11
Texas Christian University Department of Engineering Ed Kolesar
Introduction toMicroeletromechanical Systems
(MEMS)Lecture 7 Topics
• Thermal TransducersThermal SensorsThermal Actuators
• Thermoresistive Effect• Shape Memory Alloy (SMA)• Thermal Actuator Devices and
Applications
Texas Christian University Department of Engineering Ed Kolesar
MEMS Overview
Micromachining: lithography, deposition, etching, …
Processes & Foundries
Devices & Structures
Methodology
History & Market
Introduction &
Background
22
Texas Christian University Department of Engineering Ed Kolesar
Thermal Transducers
• Thermal sensor: heat causes measurable (electric) effect
Change in resistivityThermocouple …
• Thermal actuator: heat causes motionThermal expansion…
Advantages: small size, high force/volume, … Disadvantages: high power, …
thermalsensor
Vout
thermalactuator
Vin
Texas Christian University Department of Engineering Ed Kolesar
1kg+1K
Thermal TransducersHeat capacity, c : ability to hold thermal energy; energy needed to
change unit mass by unit temperature. Usually measured in J/kg·K. Analogous to electrical capacitor.
Thermal conductivity, κ : ability to transfer heat, usually measured in W/m·K. Analogous to electrical conductivity.
Coefficient of thermal expansion (CTE), α : relative change in length per unit temperature change, usually measured in 1/K.
∆l = l0 α ∆T
Note: thermal strain εT= α ∆Tthermal stress σT = E εT= E α ∆T
1W +1K
1m
+1K
l0 ∆l
1J
33
Texas Christian University Department of Engineering Ed Kolesar
Thermal Transducers
c (J/kg·K) κ (W/cm·K) α (10-6/K)
Si 712 1.49 2.6Polysilicon 920 (?) 0.34 2.33SiO2 745 0.0138 0.35Al 903 2.37 25.0Au 129 3.18 14.2GaAs 325 0.56 5.4Polyimide (Dupont PI2611D) 3.00Polyimide (Amoco Ultradel 1414) 191
c (J/kg·K) κ (W/cm·K) α (10-6/K)
Si 712 1.49 2.6Polysilicon 920 (?) 0.34 2.33SiO2 745 0.0138 0.35Al 903 2.37 25.0Au 129 3.18 14.2GaAs 325 0.56 5.4Polyimide (Dupont PI2611D) 3.00Polyimide (Amoco Ultradel 1414) 191
Texas Christian University Department of Engineering Ed Kolesar
Thermal Transducers
Note: Different coefficients of thermal expansion cause stress in bonded layers and thin films.
[After Madou 1997]
T/K
W
200 400 600 800
α/10-6K8
6
4
2 Si
SiO2
Pyrex
Ni Co Fe Alloy
44
Texas Christian University Department of Engineering Ed Kolesar
Thermal Transducers
Heat flux, q : thermal energy flowing across unit area per unit time (“heat current”)
m) (inflow heat of direction in distance K)(in etemperatur
)m (inarea sectional-crossW/mK) (inty conductivi thermal
J) (in ferredheat trans ofquantity
in measured :common more
W in measured :law sFourier'
2
2
===
==
∇−=
−==
xTA
QmWTq
dxdTA
dtdQq
κ
κ
κ
r
Texas Christian University Department of Engineering Ed Kolesar
Thermal Transducers
Temporal diffusion of thermal energy:
Thermal diffusivity: ratio of thermal conductivity to “normalized” thermal capacity
/s)m(in y diffusivit thermal
K)(in etemperaturK)J/kg(in capacity heat )kg/m(in density mass
K)(in W/mty conductivi thermal
:Equation Diffusion
2
3
2
=
=⋅=
=
⋅=
∇=
c
Tc
Tcdt
dT
ρκ
ρ
κρκ
55
Texas Christian University Department of Engineering Ed Kolesar
Thermal Pressure Sensor
Thermal conductivity changes with temperature
Want to measure pressure in volume with dimensions < λ:gas molecules collide mostly with heater/sensorlinear relationship between conductivity κ and pressure P
[Kovacs 1998, p. 596]
K)(J/kg gas ofheat specific(m/s) molecules gas ofity mean veloc
)(1/m eunit volumper molecules ofnumber (m)path freemean
(kg) massmolecular average
3
31
⋅===
==
=
V
V
cvn
mcvnm
λ
λκ
Texas Christian University Department of Engineering Ed Kolesar
Resistivity TCR
C (graphite) 1390 -500
Al 2.83 3600
Au 2.4 8300
(µΩ·cm) (10-6/K)
Resistivity TCR
C (graphite) 1390 -500
Al 2.83 3600
Au 2.4 8300
(µΩ·cm) (10-6/K)
Thermoresistive EffectResistivity changes with temperature
yresistivit oft coefficien etemperaturdifference etemperatur
tureat temperay resistivit tureat temperay resistivit
)1(
0
00
0
==−=∆
==
∆+=
R
T
RT
TTTTRTR
TRR
α
α
66
Texas Christian University Department of Engineering Ed Kolesar
Shape Memory Alloy (SMA)
Materials that exhibit a temperature-dependent phase transition (“martensite” - “austensite” crystal structure)Mostly Ni/Ti alloy, but also Au/Cu, In/Ti
Advantages:
• considerable temperature dependent expansion/contraction
• relatively linear control
• very high stress (> 200 MPa)
• arbitrary shapes
• simple actuation
• life time: millions of cycles (?)
Advantages:
• considerable temperature dependent expansion/contraction
• relatively linear control
• very high stress (> 200 MPa)
• arbitrary shapes
• simple actuation
• life time: millions of cycles (?)
Disadvantages:
• special alloy material
• high temperature annealing
• low efficiency (≈3%)
• thermal: long time constants (?)
Disadvantages:
• special alloy material
• high temperature annealing
• low efficiency (≈3%)
• thermal: long time constants (?)
Texas Christian University Department of Engineering Ed Kolesar
02
21 µB
221
bE⋅ε
2PZT2
1bE⋅ε
Energy densities for different actuation principles
Actuation Maximum Energy Density Equation
Magnetic 0.95
SMA 10.4 -
Electrostatic 0.4
Piezoelectric 52[106J/m3]
Shape Memory Alloy (SMA)
[After Madou 1997]
77
Texas Christian University Department of Engineering Ed Kolesar
“Chevron” Actuator
• Current flows through expansion members (legs)
• Legs expand and push shuttle towards right
Anchor to Substrate
Anchor to Substrate
Shuttle
Legs
[Cragun and Howell, ASME IMECE, 1999]
Texas Christian University Department of Engineering Ed Kolesar
Thermal Actuator
• “Pseudo bimorph”• Current runs through
loop of released structure (usually polysilicon, “cold arm” with gold layer)
• Regions with higher current density (“hot arm”) are heated up more and expand more (“pseudo bimorph”)
Substrate
Hot Arm
Cold Arm
[Allen, Howard, Ruff and Kolesar, TCU, 1997]
88
Texas Christian University Department of Engineering Ed Kolesar
Thermal Bimorph
Two layers of materials with different coefficients of thermal expansion
Approximation: t1
t2
R
iit
ttR
i
i
layer ofstrain thermallayer of thickness
12
21
==
−+
=
ε
εε
Texas Christian University Department of Engineering Ed Kolesar
Polyimide Cilia Array
Thermobimorph and electrostatic actuator:length: 430 µm, height (out-of-plane) ≈ 120 µmrange of motion ≈ 10 µm in plane, 40 µm out of planelifting capacity ≈ 76 µN (250 µ N/mm2)
Thermobimorph and electrostatic actuator:length: 430 µm, height (out-of-plane) ≈ 120 µmrange of motion ≈ 10 µm in plane, 40 µm out of planelifting capacity ≈ 76 µN (250 µ N/mm2) [From Suh et al., 1996]
99
Texas Christian University Department of Engineering Ed Kolesar
Polyimide Bimorph MEMS
• Surface Micromachining of two polyimides with very different CTE
• Al connectors, bonding pads, and sacrificial layer for release of structures
• TiW heater loop• Nitride masking material• Spin-on deposition of
polyimide layers• Lower temperature → higher
curvature
[Figure: Suh et al., 1996]
Texas Christian University Department of Engineering Ed Kolesar
Polyimide Cilia Array
• 4 actuators per “motion pixel”
• 8 x 8 motion pixels per chip
• Up to 4 chips per array:1024 cilia total
Right: MEMS cilia array moving an ADXL50 accelerometer chip.
[From Suh et al., 1996]
• 4 actuators per “motion pixel”
• 8 x 8 motion pixels per chip
• Up to 4 chips per array:1024 cilia total
Right: MEMS cilia array moving an ADXL50 accelerometer chip.
[From Suh et al., 1996]
1010
Texas Christian University Department of Engineering Ed Kolesar
Fully Addressable Cilia Array
• Polyimide on top of CMOS
• 8 x 8 motion pixels
• Serial / parallel PC interface
• Interactive control software
• Polyimide on top of CMOS
• 8 x 8 motion pixels
• Serial / parallel PC interface
• Interactive control software
[From J. Suh et al., 1999]
Texas Christian University Department of Engineering Ed Kolesar
Related Work• Micro Actuator Arrays
Pister, Fearing and Howe (UC Berkeley) 1990 micro-air bearing
Ataka, Omodaka, Konishi and Fujita (U. Tokyo) 1993 and 1995 various µ-actuators
Böhringer, Donald, Mihailovich and MacDonald (Cornell) 1994 and 1996 electrostatic
Liu and Will (USC ISI) 1995 magnetic
Suh, Glander, Darling, Storment and Kovacs (Stanford) 1996 thermobimorph
Yonezawa et al., 1997 magnetic
Hirata et al. (Neuchâtel) 1998 micro-pneumatics
Ebefors et al. (RIT Stockholm) 1999polyimide thermal actuator
• Macroscopic (Robotic) Actuator ArraysLuntz, Messner and Choset (CMU) 1997
macro-actuator arrayBerlin et al. (PARC) 1997
air jets for paper transport Frei, Wiesendanger, Büchi and Ruf (ETH Zurich) 1998
electromotors
1111
Texas Christian University Department of Engineering Ed Kolesar
Kinematics of Actuation
Texas Christian University Department of Engineering Ed Kolesar
2-Phase and 4-Phase Gaits
4 cilia per pixel: n,e,w,sN = north cilium is upn = north cilium is down
4 cilia per pixel: n,e,w,sN = north cilium is upn = north cilium is down
1212
Texas Christian University Department of Engineering Ed Kolesar
Diagonal (Virtual) Gaits
Motion in diagonal direction:couple two neighboring cilia of each pixel to implement “virtual cilia”(by analogy to virtual robot legs [Raibert’93])
“virtual” 4-phase gait: news neWS NEWS NEws
Motion in arbitrary direction:alternate principal and diagonal gaits(by analogy to Bresenham line scan algorithm)
Texas Christian University Department of Engineering Ed Kolesar
Orienting with Cilia Array
squeeze force
Automatic positioning and orienting with a“squeeze field”
Force at location (x,y): f(x,y) = - (sign(x),0)
1313
Texas Christian University Department of Engineering Ed Kolesar
Rotating with Cilia Array
squeeze line
squeeze force
Automatic Centering and rotating with a skewed “squeeze field”
Force at location (x,y): f(x,y) = sign(x) (-1,ε)
Texas Christian University Department of Engineering Ed Kolesar
MEMS Cilia Array in Action
MEMS cilia moving a 3 x 3 silicon chip• Translation• Rotation • Centering and aligning
[Suh et al., S&A-A, 1997; JMEMS 12, 1999][Böhringer et al., IJRR 2, 1999]
1414
Texas Christian University Department of Engineering Ed Kolesar
Cilia Summary• MEMS cilia actuator array with integrated control
circuitry
• Electrostatic and thermal bimorph sensing and actuation
• High lifting capacity (≈ 70 µN / per cilium)
• Large motion range (up to ≈ 25 µm horizontal and ≈ 125 µm vertical)
• Implementation of open-loop manipulation strategies (transport, rotate, position, orient, ...)
• Interactive PC control interface
• Abstraction: force vector field
Texas Christian University Department of Engineering Ed Kolesar
Omni-Directional WalkingMicrorobot
• Cilia chips turned upside down• Matthew H. Mohebbi and Mason L. Terry
(UW MEMS Lab)
• Two cilia chip designStabilityEnables rotation
• Wiring9 Wires - Chip 1 (N,S,E,W), Chip 2 (N,S,E,W) and ground0.05 A per channel25 µm insulated Cu 99.99%
[Mohebbi et al., ASME IMECE 2001]
1515
Texas Christian University Department of Engineering Ed Kolesar
Video Motion
• Side view of MEMS robot moving in NW direction
Texas Christian University Department of Engineering Ed Kolesar
• Displacement of micro robot:
• Velocity of micro robot:
• Constants dependent on the properties of the cilia and operating conditions
• Constants found using Monte Carlo search method to minimizing the least squares fit
Data Modeling
ctbceatf −+=)('
)1()( ctebattf −−−=
0,, ≥cbabcafvo +== )0(' atfv =∞→=∞ )('
1616
Texas Christian University Department of Engineering Ed Kolesar
Experimental Results• Maximum velocity of 635µm/s (Northwest at 110 Hz)• Carrying capacity of 1.448 g
[Mohebbi et al., ASME IMECE 2001]
Texas Christian University Department of Engineering Ed Kolesar
• Larger displacement is seen with the northwest motion• Near linear correspondence between control frequency
and velocity below 60 Hz
Experimental Results
1717
Texas Christian University Department of Engineering Ed Kolesar
MEMS Robot Summary
• Thermal bimorph cilia arrays work for micro robot locomotion
• Maximum velocity of 635µm/s and carrying capacity of 1.4 g observed
• Each cilium has undergone close to 1 million actuations with no failures
• Accurate method for tracking robot investigated• Linear plus negative exponential model closely
describes movement of the robot• Further work includes:
Temperature – velocity correspondenceWalking surface effectsActuation voltage effects