in the Figure below, in which a mass is suspended so that it can move in different directions. Its suspension is designed so that its stiffness is low in x- and y- translational motion and very high in z-translation and the three rotational directions. This gyroscope measures the angular rate about the z-axis of the frame on which it is mounted. If the gyroscope is mounted on a car to measure its gentle roll about its longitudinal axis, the car is considered as the frame. This gyroscope is a vibrating gyroscope because its mass is continuously set into vibration by applying a force. The electrostatic combs are present in the Figure below, because they are needed for applying the force. Figure 25. Micromachined gyroscope made by surface mounted method. It is sensitive in rotation with respect to z-axis. A gyroscope is a device that measures the angular rotation rate of an automobile, an aircraft, a ship, or any object on which it is mounted. The suspension of a gyroscope should have equal stiffness in the x- and y-directions. In the example of the car on which the gyroscope is mounted, the road is the fixed frame and the car is the rotating moving frame. In a Foucalt pendulum when the support of the pendulum begins to rotate about the z- axis as shown in the Figure below, the ball oscillating along the x-axis experiences Coriolis acceleration along the y-axis. In the Foucalt pendulum, the natural frequencies of the pendulum in the two perpendicular directions are the same, and this is the idea of vibrating rate gyroscope as well. The energy from one mode of oscillation is then transferred very efficiently to another mode. Figure 26. Foucalt pendulum. Many configurations are possible in order that the two frequencies corresponding to the two modes are the same. A familiar example is a tuning fork in the Figure below. Its two tines can vibrate in-phase (the first mode) and out of phase (the second mode). When the frame of the tuning fork’s base rotates while the tuning fork is vibrating in its second mode, its tines start vibrating in the plane perpendicular to their original plane of vibration - just like in the so called Foucalt pendulum. Since their original oscillations (the driven mode) are out of phase with each other, the Coriolis effect causes motion of one tine in one direction and the other in the opposite direction. This principle is used in a micromachined gyroscope shown in the Figures below. Figure 27. Gyroscopes based on fork oscillations created by Coriolis force. Just like the tuning fork, one can have a dual mass gyroscope. Here, too, the masses are constantly driven in-plane in resonance vibration but out of phase with each other. When the frame rotates about an in-plane perpendicular direction, one mass starts oscillating up and down while the other down and up because of the Coriolis effect. This motion can be sensed capacitively or piezoresistively and then correlated to the angular rate. In the Figures below is a micromachined dual mass gyroscope in drive and sense modes. Figure 28. Micromachined dual mass gyroscope in drive and sense modes. 4.0 Thin-Film Deposition Please check the file: “Thin-film-deposit.pdf” in subfolder “Thin-Film Deposition”. 4.1. Vacuum evaporation chamber Figure 29. Vacuum evaporation chamber. 4.2. DC Sputtering system 1. Positive Argon ions are generated in a plasma in a high vacuum chamber and accelerated toward a target material at a negative potential. 2. During acceleration, the Ar-ions gain momentum and strike the target. 3. The ions physically dislodge (sputter) atoms from the target. 4. The dislodged (sputtered) atoms migrate to the wafer surface. 5. The sputtered atoms condense and form a thin film on the wafer surface with essentially the same material composition as the target, following the stages for thin film growth. 6. Excess material is removed from the chamber by vacuum pump. Figure 30. Typical sputtering chamber. 4.3. CVD Reaction 1. Gas transport to deposition zone 2. Formation of film precursors 3. Film precursors at wafer 4. Precursor adsorption 5. Precursor diffusion 6. Surface reactions 7. By-product removal from surface 8. By-product removal from chamber reactor Figure 31. Typical CVD-chamber. 4.4. Plasma-Assisted CVD Reaction 1. Lower processing temperature 2. Excellent gap-fill for high aspect ratio gaps (with high density plasma) 3. Good film adhesion to the wafer 4. High deposition rates 5. High film density 6. Lower temperature offers wide range applications Figure 32. Typical plasma assisted CVD chamber. 4.5. Epitaxial Growth Figure 33. Epitaxial process. Figure 34. Vertical epitaxial deposition reactor. 4.6. Thermal oxidation of Silicon 5.0 Etching Photolithography is one way for producing high precision patterning on substrates in micro scale. Photolithography process involves the use of an optical image and a photosensitive film to produce a pattern on a substrate. In microsystems photolithography is used to set patterns for masks for cavity etching in bulk micro- manufacturing, or for thin film deposition and etching of sacrificial layers in surface micromachining, as well as for the primary circuitry of electrical signal transduction in sensors and actuators. Figure 35. Typical etching process with UV-photolithography. The general procedure of photolithography is outlined in the Figure above. The substrate is at the top left and it could be Silicon wafer, Silicon dioxide, Silicon nitride or other material. A photoresist is first coated onto the flat surface of the substrate. The substrate with photoresist is the exposed to a set of lights through a transparent mask with the desired patterns. Masks used for this purpose are often made of quartz. Patterns on the mask are photographically reduced from macro- or meso-sizes to the desired micro scale. Photoresist materials change their solubility when they are exposed to light. Photoresists that become more soluble under light are classified as positive photoresists, whereas the negative photoresists become more soluble under the shadow. The exposed substrate after development with solvents will have opposite effects in these two types of photoresists, as illustrated in the right column of the Figure above. The retained photoresist materials create the imprinted patterns after the development [see the step (a) in the Figure]. The portion of the substrate under the shadow of the photoresists is protected from the subsequent etching [see the step (b) in the Figure]. A permanent pattern is thus created in the substrate after the removal of the soluble photoresist. In MEMS fabrication photolithography needs to be performed in a class-10 clean room or better. A class-10 clean room means that the number of dust particles size of 0.50 µm or larger in a cubic foot ( 0.027 m3 ) of air is in the clean room is less than 10. Most other microfabrication processes can tolerate a clean room of class 100. Standard air means class-5 000 000 in a typical environment. Photoresists are sensitive to short wavelength light with wavelengths ranging from 300 to 500 nm. The Mercury vapor lamps are usually used in photolithography. Deep UV- light has wavelengths of 150 to 300 nm and in special applications for extremely high resolution x-rays are used (EUV), where wavelengths are 0.40 to 5.0 nm. Bulk micro scale manufacturing involves the removal of materials from the bulk substrates, usually Silicon wafers, to form the desired 3D-geometry of the microstructures. Physical or chemical techniques, either by dry or wet etching, are the only practical solutions. Etching, either the orientation-independent isotropic etching or the orientation-dependent anisotropic etching, is thus the key technology used in bulk micromachining. For deeper etching such as bulk etching of Silicon, a hard mask of other deposited material may be required. In wet etching the material is removed by chemical reaction between the etching solution and the material on the substrate surface. In dry etching, this is done by chemical/physical interaction between atoms of an ionized gas and the substrate. 5.1. In isotropic Wet etching the chemical etchants will attack the material uniformly in all directions. Isotropic etching is hardly desirable in micro scale manufacturing because the lack of control of the finished geometry of the work piece. Fortunately, most commonly used substrate materials are not isotropic in their crystalline structures. Three planes of Silicon crystals are of particular importance in micromachining. These are (100), (110), and (111) planes. The three orientations , and are the respective normal lines to the (100), (110) and (111) planes. In micromachining the orientation is the favored orientation, because in this orientation the wafer breaks or cleaves more cleanly than in the other orientations. The (110) plane is the only plane in which one can cleave the crystal in vertical edges. On the other hand the (111) plane is the toughest plane to treat. Thus the orientation is the least-used orientation in micromachining. This non uniformity in mechanical strength also reflects the degree of readiness for etching. The material on the (111) plane obviously is the hardest to etch. A 400:1 ratio in etching rates for Silicon in to orientations is possible. By referring to the arrangement of atoms in the Silicon crystals as illustrated in Figure below we will find that the (111) plane intersects the (100) plane at a steep angle of 54.74o. Thus, when a wafer whose face coincides with the (100) plane is exposed to etchants, we can expect different etching rates in different crystalline orientations. A pyramid with sidewall slope at 54.74o exists in the finished product. Figure 36. Anisotropic wet-etching. Despite the many advantages of anisotropic etching in controlling the shape of the etched substrates, there are some disadvantages: anisotropic etching is slow process (about 1 µm/min) compared to isotropic etching, etching rate is temperature sensitive and elevated temperature + 100 oC is necessary, which precludes the use of many photoresistive masking materials. Popular anisotropic etchants for Silicon are Potassium hydroxide (KOH), Ethylene- diamine and pyrocatecol (EDP), Tetramethyl ammonium hydroxide (TMAH). Etching rates for different chemicals for Silicon, Silicon dioxide and Silicon nitride are shown in the Figure below. Figure 37. Etching rates with different etchants and materials. The selectivity ratio of a material is defined as the ratio of the etching rate of Silicon to the etching rate of another material using the same etchant. The Silicon dioxide has a selectivity ratio of 103 meaning this material has an etching rate in KOH that is 103 times slower than etching rate of Silicon. Selectivity ratios for Silicon dioxide and Silicon nitride are shown in the Figure below. The high selectivity ratio of Silicon dioxide and Silicon nitride makes these materials suitable candidates for the masks for etching Silicon substrates. Figure 38. Selectivity with respect to crystalline Silicon. The timing scheduling of etching and the agitated flow patterns of the etchants over the substrate surfaces need to be carefully controlled in order to avoid serious under etching and undercutting as shown in the Figure below. Figure 39. Under etching. 5.2. Dry etching Dry etching can serve as a replacement to wet etching in several ways: 1. It offers the capability of non-isotropic (=anisotropic) etching of single-crystal, polycrystalline, and amorphous materials such as Silicon, oxides, nitrides, and metals. 2. Etch geometries with vertical side walls can be achieved even in single-crystal Silicon and other semiconductors. 3. Dry etching also eliminates the important manufacturing disadvantage of handling, consumption, and disposal of the large quantities of dangerous acids and chemical solvents used in wet etching. 4. Dry etching and resist-stripping operations use relatively small amounts of chemicals. The prohibitive cost of the equipment and other infrastructure required for dry etching usually compels many small companies to be content with wet-etching-based processes for their development activities. However, for fine geometry patterning, dry etching is indispensable.  Dry etching is synonymous with plasma etching because all the dry etching processes use plasma of either chemically inert or active species. Plasma is a fully or partially ionized gas composed of equal numbers of positive and negative charges and different unionized molecules and radicals. Plasma is produced by the collision of electrons energized by an electric field of sufficient magnitude, causing the gas to breakdown and become ionized and is considered the fourth state of matter. Several types of dry etching processes based on the etching mechanism can be attempted: 1. Physical removal a) ion milling or ion-beam etching (IBE); b) glow-discharge sputtering. 2. Chemical removal; plasma etching 3. Combination of the two first ones 1. and 2. ; reactive ion etching (RIE) 4. Deep reactive ion etching (DRIE), a special case of RIE. 5.2.1. Ion-beam etching (IBE) and glow-discharge sputtering Figure 40. Ion Beam etching (IBE). 5.2.2. Chemical removal; Plasma etching. Figure 41. Plasma etching. 5.2.3. Reactive Ion Etching RIE Figure 42. Reactive Ion etching (RIE). 5.2.4. Deep Reactive Ion Etching (DRIE), a special case of RIE. Figure 43. Deep Reactive Ion etching (DRIE). Figure 44. Example of plasma etching of Silicone. Please check the file: “Dry-etching.pdf” in subfolder “Etching”. 6.0 Anodic bonding; Silicon Fusion bonding Anodic bonding of Silicon to glass (also known as Field Assisted Thermal Bonding; FATB), is the oldest bonding technique in micro fabrication. It has many features that make it easy and useful: glass is a soft material that will conform at +400 oC to 500 oC bonding temperatures, sealing structures and irregularities of up to 50 nm hermetically. Native oxides, and thin grown or deposited oxides, do not prevent bonding. Anodic bonding can be visually checked through the glass side: bonded surfaces look black and non-bonding areas are seen as lighter. Not all glass materials are amenable to anodic bonding. Thermal expansion coefficient mismatch between Silicon and glass needs to be considered at two temperatures: bonding temperature and room temperature/operating temperature of the device. Glasses usually have higher coefficients of thermal expansion than Silicon, but the match at two temperatures is approximately met with glasses like Schott 8339 and 8329 and Corning 7070 and 7740 (Pyrex). CTE of Corning 7740 is almost constant 3.3. 10-6 1/oC from room temperature to +450 oC, and that of Silicon increases from 2.5 to 4.0. 10-6 1/oC. When glass is heated to +400 oC, Sodium oxide (Na2O) decomposes into Sodium and Oxygen ions. The bonding process uses -300 V to - 1000 V applied to the glass wafer. Sodium ions (Na+) move towards the glass top surface and Oxygen ions (O2- ) towards the Silicon wafer. This will create a depletion layer and electrostatics force pulls the Figure 45. Anodic bonding device. glass and the Silicon wafer together. The resulting electrostatic forces are very strong: if the thickness of the depletion region is 1.0 mm; electric field E is about 500 MV/m = 5.00. 108 V/m, and the electrostatic attractive force is proportional to E2. Oxygen ions react at the glass/Silicon interface according to Si + 2 O2- ----> SiO2 + 4 e- and Sodium ions are neutralized at the cathode. If higher temperatures are used, Sodium atoms will diffuse faster, and the depletion width is greater, leading to stronger atomic bonds. Bonding initiation is done by applying pressure at the wafer center, but, if bonding is done in vacuum, it is possible to bond without an initiation point. Bonding current increases rapidly at the initiation of bonding because contact area increases and then bonding current decreases exponentially as Oxygen ions react at the interface to form SiO2, and the oxide becomes thicker. When the current has dropped to 10 % of its peak value, bonding is termed finished. Typical bonding times are 10 min to 30 min. This is fairly long for a single-wafer operation, and special wafer holders have been designed so that wafer loading and unloading can be done while another wafer is being bonded. A sizable area of Silicon is needed for good bonding. At least a 200 µm ‘collar’ around a cavity or recess is necessary for hermetic sealing, but there are no standardized design rules for wafer bonding. Anodic bonding of multilayer structures is also possible: glass/Silicon/glass systems can be made in a single bonding step. Heating uniformity is important, and double side heating is usually employed. Contacting the middle wafer electrically can be difficult. Bonding two Silicon wafers or two glass wafers by anodic bonding is not possible, but deposited films between enable the bonding. Sputtered Pyrex glass on Silicon is a standard approach. Silicon nitride and Silicon carbide can be used for Silicon wafers, and deposited Silicon for glass wafers. Doped spin-on glass has also been experimented with. It is important for anodic bonding that a depletion layer is being formed at the interface, and this requires that the intermediate layer acts as an ion barrier. Figure 46. Anodic bonding process flow. Silicon fusion bonding can yield abrupt pn-junctions, when p-type and n-type wafers are bonded without oxide layers in the interface. The first step in fusion bonding is cleaning with ammonia-peroxide mixture. It is very effective in particle removal and it leaves the surface in a hydrophilic condition with Silanol groups (Si-OH). Surface are extremely smooth, roughness < 0.50 nm, which is essential for good bonding. Wafers cleaned with HF-last process result in Si-H terminated surfaces, which are rougher and prone to attract impurity particles. Deposited film surfaces are usually not smooth enough for bonding, but Chemical-Mechanical polishing (=CMP) can be done to achieve surface roughness below 1 nm required for successful bonding. Figure 47. Roughness of CVD surface with respect to Chemical-Mechanical-Polished CVD surface. Surface energy is the energy required to break a bond and create two new surfaces. Two wafers in close contact are bonded by hydrogen bonds, as shown in the first Figure below. We can get an estimate for surface energies from Silicon atom surface density 1015/cm2, and hydrogen bond energies 25 to 40 kJ/mol, which translate to about 200 to 350 mJ/m2. Measured values for room temperature-bonded Silicon wafers are between 50 to 80 mJ/cm2. This indicates that less than 100 % of the area is in contact with hydrogen bonds. The reaction that takes place during storage or anneal is Siloxane bond (Si-O-Si) formation. Si-OH + HO-Si ---> Si-O-Si + H2O Siloxane bonds are much stronger than Silanol hydrogen bonds, and measured surface energies are about 1300 mJ/m2. This surface energy is almost constant from 150 to 800 o C. Above temperature + 800 oC, the oxide becomes viscous and flows, which increases contact area and leads to higher surface energy, as shown in the Figures below. Figure 48. Bonding of Silicon surfaces and water removal. Wafer breakage will take place inside Silicon, because Si-Si bonds are weaker than the Si-O bonds. The wafer released during the formation of Si-O-Si bonds will oxide Silicon further. The thinner the oxide on the wafers, the more important is the effect of this oxide, if wafers with thick oxides are bonded. Water diffusion will be slow and the additional oxidation, minuscule. A combination of thin oxide wafer and thick oxide wafer is a compromise: oxidation will proceed according to the aforementioned equation, strengthening the bond, and hydrogen can dissolve in the oxide, preventing build-up of interfacial stress. Figure 49. Viscous flow of high temperature oxide. Figure 50. a) Single-crystal Silicon and polycrystalline Silicon. b) Two single-crystal Silicon wafers bonded by amorphous oxide layer. 7.0 Wire bonding; Wedge, Ball Wire bonding is the most common method for electrically connecting the metallic pads on the MEMS-application to the ASIC and MC; and to the package inner lead terminals on the lead frame or substrate. This high-speed operation spools and bonds a fine diameter wire from the MEMS-element bond pads to lead frame inner lead pads, with the ability to form multiple bonds per second (maximum speed is about 25 wire bonds per second). The tool bonds the wire at each die bonding pad or lead frame pad and steps to the next location. Wire bond placement accuracy is about (2 - 5) µm. The bond wire is either Au or Al (also Cu possible) wire because it bonds well to both die pads, ASIC and MC chip pads and lead frame inner pads, with common wire diameter between 25 to 75 µm (pad pitch of about 70 µm). Figure 51. Maximum electric currents in different kind of bond wires. The three basic types of wirebonding derive their names from the type of energy used during the wire termination process. The three methods of wirebonding are: Thermocompression bonding, Ultrasonic bonding and Thermosonic bonding. In Thermocompression bonding, thermal energy and pressure are used to form the wire bond of a gold wire to the die pad, ASIC chip pad, and the leadframe inner lead pad. A bonding mechanism, referred to as a capillary tip, positions the wire on a heated chip bonding pad and applies pressure. The combination of force and heat causes the gold wire and metallic pad to form a bond, referred to as a wedge bond. The capillary tip then feeds additional wire while moving to the leadframe inner pad, where another wedge bond is formed in the same manner. This wirebonding process repeats until all MEMS die pads, corresponding ASIC and MC chip pads and corresponding lead frame pads are bonded. Figure 52. Wire bonding from chip to bonding pads and connection to leadframe. Figure 53. Thermocompression bonds. Ultrasonic bonding is based on ultrasonic energy and pressure as a means to form a wedge bond between the wire and pad. It can form bonds between similar and dissimilar metals, such as Al wire/Al pad or Au wire/Al pad. The wire is fed through a groove in the bottom of the capillary tip (similar to thermocompression bond) and positioned over the die bonding pad. The capillary tip applies pressure and rapid mechanical vibration, usually at an ultrasonic frequency of 60 kHz (or as high as 100 kHz), to form a metallurgical bond. The substrate is not heated in this technique. Once the bond is formed, the tool moves to the inner lead pad of the lead frame, forms the bond, and breaks the wire. This process repeats until all die bonding pads are wire bonded to the appropriate ASIC and MC chip and lead frame inner pads. Figure 54. Ultrasonic wire bonding process. Thermosonic Ball bonding is a technique combining ultrasonic vibration, heat and pressure to form a bond referred to as a ball bond. The substrate is maintained at a temperature of approximately +150oC. The thermosonic ball bond has a capillary tip, made of tungsten carbide or ceramic material that feeds a fine diameter Au-wire vertically through a hole in its center. The protruding wire is heated by a small flame or capacitor discharge spark, causing the wire to melt and form a ball at the tip. During bonding, ultrasonic energy and pressure causes a metallurgical bond to form between the Au wire ball and the Al pad. After completion the ball bond, the bonding mechanism moves to the substrate inner lead pad and forms a thermocompression wedge bond. The wire is broken and the tool continues to the next die bonding pad. This ball bond/wedge bonding sequence has excellent control of the wire loop size between the bonding pad and inner lead pad, which is important for thinner ASIC- and MC-packages. Figure 55. Thermosonic bonding process. Wire bond quality measurement is very important part of the micro fabrication process. It is essential that wire bonding have a consistently high field. Two primary methods for assessing quality are visual inspection (AOI-devices) and pull tests. Visual inspection is done by looking at the wedge or ball and verifying that a good bond has formed. For example, a wedge bond should have a flat region where the ultrasonic vibration of the capillary tip came in contact. For a ball bond, there is deformation of the ball from the applied pressure, but it is undesirable to have excessive deformation. Wire pull tests provide a quantitative assessment of the wire bond quality. The pull test measures the strength of individual bonds and highlights where the bond fails, such as at the heel, interface between the wire and flat region. These numerical test measurements can be monitored with statistical process and quality factor control to assess process stability and trends. Customers ask this testing to be completed with different statistical standard methods. Testing results are enclosed in product database for customer usage. Figure 56. Wirebond pull testing. In Gyro + 3-axis Accelerometer product by Murata shown below might be as many as 56 wire bondings altogether. Ball wire bonding in the center ASIC + MC and 2 micromechanical sensors, and wedge bonding in the lead frame contacts. There is only about (5 -10) seconds time for one MEMS-element and it could mean about 20 wire bondings in one second. Figure 57. Wirebonding: MEMS-elements (Gyro and 3-axial accelerometer = Combo sensor) , ASIC, MC, Pads and Leadframe. 8.0 Packaging; Hermetic sealing In its basic form, a package is a protective housing with an enclosure to hold one or more dice, which forms a complete microelectromechanical device or system. The package provides, when necessary, electrical and fluid connectivity between the dice and the external world. A top Silicon cap attached, for example, by Silicon-fusion bonding can maintain a hermetic seal and hold a vacuum while protecting the sensitive microstructures from damage during saw and assembly. A top cap also allows the use of plastic molding, ubiquitous in low-cost packaging solutions. In this method, molten plastic flows under high pressure, filling the inner cavity of a mold, and encapsulating a metal lead frame upon which the die or capped microstructure rests. A crystalline Silicon cap protects the sensing elements of the Murata accelerometer during molding of the plastic package over the die. Fixed to ground potential, the cap also becomes an effective shield against electromagnetic interference (EMC). There are three general categories of widely adopted packaging approaches in MEMS. They are ceramic, metal, and plastic, each with its own merits and limitations. For instance, plastic is a low-cost and often small size (surface-mount) solution, but it is inadequate for harsh environments. The asking price for a plastic-package MEMS pressure or acceleration sensor is below 5 euro. In contrast, a similar sensor packaged in a hermetic metal housing may cost well over 30 euro. Ceramic packaging of optical MEMS can be complex and very expensive. This is true for DMDTM packages. The DMDTM type-A package for SVGA resolution displays consists of a 114-pin alumina (Al2O3) ceramic base, with metallization for electrical interconnects and Cu-Ag brazed Kovar seal ring. Wire bonds establish electrical connectivity between the die and metal conductors on the ceramic base. A transparent window, consisting of a polished Corning 7056 glass fused to a stamped gold-nickel- plated Kovar frame, covers the assembly. Resistance seam welding of the seal ring on the ceramic base to the Kovar glass frame provides a permanent hermetic seal. Two zeolite getter strips attached to the inside of the glass window ensure long-term desiccation. The particular choice of metal and glass window materials minimizes the mismatch in coefficients of thermal expansion, and reduces stresses during the high temperature (about + 1000 oC) metal-to-glass fusion process. Antireflective coatings applied to both sides of the glass window reduce reflections to less than 0.50 %. A heat sink attached to the backside of the ceramic package by means of adhesives keeps the temperature of the DMDTM within tolerable limits. Figure 58. DMD-package. Metal packages are attractive to MEMS for the same reasons the integrated circuit industry adopted the technology over 30 years ago. They satisfy the pin-count requirements of most MEMS applications; they can be prototyped in small volumes with rather short turnaround periods; and they are hermetic when sealed. Metal packages are expensive and costs are more than ten-times higher than equivalent plastic packages. Packaging solutions for harsh environments, namely those found in heavy industries and aerospace, can be complex and costly. The custom requirements of the application coupled with the lack of high volume market demand, has turned metal packaging for harsh environments into a niche art. One particularly interesting design is the metal packaging of media-isolated pressure sensors for operation in heavily industrial environments. The design immerses the Silicon pressure sensor within oil filled stainless-steel cavity that is sealed with a thin, stainless steel diaphragm. The Silicon pressure sensor measures pressure transmitted via the steel diaphragm and through the oil. The robust steel package offers hermetic protection of the sensing die and the wire bonds against adverse environmental conditions. Each stainless-steel package is individually machined to produce a cavity. Electrical pins are glass-fired in holes through the steel housing. The die is attached inside the cavity, and wire bonding to the electrical pins is completed. Welding of a stainless-steel diaphragm seals the topside of the assembly. Oil filling of the cavity occurs through a small port at the bottom, which is later plugged and sealed by welding a ball. Figure 59. Metallic pressure sensor MEMS-package. Molded plastic packages are not hermetically sealed. They dominate in the packaging of MEMS-applications because they are cost-effective solutions. Advances in plastic packaging have further improved reliability to high levels. Today’s failure rates in plastic-package MEMS-applications are extremely low. There are two general approaches to plastic packaging: Postmolding and premolding. In the first approach, the plastic housing is molded after the die is attached to a lead frame. The process subjects the die and the wire bonds to the harsh molding environments. In the premolding, the die is attached to a lead frame over which plastic was previously molded. The metal lead frame in either approach is an etched or stamped metal sheet consisting of a central platform and metal leads supported by an outer frame. The leads provide electrical connectivity and emanate from the platform in the shape of a fan. The metal is typically a Copper alloy of Alloy-42 (Ni42Fe58); the latter has a coefficient of thermal expansion which is closely matching that of Silicon. In postmolded plastic packaging, the lead frame is spot-plated with gold or silver on the platform and the lead tips to improve wire bonding. The die is then attached with adhesive or eutectic solder. Wires are bonded between the die and the lead tips. Plastic molding encapsulates the die and lead frame assembly, but leaves the outer edges of the leads exposed. These leads are later plated with tin or tin-lead, to improve wetting during soldering to printed circuit boards. Finally the outer frame is broken off and the leads are formed into an S-shape. Figure 60. Postmolded plastic package. The sequence of process steps differs for premolded plastic packages. First, a plastic body is molded onto a metal lead frame. The molded thermosetting plastic polymer encapsulates the entire lead frame, with the exception of the platform and the outer edges of the leads. Deflashing of the package removes any undesirable or residual plastic on the die-bonding areas. The molded body may contain ports or openings, which later may be used to admit a fluid (for pressure or flow sensing). The lead frame is spot-plated with gold or silver to improve wire bonding and soldering. At this point, the die is attached and wire-bonded to the lead frame. A protective encapsulant, such as Silicone gel, is then dispensed over the die and wire bonds. Finally, a premolded plastic cap is attached, using an adhesive. If necessary, the cap itself may also contain a fluid acces port. Figure 61. Molded plastic package of pressure MEMS-sensor. The molding process is a harsh process which involves melting the thermosetting plastic at approximately +175 oC, then flowing it under relatively high pressure (about 6 MPa) into the mold cavity before it is allowed to cool. The plastic material is frequently an epoxy. Novolac epoxies are preferred because of their improved resistance to heat. The temperature cycle gives rise to severe thermal stresses, due to the mismatch in coefficients of thermal expansion between the plastic, the lead frame, and the die. These material properties of the plastic, and especially its coefficient of thermal expansion, are carefully adjusted by the introduction of additives to the epoxy. Plastic packaging for MEMS is not governed by any standards yet, it often uses standard or slightly modified IC plastics packages. The development of new plastic packaging technologies for MEMS is going on and one interesting material are so called NLC-materials (Nano Liquid Crystals).  Figure 62. Process flow from Silicon wafers to final packaged products. Figure 63. Diamond saw process. Figure 64. Three wafers bonded together and individual MEMS-products separated using diamond saw. Figure 65. Leadframe and die attach. Figure 66. Die attach.  9. References 1. http://scme-nm.org/files/MEMS_Applications.pdf 2. http://www.youtube.com/watch?v=ebnpmf3kOq4 3. http://www.youtube.com/watch?v=A03AENwOVNY 4. http://www.youtube.com/user/SCME2012 View publication stats