PY5020: Nanomechanics Nanoscience MEMS and NEMS Toby Hallam hallamt@tcd.ie
Overview - MEMS (Micro-Electro-Mechanical Systems) - Some enabling technologies RIE CPD - Capacitive actuators Pull-in instability Switches First MEMs device (1965) Comb Drive Millipede - Capacitive sensors Accelerometer
Introduction MEMS are interdisciplinary in their design, fabrication, and operation. Engineering Mechanical (structures and phenomena) Electrical (electrical signals: detected, generated, processed) Chemical and Biochemical (reactions, processes, and kinetics) Science Physics and Biophysics (external world vs. materials/properties including living organisms at macro and nano scale) Biology (macro and nano effects in plants, animals, and humans observed by smart transducers) Technology Macro, Micro, Nano
Intrinsic Characteristics of MEMS Miniaturization: dimensions of MEMS structures are much larger than in VLSI ICs (µm). Further scaling leads to NEMS (nano) that are comparable/smaller than ICs (1-100 nm). Scaling laws describe how properties/behavior change with dimensions Scaling of spring constants (ex. behavior of cantilevers E- Young modulus of elasticity; l, w, t - dimensions k Ewt 3 k C 4l 3 L4 4EL 3 L Scaling Law of Area-to-Volume Ratio l L,w L,t L,C 3 Decreasing length of cantilever: smaller spring constant, higher resonance frequency (GHz) and quality factor (50,000), better sensitivity 2 area L 1 3 volume L L (important in all surface effects: forces friction, surface tension, van der Waals etc) Microelectronics Integration - the most widely used is that with CMOS
Building Blocks Major components in MEMS systems include Design Much more difficult than IC designs due to the interdisciplinary character of MEMS Design includes packaging Packaging is one of the most challenging step both in design and realization Transducers must be integrated with electronics Integration with ICs is another challenge for MEMS due to difficult issues of process compatibility Fabrication Silicon technology is widely used in MEMS with new step added Dimensions are usually much larger than those in ICs even for nanotransducers. To feel NANO you do not need to be in the nano-scale size! Other materials are included to perform required functions of transducers MEMS are frequently integrated with fluidics (polymers, glass ) Materials Materials that can perform required functions (thermo, piezo-, magnetoresististance ) Interaction with fluidics (half-cell potential, corrosion )
Micromachining - Subtractive
Reactive Ion Etching Wet etching Selective Undercutting Poor for high aspect ratio Reactive Ion Etching Some selectivity High aspect ratios Choice of chemistry Material Being Etched Deep Si trench Shallow Si trench Poly Si Al AlSiCu W TiW WSi2, TiSi2, CoSi2 Si02 Si3N4 Etching Chemistry HBr/NF3/O2/SF6 HBr/Cl2/O2 HBr/Cl2/O2, HBr/O2, BCl3/Cl2, SF6 BCl3/Cl2, SiCl4/Cl2, HBr/Cl2 BCl3/Cl2/N2 SF6 only, NF3/Cl2 SF6 only CCl2F2/NF3, CF4/Cl2, Cl2/N2/C2F6 CF4/O2, CF4/CHF3/Ar, C2F6, C3F8,C4F8/CO, C5F8, CH2F2 CF4/O2, CHF3/02, CH2F2, CH2CHF2
Deep Reactive Ion Etching 1. A standard, nearly isotropic plasma etch. Sulfur hexafluoride [SF6] Anisotropy is lost Chemical etching 2. Deposition of a chemically inert passivation layer. Octafluorocyclobutane [C 4 F 8 ] Bosch process (DRIE) Teflon-type sidewall protection Cyclic process Unlimited depth with high aspect ratio
Critical point drying Wet etching Surface tension scaling Drying from wet processing Exchange process liquid Use supercritical liquid CO 2 Continuous shift from liquid to gas MEMS/Biological/Aerogels
Simple cantilever fabrication
Capacitive actuator There are two ways of changing the energy in the capacitor: a) by changing the charge, b) by changing its geometry; hence, E(Q,x), with Q and x as state variables. x de = udq + F ext dx u: electrostatic potential [V] F ext : external net force; x : gap So then, the total energy of the system will be: E Q, x = 1 2C Q2 + 1 2 Kx 2 E Q, x = Q2 (d x) 2εA + 1 2 Kx 2 C = ε 0A d x d: separation between plates K: spring constant A: overlap area of the plates ε 0 : permittivity of the media between plates
Capacitive actuator de = E dx + E dq x Q Q x System energy a function of x and Q Differentiate the energy equation E = 1 Q 2 x Q 2 εa Kx = F ext E = xq Q x εa = Q C = u OR E Q, x = Q2 (d x) 2C + 1 2 Kx 2 Force is a combination of spring and electrostatic forces Potential in the system is just related to capacitance Electrostatic part of force shows no dependence on x, (Q constant) Also: 2 E x 2 Q = K 2 E Q 2 x = 1 C
Capacitive actuator Instead, we should attempt to reformulate the energetic relationship for potential of the system. Recall capacitance can be written Q=Cu E Q, x = 1 2C Q2 + 1 2 Kx 2 E u, x = 1 2 Cu2 + K 2 x2 = εau2 2(d x) + 1 2 Kx2 F ext = E x u = εau2 2(d x) 2 kx Electrostatic force component now has x dependence We can now find some balance point where spring and electrostatic forces balance F=0 kx = εau2 2(d x) 2
Capacitive actuator As we did before, figure out a system stiffness. 2 E x 2 u = k εau2 (d x) 3 2 E x 2 u = E x u = K Now relate this to the balance point from the last slide. 2 E x 2 u = 2kx d x k kx = εau2 2(d x) 2 Solution with F=0 x = 1 3 d Pull-in criteria u pull in = 8 kd 3 27 εa
NEMS using pull-in instability NEM switched capacitor structure Vertically aligned multiwalled carbon nanotubes mechanical movement of a nanotube relative to a carbon nanotube based capacitor defines ON and OFF states Carbon Nanotube Fabrication: Grown with controlled dimensions at pre-defined locations on a silicon substrate Compatible with existing silicon technology Vertical orientation allows for a significant decrease in cell area over conventional devices JAE EUN JANG, et. al. Nature Vol 3 January 2008 doi:10.1038/nnano.2007.417
CNT-NEMS switched Capacitor JAE EUN JANG, et. al. Nature Vol 3 January 2008 doi:10.1038/nnano.2007.417
Graphene RF NEMS switches Objectives Fabricate NEM capacitive switch with C on /C off > 100. Optimise design for operation at 2-5 GHz. Advantages of graphene for RF MEMS Low actuation voltage High speed Hallam et al., Phys. Status Solidi B, 1 4 (2015) Extracted capacitance and resistance values and their tunability with voltage
Graphene RF NEMS switches Fabrication Graphene release by wet etching of a sacrificial layer. Yield limited by the releasing step. Successfully released membranes allowed a first assessment of graphene performance for RF NEMS applications.
A different capacitive actuator Capacitive switch Capacitive drive Can use the pull-in instability to drive a microscale device Can we avoid the instability? What kind of force can we achieve?
Electrostatic comb drive Axis of motion Multiple capacitive elements Staggered fingers Capacitive gradient Fixed fingers Movable fingers
Electrostatic comb drive We can characterise one pair of fingers such that: d=finger separation a=finger depth And the work equation (from before) can be written with a modified capacitive term E(Q, x) = Q 2 d 2εa b x + K 2 x x 0 2
Electrostatic comb drive The capacitive area for the fingers A=ab is modified to a(b-x) when the structure moves in x F ext = E x Q = Q 2 d 2εa b x 2 + K x x 0 Here the force is dependent on position as the area of the capacitive element changes (even at constant charge) E(u, x) = u2 εa b x 2d + K 2 x x 0 F ext = E u x = u2 εa u 2d + K 2 x x 0 K u = 2 E x 2 u = K > 0 So, the instability is in d rather than the actuation direction (x) The electrostatic force is also proportion to the voltage!!
Electrostatic comb drive Much force, wow! F array = ε a d u2 N N: number of finger pairs Micromachining: a 2µm, b 2µm and u=15v 1000 fingers for 1 µn DRIE: a 200µm, b 2µm and u=15v 10000 fingers for 1 mn
Electrostatic comb drive - design Folded beams (movable support) Fixed comb Moving comb Anchor points Ground plate
What are comb drives used for? Large stroke (225um) Small footprint Fully integrated Journal of Micromechanics and Microengineering, Volume 23, Number 10 (2013)
Magnetic HDD head
What are comb drives used for? Multiple AFM heads Direct write on polymer storage media 1TB/inch 2 Graham Cross IBM Millipede Project 2000-2010
What are comb drives used for? Thermal indentation PMMA write media SU-8 silicon based underlayer Decreased pitch for erasure
Capacitive sensing C = ε rε 0 A d = εa d = εwl d C d = εwl d 2 = f(d) C = C εwl d = d d 2 d Big nonlinear response Gap closing capacitor Try to determine conditions for linear response C d + d = εa d + d Taylor expansion C d + d = C d 0 + C d d0 d + 1 2 2 C d 2 d 0 d 2 + o d 3 = εa d 0 1 d d 0 + d d 0 2 o d 3 For large d 0 and small d we can ignore higher order terms and use linear relation Typical C d = 1pF/m
Capacitive Sensing: Lateral Comb Drive C = εwx d C x = εw d C = C x x Constant term linearity!! Significant reduction in footprint Increase in sensitivity!!
Accelerometer from Comb sensor Test mass C = C x V sense = C C 0 V = x C = εwx d εw x d d εwx V V sense = V x x Basic laws of motion can be applied Damped Hook s law
And I want to offer a prize if I can figure out how to phrase it so that I don't get into a mess of arguments about definitions of $1,000 to the first guy who makes an operating electric motor a rotating electric motor which can be controlled from the outside and, not counting the lead-in wires, is only 1/64 inch cube. - Feynman, Plenty of room at the bottom (1959)
Next Lecture: Ink-Jet printing 12pm Mon 23 rd November