JPL led the development of US rocket technology in ...

Development of Microactuator Technologies for Space Applications Dr. Eui-Hyeok (EH) Yang Nano and Micro Systems NASA Jet Propulsion Laboratory

NASA’s JPL Operated by Caltech

National Aeronautics and Space Administration Jet Propulsion Laboratory California Institute of Technology

•JPL led the development of US rocket technology in WWII. •JPL was transferred to NASA upon its creation in 1958. •Developed the first U.S. satellite, Explorer I. •JPL spacecraft have explored all the planets of the solar system except Pluto.

•

•

Has a dual character: – A Federally-Funded Research and Development Center (FFRDC) under NASA; – A division of Caltech, staffed with 5500 employees; Is a major national research and development (R&D) capability supporting: – NASA programs; – Defense programs and civilian programs of national importance.

Seventeen JPL Spacecraft, and Three Major Instruments, Operating across the Solar System


Cassini studying Saturn

Spitzer studying stars and galaxies in the infrared

Two Voyagers on an interstellar mission Ulysses, Genesis, and ACRIMSAT studying the sun

GALEX studying UV universe

Stardust returning comet dust

Mars Global Surveyor and Mars Odyssey orbiters; “Spirit” and “Opportunity” on Mars

Topex/Poseidon, QuikSCAT, Jason 1, and GRACE (plus ASTER, MISR, and AIRS instruments) monitoring Earth

Micro and Nano Devices for Space Applications

High-pressure, leak-tight microvalve for proportional flow control


Miniaturized, lightweight, self-latched Inchworm actuator

Microactuator Technology

Electroactive polymer mirror membrane

Large-stroke, large-area membrane deformable mirror for ultra-large telescopes

Nano-manufacturing Technology

Outline

• MEMS Deformable Mirror • Inchworm Microactuator • Piezoelectric Microvalve


Adaptive and active optics for space telescopes

Micropropulsion for microspacecrafts


Adaptive Optics, Why?

•Turbulence in earth’s atmosphere spreads out light; makes it a blob rather than a point. – Temperature fluctuations in small patches of air cause changes in index of refraction. – When they reach telescope they are no longer parallel; rays can’t be focused to a point: ` Parallel light rays http://www.ucolick.org/%7Emax/289C/

Point focus

`

blur

Light rays affected by turbulence


Adaptive Optics Adaptive optics system Star

Atmosphere

Courtesy: Lick Observatory

Turbulence

No AO

With AO Telescope

Beam splitter No AO

Camera

Intensity Deformable Mirror

With AO

Wavefront Sensor

61-element deformable mirror on a ground-based 3-m telescope

Control System

Ultra Large Space Telescope


Do we still need AO in Space? Future ultra-large space telescope mirrors Ultra-lightweight, flexible materials ÎLarge, localized surface errors

The resolution of an optical system is limited by the diffraction of light waves.

(Dekany et al. “Advanced Segmented Silicon Space Telescope (ASSiST)”, Proc. of SPIE, V.4849, p.103)

> 30 m

1. Adaptive wavefront correction at tertiary optics: Requiring deformable mirrors (DMs) scalable to large-area, large-stroke. 2. Active control of mirror surface: Requiring miniaturized inchworm actuators.

Existing Deformable Mirrors (DMs) Most deformable mirrors today have thin glass face-sheets.

MEMS DMs (electrostatic)

Courtesy Xinetics


BMC (Boston Univ.), Flexible Optical BV (Delft Univ.), Agil-Optics, Inc. (Stanford Univ.)

Courtesy BMC

Courtesy OKO

;LQHWLFVSKRWRQLFVPRGXOH &RVWa.[ 6WURNHaPP

/LPLWDWLRQLQVWRNHPLUURUVL]HRULQIOXHQFHIXQFWLRQ

BMC

Glass mirror

OKO, Lucent

Piezoelectric Stack

JPL DM architecture

PZT film

Actuators for DMs


Electrostatic Actuators Advantages No hysteresis Disadvantages Limited stroke (~2 Pm) High voltage (few 100’s V) Snap-down (deflection < ~1/3 of gap) Piezoelectric Unimorph Actuators Advantages Low voltage (~50V) Stroke proportional to actuator size High bandwidth (for our design) Disadvantages 10~20% Hysteresis PZT film process needs further improvement. (many unknown parameters)

Piezoelectric Unimorph Actuator: Actuation Principle Top Electrode (Cr/Pt/Au)


V

Bottom Electrode (Ti/Pt)

PZT film (1~5μm)

Device Silicon 10~20μm thick

E SOI wafer

Cavity

• Scalable to large-stroke for large-wavefront correction. • Highly scalable actuator count, potentially up to 106 actuators. • Fast response and low power (30 Ps/cycle, 4 nF)

Example: d31 mode actuation, 2.5 mm in diameter, the Si/PZT thickness ratio of 6, the electrode size 60%.

Fabrication Process

Cr/Au deposition PZT deposition (sol-gel (sol-geel method) Ti / Pt oxidation xidatio


Fabrication Process

Evaporate & pattern Cr / Au


Fabrication Process

Etch PZT to expose&bottom electrode: Evaporate pattern Cr / Au HCl:H2O (2:1)


Fabrication Process Deep RIE backside, evaporate, Prepare Mirror wafer pattern Cr/Au & Indium bumps Start with SOI wafer

Evaporate Backside Backsid B ackand sidepattern llithography itho thog grindium aphy &bumps a D Deep e p(~2 eep RI R RIE IEμm IE height) Remove e exposed exposed posed oxide po ox


Mirror wafer

Actuator wafer

Fabrication Process

Bo Bond Actuator wafer & A Mirror Membrane wafer M



Membrane Deflection (Pm)

Preliminary DM: Testing Mirror membrane

3.0 2.5

Center

2.0 1.5 1.0

Perimeter

0.5 0.0

Actuator wafer

Actuator alone

0

10

20

30

40

PZT Voltage (V)

50

60

Actuator: I 2.5mm, Actuator membrane: 20μm thick Mirror membrane: 20μm thick

Applied voltage: 50V Mirror surface

Silicon mirror membrane

PZT

Silicon actuator membrane

Actuator: I 2.5mm

Activated

Influence function: 33%

Y. Hishinuma and E. H. Yang, "Piezoelectric Unimorph Microactuator Arrays for Single-Crystal Silicon Continuous Membrane Deformable Mirror," IEEE/ASME Journal of Microelectromechanical Systems, Vol. 15, No. 2, April 2006.

Preliminary DM: Testing


DM: Actuator: I 1.0mm, 20μm thick Si membrane 1 pixel actuation

Act. 2 Act. 1

2 pixels actuation

Act. 2 Act. 1

Unimorph Actuator Modeling Membrane deflection function


a

Ea h r 0

From Thin Plate Theory* wr

§ c1 2 r · ln O r c c ¨ ¸, 0 r E a 2 3 ° 4 a ° © ¹ ® °O §¨ d1 r 2 d ln r d ·¸, E a r a 2 3 °¯ © 4 a ¹

wr OF r

O: Lagrange multiplier Find unknown coefficients using boundary conditions

7LPRVKHQNRDQGWoinowsky.ULHJHUಯ7KHRU\RI3ODWHVDQG6KHOOVರ


Unimorph Actuator Modeling Total energy of membrane under deflection

U el U s U M

U tot

(1) Elastic energy

(3) Potential energy due to bending moment D

1 DO2 ³ 0 2

Us

1 2 a § dF · SO ³ 2Sr ¨ ¸ dr 0 2 © dr ¹

2

(2) Stretching energy due to tensile stress

Yh3 12 1 Q 2

ª§ d 2 F 1 § dF · · 2 1 dF d 2 F º » dr 2Sr «¨¨ 2 ¨ ¸ ¸¸ 21 Q 2 r © dr ¹ ¹ r dr dr » «¬© dr ¼

U el

a

UM

S

Flexural rigidity Energy minimum condition

Y = 107 GPa : Young’s Modulus of Silicon Q= 0.22 : Poisson’s ratio

§ d F 1 dF · 1 ¸¸ MO ³ 2Sr ¨¨ 2 0 2 r dr ¹ © dr 2

Ea

EV p t p ¦ V i ti i

Stretching force per unit length

wU tot wO

M

V p E3 t p

e31 = -6 C/m2 VTi, Pt= 300 MPa VPZT= 260 MPA

heff 2

Bending moment

Vp

e~31 E3

Stress due to PZT

0 Ref: P. Muralt, et al., Sensors and Actuators A 53 p398 (1996)


Unimorph Actuation 1

5

-1 -2 -3

09/02/03b 400.opd PZT_wafer_08

-4

Z (Pm)

0

4

I mm

3

Measured 2

Model

1

0

-1.0

1

PZT:2Pm thick Applied voltage: 40V E=0.6

I mm

-0.5

0.0

0.5

1.0

1.5

10

4.0

Electrode 80% Si membrane 20um thickness at 40V

-1 -2 -3 09/02/03c 400.opd PZT_wafer_08

-4 -1.0

-0.5

0.0

x (mm)

20

30

40

Membrane Thickness (Pm)

x (mm)

0.5

1.0


Z (Pm)

0


Electrode 40% Si membrane 20um thick at 40V

PZT: 2Pm thick Membrane diam. 2.5mm

Deflection at 30V

3.5

Si membrane thickness 10Pm

3.0

15Pm 20Pm

2.5 2.0

5Pm

1.5 PZT_wafer_08

1.0 30

40

50

60

70

80

90

Electrode diameter (% of membrane diam.)


Dynamic Response 100

Actuator alone Resonant Frequency=47.32kHz Actuator+Membrane Resonant Frequency=42.17kHz

Phase (deg)

10 Gain

100

1

0.1

2.5mm diam. actuator DM01 Membrane response at A-3

0.01 10

2 2

4

6 8

2 3

4

Actuator alone

0 Actuator + Mirror mem.

-100

-200

6 8

10 10 Frequency (Hz)

2 4

4

6 8

10

5

2

10

2

4

6 8

2 3

4

6 8

10 10 Frequency (Hz)

Frequency Bandwidth: ~30 kHz Bandwidth limited by mechanical response

2 4

4

6 8

10

5


Large-Stroke Unimorph Actuator Membrane Deflection (Pm)

7

The PZT layer was patterned with optimized shapes to increase the amount of stroke.

PZT:2 Pm thick Applied voltage: 40 V E=0.6 diam.= 2.5 mm

6 5 4

Experiment

3 2 Model

1 0 10

7

PZT: 2-Pm-thick Silicon membrane: 20-Pm-thick

Deflection (Pm)

6

PZT

30

40

Membrane Thickness (Pm)

Au

C

5

20

A

B

4

C

3

Oxide

A 2

A. Continuous PZT B. Patterned PZT C. Patterned PZT & Si membrane

1 0 0

10

20

30

Volt

40

50

60

Silicon

Silicon

PZT

B Au


Current Status PZT

Au electrode

Indium post Mirror membrane

Ti/Pt

Actuator wafer

SOI wafer (a)

Mirror Membrane

Electrode Pads

• Optimization of the actuator structure and PZT films. • Stress compensation (both for actuators and mirrors) • Actuator hysteresis compensation.

(b)

(c)

Fabricated DM with 20 x 20 piezoelectric unimorph. (a) Crosssectional schematic (b, c) Photographs of the actuator arrays and the DM.

• Reliable fabrication to improve the mirror quality and the actuator stroke.

E. H. Yang, Y. Hishinuma, J.-G. Cheng, S. Trolier-McKinstry, E. Bloemhof, and B. M. Levine, “Thin-Film Piezoelectric Unimorph Actuator-based Deformable Mirror with a Transferred Silicon Membrane,” IEEE/ASME Journal of Microelectromechanical Systems, 2006. in-press

Outline



For space telescopes

Inchworm Microactuator: Why?


•Is an AO system at tertiary optics sufficient? Do we still need direct correction in mirror shape in addition to AO? •Deformable mirrors at tertiary optics cannot correct wavefront errors exceeding several microns. Generic Requirements

> 30 m

1. Adaptive wavefront correction at tertiary optics: Requiring deformable mirrors with large-area, large-stroke. 2. Active control of mirror surface: Requiring miniaturized inchworm actuators.

Ultra-Large Monolithic Mirror Top layer (nanolaminate)

Passive or active face sheet

Hinged truss (flexure hinges) with beams with actuators

Facesheet, passive or active

Qi

Qi+1 Ri

Ri+1

Flexure hinge


Ultra-large, ultra-lightweight, nanolaminate-based space telescope mirror concept.

The flexures are bonded at the points {Qi} and {Ri}, and the angles between the tangents to flexures at the points remain constant. A possible actuated beam concept:

Flexure 100 mg large-stroke actuator

Tube with electronics chip

Concept of thin mirror face sheet with hinged supporting lightweight actuating truss structures. (Gullapalli et al.) S. N. Gullapalli, E. H. Yang and S. –S. Lih, “ New technologies for the actuation and control of large aperture lightweight optical quality mirrors,” IEEE Aerospace Conference, Big Sky, Montana, USA, 2003.

Existing Inchworm Actuators


• MEMS Inchworm actuators [1~3]: 50~500PN max push force. No self-latching (zero-power latching). • Conventional inchworm actuators [4]: 100g (mass) [1] R. Yeh, et al., MEMS 01, Interlaken, Switzerland, Jan. 21-25, pp.260-264, 2001. [2] H. N. Kwon, J. H. Lee, MEMS 02, Las Vegas, pp. 586-593, 2002. [3] M. P. de Boer, et al.,JMEMS, Vol. 13, pp. 63-74, 2004. [4] Q. Chen, et al., MEMS 98, Heidelberg, Germany, pp. 384-389, 1998.

Target Performance • • • • • •

Max. Freq. Stroke Resolution Force Power Mass


~1 kHz > 1 mm 30 mN 100 PW ~ 100 mg

•This actuator has been designed for correcting the surface shape of a segmented or thin-monolithic mirror after deployment in space. •If microactuators weighing a100-mg are available, a hundred such microactuators per square meter will add only about 0.01 kg/m2.

Self-Latched Inchworm Microactuator


Before slider insertion

After slider insertion

Comb drive powered, Clutches separated

R. Toda and E. H. Yang, “Zero-Power Latching, Large-stroke, High-precision Linear Microactuator for Lightweight Structures in Space,” IEEE Micro Electro Mechanical Systems (MEMS) Conference, Istanbul, Turkey, January, 2006.

2-Point Clamping to 4-Point Clamping 1

2


3

4

2-point clamp actuator is prone to slider tilt Undesirable drag possible

A 4-point clamp actuator is more stable

B

PZT

Clutch driven by comb drive

(a) Unit A released. (b) Unit B laterally Unit B clutched. moved.

(c) Unit A clutched. Unit B released.

R. Toda and E. H Yang, “Four Point Latching Microactuator,” (NPO-42041)

Comb Drive Design: Electrostatic Force


㼟

W

㻌㼘㻙㼟㻌㼘㼓

Fs

P

Fe

㼔㼟

㼣

Capacitance at comb drive: C

(Ctip Cside )

W 2( w g )

§ w l s x · hW (l s x)hW ¸¸ # H0 g ¹ w g g (w g ) ©sx

H 0 ¨¨

Electrostatic force at comb drive Fe: 1 wC 2 1 2 H 0 hW V Fe V 2 g (w g ) 2 wx

Fs: Tether beam restoring force P: friction coefficient Fe: Electrostatic force For unclamping, Fe>Fs Maximum clamp load=PFs


Modeling: Slider Force Forces applied by slider

Assuming the static friction coefficient is 0.24, the estimated push force when actuating is approximately 24 mN. In power-off mode, the estimated clamping force is approximately 48 mN.

0.12 0.10 0.08 0.06

0.12 0.10 0.08

beam width 20um beam width 15um beam widthbeam 20umwidth 10um

beam width 15um beam width 10um

Bending force from one clutch: 0.06 24 mN

0.04

0.04 0.02

Fs: > 25 mN 0

0.00 0

0.12 0.10 0.08 0.06

Fe: > 35 mN

0.04 0.02

0.00

0.02

300V 250V 200V 150V 100V 50V

0.14

Electrostatic Force [N]

Beam Bending Force [N]

0.14

Beam Bending Force [N]

0.14

Hwang et al., IEEE MEMS 06, Istanbul, Turkey, Jan. 2006, p.210.

5 10 15 Tether Beam Displacement [um]

5 10 15 Tether Beam Displacement [um] Normal Force [N]

20

0.00 20

0

1

2 3 Comb gap [um]

4

5

Fabrication Process Sequence A

A’

SOI wafer (100Pm/1.5Pm/400Pm) Metal patterning (lift-off) Deep RIE RIE Backside Deep HF release


Fabricated Silicon Components


Fabricated and Assembled


Comb-drive actuator array Slider PZT stack

Rail

Gold wire 10 mm

Actuation: Measured Stroke


95um

Clutch

Cycle 0 Slider

Cycle 100

Cycle 200

Actuation speed: 1 cycle / 2 second Slider displacement [um]

After the 500-cycle actuation, the slider was moved by approximately 450 Pm. PZT voltage was 20V. There is no conceivable limit to the maximum stroke of our actuator other than constraints imposed by length of the slider and external load.

500 400

PZT 20V

300 200

PZT 10V

100 PZT 5V 0 0

100

200 300 400 Actuation cycle [cycle]

500

Slider Motion (3rd Batch Actuator)


500 Pm

Actuated @ 20-cycle per second with 100 V to PMN-PT

Pathfinder Mirror Modeling


Structure control to the Zernike* mode 4. Actuator displacement commands were calculated and were subsequently applied, and the structure response to the commands was computed. * In optics, the aberrations are often represented as a sum of special polynomials, called Zernike polynomials. Atmospheric random aberrations can be considered in the same way; however, the coefficients of these aberrations (defocus, astigmatism, etc.) are now random functions changing in time. A Zernike polynomial is defined in polar coordinates on a circle of unit radius .

Current Status • • • •


Improve the actuator design and the fabrication process. Address the packaging issue. Study integration feasibility with a membrane-mirror face-sheet. Model a pathfinder mirror consisting of the inchworm actuators.

Outline



For space telescopes

For microspacecrafts

Small Spacecrafts Requiring Microvalves NASA Sun-Earth Connection Mag Con 20 kg, 10s - 100s of S/C Launch 2010


Microspacecraft envelope: 1~ 10 kg mass, 10x10x10 cm3 Volume, 1 W power Î Requiring microvalves capable of fast-actuation, low-leak, highpressure and low-power operation for micropropulsion

Microvalve Requirements • Conventional microvalve technologies: mass/volume, power consumption

•

Leak Rate - 0.005 sccm He

•

Inlet Pressure - ~ 1000 psi

•

Actuation Speed - > PZT film thickness) - Membrane stress no longer dominates dynamics - Larger deflection with full circle electrodes

We decided to work only with thick-silicon unimorph design.


PZT Film: Preparation Lead acetate trihydrate

Titanium iso-propoxide

2-methoxyethanol

o

Dehydrate at 120 C

2-methoxyethanol

Stir at RT

PZT 52/48films

Stir at RT

Crystallization Anneal (RTA) 700 oC for 60 sec

Refluxt 120oC for 2 hrs Distill by-products at 120oC Acetylacetone

Zirconium n-propoxide

Pyrolysis 350-500 oC Spin-coat: 1500rpm for 30 sec

Multilayering a0.2Pm/layer

(Penn State process)

0.7M PZT 52/48 solution 20 mol% excess Pb

PZT Film preparation performed at Penn State University

Large-Area DM Fabrication

Bond Remove substrate silicon R Actuator wafer & wafer of Actuator Mirror Membrane wafer


Fabricated Unimorph Actuators Actuator arrays


WYKO optical interferometer PZT: 2μm thick Si: 15μm

60% Electrode I 2.5mm

2.5mm Applied voltage: 50V 1) 2) 3) 4) 5) 6) 7) 8)

Place actuator wafer on the WYKO stage Turn on WYKO monitor and white LED lamp Apply probe needle to top & bottom electrodes Find interference fringes on WYKO Make sure voltage on the power supply is zero. Turn on the power supply Take a profile measurement Get a cross-section of the profile. Place one cursor at a reference, another cursor at the center of the membrane. 9) Find the height difference between a reference point and the center of the membrane. Record. 10) Raise the voltage, repeat 7)~10).

ȝP

WYKO interferometer image of a PZT unimorph actuator under actuation.


Preliminary DM: Testing Deflection vs. Voltage 4

Deflection (Pm)

Strain

6

3

7 4

3

2

2

5

1st ramp-up

1

1

1

E Ec (coercive field)

2

3

4

0 0

10

20

30

40 Randomly oriented no strain

Volt

PZT: 2-μm-thick Si membrane: 1-μm-thick Electrode diam. 60% Diaphragm diam. 2.5mm

Slightly polarized up small strain

6

5

Randomly oriented no strain

Polarized upward large strain

4

Polarized downward large strain

Polarized upward large strain Reference: Kenji Uchino “Piezoelectric Actuators and Ultrasonic Motors”, 1997

Polarized downward large strain


Wafer-Scale Membrane Transfer mm 1.4

[Bonding and etching process] Thermo-compression bonding Mirror membrane to be transferred

Carrier SOI wafer Substrate

1.2

Backside silicon etch Teflon protective fixture for selective chemical etch

TMAH

1.0 0.8 0.6 0.4 0.2 0.0

mm 0

(c)

0.2 0.4 0.6 0.7 0.8 1.0 1.2

(d) nm

Timed droplet etch of buried oxide HF

(e)

Si PZT Indium Oxide

4 2 0 -2 -4 -6 -8

Pm 0 100 200 300 400 500 600 700 800 900

E. H. Yang et al. "A Wafer-Scale Membrane Transfer Process for the Fabrication of Optical Quality, Large Continuous Membranes," IEEE/ASME Journal of Microelectromechanical Systems, Vol. 12, No. 6, Dec. 2003.


C

I(t)

Vs t I t R I t

R

I t

Vs(t)

1 R

1 j ZC

Frequency = 800 kHz 't = 290 ns R=5.3 Ohm C=4.3 nF

k

64S 1.9 u 1011 15 u10 6 2 12 1.25 u10 3

1

Vs t

e jT Vs t

R2

1 Z 2C 2

T

§ 1 · tan 1 ¨ ¸ Z RC ¹ ©

Vs t Vs sin Zt

Stiffness of PZT Unimorph Actuator 64S Et 3 12 r 2

1 j ZC

f

10.2 t E 2 2S r 12 w

3

Force of PZT Unimorph Actuator 1μm displacement: ~ 7 mN

6876 N

m

t = thickness r = radius of disk E = Young’s modulus w = density

f = 41 kHz with t=15Pm, r=1.25mm

PZT Film: Characterization


After deposition of a single layer, the film is pyrolyzed to remove organics and heat-treated (typically at 700 oC) to crystallize the film. Dielectric constant: ~1000 Loss tangents of about: 2-3% Remanent polarizations: !20PC/cm2 Effective transverse piezoelectric coefficients: e31,f ~ –5 to –7 C/m2 80 60

Pr=36 PC/cm2 Ec=51 kV/cm

Polarization (PC/cm2)

40 20 0 -20 -40

Polarization hysteresis loop

-60 -80 800

600

400

200

0

200

400

600

800

E (kV/cm)

PZT Film preparation & characterization performed at Penn State University


)OLSSLQJ3RODUL]DWLRQ'LUHFWLRQ

Deflection (um)

1 0.8 0.6 0.4 0.2 0 -30

-20

-10

0

10

20

30

PZT voltage (V)

-Membranes deflecting downward for both +/- PZT voltage

Meeting Requirements of Several Applications


DMs

AOptics (PZT bimorph)

BMC (MEMS Electrostatic)

Xinetics (PMN Electrostrictive)

JPL (Projected near term performa nce)

JPL (projected ultimate performance)

How to get to the projected performance?

Stroke (Pm)

16

2

0.2

6

6~16

Optimizing unimorph structure

Operating temperature

Data N/A

Data N/A

RT only. (Cryo version does not work at RT.)

-55¶C ~ 50¶C

-140¶C ~ 100¶C

Optimizing PZT film by tailoring the transition temperature (data available).

Bandwidth (KHz)

12

3

2

30

20~100

Tailoring the actuator design

Voltage (V)

400

200

100

50

28~50

Thin film PZT

DOF

35

1024

4096

400

400~10,000

Photomask-based microfabrication

Active mirror area (mm)

20 (diameter)

20 u 20

70 u 70

50 u 50

250 u 250

Membrane transfer technique in conjunction with large-format arrays

Mirror material (before coating)

Polished PZT

Poly-Silicon

Glass

Silicon

Nanolaminate or other optical materials

Membrane transfer technique

Mass production

No

Yes

No

Yes

Yes

Microfabrication

Production cost when fully developed ($K)

Data N/A

50

800

50

50~100

Microfabrication

Inchworm Actuator: Objectives


• Objective: Develop self-latched inchworm microactuators capable of large-stroke, high-precision actuation for ultra-large, ultra-lightweight space telescope mirrors. Requirements (performance goal) • • • • • •

Max. Freq. Stroke Resolution Force Power Mass

~1 kHz > 1 mm 30 mN 100 PW ~ 100 mg

Conventional Inchworm Actuator

1

4

2

5

3

6


Typical inchworm actuation sequence •Conventional inchworm actuator [1]: 100g (mass) [1] Q. Chen, D. J. Yao, C. J. Kim, G. P. Carman, “Mesoscale Actuator Device with Micro Interlocking Mechanism”, MEMS 98, Heidelberg, Germany, pp. 384-389, 1998.


1st Generation Microactuator •Slider is clutched with holder and driver by electrostatic force - This version is not of self-latching. •Driver is pushed by external PZT for vertical movement Slider moves vertically (out of plane) Holder Bistable beam

Meandered flexure Slider

SOI substrate

Driver (moves vertically by external PZT) Holder

Footprint 2mmx4mm

Actuation Sequence

0. Initial 1. 2. 3. 4. Driver Slider position clamped moves back upto holder Driver

H

S D

H


Actuator Test Setup


Wafer LabVIEW PC

WYKO

Test setup

PZT actuator under prober

Power supply

WYKO screen image


Actuation Results

•Step-by-step measurement of actuation showed pull-back behavior. •Step height change was not repeatable. •Electrostatic clutching force appeared to be weaker than estimation. photoresist

scalloping footing Buried oxide

Actuation trend

Typical surface profile of DRIE sidewall ÆWeaker electrostatic force ÆFriction at scallop notch

Bi-Stable Beam Mechanism


•Bistable beam is buckled to its 2nd stable position; electrode gap narrows from 143 Pm to approximately 1 Pm. Measured displacement= 142 Pm

Bistable beam test structure

•Electrostatic force enables pull-in clutching.


Pull-in Force 0

1

2

3

4

5

6

1.00E+00

1.00E-01

7

F= e

2

/ 2g

2

g = gap,

Electrostatic force [N] Tether force [N]

V = voltage

1.00E-02

Force [N]

0

AV

A = area

1.00E-03

d = Fl3/(3EI) 750um

1.00E-04

h=15um

1.00E-05

1.00E-06

Slider pull-in analysis

1.00E-07

Displacement [Pm]

b=50um

1.5mm

S

750um


Microfabrication •DRIEÆHF ReleaseÆOxidationÆStenciled RIE •Reduce oxide stress by oxide etch at bistable beam. •Electrode sidewall oxide not etched.

Start with wafer Front RIE pattern HF release: oxidation Oxide RIE CF C FThermal //O ODeep RIE IDeep ESOI with w ith Stencil Steetch ncil 4Front 2R

Oxide etch (RIE) with stencil


Characterization Setup-LabVIEW •LabVIEW-controlled power relays provide square-wave at arbitrary timing. •Applied DC voltage 100~300V (for electrostatic pull-in), ~50V (for PZT actuator).

Timing chart

LabVIEW VI

2nd Generation Microactuator Before slider insertion


After slider insertion

(Top view)

Top Wafer

PZT Bottom Wafer (Cross section)

Clamp Slider Prober needle

Advantages •Latching: using tether restoring force. Zero power latching. •Travel distance not limited by meander suspension beam.

Slider


Working Principle 0. Power-off steady state 1. Top clamps released

4. Bottom clamps released

2. PZT expands

5. Slider moved down one step as PZT contracts

3. Top clamps closed

6. Bottom clamps closed (Power off)

Fabricated Structures


Comb Drive Fabrication Issue

The bottom of the comb is too wide because of insufficient etching.


The bottom of the comb is overetched.

Sticking Issue


• Sticking behavior: The successfully demonstrated slider did not move when the device was tested two weeks after initial successful actuation. – A Si-O-Si chemical bond was formed in the silicon interface (between the slider and the driver plate) – The slider coated with thermal silicon dioxide to avoid the sticking. – The problem solved.

Actuator Test Setup


Image Processing Using Matlab


Actuator images taken before and after 100-cycle actuation: Image processing was performed to accurately calculate the slider movement, by comparing two different images taken before and after the actuation. The resolution from the calculation is approximately 50 nm.

Modeling: Lateral Forces Applied on Clutch


Line of contact

Fixed normal displacement: 10 Pm at the line of contact

Lateral Displacement vs. Lateral Force with a Fixed Normal Displacement

Maximum-Stress vs. Lateral Displacement with a Fixed Normal Displacement

Assembly Sequence Driver fabrication

Rail fabrication


Slider fabrication

Slider insertion Individual comb drive actuation test

1. Fabricate rail guides.

Epoxy attachment with PMN-PT Wire bonding to Rail Mount on PCB, wiring with Ag-epoxy Actuation test

4. Bond Au wires.

2. Attach driver.

5. Attach the lid.

3. Insert slider.

6. Attach the PMN-PT to driver and rail.

Non-linearity: Why?


30

Translation [micron]

25

20

15

10

• 10V-PZT • 1-cycle/s

5

0 -50

0

50 Cycle

100

Equilibrium position shifts by 1.68 Pm due to the broken tether 150


Non-linearity: Why? 1U

2U

3U

Slider The clutches with a broken tether

1D

2D

3D

4U

4D

It is more difficult for the clutches to move to the right, because both 1U and 3U are against the movement. In the mean time, there is no resistance against the clutches to move to the left. The clutches with a broken tether

1U

2U

3U

4U

2D

3D

4D

Fixed clutches

Moving clutches

Slider 1D Fixed clutches

Moving clutches

•When the moving clutches are in the clamping position, they have uneven resistance for moving to the left or to the right. •When the moving clutches are in the releasing position, they have about the same resistance for moving in both directions.

It is about the same for the clutches to move to the right or left, because 2D is against the movement to the right, while 4U is against the movement to the left.

Comparison with SOAs


Requirements

Buleigh (conventional)

Sandia (MEMS)

JPL Demonstrated Performance

Push force, N

15

0.00005

0.05

Speed, mm/s

1

4

0.08 Æ 3

Travel, mm

100

0.1

0.5 / 130-cycle

Size, mm3

25x25x70

0.6x0.2x?

14x7x0.6

Mass, g

10

-

0.1

Position resolution, nm

1

10

50

Maximum motor frequency, KHz

1

Failed at 46

Force density (force x speed/mass) W/Kg

1

-

0.04 Æ 1.5

Glitch, nm

< Thermal SiO2 >

- rms roughness: 2.6 nm

- rms roughness: 0.3 nm

[Pm] 5 4


Test Setup Pressure gauge

Valve to be tested Relief valve Exhaust Gauge Needle valve

He-tank

Toggle valve

Dynamic Power Consumption -


Va +

P=1/2VIcosT=VrmsIrmscos T

4.9 :

T: phase difference between V and I

Vs I

Example; @ 50 Hz

Power [mW]

800 GND

600

Vs(peak)=10V

Vp-p=20 V

400

't=4.6 msec ? T=2Sf't =1.445

200 100 psi

0

0

100

200

300

Frequency [Hz]

400

Va(peak)=400 mV ? I(peak)=Va(peak)/R = 81.6 mV

? PZT power consumption: ~61 mW